From uche.ogbuji@fourthought.com  Sun Apr  1 00:16:06 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Sat, 31 Mar 2001 17:16:06 -0700
Subject: [XML-SIG] Metadata in XBEL
In-Reply-To: Message from "Fred L. Drake, Jr." <fdrake@acm.org>
 of "Sat, 31 Mar 2001 14:21:47 EST." <15046.11851.341533.770037@beowolf.pythonlabs.org>
Message-ID: <200104010016.RAA07490@localhost.localdomain>

> Uche Ogbuji writes:
>  > Yes.  I actually implemented an off-line merge earlier, but I think a 
>  > standardized merge indicator would be useful.
> 
>   To make this meaningful, do we need more discussion of what "merge"
> means, or should this be left entirely to clients?  I'm inclined to
> think we need a good description of the expected range of application
> and motivation, and the rest can be left to specific applications.

OK.  I agree as you say later that a merge can be at top level or folder 
level.  In either case, I'd use the following guidelines:

1. A merge element is of the form "<merge xinclude:href='[target uri]'/>.  The 
source is an XBEL file in which the merge element appears.  The target is the 
XBEL file referenced by the URI in the xinclude:href attribute of the merge 
element.

2. the current folder of the source is the folder in which the relevant merge 
element appears.  This can be the top-level "folder".

3. All bookmarks at the top level of the target are added to the source as if 
directly specified at the location of the merge element in the current folder.

4. If any bookmark element in the current folder of the source has an 
identical href attribute to a bookmark in the target, the bookmark in the 
target is ignored.

5. All folders at the top level of the target are added to the source as if 
directly specified at the location of the merge element (whether top-level or 
within a folder).  This addition involves a recursive copying of all the 
bookmarks and sub-folders contained in the target folder.

6. If any folder in the current folder of the source has an identical title 
child element to a folder in the target, the folder in the target is merged 
into the folder in the source according to this process as if he target folder 
and the matching source folder were both top-level xbel elements.

7. The expanded bookmark file is the result of applying this process to each 
merge element in document order of the merge source.  All merge elements in a 
merge target are first processed before incorporation into the merge source.


>  > That should instead be spelled
>  > 
>  > <merge include:href="file:/path/to/bookmarks/collection.xml"/>
>  > 
>  > Or such, so that processors that don't have first-class merge support can 
>  > still include the other file through xinclude.
> 
>   This syntax seems reasonable; I presume we'll want to include some
> way to mark multiple <merge/> sources with priorities to determine
> "who wins" in the presence of multiple sources for a bookmark; some
> applications will present all versions of a bookmark and others will
> only want to present one but make the determination based on the
> bookmark data.

The process above implicitly specifies that the priority is according to order 
of appearance of each merge element in the source, by document order.  Earlier 
merge elements take precedence.

>   I presume this <merge/> element should be allowed in both <xbel/>
> and <folder/> elements.  Do we want to do this in XBEL 1.1 or wait for
> more experiance before adding it?

Probably a question for the browser implementors.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From uche.ogbuji@fourthought.com  Sun Apr  1 00:18:26 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Sat, 31 Mar 2001 17:18:26 -0700
Subject: [XML-SIG] Metadata in XBEL
In-Reply-To: Message from David Faure <david@mandrakesoft.com>
 of "Sat, 31 Mar 2001 23:21:13 +0100." <200103312221.f2VMLEX02984@faure.worldonline.co.uk>
Message-ID: <200104010018.RAA07501@localhost.localdomain>

> On Saturday 31 March 2001 20:21, Fred L. Drake, Jr. wrote:
> > Uche Ogbuji writes:
> >  > Yes.  I actually implemented an off-line merge earlier, but I think a 
> >  > standardized merge indicator would be useful.
> 
> What's off-line merge ?

Basically, merging two XBEL files on the command line into a single XBEL file 
for re-import into browser format.  By on-line merge, I would imagine the 
browser starts up, and reads the master XBEL file, and processes any merge 
elements right then, therefore merging for the duration of the browser 
session.  This would provide the user max flexibility.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From stuff4gary@hotmail.com  Mon Apr  2 00:52:18 2001
From: stuff4gary@hotmail.com (gary cor)
Date: Sun, 01 Apr 2001 23:52:18
Subject: [XML-SIG] I am confused!.. can I pick someones brains on parsers?
Message-ID: <F5JwnzsLqIfQdmEB4HG000071a2@hotmail.com>

I am just starting doing my first bits of XML application development in 
python and for the large part am rushing around trying to get to grips with 
the concepts involved to create applications in publishing for editors.

It seems to me that there are loads of parsers in python for XML that do the 
similar things, I wonder if someone could try and explain to me how parsers 
differ from Dynamic Linking Extensions on a PC or a Macintosh extension 
files and developing for these systems.  Is it not possible to have one 
parser for image files (like a image gear DLL), another for dealing with 
text data, another for XHTML, another for SVG, another for movies, etc...  
My point being can't parsers be more application specific?  And if they are 
application specific how do I find which ones I should use where and the way 
each is to be used for its task within its application?

Kind Regards

Gary C
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.



From jtauber@bowstreet.com  Mon Apr  2 01:49:51 2001
From: jtauber@bowstreet.com (James Tauber)
Date: Sun, 1 Apr 2001 20:49:51 -0400
Subject: [XML-SIG] I am confused!.. can I pick someones brains on pars
 ers?
Message-ID: <C4EE90263CBFD411A0C800B0D0490095CCC5B7@bst-mail02>

One of the characteristics of XML is that it provides a common syntax
*regardless* of application and so a certain level of processing can be done
that is *not* application specific.

An XML parser or XSLT engine are examples of this.

Typically, however, you want to build application specific objects from your
XML. This can be done in a layered approach with a generic event-firing XML
parser connected to an application specific object builder.

If you look, for example, at the code for PyTREX (at
http://cvs.sourceforge.net/cgi-bin/cvsweb.cgi/pytrex/pytrex.py?rev=5.0&conte
nt-type=text/x-cvsweb-markup&cvsroot=pytrex) you'll see a series of handlers
that respond to events from the parser expat by building up application
specific objects.

Hope this helps.

James

> -----Original Message-----
> From: gary cor [mailto:stuff4gary@hotmail.com]
> Sent: Sunday, April 01, 2001 7:52 PM
> To: xml-sig@python.org
> Subject: [XML-SIG] I am confused!.. can I pick someones brains on
> parsers?
> 
> 
> I am just starting doing my first bits of XML application 
> development in 
> python and for the large part am rushing around trying to get 
> to grips with 
> the concepts involved to create applications in publishing 
> for editors.
> 
> It seems to me that there are loads of parsers in python for 
> XML that do the 
> similar things, I wonder if someone could try and explain to 
> me how parsers 
> differ from Dynamic Linking Extensions on a PC or a Macintosh 
> extension 
> files and developing for these systems.  Is it not possible 
> to have one 
> parser for image files (like a image gear DLL), another for 
> dealing with 
> text data, another for XHTML, another for SVG, another for 
> movies, etc...  
> My point being can't parsers be more application specific?  
> And if they are 
> application specific how do I find which ones I should use 
> where and the way 
> each is to be used for its task within its application?
> 
> Kind Regards
> 
> Gary C
> ______________________________________________________________
> ___________
> Get Your Private, Free E-mail from MSN Hotmail at 
http://www.hotmail.com.


_______________________________________________
XML-SIG maillist  -  XML-SIG@python.org
http://mail.python.org/mailman/listinfo/xml-sig


From ken@bitsko.slc.ut.us  Mon Apr  2 02:23:23 2001
From: ken@bitsko.slc.ut.us (Ken MacLeod)
Date: 01 Apr 2001 20:23:23 -0500
Subject: [XML-SIG] I am confused!.. can I pick someones brains on parsers?
In-Reply-To: "gary cor"'s message of "Sun, 01 Apr 2001 23:52:18"
References: <F5JwnzsLqIfQdmEB4HG000071a2@hotmail.com>
Message-ID: <x5zoe0hz78.fsf@bitsko.slc.ut.us>

"gary cor" <stuff4gary@hotmail.com> writes:

> It seems to me that there are loads of parsers in python for XML
> that do the similar things, I wonder if someone could try and
> explain to me how parsers differ from Dynamic Linking Extensions on
> a PC or a Macintosh extension files and developing for these
> systems.  Is it not possible to have one parser for image files
> (like a image gear DLL), another for dealing with text data, another
> for XHTML, another for SVG, another for movies, etc...  My point
> being can't parsers be more application specific?

Yes they could, and some (grovesy folks*) would argue that there
should be lots of them.  That's not the current, common practice
though.

> And if they are application specific how do I find which ones I
> should use where and the way each is to be used for its task within
> its application?

Like the mapping from mime-types to applications in browsers, there
would also be some kind of configuration that would map mime-types to
their respective parsers.  All this is just brainstorming, though,
since no such system exists as yet.

There recently has been a new specification out called Resource
Directory Description Language (RDDL) <http://RDDL.org/> that may be a
great help in mapping data types to resources that apply to those data
types (parsers, style sheets, documentation, validators, services,
etc.).  It's currently focused on providing resources for XML
Namespaces, but the parallel to any type of data is very clear.

  -- Ken

* The term "groves" refers to an abstract way of modeling data as
nodes and properties (using objects and attributes in Py), with the
ability to do processing on any set of nodes in a common way.
Resource Description Format (RDF) is a similar abstract model with
roughly the same characteristics.


From ken@bitsko.slc.ut.us  Mon Apr  2 02:51:48 2001
From: ken@bitsko.slc.ut.us (Ken MacLeod)
Date: 01 Apr 2001 20:51:48 -0500
Subject: [XML-SIG] I am confused!.. can I pick someones brains on pars 	ers?
In-Reply-To: James Tauber's message of "Sun, 1 Apr 2001 20:49:51 -0400"
References: <C4EE90263CBFD411A0C800B0D0490095CCC5B7@bst-mail02>
Message-ID: <x5wv94hxvv.fsf@bitsko.slc.ut.us>

James Tauber <jtauber@bowstreet.com> writes:

> One of the characteristics of XML is that it provides a common
> syntax *regardless* of application and so a certain level of
> processing can be done that is *not* application specific.
> 
> An XML parser or XSLT engine are examples of this.
> 
> Typically, however, you want to build application specific objects
> from your XML. This can be done in a layered approach with a generic
> event-firing XML parser connected to an application specific object
> builder.

Rephrasing what James is saying, tying it to what I just posted,
Groves and RDF provide a common data model, and again, "a certain
level of processing can be done that is *not* application specific",
but also at a higher level than XML elements and attributes.

The RDF IG folks, for example, have recently been talking about
RDFTransform (XSLT mapped to RDF), RDFPath (XPath for RDF), etc.

This is all mostly academic at this point, so I hope no one is
bothered by my rambling about it!

  -- Ken


From david@mandrakesoft.com  Mon Apr  2 15:18:41 2001
From: david@mandrakesoft.com (David Faure)
Date: Mon, 2 Apr 2001 15:18:41 +0100
Subject: [XML-SIG] Metadata in XBEL
In-Reply-To: <200104010016.RAA07490@localhost.localdomain>
References: <200104010016.RAA07490@localhost.localdomain>
Message-ID: <200104021418.f32EIgM08766@faure.worldonline.co.uk>

On Sunday 01 April 2001 01:16, Uche Ogbuji wrote:
> > Uche Ogbuji writes:
> >  > Yes.  I actually implemented an off-line merge earlier, but I think a 
> >  > standardized merge indicator would be useful.
> > 
> >   To make this meaningful, do we need more discussion of what "merge"
> > means, or should this be left entirely to clients?  I'm inclined to
> > think we need a good description of the expected range of application
> > and motivation, and the rest can be left to specific applications.
> 
> OK.  I agree as you say later that a merge can be at top level or folder 
> level.  In either case, I'd use the following guidelines:
> 
> 1. A merge element is of the form "<merge xinclude:href='[target uri]'/>.  The 
> source is an XBEL file in which the merge element appears.  The target is the 
> XBEL file referenced by the URI in the xinclude:href attribute of the merge 
> element.
> 
> 2. the current folder of the source is the folder in which the relevant merge 
> element appears.  This can be the top-level "folder".
> 
> 3. All bookmarks at the top level of the target are added to the source as if 
> directly specified at the location of the merge element in the current folder.
> 
> 4. If any bookmark element in the current folder of the source has an 
> identical href attribute to a bookmark in the target, the bookmark in the 
> target is ignored.
> 
> 5. All folders at the top level of the target are added to the source as if 
> directly specified at the location of the merge element (whether top-level or 
> within a folder).  This addition involves a recursive copying of all the 
> bookmarks and sub-folders contained in the target folder.
> 
> 6. If any folder in the current folder of the source has an identical title 
> child element to a folder in the target, the folder in the target is merged 
> into the folder in the source according to this process as if he target folder 
> and the matching source folder were both top-level xbel elements.
> 
> 7. The expanded bookmark file is the result of applying this process to each 
> merge element in document order of the merge source.  All merge elements in a 
> merge target are first processed before incorporation into the merge source.

Wow, that's a very precise definition of a #include mechanism, with
all details fleshed out.
I'm fine with this specification, I think it's a simple one to implement, but
at the same time it gives the user what we want.

Thanks for this detailed spec.

-- 
David FAURE, david@mandrakesoft.com, faure@kde.org
http://perso.mandrakesoft.com/~david/, http://www.konqueror.org/
KDE, Making The Future of Computing Available Today


From Alexandre.Fayolle@logilab.fr  Mon Apr  2 18:12:34 2001
From: Alexandre.Fayolle@logilab.fr (Alexandre Fayolle)
Date: Mon, 2 Apr 2001 19:12:34 +0200 (CEST)
Subject: [XML-SIG] Question about xml.dom.ext.reader.Sax2
Message-ID: <Pine.LNX.4.21.0104021909001.1982-100000@leo.logilab.fr>

Hello,

I've noticed the following lines of codes in Sax2.py, class Reader (this
is line 288 to 293, from the file shipping with 4Suite 0.10.2, in method
Reader.__init__):

        self.parser.setProperty(handler.property_lexical_handler,
self.handler)
        try:
            self.parser.setProperty(handler.property_declaration_handler,
self.handler)
        except:
            pass


Am I missing something obvious, or should the first line be deleted (I
mean, if there is a chance that the third line raise an exception, it will
presumably have been thrown on the first time the call is made, or are
there some weird side effects that I'm not aware of ?)

Cheers,

Alexandre Fayolle
-- 
http://www.logilab.com 
Narval is the first software agent available as free software (GPL).
LOGILAB, Paris (France).



From Alexandre.Fayolle@logilab.fr  Mon Apr  2 18:15:41 2001
From: Alexandre.Fayolle@logilab.fr (Alexandre Fayolle)
Date: Mon, 2 Apr 2001 19:15:41 +0200 (CEST)
Subject: [XML-SIG] Re: Question about xml.dom.ext.reader.Sax2
Message-ID: <Pine.LNX.4.21.0104021914090.1982-100000@leo.logilab.fr>

OK, forget it, I'm obviously getting tired. The parameters are not the
same. 

Sorry for the silly question.

Alexandre Fayolle
-- 
http://www.logilab.com 
Narval is the first software agent available as free software (GPL).
LOGILAB, Paris (France).



From uche.ogbuji@fourthought.com  Mon Apr  2 22:01:31 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Mon, 02 Apr 2001 15:01:31 -0600
Subject: [XML-SIG] Question about xml.dom.ext.reader.Sax2
In-Reply-To: Message from Alexandre Fayolle <Alexandre.Fayolle@logilab.fr>
 of "Mon, 02 Apr 2001 19:12:34 +0200." <Pine.LNX.4.21.0104021909001.1982-100000@leo.logilab.fr>
Message-ID: <200104022101.PAA20945@localhost.localdomain>

> I've noticed the following lines of codes in Sax2.py, class Reader (this
> is line 288 to 293, from the file shipping with 4Suite 0.10.2, in method
> Reader.__init__):
> 
>         self.parser.setProperty(handler.property_lexical_handler,
> self.handler)
>         try:
>             self.parser.setProperty(handler.property_declaration_handler,
> self.handler)
>         except:
>             pass
> 
> 
> Am I missing something obvious, or should the first line be deleted (I
> mean, if there is a chance that the third line raise an exception, it will
> presumably have been thrown on the first time the call is made, or are
> there some weird side effects that I'm not aware of ?) python.org/mailman/listinfo/xml-sig

It's a cut-n-paste bug.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From swanand75@yahoo.com  Mon Apr  2 22:42:02 2001
From: swanand75@yahoo.com (Swanand Bhorkar)
Date: Mon, 2 Apr 2001 14:42:02 -0700 (PDT)
Subject: [XML-SIG] help needed
Message-ID: <20010402214202.31623.qmail@web11304.mail.yahoo.com>

Hi,

I am trying to install PyXML package.
I have python 2.0.

Whenever i do either "python setup.py install" or
"python setup.py build" on the command prompt, it
exits giving an error that "distutils.core not found".

I think the PyXML tar.zip file should have had it.But
its not there.What should i do now ? 

thanks and regards,
swanand

__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail. 
http://personal.mail.yahoo.com/?.refer=text


From rsalz@zolera.com  Tue Apr  3 03:42:16 2001
From: rsalz@zolera.com (Rich Salz)
Date: Mon, 2 Apr 2001 22:42:16 -0400
Subject: [XML-SIG] xpath (4xpath) and CDATA
Message-ID: <200104030242.WAA12192@os390.zolera.com>

xpath (which Martin has merged in from 4Xpath) doesn't recognize CDATA
nodes.  According to the spec, http://www.w3.org/TR/xpath#NT-NodeType ,
"a CDATA section is treated as if the <![CDATA[ and ]]> were removed and
every occurrence of < and & were replaced by &lt; and &amp; respectively."

I think the way to fix this is to add CDATA_SECTION_NODE next to every
place that TEXT_NODE appears, but to do the above conversion whenever the
*value* of the node is needed.

Make sense?  I just want to double-check before I (ab)use my new status. :)
	/r$


From Nicolas.Chauvat@logilab.fr  Tue Apr  3 09:27:25 2001
From: Nicolas.Chauvat@logilab.fr (Nicolas Chauvat)
Date: Tue, 3 Apr 2001 10:27:25 +0200 (CEST)
Subject: [XML-SIG] help needed
In-Reply-To: <20010402214202.31623.qmail@web11304.mail.yahoo.com>
Message-ID: <Pine.LNX.4.21.0104031025270.24814-100000@aries>

Hi,

> I am trying to install PyXML package. I have python 2.0. Whenever i do
> either "python setup.py install" or "python setup.py build" on the
> command prompt, it exits giving an error that "distutils.core not
> found".
>=20
> I think the PyXML tar.zip file should have had it.But its not
> there.What should i do now ?

I don't use python 2.0, but looking at the standard library doc for python
2.0 would tend to prove distutil is not a standard package as it is not
there. But life is not that bad as you can easily download it from
http://www.python.org/sigs/distutils-sig/download.html

--=20
Nicolas Chauvat

http://www.logilab.com - "Mais o=F9 est donc Ornicar ?" - LOGILAB, Paris (F=
rance)



From jfk@informaticon.dk  Tue Apr  3 09:46:59 2001
From: jfk@informaticon.dk (=?iso-8859-1?Q?J=F8rgen=20Fr=F8jk=20Kj=E6rsgaard?=)
Date: Tue, 03 Apr 2001 10:46:59 +0200
Subject: [XML-SIG] Python binding for Xalan C++ ?
Message-ID: <3AC98E03.104942BF@informaticon.dk>

Is there a Python binding for the Xalan C++ XSLT processor
(http://xml.python.org/xalan-c) or is anybody working on creating one?
If not, would it be worth doing, or should I instead use Jython with
Xalan-j? I'd like to use Xalan because 4Suite XSLT doesn't perform
adequately.

In other words, what I'm looking for are pros and cons of
cPython+Xalan-c and Jython+Xalan-j.

regards, jfk

-- =

J=F8rgen Fr=F8jk Kj=E6rsgaard, Systemkonsulent (Systems Consultant)
Inform@ticon ApS * Web: www.informaticon.dk * Tlf: 8672 0093
Internet programmering * Systemudvikling p=E5 Linux, FreeBSD og PalmOS


From fdrake@acm.org  Tue Apr  3 15:42:57 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Tue, 3 Apr 2001 10:42:57 -0400 (EDT)
Subject: [XML-SIG] help needed
In-Reply-To: <Pine.LNX.4.21.0104031025270.24814-100000@aries>
References: <20010402214202.31623.qmail@web11304.mail.yahoo.com>
 <Pine.LNX.4.21.0104031025270.24814-100000@aries>
Message-ID: <15049.57713.274634.4860@cj42289-a.reston1.va.home.com>

Nicolas Chauvat writes:
 > I don't use python 2.0, but looking at the standard library doc for python
 > 2.0 would tend to prove distutil is not a standard package as it is not
 > there. But life is not that bad as you can easily download it from
 > http://www.python.org/sigs/distutils-sig/download.html

  Distutils has been a part of the standard library starting with
release 1.6.  The documentation for Distutils is in two separate
documents that are part of the standard documentation:

    "Installing Python Modules"
    http://www.python.org/doc/current/inst/inst.html

    "Distributing Python Modules"
    http://www.python.org/doc/current/dist/dist.html

  The separate package distributed by the Distutils-SIG is needed for
Python 1.5.2 users.


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From fdrake@acm.org  Tue Apr  3 15:46:21 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Tue, 3 Apr 2001 10:46:21 -0400 (EDT)
Subject: [XML-SIG] xpath (4xpath) and CDATA
In-Reply-To: <200104030242.WAA12192@os390.zolera.com>
References: <200104030242.WAA12192@os390.zolera.com>
Message-ID: <15049.57917.459340.366669@cj42289-a.reston1.va.home.com>

Rich Salz writes:
 > xpath (which Martin has merged in from 4Xpath) doesn't recognize CDATA
 > nodes.  According to the spec, http://www.w3.org/TR/xpath#NT-NodeType ,
 > "a CDATA section is treated as if the <![CDATA[ and ]]> were removed and
 > every occurrence of < and & were replaced by &lt; and &amp; respectively."
 > 
 > I think the way to fix this is to add CDATA_SECTION_NODE next to every
 > place that TEXT_NODE appears, but to do the above conversion whenever the
 > *value* of the node is needed.

  It's not at all clear what conversion is needed.  Do you propose
converting it to a text node and replacing the '<' and '&' characters
with EntityReference nodes?  It sounds to me like the XPath spec is
referring to the serialization format, which is just confusing when
talking about nodes.
  I think adding the CDATA_SECTION_NODE to the nodeType check is
sufficient.  So don't let me stop you!  ;-)


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From noreply@sourceforge.net  Wed Apr  4 16:42:10 2001
From: noreply@sourceforge.net (noreply@sourceforge.net)
Date: Wed, 04 Apr 2001 08:42:10 -0700
Subject: [XML-SIG] [ pyxml-Patches-413722 ] Sax2 reader for pDomlette
Message-ID: <E14kpQA-0000Nu-00@usw-sf-web3.sourceforge.net>

Patches item #413722, was updated on 2001-04-04 08:42
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=306473&aid=413722&group_id=6473

Category: 4Suite
Group: None
Status: Open
Priority: 5
Submitted By: Alexandre Fayolle (afayolle)
Assigned to: Nobody/Anonymous (nobody)
Summary: Sax2 reader for pDomlette

Initial Comment:
I need to build a pDomlette document from a SAX2
parser. Currently, only SAX1 parsers are supported. The
attached patch is a quick hack providing support for
SAX2 parsers. 

Alexandre Fayolle

----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=306473&aid=413722&group_id=6473


From sam@webslingerZ.com  Wed Apr  4 18:16:25 2001
From: sam@webslingerZ.com (Sam Brauer)
Date: Wed, 4 Apr 2001 13:16:25 -0400 (EDT)
Subject: [XML-SIG] Announcing maki
In-Reply-To: <E14kpiX-0003NU-00@mail.python.org>
Message-ID: <Pine.LNX.4.31.0104041251200.6416-100000@localhost.localdomain>

I have just made the first release of a project I have been working on for
the past couple of months.
The project's name is 'maki'.  To quote from the manual:

"The purpose of maki is to serve XML files via the web. A web developer
can specify that the XML data be processed or transformed through any
number of steps. Each step is either a stylesheet transformation or a
custom process. A processor that evaluates embedded Python code is
included. The output of each step is passed as the input to the next step
(similar to a Unix pipe). Additionally, the output of each processor step
can be cached for a user-specified time period. All configuration is in a
central XML file that specifies rules based on matching paths against
regular expressions."

maki relies on Python2, Apache, 4Suite, and mod_python.  It also
supports the use of Sablotron and Sab-pyth for XSLT (these are optional,
since maki uses 4xslt as the default XSLT transformer).

maki is currently in the alpha stage (fully functional, but in need of
community criticism and real-world testing).  I would very much welcome
any feedback.

If you're interested, take a look at http://maki.sourceforge.net/

Thanks,
Sam

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Sam Brauer : Systems Programmer : sam@webslingerZ.com




From sam@webslingerZ.com  Wed Apr  4 19:36:54 2001
From: sam@webslingerZ.com (Sam Brauer)
Date: Wed, 4 Apr 2001 14:36:54 -0400 (EDT)
Subject: [XML-SIG] Announcing maki
In-Reply-To: <Pine.LNX.4.21.0104041328560.15234-100000@clarkevans.com>
Message-ID: <Pine.LNX.4.31.0104041429570.6416-100000@localhost.localdomain>

It does caching already.  On my test system (a 400 mhz PII)
static XML -> XSLT -> HTML is served out of cache in about 5
milliseconds.  With every request, maki tests whether any of the input
files have been modified (thereby invalidating the cache).
Each step is cached, so if your XML goes through two XSLT passes and only
the last stylesheet has changed, the cached output of the first XSLT pass
is used, instead of starting over from scratch.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Sam Brauer : Systems Programmer : sam@webslingerZ.com

On Wed, 4 Apr 2001, Clark C. Evans wrote:

>
> Looks neat.  I'll dig into it in about 2-3 weeks.
> I'd like to see a simple caching mechanism as well,
> so that HTML pages arn't re-generated all the time
> for relatively static content.
>
> Clark
>
> On Wed, 4 Apr 2001, Sam Brauer wrote:
>
> > Date: Wed, 4 Apr 2001 13:16:25 -0400 (EDT)
> > From: Sam Brauer <sam@webslingerZ.com>
> > To: xml-sig@python.org
> > Subject: [XML-SIG] Announcing maki
> >
> > I have just made the first release of a project I have been working on for
> > the past couple of months.
> > The project's name is 'maki'.  To quote from the manual:
> >
> > "The purpose of maki is to serve XML files via the web. A web developer
> > can specify that the XML data be processed or transformed through any
> > number of steps. Each step is either a stylesheet transformation or a
> > custom process. A processor that evaluates embedded Python code is
> > included. The output of each step is passed as the input to the next step
> > (similar to a Unix pipe). Additionally, the output of each processor step
> > can be cached for a user-specified time period. All configuration is in a
> > central XML file that specifies rules based on matching paths against
> > regular expressions."
> >
> > maki relies on Python2, Apache, 4Suite, and mod_python.  It also
> > supports the use of Sablotron and Sab-pyth for XSLT (these are optional,
> > since maki uses 4xslt as the default XSLT transformer).
> >
> > maki is currently in the alpha stage (fully functional, but in need of
> > community criticism and real-world testing).  I would very much welcome
> > any feedback.
> >
> > If you're interested, take a look at http://maki.sourceforge.net/
> >
> > Thanks,
> > Sam
> >
> > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > Sam Brauer : Systems Programmer : sam@webslingerZ.com
> >
> >
> >
> > _______________________________________________
> > XML-SIG maillist  -  XML-SIG@python.org
> > http://mail.python.org/mailman/listinfo/xml-sig
> >
>
>



From Eugene.Leitl@lrz.uni-muenchen.de  Wed Apr  4 21:43:22 2001
From: Eugene.Leitl@lrz.uni-muenchen.de (Eugene.Leitl@lrz.uni-muenchen.de)
Date: Wed, 04 Apr 2001 22:43:22 +0200
Subject: [XML-SIG] rendering XML trees to HTML
Message-ID: <3ACB876A.DEA7650B@lrz.uni-muenchen.de>

I'm writing a web app that renders trees of chemical
reactions (spit out as XML) as (.png's embedded in)
HTML <TABLE>s. I don't know exactly how to encode 
them yet, but it will be something like this (but 
minus length, since this is phylogenetic tree):

	<!ELEMENT trees (tree+)>  
	<!ELEMENT tree (branch,branch,branch?,length?)>  
	<!ELEMENT branch (node,length?)>  
	<!ELEMENT node ((branch,branch)|specie)>
	<!ELEMENT length (#PCDATA)>
	<!ELEMENT specie (#PCDATA)>

Do you know of any code that renders XML as a 
HTML table? Maybe even something that will do a
a fair liking of Tk tree widget, producing

	http://archive.eso.org/~abrighto/tree/tree.html

a server-side bitmap?

Being very lazy, I don't feel like reinventing the wheel, 
and XML + trees is a rather frequent combination (30+ k 
hits on Google). Have any of you ran across any such beast?

TIA,
Eugene Leitl


From Nicolas.Chauvat@logilab.fr  Wed Apr  4 19:52:05 2001
From: Nicolas.Chauvat@logilab.fr (Nicolas Chauvat)
Date: Wed, 4 Apr 2001 20:52:05 +0200 (CEST)
Subject: [XML-SIG] rendering XML trees to HTML
In-Reply-To: <3ACB876A.DEA7650B@lrz.uni-muenchen.de>
Message-ID: <Pine.LNX.4.21.0104042050370.26321-100000@aries>

> Being very lazy, I don't feel like reinventing the wheel, and XML +
> trees is a rather frequent combination (30+ k hits on Google). Have
> any of you ran across any such beast?

It's no HTML rendering, but maybe http://www.logilab.org/xmltools/ and
http://www.logilab.org/xmltools/img/xmltree.png is something alike what
you're for?

Hope this helps.

--=20
Nicolas Chauvat

http://www.logilab.com - "Mais o=F9 est donc Ornicar ?" - LOGILAB, Paris (F=
rance)



From Eugene.Leitl@lrz.uni-muenchen.de  Wed Apr  4 21:57:05 2001
From: Eugene.Leitl@lrz.uni-muenchen.de (Eugene.Leitl@lrz.uni-muenchen.de)
Date: Wed, 04 Apr 2001 22:57:05 +0200
Subject: [XML-SIG] rendering XML trees to HTML
References: <Pine.LNX.4.21.0104042050370.26321-100000@aries>
Message-ID: <3ACB8AA1.B03B82FC@lrz.uni-muenchen.de>

Nicolas Chauvat wrote:

> It's no HTML rendering, but maybe http://www.logilab.org/xmltools/ and
> http://www.logilab.org/xmltools/img/xmltree.png is something alike what
> you're for?

Thanks. I've seen it. No, I'd rather not use gtk. And Tk tree widget
does essentially all I need (unfortunately it is a bit heavy on 
bandwidth, and evaluating clicks on nodes is not quite trivial).

I guess I have to roll my own...

Thanks again!

-- Eugene


From gianni@postino.it
     Indirizzo primario
                                       gianni@postino.it  Wed Apr  4 22:28:22 2001
From: gianni@postino.it
     Indirizzo primario
                                       gianni@postino.it (Gianni Rubagotti)
Date: Thu, 5 Apr 2001 1:58:22 +0430
Subject: [XML-SIG] XML for humanities
Message-ID: <20010404212822.RGIQ21859.sirio@[213.145.3.20]>

I'm glad to announce to you that 2 mailing lists are borning, now on yahoogroups 
to be soon moved to a university server (University of Milan through Professor 
Degli Antoni is preparing to to do it)

x-humanities (in English)
Mailing list about the utility of XML for Humanist applications 
http://groups.yahoo.com/group/x-humanities
to Subscribe send a mail to 
x-humanities-subscribe@yahoogroups.com 

x-umanisti (in Italian)
Applicazioni XML per letteratura, filosofie e scienze umane 
http://groups.yahoo.com/group/x-umanisti
to Subscribe send a mail to
x-umanisti-subscribe@yahoogroups.com 

All the people interested in this topic is invited to join!

Best regards,
Gianni Rubagotti





From a.kellett@lancaster.ac.uk  Wed Apr  4 23:30:02 2001
From: a.kellett@lancaster.ac.uk (Alexandertje)
Date: Wed, 4 Apr 2001 23:30:02 +0100
Subject: [XML-SIG] opera bookmark converter
Message-ID: <20010404233002.B7735@unix.lancs.ac.uk>

--G4iJoqBmSsgzjUCe
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

Included is a perl script to convert from opera bookmark files into
konqueror/xbel bookmarks.

btw, unlike the one to kde-cvs@ this version actually works ;)

thanks,
Alex

--G4iJoqBmSsgzjUCe
Content-Type: application/x-perl
Content-Disposition: attachment; filename="opera2xbel.pl"

#!/usr/bin/perl
use HTML::Entities;
print "<!DOCTYPE BOOKMARKS ><xbel folded=\"no\" OPEN=\"1\" hide_nsbk=\"yes\" >\n";
while (<>) {
   if (/^\#/) {
      chomp;
      $type = $_;
      $entry=();
      ENTRY: while(<>) {
         chomp;
         last ENTRY if /^$/;
         next ENTRY unless /^[ \t](URL|NAME)=(.*)/;
         $entry{$1} = $2;
      }
      if ($type eq "#URL") {
         print "  <bookmark icon=\"www\" href=\"" . encode_entities($entry{"URL"}) . "\" >\n";
         print "    <title>" . $entry{"NAME"} . "</title>\n";
         print "  </bookmark>\n";
      } elsif ($type eq "#FOLDER") {
         print "<folder folded=\"yes\" toolbar=\"no\" >\n";
         print "  <title>" . $entry{"NAME"} . "</title>\n";
      }
   } elsif (/^\-$/) {
      print "</folder>\n";
   }
}
print "</xbel>\n"

--G4iJoqBmSsgzjUCe--


From rsalz@zolera.com  Thu Apr  5 14:51:33 2001
From: rsalz@zolera.com (Rich Salz)
Date: Thu, 5 Apr 2001 09:51:33 -0400
Subject: [XML-SIG] Documenting xpath
Message-ID: <200104051351.JAA20218@os390.zolera.com>

Next week I want to start documenting the xpath (nee 4XPath) classes.
I'll work bottom-up.

First, is there documentation on the Node class, other than the DOM and
Python mapping?  If not, what do folks think should be documented?

General model: scanner (currently only pyxpath) uses a parser factory
(currently only FtFactory) to build up its "compiled scanner" object.
This compiled scanner must support an evaluate(context) method which
returns a nodeset (list of nodes).

Parsing, pyxpath, parsed expressions
	pyxpath.Compile -- module-level method for compiling a query

Context class

Package-level:
    Errors: SyntaxException(pos,msg), GeneralException, Error class(?)

    def Evaluate(expr, contextNode=None, context=None):
	registers $EXTMODULES, creates context, compiles, evaluates

    def Compile(expr):
	calls "a" compiler (currently only pyxpath)
    def CreateContext(contextNode):

Writing extensions:
    if os.environ.has_key('EXTMODULES'):
    def RegisterExtensionModules(moduleNames):
	(a dictionary of (namespace-string localname-string):method)

Comments?


From noreply@sourceforge.net  Thu Apr  5 15:00:45 2001
From: noreply@sourceforge.net (noreply@sourceforge.net)
Date: Thu, 05 Apr 2001 07:00:45 -0700
Subject: [XML-SIG] [ pyxml-Bugs-414001 ] _checkversion.py obsolete
Message-ID: <E14lAJZ-0002B7-00@usw-sf-web3.sourceforge.net>

Bugs item #414001, was updated on 2001-04-05 07:00
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=414001&group_id=6473

Category: None
Group: None
Status: Open
Priority: 5
Submitted By: Rich Salz (rsalz)
Assigned to: Martin v. Löwis (loewis)
Summary: _checkversion.py obsolete

Initial Comment:
Should _checkversion.py be removed?
; python _checkversion.py 
Traceback (most recent call last):
  File "_checkversion.py", line 5, in ?
    import pyversioncheck
ImportError: No module named pyversioncheck
; python -V
Python 2.0


----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=414001&group_id=6473


From krussll@cc.UManitoba.CA  Thu Apr  5 23:26:12 2001
From: krussll@cc.UManitoba.CA (Kevin Russell)
Date: Thu, 5 Apr 2001 17:26:12 -0500 (CDT)
Subject: [XML-SIG] How to leave character entities alone
Message-ID: <Pine.GSO.4.20.0104051702040.27170-100000@wisahkecahk.cc.umanitoba.ca>

I'm using the DOM from PyXML 0.6.5 to manipulate documents with the Text
Encoding Initiative DTD (teixlite.dtd).  But all of the character
entities, like &ndash; and &semi;, vanish into thin air when read into a
DOM object. I've tried with minidom and 4DOM; I've tried with validation
under 4DOM (where it gags on the DTD itself, which will probably be the
subject of my next frantic query) and without; I've tried all the standard
readers (please tell me I don't have to build my own).  I'm obviously
thrashing around clueless.

So, using any combination of stuff in PyXML, how can I achieve *any* of
the following (in decreasing order of niceness for me):

- leave all character entity references unexpanded, sitting in the tree as
  well-behaved little EntityReference objects.

- expand them into raw text if necessary, but trick the Printer into
  turning them back into &ndash;, &semi;, etc., when it's time to output
  the mangled DOM back into XML.

- expand them into raw text and leave them that way.  Indeed, anything
  short of having them vanish into thin air would be tolerable.

Sorry for such a braindead question, but several wee-hours of squinting at
documentation and source-code have left me unable to see the answer that's
undoubtedly sitting out there in plain sight.

-- Kevin Russell



From uche.ogbuji@fourthought.com  Fri Apr  6 05:37:52 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Thu, 05 Apr 2001 22:37:52 -0600
Subject: [XML-SIG] "Borrowed" tests
Message-ID: <200104060438.f364bqg16888@borgia.local>

I just noticed that in the transition from 4Suite to PyXML, the "borrowed" 
test for 4DOM went missing.

I've long had the practice of placing test cases based on bug reports or 
examples provided by others into a "test_suite/borrowed" directory.  I think 
this is useful, and important to distinguish from the "canned" tests.

If these are worth keeping, where should we put them?

I can find the old "borrowed" test for 4DOM, but should I check them in to 
PyXML?  If so, where?  "test/dom/borrowed"? "test/dom/use_case"?

These are especially important for when 4XPath/4XSLT migrate into PyXML, since 
some of the best testing of 4XSLT comes from examples culled from the twisted 
minds on xml-sig.

Thoughts?


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From uche.ogbuji@fourthought.com  Fri Apr  6 06:15:48 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Thu, 05 Apr 2001 23:15:48 -0600
Subject: [XML-SIG] xpath (4xpath) and CDATA
In-Reply-To: Message from Rich Salz <rsalz@zolera.com>
 of "Mon, 02 Apr 2001 22:42:16 EDT." <200104030242.WAA12192@os390.zolera.com>
Message-ID: <200104060516.f365FmS16995@borgia.local>

> xpath (which Martin has merged in from 4Xpath) doesn't recognize CDATA
> nodes.  According to the spec, http://www.w3.org/TR/xpath#NT-NodeType ,
> "a CDATA section is treated as if the <![CDATA[ and ]]> were removed and
> every occurrence of < and & were replaced by &lt; and &amp; respectively."

???  Do you mean when passing 4DOM nodes to 4XPath?  The Domlettes 
transparently turn CDATASections into contiguous text nodes.

If you do mean the situation with 4DOM, there is no easy fix for this.  There 
was some discussion on xsl-list about the problems of efficiently mapping DOM 
to XPath.  It seems some XPath implementors have simply declined DOM support 
except through complete export to a specialized format.

> I think the way to fix this is to add CDATA_SECTION_NODE next to every
> place that TEXT_NODE appears, but to do the above conversion whenever the
> *value* of the node is needed.

This could be a performance nightmare (you'll be amazed how often 
string-value() conversion is done in typical XSLT processing).  Besides, it 
just papers over the one problem, but what about adjacent text/CDATA nodes?  
What about entity reference nodes?

> Make sense?  I just want to double-check before I (ab)use my new status. :)

Nice to have you chipping in.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From uche.ogbuji@fourthought.com  Fri Apr  6 06:27:47 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Thu, 05 Apr 2001 23:27:47 -0600
Subject: [XML-SIG] Documenting xpath
In-Reply-To: Message from Rich Salz <rsalz@zolera.com>
 of "Thu, 05 Apr 2001 09:51:33 EDT." <200104051351.JAA20218@os390.zolera.com>
Message-ID: <200104060528.f365Rlr17027@borgia.local>

> Next week I want to start documenting the xpath (nee 4XPath) classes.
> I'll work bottom-up.

Unfortunately, 4XPath is not fully merged in yet.  It's still actively 
developed as part of 4Suite and I try to merge in at stable points.

If you let me know when you want to start your work, I'll try to sync up for 
you, but I'd be careful.

Normally it wouldn't be a big deal, but I'm currently whipping 4Suite into 
shape for the 0.11.0 release so changes are a-brewing.

4XSLT and 4XPath will be moving completely into PyXML immediately following 
the 4Suite 1.0 release ca. June 1.  Note that they'll keep their names while 
in PyXML.

> Package-level:
>     Errors: SyntaxException(pos,msg), GeneralException, Error class(?)

I just added a RuntimeException class which deals with such things as 
undefined variable and function refs during eval.

>     def Evaluate(expr, contextNode=None, context=None):
> 	registers $EXTMODULES, creates context, compiles, evaluates
> 
>     def Compile(expr):
> 	calls "a" compiler (currently only pyxpath)

No, same with 4XPath.

>     def CreateContext(contextNode):
> 
> Writing extensions:
>     if os.environ.has_key('EXTMODULES'):
>     def RegisterExtensionModules(moduleNames):
> 	(a dictionary of (namespace-string localname-string):method)

Yes.  The user can either set the environment or just directly call 
RegisterExtensionModules(['mymodulename'])

You'll want to have a look at

http://services.4Suite.org/documents/4Suite/4XPath-Api

to get a start.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From uche.ogbuji@fourthought.com  Fri Apr  6 06:34:47 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Thu, 05 Apr 2001 23:34:47 -0600
Subject: [XML-SIG] How to leave character entities alone
In-Reply-To: Message from Kevin Russell <krussll@cc.UManitoba.CA>
 of "Thu, 05 Apr 2001 17:26:12 CDT." <Pine.GSO.4.20.0104051702040.27170-100000@wisahkecahk.cc.umanitoba.ca>
Message-ID: <200104060535.f365Ylg17064@borgia.local>

> I'm using the DOM from PyXML 0.6.5 to manipulate documents with the Text
> Encoding Initiative DTD (teixlite.dtd).  But all of the character
> entities, like &ndash; and &semi;, vanish into thin air when read into a
> DOM object. I've tried with minidom and 4DOM; I've tried with validation
> under 4DOM (where it gags on the DTD itself, which will probably be the
> subject of my next frantic query) and without; I've tried all the standard
> readers (please tell me I don't have to build my own).  I'm obviously
> thrashing around clueless.

There was a bug in 4DOM where entities weren't being handled properly, but 
this should now be sorted out.

If you can provide a test case, I'll double-check that this is the case.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From noreply@sourceforge.net  Fri Apr  6 11:34:40 2001
From: noreply@sourceforge.net (noreply@sourceforge.net)
Date: Fri, 06 Apr 2001 03:34:40 -0700
Subject: [XML-SIG] [ pyxml-Bugs-414263 ] pDomlette: appending empty DocFrag
Message-ID: <E14lTZg-0003fR-00@usw-sf-web3.sourceforge.net>

Bugs item #414263, was updated on 2001-04-06 03:34
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=414263&group_id=6473

Category: 4Suite
Group: None
Status: Open
Priority: 5
Submitted By: Alexandre Fayolle (afayolle)
Assigned to: Nobody/Anonymous (nobody)
Summary: pDomlette: appending empty DocFrag

Initial Comment:
When using appendChild with an empty DocumentFragment,
the fragment is appended to the node. Same thing with
insertBefore.

Demo code :

>>> from Ft.Lib.pDomlette import PyExpatReader
>>> d =
PyExpatReader().fromString('<doc><whatever/></doc>')
>>> df = d.createDocumentFragment()
>>> d.documentElement.appendChild(df)
<Domlette DocumentFragment Node at 8145a20 with 0
children>
>>> d.documentElement.childNodes # We expect to see
only the <whatever> child here
[<Domlette Element Node at 813ec78: name='whatever'
with 0 attributes and 0 children>, <Domlette
DocumentFragment Node at 8145a20 with 0 children>]


The problem is that the new node is tested both for
DocumentFragmentitude and non emptyness in the code,
and the the else clause catches the empty doc frag and
does the default processing with it.

Attached to this report is a patch to pDomlette.py that
fixes the problem. It was generated on top of the other
patch I submitted this week, but they do not overlap,
so this should not be problem. 

Cheers

Alexandre Fayolle

----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=414263&group_id=6473


From rsalz@zolera.com  Fri Apr  6 11:43:20 2001
From: rsalz@zolera.com (Rich Salz)
Date: Fri, 06 Apr 2001 06:43:20 -0400
Subject: [XML-SIG] "Borrowed" tests
References: <200104060438.f364bqg16888@borgia.local>
Message-ID: <3ACD9DC8.96D30286@zolera.com>

How about test/dom/regress -- regression tests seems like a reasonable
definition.
	/r$


From uche.ogbuji@fourthought.com  Fri Apr  6 14:46:03 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Fri, 06 Apr 2001 07:46:03 -0600
Subject: [XML-SIG] "Borrowed" tests
In-Reply-To: Message from Rich Salz <rsalz@zolera.com>
 of "Fri, 06 Apr 2001 06:43:20 EDT." <3ACD9DC8.96D30286@zolera.com>
Message-ID: <200104061346.f36Dk3S20089@borgia.local>

> How about test/dom/regress -- regression tests seems like a reasonable
> definition.

Well, that doesn't properly differentiate them from the "canned" test, since 
those are also intended to be regression tests.

What about "real-world" tests?

I also considered calling them "black-box" tests, except that again, some of 
the canned tests are black-box (and some are white-box).


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From cce@clarkevans.com  Fri Apr  6 15:18:24 2001
From: cce@clarkevans.com (Clark C. Evans)
Date: Fri, 6 Apr 2001 09:18:24 -0500 (EST)
Subject: [XML-SIG] "Borrowed" tests
In-Reply-To: <200104060438.f364bqg16888@borgia.local>
Message-ID: <Pine.LNX.4.21.0104060916300.28660-100000@clarkevans.com>

On Thu, 5 Apr 2001, Uche Ogbuji wrote:
> I've long had the practice of placing test cases based on bug reports or 
> examples provided by others into a "test_suite/borrowed" directory.  I think 
> this is useful, and important to distinguish from the "canned" tests.

Your logic here escapes me.  Why would this distinction 
be necessary?  It just seems like extra work with no
additional benifit.  As I find bugs in my software I 
add it to my regression test, and usually name the
test after the bug.  You could opt for a naming
convention, where "canned" tests start with a "c"
and borrowed tests start with a "b" (for bug)

Clark




From rsalz@zolera.com  Fri Apr  6 15:19:17 2001
From: rsalz@zolera.com (Rich Salz)
Date: Fri, 06 Apr 2001 10:19:17 -0400
Subject: [XML-SIG] Documenting xpath
References: <200104060528.f365Rlr17027@borgia.local>
Message-ID: <3ACDD065.907F3D8C@zolera.com>

> Unfortunately, 4XPath is not fully merged in yet.  It's still actively
> developed as part of 4Suite and I try to merge in at stable points.
> 
> If you let me know when you want to start your work, I'll try to sync up for
> you, but I'd be careful.

No prob, I'll be careful.  And I'll watch the checkin messages.

> Normally it wouldn't be a big deal, but I'm currently whipping 4Suite into
> shape for the 0.11.0 release so changes are a-brewing.

Good luck.

> http://services.4Suite.org/documents/4Suite/4XPath-Api

Yeah, somehow I missed that one.

Hmm, perhaps it makes sense to work on something else first? 
Suggestions, anyone (fdrake?)
	/r$


From fdrake@acm.org  Fri Apr  6 16:01:54 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Fri, 6 Apr 2001 11:01:54 -0400 (EDT)
Subject: [XML-SIG] Documenting xpath
In-Reply-To: <3ACDD065.907F3D8C@zolera.com>
References: <200104060528.f365Rlr17027@borgia.local>
 <3ACDD065.907F3D8C@zolera.com>
Message-ID: <15053.55906.109080.617192@cj42289-a.reston1.va.home.com>

Rich Salz writes:
 > Hmm, perhaps it makes sense to work on something else first? 
 > Suggestions, anyone (fdrake?)

  Given that we expect Python 2.1 to be released in a week, I'd like
someone to review and, if appropriate, elaborate on the DOM and SAX
documentation in the Python Library Reference.  It definately needs
some work, but I don't know how much I'll be able to do over the next
week -- between other projects and building up the documentation on
testing facilities, I'm pretty swamped.  (Look for real unittest docs
to land in the Python CVS this weekend!)


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From fdrake@acm.org  Fri Apr  6 16:09:04 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Fri, 6 Apr 2001 11:09:04 -0400 (EDT)
Subject: [XML-SIG] "Borrowed" tests
In-Reply-To: <200104061346.f36Dk3S20089@borgia.local>
References: <rsalz@zolera.com>
 <3ACD9DC8.96D30286@zolera.com>
 <200104061346.f36Dk3S20089@borgia.local>
Message-ID: <15053.56336.313068.591813@cj42289-a.reston1.va.home.com>

Uche Ogbuji writes:
 > Well, that doesn't properly differentiate them from the "canned" test, since 
 > those are also intended to be regression tests.

  I agree with the sentiment that there's no need to separate them
from the rest of the tests -- they are needed to ensure coverage, or
the bug would have been caught earlier anyway.


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From rsalz@zolera.com  Fri Apr  6 16:08:32 2001
From: rsalz@zolera.com (Rich Salz)
Date: Fri, 06 Apr 2001 11:08:32 -0400
Subject: [XML-SIG] xpath (4xpath) and CDATA
References: <200104060516.f365FmS16995@borgia.local>
Message-ID: <3ACDDBF0.E39905C7@zolera.com>

> This could be a performance nightmare (you'll be amazed how often
> string-value() conversion is done in typical XSLT processing).  Besides, it
> just papers over the one problem, but what about adjacent text/CDATA nodes?
> What about entity reference nodes?

I don't care about entity references right now. :)

Fred and I disagree about the interepretation of the Xpath spec, so for
now my changes just treat a CDATA identically to a text node.

	/r$


From uche.ogbuji@fourthought.com  Fri Apr  6 16:39:36 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Fri, 06 Apr 2001 09:39:36 -0600
Subject: [XML-SIG] Documenting xpath
In-Reply-To: Message from "Fred L. Drake, Jr." <fdrake@acm.org>
 of "Fri, 06 Apr 2001 11:01:54 EDT." <15053.55906.109080.617192@cj42289-a.reston1.va.home.com>
Message-ID: <200104061539.f36Fda721122@borgia.local.dhcp.fourthought.com>

> 
> Rich Salz writes:
>  > Hmm, perhaps it makes sense to work on something else first? 
>  > Suggestions, anyone (fdrake?)
> 
>   Given that we expect Python 2.1 to be released in a week, I'd like
> someone to review and, if appropriate, elaborate on the DOM and SAX
> documentation in the Python Library Reference.

Yes, since this is probably source #1 of FAQs.

I know Greg Wilson would be happy, since I haven't been able to help him with 
civilized DOM docs for Python.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From fdrake@acm.org  Fri Apr  6 16:42:04 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Fri, 6 Apr 2001 11:42:04 -0400 (EDT)
Subject: [XML-SIG] Documenting xpath
In-Reply-To: <200104061539.f36Fda721122@borgia.local.dhcp.fourthought.com>
References: <fdrake@acm.org>
 <15053.55906.109080.617192@cj42289-a.reston1.va.home.com>
 <200104061539.f36Fda721122@borgia.local.dhcp.fourthought.com>
Message-ID: <15053.58316.100972.112192@cj42289-a.reston1.va.home.com>

Uche Ogbuji writes:
 > I know Greg Wilson would be happy, since I haven't been able to help him with 
 > civilized DOM docs for Python.

  Heck, I'm having a hard time imagining the DOM as civilized!


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From uche.ogbuji@fourthought.com  Fri Apr  6 16:45:22 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Fri, 06 Apr 2001 09:45:22 -0600
Subject: [XML-SIG] "Borrowed" tests
In-Reply-To: Message from "Clark C. Evans" <cce@clarkevans.com>
 of "Fri, 06 Apr 2001 09:18:24 CDT." <Pine.LNX.4.21.0104060916300.28660-100000@clarkevans.com>
Message-ID: <200104061545.f36FjMW21151@borgia.local.dhcp.fourthought.com>

> On Thu, 5 Apr 2001, Uche Ogbuji wrote:
> > I've long had the practice of placing test cases based on bug reports or 
> > examples provided by others into a "test_suite/borrowed" directory.  I think 
> > this is useful, and important to distinguish from the "canned" tests.
> 
> Your logic here escapes me.  Why would this distinction 
> be necessary?

It's not necessary.  My main goal is to get the tests back in regardless of 
where they go.  I was just suggesting a course based on what I'd done before.

The reason why I made the distinction before is that the "canned" tests were 
grouped according to speficication fiat, so there'd be, say a test_variable.py 
which tests the various diktats of section 11.4 of the XSLT spec.

The tests that come from bug-reports and borrowed code tend not to be so easy 
to neatly categorize.  Therefore I placed them in a separate directory just to 
provide a separate axis of grouping.  Not a big deal, except that this 
directory got lost for 4DOM as it moved to PyXML.

> It just seems like extra work with no
> additional benifit.

There's no extra work whatsoever.  Why do you think there is?

> As I find bugs in my software I 
> add it to my regression test, and usually name the
> test after the bug.  You could opt for a naming
> convention, where "canned" tests start with a "c"
> and borrowed tests start with a "b" (for bug)

This is exactly what I do.  The *only* difference being that I differentiate 
by directory placement rather than prefix.  I actually think the directory 
approach is less work than the prefix approach, but perhaps we're 
misunderstanding each other.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From uche.ogbuji@fourthought.com  Fri Apr  6 16:49:08 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Fri, 06 Apr 2001 09:49:08 -0600
Subject: [XML-SIG] xpath (4xpath) and CDATA
In-Reply-To: Message from Rich Salz <rsalz@zolera.com>
 of "Fri, 06 Apr 2001 11:08:32 EDT." <3ACDDBF0.E39905C7@zolera.com>
Message-ID: <200104061549.f36Fn8w21159@borgia.local.dhcp.fourthought.com>

> > This could be a performance nightmare (you'll be amazed how often
> > string-value() conversion is done in typical XSLT processing).  Besides, it
> > just papers over the one problem, but what about adjacent text/CDATA nodes?
> > What about entity reference nodes?
> 
> I don't care about entity references right now. :)
> 
> Fred and I disagree about the interepretation of the Xpath spec, so for
> now my changes just treat a CDATA identically to a text node.

Well, we don't know who else is out there: we just got a query about TEI lite. 
Woe betide anyone who tries to use TEI as 4DOM in XSLT.

But I guess going with the XP rule of solving the most tractable problem might 
be OK here.

I'd once though about adding a node.naziNormalize() sort of method, which 
would savagely go through the DOM expanding entity refs, merging adjacent text 
and CDATA nodes, and converting CDATA nodes to text.  I also considered 
recommending TreeWalker for using 4DOM with 4XPath.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From fdrake@acm.org  Fri Apr  6 17:56:39 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Fri, 6 Apr 2001 12:56:39 -0400 (EDT)
Subject: [XML-SIG] XBEL 1.1 docs up for review
Message-ID: <15053.62791.21353.257128@cj42289-a.reston1.va.home.com>

  I've posted the docs for XBEL 1.1 for review:

	http://starship.python.net/crew/fdrake/xbel-1.1/xbel.html

  Please send comments to this forum.
  Note that something has broken in the formatter's support for
bibTeX; I'll fix that as soon as I can, but the references are
seriously broken in the preview copy.


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From martin@loewis.home.cs.tu-berlin.de  Mon Apr  9 02:21:28 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Mon, 9 Apr 2001 03:21:28 +0200
Subject: [XML-SIG] "Borrowed" tests
In-Reply-To: <200104060438.f364bqg16888@borgia.local> (message from Uche
 Ogbuji on Thu, 05 Apr 2001 22:37:52 -0600)
References: <200104060438.f364bqg16888@borgia.local>
Message-ID: <200104090121.f391LSj02038@mira.informatik.hu-berlin.de>

> If these are worth keeping, where should we put them?

I think it is always worth to keep test cases. The other question is
whether it is worth distributing them with every PyXML/4Suite
installation. On cvs.pyxml.sourceforge.net, a new test module was
created to carry test cases which are not meant to be shipped - so
that would be another option.

One of the problems I have with your "borrowed" test cases is that it
is not always easy to tell PASS from FAIL; they just produce a lot of
output. To be useful for a user, a clear pass/fail indication is
necessary.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Mon Apr  9 01:50:59 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Mon, 9 Apr 2001 02:50:59 +0200
Subject: [XML-SIG] xpath (4xpath) and CDATA
In-Reply-To: <3ACDDBF0.E39905C7@zolera.com> (message from Rich Salz on Fri, 06
 Apr 2001 11:08:32 -0400)
References: <200104060516.f365FmS16995@borgia.local> <3ACDDBF0.E39905C7@zolera.com>
Message-ID: <200104090050.f390oxV01987@mira.informatik.hu-berlin.de>

> Fred and I disagree about the interepretation of the Xpath spec, so for
> now my changes just treat a CDATA identically to a text node.

Does that really work? Shouldn't multiple text nodes be combined with
any CDATA sections before doing any processing?

Regards,
Martin



From martin@loewis.home.cs.tu-berlin.de  Mon Apr  9 02:26:16 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Mon, 9 Apr 2001 03:26:16 +0200
Subject: [XML-SIG] "Borrowed" tests
In-Reply-To: <Pine.LNX.4.21.0104060916300.28660-100000@clarkevans.com>
 (cce@clarkevans.com)
References: <Pine.LNX.4.21.0104060916300.28660-100000@clarkevans.com>
Message-ID: <200104090126.f391QGZ02040@mira.informatik.hu-berlin.de>

> Your logic here escapes me.  Why would this distinction 
> be necessary?  It just seems like extra work with no
> additional benifit.  As I find bugs in my software I 
> add it to my regression test, and usually name the
> test after the bug.

Outright doing so with bugs found by others might violate the
copyright of the others; there should be an indication of authorship,
and there must be, of course, a permission of the author of the test
case that copying it is allowed. It is often hard to determine whether
the author would agree to redistribution of a test case, so it is
clearly better to keep them separate from the cases with clear
licensing conditions.

The GCC team could not ship its entire regression test suite with gcc
2.95 because authorship of many of the tests cannot be established.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Mon Apr  9 10:23:00 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Mon, 9 Apr 2001 11:23:00 +0200
Subject: [XML-SIG] "Borrowed" tests
In-Reply-To: <200104090225.f392Pkr01687@borgia.local> (message from Uche
 Ogbuji on Sun, 08 Apr 2001 20:25:46 -0600)
References: <200104090225.f392Pkr01687@borgia.local>
Message-ID: <200104090923.f399N0500930@mira.informatik.hu-berlin.de>

> > One of the problems I have with your "borrowed" test cases is that it
> > is not always easy to tell PASS from FAIL; they just produce a lot of
> > output. To be useful for a user, a clear pass/fail indication is
> > necessary.
> 
> Hmm.  All the 4DOM and 4XSLT tests do indicate pass/fail using the "[OK]" or 
> "[FAILED]" notation we borrowed from Red Hat start-up scripts.

Does that include the files which currently live in /xml/test/dom? I
run that as 'python test.py', and see no such output. Some of the
tests end with printing 'foo works', some only write the test name
(such as

********** Node **********
********** NodeList **********
********** NamedNodeMap **********
********** NodeIterator **********
********** TreeWalker **********
********** Attr **********
...

The entire test run ends with

********** HTML HTML_DOM_IMPLEMENTATION **********
testing source code syntax
 
<HTML>
  <HEAD>
    <TITLE>The Title</TITLE>
  </HEAD>
  <BODY></BODY>
</HTML>
Test Time - 1.782 secs

That probably means 'pass', although I'm not certain whether I would
recognize a failure.

Regards,
Martin


From jeremy.kloth@fourthought.com  Mon Apr  9 17:42:30 2001
From: jeremy.kloth@fourthought.com (Jeremy Kloth)
Date: Mon, 09 Apr 2001 10:42:30 -0600
Subject: [XML-SIG] "Borrowed" tests
References: <200104090225.f392Pkr01687@borgia.local> <200104090923.f399N0500930@mira.informatik.hu-berlin.de>
Message-ID: <3AD1E676.1BD82F9B@fourthought.com>

"Martin v. Loewis" wrote:
> 
> > > One of the problems I have with your "borrowed" test cases is that it
> > > is not always easy to tell PASS from FAIL; they just produce a lot of
> > > output. To be useful for a user, a clear pass/fail indication is
> > > necessary.
> >
> > Hmm.  All the 4DOM and 4XSLT tests do indicate pass/fail using the "[OK]" or
> > "[FAILED]" notation we borrowed from Red Hat start-up scripts.
> 
> Does that include the files which currently live in /xml/test/dom? I
> run that as 'python test.py', and see no such output. Some of the
> tests end with printing 'foo works', some only write the test name
> (such as
> 

That was a bug in the TestSuite module which was checked into there.
The tests now properly display their status for each test.  One
exception,
the HTML tests were NOT updated to use this method.  They do validate
the
results, however.

-- 
Jeremy Kloth                             Consultant
jeremy.kloth@fourthought.com             (303)583-9900 x 105
Fourthought, Inc.                        http://www.fourthought.com
Software-engineering, knowledge-management, XML, CORBA, Linux, Python


From pyxml@xhaus.com  Mon Apr  9 18:56:58 2001
From: pyxml@xhaus.com (Alan Kennedy)
Date: Mon, 9 Apr 2001 18:56:58 +0100
Subject: [XML-SIG] Using character entities in external DTD without validating.
Message-ID: <LPBBKIJMCOEOJAFDNMFEOEILCEAA.pyxml@xhaus.com>

Hi all,

Firstly, thanks for the great XML software.

I have a small problem. I have several hundred xml files, which contain data
for the members of a scientific association, as well as research abstracts.
I have used a variety of XML tools over the years to process them, including
Java tools with Jpython.

But now I want to use Cpython, for speed and memory reasons.

The problem I have is this. Many of the files contain character entity
references, which refer to ISO-8859-1 characters, as well as other
characters such greek letters (alpha, mu, etc).

The entity references are defined in one central DTD file, which every
single XML refers to using a DOCTYPE declaration. But I do not have an
actual structure for the XML files themselves, they are a pretty random
structure that has grown over the years.

I first started my Cpython/PyXML port by trying to use PyExpat. However,
since PyExpat doesn't read the external subset, it dropped all my character
entities.

Then I tried to the Sax2.Reader from xml.dom.ext.reader. This reads the
external subset, when the vaidate flag is turned on (i.e. the reader is
instantiated like so "reader=Sax2.Reader(validate=1)".

But now the Sax2.Reader is, understandably, insisting that my documents
conform to a structure, which they don't, so I get errors such as "Element
<XYZ> not declared".

Can anyone suggest a way that I can keep the character entity definitions in
an external file, AND read the documents without validating them?

I considered converting all of the documents to ISO-8859-1 encoding, but
doesn't solve the problem of the Greek letters in paper abstracts. I really
don't want to have to define those character entities in the internal subset
of all these documents.

Thanks in advance for any help,

Regards,

Alan.



From brunson@level3.net  Mon Apr  9 22:29:24 2001
From: brunson@level3.net (Eric Brunson)
Date: Mon, 9 Apr 2001 15:29:24 -0600
Subject: [XML-SIG] FileReader?
Message-ID: <20010409152924.A8363@level3.net>

I feel like this is a terribly stupid question, but I can't seen to
figure out where I should be looking.

Where is the xml.dom.util module?  Is it supposed to be part of PyXML?
I've checked the latest source distro and CVS, but I can't locate it.
Most of the example code seems to depend on it, so if anyone could
point me to the source, I'd appreciate it.

Thanks,
e.

-- 
Eric Brunson - brunson@level3.net - page-eric@level3.net  

"When governments fear the people there is liberty. When the people
fear the government there is tyranny." - Thomas Jefferson


From martin@loewis.home.cs.tu-berlin.de  Mon Apr  9 22:55:00 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Mon, 9 Apr 2001 23:55:00 +0200
Subject: [XML-SIG] Using character entities in external DTD without validating.
In-Reply-To: <LPBBKIJMCOEOJAFDNMFEOEILCEAA.pyxml@xhaus.com>
References: <LPBBKIJMCOEOJAFDNMFEOEILCEAA.pyxml@xhaus.com>
Message-ID: <200104092155.f39Lt0u03361@mira.informatik.hu-berlin.de>

> Can anyone suggest a way that I can keep the character entity definitions in
> an external file, AND read the documents without validating them?
> 
> I considered converting all of the documents to ISO-8859-1 encoding, but
> doesn't solve the problem of the Greek letters in paper abstracts. I really
> don't want to have to define those character entities in the internal subset
> of all these documents.

Did you consider using character references, instead of external
entities?

If that is also not feasible, I believe that none of the existing
parsers will exactly fit your need. You cannot talk pyexpat into
reading the external subset. With some efforts, you might manage to
talk xmlproc (the validating parser) into not producing validation
errors.

The most promising approach might be to use sgmlop. There is currently
no SAX2 sgmlop driver, but there is a SAX1 one; this does not support
entity references, though.

So here is a rough outline of what might succeed:
- extend drv_sgmlop.py to also support entity references. To do that,
  you best inherit from xml.sax.drivers.drv_sgmlop.Parser and add a
  handle_entityref method. In your code, this method should magically
  know your DTD; off-hand, I don't see a way to have sgmlop actually
  parse the external subset as well. Whenever you see an entity
  reference, invoke

    self.doc_handler.characters(<replacement>,0,len(<replacement>))

- Create an instance of your SAX driver.

- Pass that to Sax.From*, as the parser= parameter.

Hope this helps,
Martin


From swu@sybase.com  Mon Apr  9 23:05:39 2001
From: swu@sybase.com (Sonny Wu)
Date: Mon, 09 Apr 2001 15:05:39 -0700
Subject: [XML-SIG] XML howto update for Python 2.0?
Message-ID: <5.0.2.1.0.20010409150403.01e80540@olympus.sybase.com>

Hello,

Is there an expected version of document for XML.SAX in the near future?
Right now it seems to be insync with what is in 2.0.


Thanks,
Sonny



From martin@loewis.home.cs.tu-berlin.de  Tue Apr 10 06:23:13 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 10 Apr 2001 07:23:13 +0200
Subject: [XML-SIG] FileReader?
In-Reply-To: <20010409152924.A8363@level3.net> (message from Eric Brunson on
 Mon, 9 Apr 2001 15:29:24 -0600)
References: <20010409152924.A8363@level3.net>
Message-ID: <200104100523.f3A5NDX00867@mira.informatik.hu-berlin.de>

> I feel like this is a terribly stupid question, but I can't seen to
> figure out where I should be looking.
> 
> Where is the xml.dom.util module?  Is it supposed to be part of PyXML?

Not anymore, no. What made you think there should be such a module?

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Tue Apr 10 06:22:11 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 10 Apr 2001 07:22:11 +0200
Subject: [XML-SIG] XML howto update for Python 2.0?
In-Reply-To: <5.0.2.1.0.20010409150403.01e80540@olympus.sybase.com> (message
 from Sonny Wu on Mon, 09 Apr 2001 15:05:39 -0700)
References: <5.0.2.1.0.20010409150403.01e80540@olympus.sybase.com>
Message-ID: <200104100522.f3A5MBk00865@mira.informatik.hu-berlin.de>

> Is there an expected version of document for XML.SAX in the near future?
> Right now it seems to be insync with what is in 2.0.

I'm not sure I understand the question. The answer seems to be "no",
nobody plans to modify the XML howto. Since it is in sync with what is
in 2.0, there is no need for that, either...

Regards,
Martin



From walter@livinglogic.de  Tue Apr 10 10:16:12 2001
From: walter@livinglogic.de (=?us-ascii?Q?=22Walter_D=F6rwald=22?=)
Date: Tue, 10 Apr 2001 11:16:12 +0200
Subject: [XML-SIG] Using character entities in external DTD
 without validating.
In-Reply-To: <200104092155.f39Lt0u03361@mira.informatik.hu-berlin.de>
References: <LPBBKIJMCOEOJAFDNMFEOEILCEAA.pyxml@xhaus.com>
 <200104092155.f39Lt0u03361@mira.informatik.hu-berlin.de>
Message-ID: <200104101116120984.0032FAC8@mail.livinglogic.de>

On 09.04.01 at 23:55 Martin v. Loewis wrote:

> > Can anyone suggest a way that I can keep the character entity
> definitions in
> > an external file, AND read the documents without validating them?
> > 
> > I considered converting all of the documents to ISO-8859-1 encoding,=
 but
> > doesn't solve the problem of the Greek letters in paper abstracts. I
> really
> > don't want to have to define those character entities in the internal
> subset
> > of all these documents.
> 
> Did you consider using character references, instead of external
> entities?
> 
> If that is also not feasible, I believe that none of the existing
> parsers will exactly fit your need. You cannot talk pyexpat into
> reading the external subset. With some efforts, you might manage to
> talk xmlproc (the validating parser) into not producing validation
> errors.
> 
> The most promising approach might be to use sgmlop. There is currently
> no SAX2 sgmlop driver, 

I do have a rough, untested SAX2 driver for sgmlop, which could
be used as the base for a real SAX2 driver. If there is interest
I can post it.

> but there is a SAX1 one; this does not support
> entity references, though.
> 
> So here is a rough outline of what might succeed:
> - extend drv_sgmlop.py to also support entity references. To do that,
>   you best inherit from xml.sax.drivers.drv_sgmlop.Parser and add a
>   handle_entityref method. In your code, this method should magically
>   know your DTD; off-hand, I don't see a way to have sgmlop actually
>   parse the external subset as well. Whenever you see an entity
>   reference, invoke
> 
>     self.doc_handler.characters(<replacement>,0,len(<replacement>))
> 
> - Create an instance of your SAX driver.
> 
> - Pass that to Sax.From*, as the parser=3D parameter.

Alternatively, you could try using XIST 
(ftp://titan.bnbt.de/pub/livinglogic/xist/),
which is based on sgmlop and does exactly what Martin suggested. It
"automagically" knows the character entities, so you can type 
	&Alpha;
to get the character:
	greek capital letter alpha, U+0391

And if you need a new entity you can simple add one Python class to
define it:

class Spam(xsc.Entity):
	"the spam character, U+4242"
	codepoint =3D 0x4242

HTH

Bye,
   Walter D=F6rwald

-- 
Walter D=F6rwald =B7 LivingLogic AG =B7 Bayreuth, Germany =B7=
 www.livinglogic.de



From swu@sybase.com  Tue Apr 10 18:15:53 2001
From: swu@sybase.com (Sonny Wu)
Date: Tue, 10 Apr 2001 10:15:53 -0700
Subject: [XML-SIG] XML howto update for Python 2.0?
In-Reply-To: <200104100522.f3A5MBk00865@mira.informatik.hu-berlin.de>
References: <5.0.2.1.0.20010409150403.01e80540@olympus.sybase.com>
 <5.0.2.1.0.20010409150403.01e80540@olympus.sybase.com>
Message-ID: <5.0.2.1.0.20010410100424.01df22e0@olympus.sybase.com>

Hi Martin,

Sorry for an unclear question.

My question is about the XML howto document, as someone has posted on the 
news group,
appears to be outdated.  (please refer to this thread:
http://groups.google.com/groups?q=saxlib&hl=en&lr=&safe=off&rnum=1&seld=970271319&ic=1

The XML howto (http://py-howto.sourceforge.net/xml-howto/xml-howto.html)
still documents the usage of xml.sax.saxlib.  When I downloaded Python 2.0 
and began to read the online document, I could not find any related info 
about this package saxlib.

To that end, I am hoping to locate the new howto that utilizes
xml.sax and xml.sax.xmlreader packages.

The XML howto demonstrates the usage of
At 07:22 AM 4/10/2001, Martin v. Loewis wrote:
> > Is there an expected version of document for XML.SAX in the near future?
> > Right now it seems to be insync with what is in 2.0.
>
>I'm not sure I understand the question. The answer seems to be "no",
>nobody plans to modify the XML howto. Since it is in sync with what is
>in 2.0, there is no need for that, either...
>
>Regards,
>Martin

Thanks,
Sonny



From fdrake@acm.org  Tue Apr 10 18:36:49 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Tue, 10 Apr 2001 13:36:49 -0400 (EDT)
Subject: [XML-SIG] XML howto update for Python 2.0?
In-Reply-To: <5.0.2.1.0.20010410100424.01df22e0@olympus.sybase.com>
References: <5.0.2.1.0.20010409150403.01e80540@olympus.sybase.com>
 <5.0.2.1.0.20010410100424.01df22e0@olympus.sybase.com>
Message-ID: <15059.17585.844156.112748@cj42289-a.reston1.va.home.com>

Sonny Wu writes:
 > The XML howto (http://py-howto.sourceforge.net/xml-howto/xml-howto.html)
 > still documents the usage of xml.sax.saxlib.  When I downloaded Python 2.0 
 > and began to read the online document, I could not find any related info 
 > about this package saxlib.

  This document hasn't been updated for a while; I'm not sure if
anyone is actively maintaining it.


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From akuchlin@mems-exchange.org  Tue Apr 10 18:59:10 2001
From: akuchlin@mems-exchange.org (Andrew Kuchling)
Date: Tue, 10 Apr 2001 13:59:10 -0400
Subject: [XML-SIG] XML howto update for Python 2.0?
In-Reply-To: <200104100522.f3A5MBk00865@mira.informatik.hu-berlin.de>; from martin@loewis.home.cs.tu-berlin.de on Tue, Apr 10, 2001 at 07:22:11AM +0200
References: <5.0.2.1.0.20010409150403.01e80540@olympus.sybase.com> <200104100522.f3A5MBk00865@mira.informatik.hu-berlin.de>
Message-ID: <20010410135910.C15085@ute.cnri.reston.va.us>

On Tue, Apr 10, 2001 at 07:22:11AM +0200, Martin v. Loewis wrote:
>I'm not sure I understand the question. The answer seems to be "no",
>nobody plans to modify the XML howto. Since it is in sync with what is
>in 2.0, there is no need for that, either...

I'd like to revisit the HOWTO eventually, but have no idea when I'll
be able to find the time.

--amk



From dieter@handshake.de  Tue Apr 10 19:57:34 2001
From: dieter@handshake.de (Dieter Maurer)
Date: Tue, 10 Apr 2001 20:57:34 +0200 (CEST)
Subject: [XML-SIG] Using character entities in external DTD without validating.
In-Reply-To: <215095255@toto.iv>
Message-ID: <15059.22430.758810.339497@lindm.dm>

Alan Kennedy writes:
 > how about defining the entity file in the internal subset
 > and including it there?


Dieter


From krussll@cc.UManitoba.CA  Tue Apr 10 22:30:29 2001
From: krussll@cc.UManitoba.CA (Kevin Russell)
Date: Tue, 10 Apr 2001 16:30:29 -0500 (CDT)
Subject: [XML-SIG] Re: using character entities in external DTD without validating
In-Reply-To: <E14n0Zy-0006I6-00@mail.python.org>
Message-ID: <Pine.GSO.4.20.0104101613240.5604-100000@wisahkecahk.cc.umanitoba.ca>

Instead of trying to coerce a single set of tools to work on your files
without validation, it might be a better use of your time to come up with
a maximally vacuous DTD for your documents.  That way you'd still be able
to use your character references, but because you now have validatable
documents you wouldn't be tying yourself to any single toolset.

I'm thinking of a trivial DTD that's little more than a reference to the
external character sets plus the following for every element you've ever
used.

   <!ELEMENT xyz ANY>
   <!ATTLIST xyz
       ...stock list of every possible attribute you've ever used
       in any document... >

It probably wouldn't be too hard to write a Python program to read all
your files using a non-validating parser and generate such a DTD based
on the elements and attributes it encountered.

-- Kevin Russell



From brunson@level3.net  Tue Apr 10 23:23:54 2001
From: brunson@level3.net (Eric Brunson)
Date: Tue, 10 Apr 2001 16:23:54 -0600
Subject: [XML-SIG] FileReader?
In-Reply-To: <200104100523.f3A5NDX00867@mira.informatik.hu-berlin.de>; from martin@loewis.home.cs.tu-berlin.de on Tue, Apr 10, 2001 at 07:23:13AM +0200
References: <20010409152924.A8363@level3.net> <200104100523.f3A5NDX00867@mira.informatik.hu-berlin.de>
Message-ID: <20010410162353.A13906@level3.net>

* Martin v. Loewis (martin@loewis.home.cs.tu-berlin.de) [010409 23:27]:
> > I feel like this is a terribly stupid question, but I can't seen to
> > figure out where I should be looking.
> > 
> > Where is the xml.dom.util module?  Is it supposed to be part of PyXML?
> 
> Not anymore, no. What made you think there should be such a module?
> 

Only that most of the example code I came across relied on the
FileReader object, even code posted to this mailing list as recently
as 29 Jan 2001.  I see that all but one reference to FileReader has
made it's way out of the PyXML demo code, except for
demo/dom/benchmark.py.

But you've definitely answered my question, utils.FileReader is no
longer part of PyXML.

Thanks,
e.

-- 
Eric Brunson - brunson@level3.net - page-eric@level3.net  

"When governments fear the people there is liberty. When the people
fear the government there is tyranny." - Thomas Jefferson


From martin@loewis.home.cs.tu-berlin.de  Wed Apr 11 08:49:02 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Wed, 11 Apr 2001 09:49:02 +0200
Subject: [XML-SIG] XML howto update for Python 2.0?
In-Reply-To: <5.0.2.1.0.20010410100424.01df22e0@olympus.sybase.com> (message
 from Sonny Wu on Tue, 10 Apr 2001 10:15:53 -0700)
References: <5.0.2.1.0.20010409150403.01e80540@olympus.sybase.com>
 <5.0.2.1.0.20010409150403.01e80540@olympus.sybase.com> <5.0.2.1.0.20010410100424.01df22e0@olympus.sybase.com>
Message-ID: <200104110749.f3B7n2w01620@mira.informatik.hu-berlin.de>

> My question is about the XML howto document, as someone has posted on the 
> news group,
> appears to be outdated.  (please refer to this thread:
> http://groups.google.com/groups?q=saxlib&hl=en&lr=&safe=off&rnum=1&seld=970271319&ic=1

The article claims that the howto mentions xml.sax.saxlib.HandlerBase,
but I cannot find any such mentioning. Instead, it uses ContentHandler
in most places, as it should.

> The XML howto (http://py-howto.sourceforge.net/xml-howto/xml-howto.html)
> still documents the usage of xml.sax.saxlib.  When I downloaded Python 2.0 
> and began to read the online document, I could not find any related info 
> about this package saxlib.

Can you please give the exact URL of the page where saxlib is
mentioned? I cannot find that, either.

> To that end, I am hoping to locate the new howto that utilizes
> xml.sax and xml.sax.xmlreader packages.

How about

http://py-howto.sourceforge.net/xml-howto/node7.html

from xml.sax import make_parser
from xml.sax.handler import feature_namespaces
...

    parser = make_parser()
    parser.setFeature(feature_namespaces, 0)
...

The howto does not mention xmlreader, as you'd normally would not need
to use that module.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Wed Apr 11 08:57:16 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Wed, 11 Apr 2001 09:57:16 +0200
Subject: [XML-SIG] Re: using character entities in external DTD without validating
In-Reply-To: <Pine.GSO.4.20.0104101613240.5604-100000@wisahkecahk.cc.umanitoba.ca>
 (message from Kevin Russell on Tue, 10 Apr 2001 16:30:29 -0500 (CDT))
References: <Pine.GSO.4.20.0104101613240.5604-100000@wisahkecahk.cc.umanitoba.ca>
Message-ID: <200104110757.f3B7vGm01680@mira.informatik.hu-berlin.de>

> Instead of trying to coerce a single set of tools to work on your files
> without validation, it might be a better use of your time to come up with
> a maximally vacuous DTD for your documents.  That way you'd still be able
> to use your character references, but because you now have validatable
> documents you wouldn't be tying yourself to any single toolset.

The key point to notice here is that he wanted to use CPython for
speed. Now, if he had to produce a DTD and use a validating parser to
get the external entities expanded, then the parser would be
xmlproc. I'm not all that certain that this would be faster than using
a Java parser from JPython, since xmlproc isn't the fastest thing on
earth.

OTOH, sgmlop *is* the fastest parser that PyXML can offer, so it might
be the right tool.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Wed Apr 11 08:51:43 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Wed, 11 Apr 2001 09:51:43 +0200
Subject: [XML-SIG] XML howto update for Python 2.0?
In-Reply-To: <15059.17585.844156.112748@cj42289-a.reston1.va.home.com>
 (fdrake@acm.org)
References: <5.0.2.1.0.20010409150403.01e80540@olympus.sybase.com>
 <5.0.2.1.0.20010410100424.01df22e0@olympus.sybase.com> <15059.17585.844156.112748@cj42289-a.reston1.va.home.com>
Message-ID: <200104110751.f3B7phJ01677@mira.informatik.hu-berlin.de>

>  > The XML howto (http://py-howto.sourceforge.net/xml-howto/xml-howto.html)
>  > still documents the usage of xml.sax.saxlib.  When I downloaded Python 2.0 
>  > and began to read the online document, I could not find any related info 
>  > about this package saxlib.
> 
>   This document hasn't been updated for a while; I'm not sure if
> anyone is actively maintaining it.

The *HOWTO* was updated together with PyXML 0.6.1. It does explain
things that work only with PyXML (e.g. 4DOM), but the SAX part works
almost unmodified with stock Python 2.0 (except for a single
mentioning of DefaultHandler).

The *reference manual* is the one that has not been updated for a while.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Wed Apr 11 09:00:59 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Wed, 11 Apr 2001 10:00:59 +0200
Subject: [XML-SIG] FileReader?
In-Reply-To: <20010410162353.A13906@level3.net> (message from Eric Brunson on
 Tue, 10 Apr 2001 16:23:54 -0600)
References: <20010409152924.A8363@level3.net> <200104100523.f3A5NDX00867@mira.informatik.hu-berlin.de> <20010410162353.A13906@level3.net>
Message-ID: <200104110800.f3B80xb01754@mira.informatik.hu-berlin.de>

> Only that most of the example code I came across relied on the
> FileReader object, even code posted to this mailing list as recently
> as 29 Jan 2001.

In case you've wondered what to use instead: xml.dom.ext.reader is the
package that supports reading streams (strings, files, URLs) into DOM
trees; that gives you a 4DOM tree. Alternatively, you can use
xml.dom.minidom.parse to read a file into a minidom tree.

Eventually, the DOM level 3 interfaces should be implemented to allow
"portable" access to a DOM parser.

Regards,
Martin


From Mario.Ruggier@softplumbers.com  Wed Apr 11 13:29:47 2001
From: Mario.Ruggier@softplumbers.com (Ruggier, Mario)
Date: Wed, 11 Apr 2001 14:29:47 +0200
Subject: [XML-SIG] XML howto update for Python 2.0?
Message-ID: <7FDE48022CEFA544878EA148687723DF0FD965@mqgenevaex01.myqube.com>

Hi,=20

if I may pipe in... i was looking for some convenient=20
sample code to build a doc using DOM, using (preferably)
minidom or the full DOM implementation. But did not find=20
anything useful, e.g.:=20
http://py-howto.sourceforge.net/xml-howto/node17.html

Any pointers appreciated. Many thanks & Best Regards.

Mario Ruggier
--
SoftPlumbers SA
26, rue Maunoir
CH-1207 Gen=E8ve
+41.22.849.1002



From fdrake@acm.org  Wed Apr 11 13:46:16 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Wed, 11 Apr 2001 08:46:16 -0400 (EDT)
Subject: [XML-SIG] XML howto update for Python 2.0?
In-Reply-To: <7FDE48022CEFA544878EA148687723DF0FD965@mqgenevaex01.myqube.com>
References: <7FDE48022CEFA544878EA148687723DF0FD965@mqgenevaex01.myqube.com>
Message-ID: <15060.21016.341366.370395@cj42289-a.reston1.va.home.com>

Ruggier, Mario writes:
 > if I may pipe in... i was looking for some convenient 
 > sample code to build a doc using DOM, using (preferably)
 > minidom or the full DOM implementation. But did not find 

  I don't know if this fits "convenient", but there is some minidom
code in the Doc/tools/sgmlconv/docfixer.py file of the Python source
distribution, and some SAX code (an alternate reader) in
Doc/tools/sgmlconv/esistools.py.


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From stuff4gary@hotmail.com  Wed Apr 11 13:46:46 2001
From: stuff4gary@hotmail.com (gary cor)
Date: Wed, 11 Apr 2001 12:46:46
Subject: [XML-SIG] Missing core!
Message-ID: <F203gX2XdoYfYoiB5AN00006f49@hotmail.com>

Hi I wonder if anyone can with help these problems?

I am trying to learn how to parse XML in python using SAX and DOM... and 
want to end up doing this through a cgi.  Problems are that (other than not 
being a very good programmer hehe!):-
• I am having difficulties finding some good tuturials
• The tutorials I do have don't seem to work and I it keeps saying missing 
'core'  when I try anything with DOM.
I will have to go through the Python/XML how to as well again later and try 
that again!..  I am using the Python 2.0 with pyPyXML-0.6.5.win32-py2.0.exe 
should I use any other downloads as well?

Kind Regards

Gary C
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.



From fdrake@acm.org  Wed Apr 11 13:51:57 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Wed, 11 Apr 2001 08:51:57 -0400 (EDT)
Subject: [XML-SIG] Missing core!
In-Reply-To: <F203gX2XdoYfYoiB5AN00006f49@hotmail.com>
References: <F203gX2XdoYfYoiB5AN00006f49@hotmail.com>
Message-ID: <15060.21357.225082.295660@cj42289-a.reston1.va.home.com>

gary cor writes:
 > =95 The tutorials I do have don't seem to work and I it keeps saying=
 missing=20
 > 'core'  when I try anything with DOM.
 > I will have to go through the Python/XML how to as well again later =
and try=20
 > that again!..  I am using the Python 2.0 with pyPyXML-0.6.5.win32-py=
2.0.exe=20
 > should I use any other downloads as well?

  This sounds like you're using old documentation -- the Python/XML
Reference is out of date at this point.
  xml.dom.core is no longer part of PyXML.  You can get information
about the DOM interface in the development version of the Python
documentation:

    http://python.sourceforge.net/devel-docs/lib/markup.html

  These documents will be released for Python 2.1 next Monday, at
which point they will be published on python.org with a reliable URL.


  -Fred

--=20
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From uche.ogbuji@fourthought.com  Mon Apr  9 13:23:24 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Mon, 09 Apr 2001 06:23:24 -0600
Subject: [XML-SIG] "Borrowed" tests
In-Reply-To: Message from "Martin v. Loewis" <martin@loewis.home.cs.tu-berlin.de>
 of "Mon, 09 Apr 2001 11:23:00 +0200." <200104090923.f399N0500930@mira.informatik.hu-berlin.de>
Message-ID: <200104091223.f39CNOW03570@borgia.local>

> > > One of the problems I have with your "borrowed" test cases is that it
> > > is not always easy to tell PASS from FAIL; they just produce a lot of
> > > output. To be useful for a user, a clear pass/fail indication is
> > > necessary.
> > 
> > Hmm.  All the 4DOM and 4XSLT tests do indicate pass/fail using the "[OK]" or 
> > "[FAILED]" notation we borrowed from Red Hat start-up scripts.
> 
> Does that include the files which currently live in /xml/test/dom? I
> run that as 'python test.py', and see no such output. Some of the
> tests end with printing 'foo works', some only write the test name
> (such as
> 
> ********** Node **********
> ********** NodeList **********
> ********** NamedNodeMap **********
> ********** NodeIterator **********
> ********** TreeWalker **********
> ********** Attr **********
> ...
> 
> The entire test run ends with
> 
> ********** HTML HTML_DOM_IMPLEMENTATION **********
> testing source code syntax
>  
> <HTML>
>   <HEAD>
>     <TITLE>The Title</TITLE>
>   </HEAD>
>   <BODY></BODY>
> </HTML>
> Test Time - 1.782 secs
> 
> That probably means 'pass', although I'm not certain whether I would
> recognize a failure.

Odd.  I wonder if the checked-in version is out of date.

Jeremy, any ideas?


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python






From uche.ogbuji@fourthought.com  Mon Apr  9 03:25:46 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Sun, 08 Apr 2001 20:25:46 -0600
Subject: [XML-SIG] "Borrowed" tests
In-Reply-To: Message from "Martin v. Loewis" <martin@loewis.home.cs.tu-berlin.de>
 of "Mon, 09 Apr 2001 03:21:28 +0200." <200104090121.f391LSj02038@mira.informatik.hu-berlin.de>
Message-ID: <200104090225.f392Pkr01687@borgia.local>

> > If these are worth keeping, where should we put them?
> 
> I think it is always worth to keep test cases. The other question is
> whether it is worth distributing them with every PyXML/4Suite
> installation. On cvs.pyxml.sourceforge.net, a new test module was
> created to carry test cases which are not meant to be shipped - so
> that would be another option.

OK.  I'll put them here to start.

> One of the problems I have with your "borrowed" test cases is that it
> is not always easy to tell PASS from FAIL; they just produce a lot of
> output. To be useful for a user, a clear pass/fail indication is
> necessary.

Hmm.  All the 4DOM and 4XSLT tests do indicate pass/fail using the "[OK]" or 
"[FAILED]" notation we borrowed from Red Hat start-up scripts.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From uche.ogbuji@fourthought.com  Mon Apr  9 03:30:11 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Sun, 08 Apr 2001 20:30:11 -0600
Subject: [XML-SIG] xpath (4xpath) and CDATA
In-Reply-To: Message from "Martin v. Loewis" <martin@loewis.home.cs.tu-berlin.de>
 of "Mon, 09 Apr 2001 02:50:59 +0200." <200104090050.f390oxV01987@mira.informatik.hu-berlin.de>
Message-ID: <200104090230.f392UBe01701@borgia.local>

> > Fred and I disagree about the interepretation of the Xpath spec, so for
> > now my changes just treat a CDATA identically to a text node.
> 
> Does that really work? Shouldn't multiple text nodes be combined with
> any CDATA sections before doing any processing?

It doesn't really work.  Not only is there the problem you mention, but 
there's also the problem of expanding and merging EntiryRefs.

Rich says all he needs to solve is the CDATA-as-sole-child problem.  I say 
that this problem is so non-general that I'd hate to see specialized code to 
solve it withough at least some discussion of how to solve the general problem.

In other words, I'd not like to see code checked in to 4XPath that blindly 
checks for CDATASection nodes wherever it's looking for TEXT nodes.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From fdrake@acm.org  Wed Apr 11 17:02:44 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Wed, 11 Apr 2001 12:02:44 -0400 (EDT)
Subject: [XML-SIG] xpath (4xpath) and CDATA
In-Reply-To: <200104090050.f390oxV01987@mira.informatik.hu-berlin.de>
References: <200104060516.f365FmS16995@borgia.local>
 <3ACDDBF0.E39905C7@zolera.com>
 <200104090050.f390oxV01987@mira.informatik.hu-berlin.de>
Message-ID: <15060.32804.290270.984415@beowolf.digicool.com>

Rich Salz said:
 > Fred and I disagree about the interepretation of the Xpath spec, so for
 > now my changes just treat a CDATA identically to a text node.

Martin v. Loewis responded:
 > Does that really work? Shouldn't multiple text nodes be combined with
 > any CDATA sections before doing any processing?

  I think it's a matter of "better than it was before", and I think
better than the change that Rich described to me in email.  I'm not
sufficiently familiar with the code to the "right thing", and don't
have any particular expectation that I'll have time to pursue this
myself.


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From fdrake@acm.org  Wed Apr 11 15:01:02 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Wed, 11 Apr 2001 10:01:02 -0400 (EDT)
Subject: [XML-SIG] XML howto update for Python 2.0?
In-Reply-To: <200104110751.f3B7phJ01677@mira.informatik.hu-berlin.de>
References: <5.0.2.1.0.20010409150403.01e80540@olympus.sybase.com>
 <5.0.2.1.0.20010410100424.01df22e0@olympus.sybase.com>
 <15059.17585.844156.112748@cj42289-a.reston1.va.home.com>
 <200104110751.f3B7phJ01677@mira.informatik.hu-berlin.de>
Message-ID: <15060.25502.128525.25667@cj42289-a.reston1.va.home.com>

Martin v. Loewis writes:
 > The *HOWTO* was updated together with PyXML 0.6.1. It does explain
 > things that work only with PyXML (e.g. 4DOM), but the SAX part works
 > almost unmodified with stock Python 2.0 (except for a single
 > mentioning of DefaultHandler).

  My bad:  I hadn't realized that either had been updated.

 > The *reference manual* is the one that has not been updated for a while.

  As I mentioned about a week ago, I'm planning to work on isolating
the Python bindings from the other reference documentation sometime
after Python 2.1 is out.  I'm not sure exactly when this will happen
(since I have some Expat time scheduled, that really needs some work),
but I expect I'll have preliminary documents available fairly quickly
since they'll be based on the existing documentation.
  The intent is to avoid the relationship with Python's release cycle,
and allow the bindings documents to be more meaningfully referenced
independently.


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From lucas7@home.com  Wed Apr 11 22:47:26 2001
From: lucas7@home.com (Lucas Vogel)
Date: Wed, 11 Apr 2001 14:47:26 -0700
Subject: [XML-SIG] web page broken?
Message-ID: <004301c0c2d0$ffc328a0$01000001@cx229913e>

-----BEGIN PGP SIGNED MESSAGE-----

The XML-SIG status page and Resources page links on the mailman
page(http://mail.python.org/mailman/listinfo/xml-sig) appear to be
broken...maybe they just need to be updated? 

Thought I'd let someone know.

Lucas Vogel

-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 7.0.3 for non-commercial use <http://www.pgp.com>

iQCVAwUBOtTQ6SlA8LjISNKLAQFFpgP/eHAyucaN3ASej/LfOAj7iSdtP+FQx9DY
+rSWgpQfVx/1zmSKnkNc2XIrwYwIasFsWVy9VABS2yEQ8UXTXvbvskUVywhV1NL2
4EVgNYv+/fjoaHjMx3qRyoDJIHg06edFUhqHpFcjH3r8re/e+obyV0bhLB1zGrPP
O7SJAfoDKEY=
=2JLb
-----END PGP SIGNATURE-----




From eliot@isogen.com  Thu Apr 12 03:19:24 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Wed, 11 Apr 2001 21:19:24 -0500
Subject: [XML-SIG] One-Line Enhancement to catalog.py
Message-ID: <3AD510AC.7782ED7A@isogen.com>

Hi,

In order to use the pyXML catalog stuff with SGML-specific catalogs, I
had to enable the processing of SGMLDECL entries. This required adding
this line to catalog.py:

                         "SGMLDECL": ("s")}
 
Making the new code:

class CatalogParser(AbstrCatalogParser,xmlutils.EntityParser):
    "A parser for SGML Open catalog files."

    def __init__(self,error_lang=None):
        AbstrCatalogParser.__init__(self,error_lang)
        xmlutils.EntityParser.__init__(self)

        # p=pubid (or prefix)
        # s=sysid (to be resolved)
        # o=other
        self.entry_hash={ "PUBLIC": ("p","s"), "DELEGATE": ("p","s"),
                          "CATALOG": ("s"), "DOCUMENT": ("s"),
                          "BASE": ("o"), "SYSTEM": ("o","s"), 
                          "OVERRIDE": ("o"),
                          "SGMLDECL": ("s")}
 

Can someone make this change to the main distribution for me? 

Thanks,

Eliot
-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From martin@loewis.home.cs.tu-berlin.de  Thu Apr 12 10:20:49 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 12 Apr 2001 11:20:49 +0200
Subject: [XML-SIG] One-Line Enhancement to catalog.py
In-Reply-To: <3AD510AC.7782ED7A@isogen.com> (eliot@isogen.com)
References: <3AD510AC.7782ED7A@isogen.com>
Message-ID: <200104120920.f3C9KnJ06117@mira.informatik.hu-berlin.de>

> In order to use the pyXML catalog stuff with SGML-specific catalogs, I
> had to enable the processing of SGMLDECL entries. This required adding
> this line to catalog.py:
> 
>                          "SGMLDECL": ("s")}

It appears that just adding this line is an incomplete change. At a
minimum, the application should be informed using, say,
handle_sgmldecl. Can you come up with a patch that does that?

If so, please submit the patch as a unified (-u) or context (-c) diff.
Posting the diff to the list is fine, although submitting it to
sourceforge is preferred.

Thanks,
Martin


From noreply@sourceforge.net  Thu Apr 12 12:54:53 2001
From: noreply@sourceforge.net (noreply@sourceforge.net)
Date: Thu, 12 Apr 2001 04:54:53 -0700
Subject: [XML-SIG] [ pyxml-Bugs-415640 ] pDomlette does not normalize
Message-ID: <E14nfgb-0001kZ-00@usw-sf-web1.sourceforge.net>

Bugs item #415640, was updated on 2001-04-12 04:54
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=415640&group_id=6473

Category: None
Group: None
Status: Open
Priority: 5
Submitted By: Alexandre Fayolle (afayolle)
Assigned to: Nobody/Anonymous (nobody)
Summary: pDomlette does not normalize

Initial Comment:
Affected version : 0.10.2

The header of the pDomlette.py file states that
"Domlette is also automatically normalized". (it also
incorrectly states that Domlette is read only).
Therefore, the implementation of normalize in pDomlette
is as follow:

    def normalize(self):
        pass


However, pDomlette does not automatically normalize.
Here's a demonstration:

>>> from Ft.Lib.pDomlette import PyExpatReader
>>> from xml.xpath import Evaluate
>>> 
>>> d=PyExpatReader().fromString('<doc>glou</doc>')
>>> print d.documentElement.childNodes
[<Domlette Text Node at 81c4988: data='glou'>]
>>> t = d.createTextNode('blaaaaa')
>>> d.documentElement.appendChild(t)
<Domlette Text Node at 81d4438: data='blaaaaa'>
>>> print d.documentElement.childNodes
[<Domlette Text Node at 81c4988: data='glou'>,
<Domlette Text Node at 81d4438: data='blaaaaa'>]
>>> 
>>> Evaluate('/doc/text()',d)
[<Domlette Text Node at 81c4988: data='glou'>,
<Domlette Text Node at 81d4438: data='blaaaaa'>]


I'll be working on a patch. 

----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=415640&group_id=6473


From Juergen Hermann" <jh@web.de  Thu Apr 12 16:02:06 2001
From: Juergen Hermann" <jh@web.de (Juergen Hermann)
Date: Thu, 12 Apr 2001 17:02:06 +0200
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
Message-ID: <m14nibn-007NM2C@smtp.web.de>

Hi!

We consider doing a wrapper for Xerces and Xalan as an OS project. One 
question is where to host it:

	1. Own SourceForge project
	2. Part of PyXML
	3. Part of Xerces / Xalan projects

3 is unlikely. Option 2 seems not too clever, since installing Xerces 
and Xalan are extra steps anyway, and I doubt anyone seriously wants to =

include them in a standard PyXML distribution.

Working title for the project would be "pirxx" (Python InteRface to 
Xalan and Xerces, anyone who has read Stanislaw Lem should know where 
the name comes from).

Before we really launch this, I also would like to know who's 
interested into such a wrapper (I've seen some requests in the past) 
and also who would be able and willing to contribute (developers, 
documenters, beta-testers, ...).


Ciao, J=FCrgen

--
J=FCrgen Hermann, Developer (jhe@webde-ag.de)
WEB.DE AG, http://webde-ag.de/




From nuno.simoes@ruido-visual.pt  Thu Apr 12 16:21:56 2001
From: nuno.simoes@ruido-visual.pt (Nuno Simoes)
Date: Thu, 12 Apr 2001 16:21:56 +0100
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
References: <m14nibn-007NM2C@smtp.web.de>
Message-ID: <010901c0c364$4fc30260$8040a8c0@localdomain>

From: "Juergen Hermann" <jh@web.de>
Sent: Thursday, April 12, 2001 4:02 PM


We consider doing a wrapper for Xerces and Xalan as an OS project. One
question is where to host it:

1. Own SourceForge project
2. Part of PyXML
3. Part of Xerces / Xalan projects

3 is unlikely. Option 2 seems not too clever, since installing Xerces
and Xalan are extra steps anyway, and I doubt anyone seriously wants to
include them in a standard PyXML distribution.

--> For the startup, probably the first option will be the better one.

Working title for the project would be "pirxx" (Python InteRface to
Xalan and Xerces, anyone who has read Stanislaw Lem should know where
the name comes from).

Before we really launch this, I also would like to know who's
interested into such a wrapper (I've seen some requests in the past)
and also who would be able and willing to contribute (developers,
documenters, beta-testers, ...).

--> Well, iam, and i believe that more ppl are, because with a python
wrapper, you can easilly use Xalan + Xerces in Zope, for example.
    Yesterday i was trying to do a module for python that use Xalan to
process some XML, but i had no skills in C++, so i first have to look to c++
to do it. Its kinda simple, if you look into the Python Manual.
    About your project, i could do some beta-testing. :-)

Nuno Simões,
rvti




From mal@lemburg.com  Thu Apr 12 16:21:28 2001
From: mal@lemburg.com (M.-A. Lemburg)
Date: Thu, 12 Apr 2001 17:21:28 +0200
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
References: <m14nibn-007NM2C@smtp.web.de>
Message-ID: <3AD5C7F8.A60C9180@lemburg.com>

Juergen Hermann wrote:
> 
> Hi!
> 
> We consider doing a wrapper for Xerces and Xalan as an OS project. One
> question is where to host it:
> 
>         1. Own SourceForge project
>         2. Part of PyXML
>         3. Part of Xerces / Xalan projects
> 
> 3 is unlikely. Option 2 seems not too clever, since installing Xerces
> and Xalan are extra steps anyway, and I doubt anyone seriously wants to
> include them in a standard PyXML distribution.

I'd suggest 1 and using a distutils setup so that installation
becomes a breeze.
 
> Working title for the project would be "pirxx" (Python InteRface to
> Xalan and Xerces, anyone who has read Stanislaw Lem should know where
> the name comes from).

Isn't that name already taken by Christian Tismer ;-)
 
> Before we really launch this, I also would like to know who's
> interested into such a wrapper (I've seen some requests in the past)
> and also who would be able and willing to contribute (developers,
> documenters, beta-testers, ...).

Could you provide pointers to Xerces and Xalan ?

Thanks,
-- 
Marc-Andre Lemburg
______________________________________________________________________
Company & Consulting:                           http://www.egenix.com/
Python Pages:                           http://www.lemburg.com/python/


From Alexandre.Fayolle@logilab.fr  Thu Apr 12 16:32:29 2001
From: Alexandre.Fayolle@logilab.fr (Alexandre Fayolle)
Date: Thu, 12 Apr 2001 17:32:29 +0200 (CEST)
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
In-Reply-To: <3AD5C7F8.A60C9180@lemburg.com>
Message-ID: <Pine.LNX.4.21.0104121730260.16943-100000@leo.logilab.fr>

On Thu, 12 Apr 2001, M.-A. Lemburg wrote:

> Could you provide pointers to Xerces and Xalan ?

The starting point is http://xml.apache.org/

You'll get http://xml.apache.org/xerces-c/index.html and
http://xml.apache.org/xalan-c/index.html quite rapidly. These are the C++
implementations. There are also Java implementations, that could be
interesting for users of JPython, since I believe 4Suite is not available
on this platform.

Alexandre Fayolle
-- 
http://www.logilab.com 
Narval is the first software agent available as free software (GPL).
LOGILAB, Paris (France).



From nuno.simoes@ruido-visual.pt  Thu Apr 12 16:32:53 2001
From: nuno.simoes@ruido-visual.pt (Nuno Simoes)
Date: Thu, 12 Apr 2001 16:32:53 +0100
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
References: <m14nibn-007NM2C@smtp.web.de> <3AD5C7F8.A60C9180@lemburg.com>
Message-ID: <011601c0c365$d7bbdba0$8040a8c0@localdomain>

From: "M.-A. Lemburg" <mal@lemburg.com>
Sent: Thursday, April 12, 2001 4:21 PM

[...]

> > Before we really launch this, I also would like to know who's
> > interested into such a wrapper (I've seen some requests in the past)
> > and also who would be able and willing to contribute (developers,
> > documenters, beta-testers, ...).
>
> Could you provide pointers to Xerces and Xalan ?

    Well, i can answer that right now.. :-)

    http://xml.apache.org/

Nuno Simões,
rvti




From Nicolas.Chauvat@logilab.fr  Thu Apr 12 16:49:38 2001
From: Nicolas.Chauvat@logilab.fr (Nicolas Chauvat)
Date: Thu, 12 Apr 2001 17:49:38 +0200 (CEST)
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
In-Reply-To: <m14nibn-007NM2C@smtp.web.de>
Message-ID: <Pine.LNX.4.21.0104121745580.9697-100000@aries>

Hi Juergen,

> Before we really launch this, I also would like to know who's
> interested into such a wrapper (I've seen some requests in the past)
> and also who would be able and willing to contribute (developers,
> documenters, beta-testers, ...).

Count us in as beta-testers. 4xslt works fine for us, but we'll be glad to
help make available another XSL processing engine for python.

--=20
Nicolas Chauvat

http://www.logilab.com - "Mais o=F9 est donc Ornicar ?" - LOGILAB, Paris (F=
rance)



From mal@lemburg.com  Thu Apr 12 17:04:40 2001
From: mal@lemburg.com (M.-A. Lemburg)
Date: Thu, 12 Apr 2001 18:04:40 +0200
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
References: <Pine.LNX.4.21.0104121730260.16943-100000@leo.logilab.fr>
Message-ID: <3AD5D218.7B22866@lemburg.com>

Alexandre Fayolle wrote:
> 
> On Thu, 12 Apr 2001, M.-A. Lemburg wrote:
> 
> > Could you provide pointers to Xerces and Xalan ?
> 
> The starting point is http://xml.apache.org/
> 
> You'll get http://xml.apache.org/xerces-c/index.html and
> http://xml.apache.org/xalan-c/index.html quite rapidly. These are the C++
> implementations. There are also Java implementations, that could be
> interesting for users of JPython, since I believe 4Suite is not available
> on this platform.

Thanks. Those tools look very interesting, but doesn't FourThought
offer pretty much the same or even a superset of these tools ?

BTW, the ICU OSS project (*) hosted by IBM seems an interesting target
for Python as well -- now that we have Unicode-support in the
core, I guess hooking up Python with ICU should be easy ;-)

(*) http://oss.software.ibm.com/icu/ 
    "International Components for Unicode"

-- 
Marc-Andre Lemburg
______________________________________________________________________
Company & Consulting:                           http://www.egenix.com/
Python Pages:                           http://www.lemburg.com/python/


From akuchlin@mems-exchange.org  Thu Apr 12 17:07:28 2001
From: akuchlin@mems-exchange.org (Andrew Kuchling)
Date: Thu, 12 Apr 2001 12:07:28 -0400
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
In-Reply-To: <m14nibn-007NM2C@smtp.web.de>; from jh@web.de on Thu, Apr 12, 2001 at 05:02:06PM +0200
References: <m14nibn-007NM2C@smtp.web.de>
Message-ID: <20010412120728.C6058@ute.cnri.reston.va.us>

On Thu, Apr 12, 2001 at 05:02:06PM +0200, Juergen Hermann wrote:
>	1. Own SourceForge project
>	2. Part of PyXML
>	3. Part of Xerces / Xalan projects
>3 is unlikely. Option 2 seems not too clever, since installing Xerces 

Why is 3 unlikely?  There are already Xerces/Perl bindings; why should
Python be a second-class citizen in this respect?  Initially it'll
probably be easier to start a separate development effort, but once
you've demonstrated some working code, adding it to xml.apache.org
should be considered.  

--amk


From brunson@level3.net  Thu Apr 12 17:09:13 2001
From: brunson@level3.net (Eric Brunson)
Date: Thu, 12 Apr 2001 10:09:13 -0600
Subject: [XML-SIG] FileReader?
In-Reply-To: <200104110800.f3B80xb01754@mira.informatik.hu-berlin.de>; from martin@loewis.home.cs.tu-berlin.de on Wed, Apr 11, 2001 at 10:00:59AM +0200
References: <20010409152924.A8363@level3.net> <200104100523.f3A5NDX00867@mira.informatik.hu-berlin.de> <20010410162353.A13906@level3.net> <200104110800.f3B80xb01754@mira.informatik.hu-berlin.de>
Message-ID: <20010412100912.A2684@level3.net>

* Martin v. Loewis (martin@loewis.home.cs.tu-berlin.de) [010411 02:34]:
> > Only that most of the example code I came across relied on the
> > FileReader object, even code posted to this mailing list as recently
> > as 29 Jan 2001.
> 
> In case you've wondered what to use instead: xml.dom.ext.reader is the
> package that supports reading streams (strings, files, URLs) into DOM
> trees; that gives you a 4DOM tree. Alternatively, you can use
> xml.dom.minidom.parse to read a file into a minidom tree.
> 
> Eventually, the DOM level 3 interfaces should be implemented to allow
> "portable" access to a DOM parser.

Oh, excellent.  Thanks for the pointer.  :-)

-- 
Eric Brunson - brunson@level3.net - page-eric@level3.net  

"When governments fear the people there is liberty. When the people
fear the government there is tyranny." - Thomas Jefferson


From rsalz@zolera.com  Thu Apr 12 17:15:35 2001
From: rsalz@zolera.com (Rich Salz)
Date: Thu, 12 Apr 2001 12:15:35 -0400
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
References: <m14nibn-007NM2C@smtp.web.de> <20010412120728.C6058@ute.cnri.reston.va.us>
Message-ID: <3AD5D4A7.E28B9683@zolera.com>

I agree with Andrew.  In fact, I'd encourage you to participate in
xalan/xerces and get the code into their sourcebase as early as possible
-- perhaps as soon as you've got something that "works"
	/r$


From Juergen Hermann" <jh@web.de  Thu Apr 12 17:27:00 2001
From: Juergen Hermann" <jh@web.de (Juergen Hermann)
Date: Thu, 12 Apr 2001 18:27:00 +0200
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
In-Reply-To: <20010412120728.C6058@ute.cnri.reston.va.us>
Message-ID: <m14njvw-007OXVC@smtp.web.de>

On Thu, 12 Apr 2001 12:07:28 -0400, Andrew Kuchling wrote:

>Why is 3 unlikely? 

OK, add an "initially" there. As the next step, I plan to ask the 
Xerces and Xalan people for their opinion on the project too, just like 
here.




From Alexandre.Fayolle@logilab.fr  Thu Apr 12 17:53:36 2001
From: Alexandre.Fayolle@logilab.fr (Alexandre Fayolle)
Date: Thu, 12 Apr 2001 18:53:36 +0200 (CEST)
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
In-Reply-To: <3AD5D218.7B22866@lemburg.com>
Message-ID: <Pine.LNX.4.21.0104121845350.17194-100000@leo.logilab.fr>

On Thu, 12 Apr 2001, M.-A. Lemburg wrote:

> Thanks. Those tools look very interesting, but doesn't FourThought
> offer pretty much the same or even a superset of these tools ?

Yes, of course. Some of the great strength of 4Suite is 4RDF and DbDom
(and of course 4SuiteServer). And of course, writing XSLT extensions in
python. 

However, having used Xalan as a simple command line XSLT engine, when
it comes to pure speed, things are just not the same. Xalan is several
orders of magnitude faster, so I guess that for a number of applications,
having the choice is a Good Thing. 

Alexandre Fayolle
-- 
http://www.logilab.com 
Narval is the first software agent available as free software (GPL).
LOGILAB, Paris (France).



From martin@loewis.home.cs.tu-berlin.de  Thu Apr 12 17:58:12 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 12 Apr 2001 18:58:12 +0200
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
In-Reply-To: <m14nibn-007NM2C@smtp.web.de> (jh@web.de)
References: <m14nibn-007NM2C@smtp.web.de>
Message-ID: <200104121658.f3CGwCX00922@mira.informatik.hu-berlin.de>

> We consider doing a wrapper for Xerces and Xalan as an OS project. One 
> question is where to host it:
> 
> 	1. Own SourceForge project
> 	2. Part of PyXML
> 	3. Part of Xerces / Xalan projects
> 
> 3 is unlikely. Option 2 seems not too clever, since installing Xerces 
> and Xalan are extra steps anyway, and I doubt anyone seriously wants to 
> include them in a standard PyXML distribution.

I would not mind including it in PyXML, if setup.py could be talking
into auto-detecting a Xerces installation, and building the wrapper
only if Xerces is available. Of course, creating a new project is just
fine.

In any case, it would be desirable if the wrappers register themselves
at the relevant SAX/DOM parser factories, so that users can make use
of them without even knowing that they are there. That, of course,
assumes that you will follow the Python SAX and DOM mappings.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Thu Apr 12 18:08:10 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 12 Apr 2001 19:08:10 +0200
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
In-Reply-To: <3AD5D218.7B22866@lemburg.com> (mal@lemburg.com)
References: <Pine.LNX.4.21.0104121730260.16943-100000@leo.logilab.fr> <3AD5D218.7B22866@lemburg.com>
Message-ID: <200104121708.f3CH8AO00945@mira.informatik.hu-berlin.de>

> Thanks. Those tools look very interesting, but doesn't FourThought
> offer pretty much the same or even a superset of these tools ?

Yes, Xerces is pretty much the same functionality as PyXML, and Xalan
is what 4XSLT does (unless I mix them up right now). However, the hope
of people who want such integration typically is that:

- it will be faster than the pure Python code. That should be true
  for the validating parser (i.e. I'd hope that Xerces beats xmlproc
  in parsing speed), and might be true for the DOM implementation
  (although I'd wait for completion of the project to see whether this
  is actually the case).

- it will be more correct than PyXML, since there is more developer
  power behind Xerces - especially since that got supported by the
  Apache foundation. I'm not sure here; AFAIK, xmlproc is still one of
  the most complete parsers with regard to catalogs and such stuff, and
  4DOM supports almost all of DOM Level 2 - which Xerces didn't, last
  I checked.

> BTW, the ICU OSS project (*) hosted by IBM seems an interesting target
> for Python as well -- now that we have Unicode-support in the
> core, I guess hooking up Python with ICU should be easy ;-)

I think one of the problems with ICU is that you have to use their
string types, so quite some copying forth and back might go on when
you try to integrate that into Python. Also, it will be tricky to make
the integration seemless - users might have to use ICU functions,
instead of getting transparent access to codeset converters and locale
information.

As for codeset converters, this, again, is an area where some speed
advantage might be gained from using ICU - if you can avoid too much
copying.

Regards,
Martin


From eliot@isogen.com  Thu Apr 12 18:14:58 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Thu, 12 Apr 2001 12:14:58 -0500
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
References: <m14nibn-007NM2C@smtp.web.de> <3AD5C7F8.A60C9180@lemburg.com>
Message-ID: <3AD5E292.73F2ACA3@isogen.com>

> Juergen Hermann wrote:

> > Before we really launch this, I also would like to know who's
> > interested into such a wrapper (I've seen some requests in the past)
> > and also who would be able and willing to contribute (developers,
> > documenters, beta-testers, ...).

We (DataChannel Austin) would definitely be interested--we are building
a Python-based content management system that, among other things, lets
us apply XSLT styles to hyperdocuments (of any sort, not just XLink).
Putting Xalan as a back end is something we were going to do (just be
lauching a separate Java process), but being able to more tightly
integrate it with our overall Python framework would be very handy.

Also, while I'm thinking about it, we have modified the pyXML and the
4Suite XSLT and XPath stuff to allow us to process arbitrary groves and
arbitrary Python objects with XSL. We would like to contribute our
changes back but haven't had the bandwidth yet to properly package
things up. But we will as soon as we can.

Cheers,

Eliot

-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m




From eliot@isogen.com  Thu Apr 12 18:09:33 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Thu, 12 Apr 2001 12:09:33 -0500
Subject: [XML-SIG] One-Line Enhancement to catalog.py
References: <3AD510AC.7782ED7A@isogen.com> <200104120920.f3C9KnJ06117@mira.informatik.hu-berlin.de>
Message-ID: <3AD5E14D.B650FBA@isogen.com>

"Martin v. Loewis" wrote:
> 
> > In order to use the pyXML catalog stuff with SGML-specific catalogs, I
> > had to enable the processing of SGMLDECL entries. This required adding
> > this line to catalog.py:
> >
> >                          "SGMLDECL": ("s")}
> 
> It appears that just adding this line is an incomplete change. At a
> minimum, the application should be informed using, say,
> handle_sgmldecl. Can you come up with a patch that does that?

I can do that--the change I made was sufficient to allow my test cases
to pass, but I can do the more complete change if more is required. 

Cheers,

E.

-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m




From Nicolas.Chauvat@logilab.fr  Thu Apr 12 19:29:21 2001
From: Nicolas.Chauvat@logilab.fr (Nicolas Chauvat)
Date: Thu, 12 Apr 2001 20:29:21 +0200 (CEST)
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
In-Reply-To: <3AD5E292.73F2ACA3@isogen.com>
Message-ID: <Pine.LNX.4.21.0104122024490.593-100000@aries>

> Also, while I'm thinking about it, we have modified the pyXML and the
> 4Suite XSLT and XPath stuff to allow us to process arbitrary groves and
> arbitrary Python objects with XSL. We would like to contribute our
> changes back but haven't had the bandwidth yet to properly package
> things up. But we will as soon as we can.

Are you talking about the same groves that jade, SGML and DSSSL use?

What do you mean by "XSL and XPath to process arbitrary python objects"?

If that's interesting enough, you may quickly find people that will help
you test the merging of your changes with the main tree before it goes
mainstream... :-)

--=20
Nicolas Chauvat

http://www.logilab.com - "Mais o=F9 est donc Ornicar ?" - LOGILAB, Paris (F=
rance)




From Juergen Hermann" <jh@web.de  Thu Apr 12 20:02:27 2001
From: Juergen Hermann" <jh@web.de (Juergen Hermann)
Date: Thu, 12 Apr 2001 21:02:27 +0200
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
In-Reply-To: <200104121658.f3CGwCX00922@mira.informatik.hu-berlin.de>
Message-ID: <m14nmL8-007NnFC@smtp.web.de>

On Thu, 12 Apr 2001 18:58:12 +0200, Martin v. Loewis wrote:

>In any case, it would be desirable if the wrappers register themselves
>at the relevant SAX/DOM parser factories, so that users can make use
>of them without even knowing that they are there. That, of course,
>assumes that you will follow the Python SAX and DOM mappings.

That is one of the first goals. Also, we want to handle a C++ SAX 
stream with Python, and vice versa (feed a Python SAX stream into 
Xalan). Bi-SAXuality, in a sense. :)

Ciao, J=FCrgen




From martin@loewis.home.cs.tu-berlin.de  Thu Apr 12 21:24:49 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 12 Apr 2001 22:24:49 +0200
Subject: [XML-SIG] XML howto update for Python 2.0?
In-Reply-To: <7FDE48022CEFA544878EA148687723DF0FD965@mqgenevaex01.myqube.com>
 (Mario.Ruggier@softplumbers.com)
References: <7FDE48022CEFA544878EA148687723DF0FD965@mqgenevaex01.myqube.com>
Message-ID: <200104122024.f3CKOnU01804@mira.informatik.hu-berlin.de>

> if I may pipe in... i was looking for some convenient 
> sample code to build a doc using DOM, using (preferably)
> minidom or the full DOM implementation. But did not find 
> anything useful, e.g.: 
> http://py-howto.sourceforge.net/xml-howto/node17.html

Not sure what you are asking, probably: How can I create a document by
building a DOM tree, and then linearising the tree into an XML document?

If so, there is an example for building DOM trees in
test/test_minidom.py; the method to generate the document from the
tree is .toxml. If you use 4DOM, the methods to build the tree are the
same; linearisation then uses xml.dom.ext.PrettyPrint.

Regards,
Martin



From info@webb2e.com  Fri Apr 13 01:56:53 2001
From: info@webb2e.com (info@webb2e.com)
Date: Thu, 12 Apr 2001 17:56:53 -0700
Subject: [XML-SIG] Free register of online company's profile
Message-ID: <0d2285356000d41MAIL@mail3.chinainfoland.com>

This is a multi-part message in MIME format.

------=_NextPart_000_DA48_01C0C379.F4BA67D0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit

How much are you paying to advertise your business to the world? 

Expose Your service to the world with bold register of online business
profile. Sign up today! 

Introducing WebB2E.com -- your direct link to global information; source
of business, products, education/research, social/culture, entertainment
and travel... 
Additionally you can BUY, SELL or PROMOTE your products and services 
At www.webb2e.com <http://webb2e.com/promo1.asp>  you'll get: 

--Message center (open to the public). 
--Employment center. 
--Sponsorship center. 
--Bulletin board (business and service issue). 
--Flexible Online Office (Business Online Report). 
--Economic news. 
--View thousands of trade leads. 
--Post business propositions. 
--Merchandise marketing (Vast advertising at a low cost). 
--World shopping center. 

.. and much more. Please visit www.webb2e.com
<http://webb2e.com/promo1.asp>  

If you do not want to recieve any more e-mails from WebB2E.com and wish
to be removed from e-mail list please click here
<http://webb2e.net/remove.asp?email=xml-sig@python.org> . 



------=_NextPart_000_DA48_01C0C379.F4BA67D0
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<html><head><title>www.webb2e.com</title><meta =
http-equiv=3D'Content-Type' content=3D'text/html; =
charset=3Diso-8859-1'></head><body bgcolor=3D'#FFFFFF'>How much are you =
paying to advertise your business to the world? <br><br>Expose Your =
service to the world with bold register of online business profile. Sign =
up today! <br><br>Introducing WebB2E.com -- your direct link to global =
information; source of business, products, education/research, =
social/culture, entertainment and travel... <br>Additionally you can =
BUY, SELL or PROMOTE your products and services  <br>At <a =
href=3D'http://webb2e.com/promo1.asp'>www.webb2e.com</a> you'll get:  =
<br><br>--Message center (open to the public).  <br>--Employment center. =
<br>--Sponsorship center. <br>--Bulletin board (business and service =
issue). <br>--Flexible Online Office (Business Online Report).  =
<br>--Economic news. <br>--View thousands of trade leads.  <br>--Post =
business propositions.  <br>--Merchandise marketing (Vast advertising at =
a low cost).  <br>--World shopping center. <br><br>... and much more. =
Please visit <a href=3D'http://webb2e.com/promo1.asp'>www.webb2e.com</a> =
 <br><br>If you do not want to recieve any more e-mails from WebB2E.com =
and wish to be removed from e-mail list please <a =
href=3D'http://webb2e.net/remove.asp?email=3Dxml-sig@python.org'>click =
here</a>. <br><br></body></html>
------=_NextPart_000_DA48_01C0C379.F4BA67D0--


From frank63@ms5.hinet.net  Fri Apr 13 01:15:40 2001
From: frank63@ms5.hinet.net (Frank Chen)
Date: Fri, 13 Apr 2001 08:15:40 +0800
Subject: [XML-SIG] Re:Python Wrapper for Xerces/Xalan
References: <E14njX2-0005H5-00@mail.python.org>
Message-ID: <006e01c0c3b7$b6d1ad00$efa01ea3@MiTACUser>

> From: "Juergen Hermann" <jh@web.de>
> To: "Python XML SIG" <xml-sig@python.org>
> Date: Thu, 12 Apr 2001 17:02:06 +0200
> Reply-To: "Juergen Hermann" <jh@web.de>
> Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
>
> We consider doing a wrapper for Xerces and Xalan as an OS project. One
> question is where to host it:
>
>
> Before we really launch this, I also would like to know who's
> interested into such a wrapper (I've seen some requests in the past)
> and also who would be able and willing to contribute (developers,
> documenters, beta-testers, ...).
>

If there is such a prxx project, I would like to be a documenter and
beta-tester.

Frank






From brian@sweetapp.com  Fri Apr 13 02:27:12 2001
From: brian@sweetapp.com (Brian Quinlan)
Date: Thu, 12 Apr 2001 18:27:12 -0700
Subject: [XML-SIG] RE: Python Wrapper for Xerces/Xalan
In-Reply-To: <3AD62809.447F386C@ActiveState.com>
Message-ID: <002101c0c3b8$ddeed6f0$b503a8c0@activestate.ca>

I have actually already started a similar project. Currently, I provide a
loose wrapping around Xalan i.e.

transform( <source:file|URI|buffer> [, <stylesheet:file|URI|buffer>], [,
<toplevelparams:dict>] )
	=3D> string containing transformed source file

I was planning on making the first run available soon.

-------- Original Message --------
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
Date: Thu, 12 Apr 2001 17:02:06 +0200
From: "Juergen Hermann" <jh@web.de>
Reply-To: "Juergen Hermann" <jh@web.de>
To: "Python XML SIG" <xml-sig@python.org>

Hi!



We consider doing a wrapper for Xerces and Xalan as an OS project. One

question is where to host it:



	1. Own SourceForge project

	2. Part of PyXML

	3. Part of Xerces / Xalan projects



3 is unlikely. Option 2 seems not too clever, since installing Xerces

and Xalan are extra steps anyway, and I doubt anyone seriously wants to

include them in a standard PyXML distribution.



Working title for the project would be "pirxx" (Python InteRface to

Xalan and Xerces, anyone who has read Stanislaw Lem should know where

the name comes from).



Before we really launch this, I also would like to know who's

interested into such a wrapper (I've seen some requests in the past)

and also who would be able and willing to contribute (developers,

documenters, beta-testers, ...).





Ciao, J=FCrgen



--

J=FCrgen Hermann, Developer (jhe@webde-ag.de)

WEB.DE AG, http://webde-ag.de/







_______________________________________________

XML-SIG maillist  -  XML-SIG@python.org

http://mail.python.org/mailman/listinfo/xml-sig



From eliot@isogen.com  Fri Apr 13 03:02:12 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Thu, 12 Apr 2001 21:02:12 -0500
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
References: <Pine.LNX.4.21.0104122024490.593-100000@aries>
Message-ID: <3AD65E24.E9864477@isogen.com>

Nicolas Chauvat wrote:
> 
> > Also, while I'm thinking about it, we have modified the pyXML and the
> > 4Suite XSLT and XPath stuff to allow us to process arbitrary groves and
> > arbitrary Python objects with XSL. We would like to contribute our
> > changes back but haven't had the bandwidth yet to properly package
> > things up. But we will as soon as we can.
> 
> Are you talking about the same groves that jade, SGML and DSSSL use?

The very same. We are building a generic grove-based hyperdocument
management system. Given a grove implementation, we can both hyperlink
to it using a generic hyperdocument API (that looks a lot like HyTime
but is not dependent on the use of HyTime syntax--it can be bound to any
reasonable way of representing hyperdocuments, from HTML to Micrsoft
Project).
 
> What do you mean by "XSL and XPath to process arbitrary python objects"?

We discovered that with just a couple of lines of code, that we could
get the 4Suite XSLT processor to happily apply XSLT templates and XPath
expressions to arbitrary Python objects if we treat the object class as
the "tag name" and all member variables and parameter-less methods as
attributes. For example, our Hyperdocument object class has a
"getBosMembers()" method that returns the list of groves from which the
hyperdocument was constructed (the "bounded object set" of input
documents, converted to groves). 

Given our hack, I can do this in an XSLT style sheet (note: I am not by
any stretch an XSL expert, so please excuse any XSL errors or stupidity
in the following example):

<!-- "ext" namespace is our extension package for interrogating the
hyperdocument, which
     is passed in to the XSLT processor as part of its startup
invocation. -->
<xsl:template select=".">
  <xsl:if test="ext:isAnchorMember()">
    <!-- Given that the current node is a member of one or more
hyperlink anchors,
         get the Hyperdocument object from it and do something with it. 
      -->
    <xsl:param name="hydoc" select="ext:getHyDoc()"/>
    <xsl:for-each select="$hydoc"><!-- Probably a better way to do this
-->
      <p>Documents in this hyperdocument:</p>
      <ol>
      <xsl:for-each select="@getBosMembers">
        <li>Document of type: <xsl:value-of select="@ClassName"/></li>
      </xsl:for-each>
      </ol>
   </xsl:for-each>
</xsl:template>

Where the "ext:getHyDoc" extention method returns one of our
Hyperdocument objects. The expression "@getBosMembers" is internally
translated to a call to the getBosMembers() method of the Python object,
which returns a node list of grove nodes. As defined in the grove
standard, every grove node has a "ClassName" property that is the string
value of the class name, as defined in the grove's property set.

So, for example, say my hyperdoc consisted of one XML doc, one Word doc,
and one Excel doc, my output would look like this:

<p>Documents in this hyperdocument:</p>
<ol>
<li>Document of type: SgmlDocument</li>
<li>Document of type: WordDocument</li>
<li>Document of type: ExcelDocument</li>
</ol>

Where "WordDocument" and ExcelDocument are names defined in the Word and
Excel property sets that we have privately defined--I don't really
expect Microsoft to publish formal grove property sets for their
proprietary formats any time soon.

This hack required relatively little modification to the 4Suite code
base, although we haven't fully filled out the XPath expressions for
operating on groves.

I have submitted a paper on this work for the Extreme Markup conference
in August.

Cheers,

Eliot

-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From Nicolas.Chauvat@logilab.fr  Fri Apr 13 14:27:26 2001
From: Nicolas.Chauvat@logilab.fr (Nicolas Chauvat)
Date: Fri, 13 Apr 2001 15:27:26 +0200 (CEST)
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
In-Reply-To: <3AD65E24.E9864477@isogen.com>
Message-ID: <Pine.LNX.4.21.0104131519400.1844-100000@aries>

> > Are you talking about the same groves that jade, SGML and DSSSL use?
>=20
> The very same. We are building a generic grove-based hyperdocument
> management system. Given a grove implementation, we can both hyperlink
> to it using a generic hyperdocument API (that looks a lot like HyTime
> but is not dependent on the use of HyTime syntax--it can be bound to any
> reasonable way of representing hyperdocuments, from HTML to Micrsoft
> Project).

Cool. As XML is sometimes limited for what we're trying to achieve (that
is hypertext view of the outer world for http://www.logilab.org/narval/),
we've been thinking about implementing groves and HyTime-related things by
ourselves, but haven't got the needed bandwith yet.

Is that available as (pythonic) free software?

> > What do you mean by "XSL and XPath to process arbitrary python objects"=
?
>=20
> We discovered that with just a couple of lines of code, that we could
> ...

Nice too. I suppose a lot of people could find uses for it if you were to
contribute it back.

> Where "WordDocument" and ExcelDocument are names defined in the Word and
> Excel property sets that we have privately defined--I don't really
> expect Microsoft to publish formal grove property sets for their
> proprietary formats any time soon.

What makes you think so? ;-)
=20
--=20
Nicolas Chauvat

http://www.logilab.com - "Mais o=F9 est donc Ornicar ?" - LOGILAB, Paris (F=
rance)



From uche.ogbuji@fourthought.com  Fri Apr 13 18:39:57 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Fri, 13 Apr 2001 11:39:57 -0600
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
In-Reply-To: Message from "M.-A. Lemburg" <mal@lemburg.com>
 of "Thu, 12 Apr 2001 18:04:40 +0200." <3AD5D218.7B22866@lemburg.com>
Message-ID: <200104131740.f3DHdvj08824@borgia.local.dhcp.fourthought.com>

> Alexandre Fayolle wrote:
> > 
> > On Thu, 12 Apr 2001, M.-A. Lemburg wrote:
> > 
> > > Could you provide pointers to Xerces and Xalan ?
> > 
> > The starting point is http://xml.apache.org/
> > 
> > You'll get http://xml.apache.org/xerces-c/index.html and
> > http://xml.apache.org/xalan-c/index.html quite rapidly. These are the C++
> > implementations. There are also Java implementations, that could be
> > interesting for users of JPython, since I believe 4Suite is not available
> > on this platform.
> 
> Thanks. Those tools look very interesting, but doesn't FourThought
> offer pretty much the same or even a superset of these tools ?

Yes, but I'd be the first to say "the more, the merrier".

I should also note that probably later this year, there will be a new XSLT 
processor for Python, anyway.  We have a proof-of-concept on a 
compile-to-python processor that simply screams, and we plan to work on it 
once we release 4Suite 1.0 and then take care of some business obligations.

4XSLT will move into PyXML as a respectable and very complete processor (if of 
mediocre performance), and the "next-gen" 4XSLT (to be renamed) will be 
developed in 4Suite 1.1.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From brian@sweetapp.com  Sun Apr 15 02:26:27 2001
From: brian@sweetapp.com (Brian Quinlan)
Date: Sat, 14 Apr 2001 18:26:27 -0700
Subject: [XML-SIG] Python extension module for Xerces/Xalan
In-Reply-To: <200104131740.f3DHdvj08824@borgia.local.dhcp.fourthought.com>
Message-ID: <HPEJKHHCMJHPMNLDAPLMIEKJCGAA.brian@sweetapp.com>

I've just made public my first run at a Python extension module providing
XSLT processing capabilities using Xerces/Xalan. The source and Windows
binaries for Python 2.0 and Python 2.1 are available. Unfortunately, there
is no Disutils setup file so you will have to manually place a few DLLs (or
change sys.path). Feedback would be much appreciated.

The URL is http://www.sweetapp.com/Pyana/

It is pretty early stage but I hope to add significant improvements soon.
Below is a sample usage:

>>> import Pyana
>>> params = { 'p1' : "'test'", 'p2' : 'count(//*)', 'p3' : '.' }
>>> sourceDoc = '<?xml version="1.0" ?><A>This is a test</A>'
>>> styleDoc =
Pyana.URI( 'http://www.sweetapp.com/Pyana/samples/params.xsl' )
>>> print Pyana.transform( sourceDoc, styleDoc, params )
<?xml version="1.0" encoding="UTF-8"?>
<params>
<p1>test</p1>
<p2>1</p2>
<p3>This is a test</p3>
<p4>p4 not provided</P4>
<p5>p5 not provided</p5>
</params>



From eliot@isogen.com  Sat Apr 14 16:46:48 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Sat, 14 Apr 2001 10:46:48 -0500
Subject: [XML-SIG] Python Wrapper for Xerces/Xalan
References: <Pine.LNX.4.21.0104131519400.1844-100000@aries>
Message-ID: <3AD870E8.5EEAAE70@isogen.com>

Nicolas Chauvat wrote:
> 
> Is that available as (pythonic) free software?

Unfortunately, at the moment we can only contribute back the changes we
made to PyXML and 4Suite--our own stuff is still proprietary. However I
spoke with our CEO just yesterday about providing an open source version
of what we've got, which is all entirely Python based. She said she
would give it serious consideration.

One additional problem is that a key component of our stuff is a system
provided by another company and they have no plans to make *that* open
source, unfortunately. This is the grove and HyTime engine
implementation, which gives us a lot of functionality. The grove
component can be replaced with PyGrove, but the HyTime engine would be
non-trivial to replace (although not impossible--I've implemented a
HyTime engine in VB, so I know it can be done, I'm just not keen to do
it at the moment).

It's frustrating to me because I would really like to make all the stuff
we've done available to everyone, but current business realities (or
perceived realities) disallow it for now. I'm trying to sell my
employer, DataChannel, on the benefit of other integrators having easy
access to our software so they can develop additional add ons. I think
the 4Suite/Zope model is a good one and is appropriate for what we're
building, but it can be a tough sell in the company wasn't founded with
that model, which DataChannel was not (although we have contributed some
stuff to the Apache project, so there is some hope).

What we're building is the core components that are needed to build a
completely link-based, versioning information management system, where
the goal is to be able to manage arbitrarily complex versioned
hyperdocuments. We have explicitly designed it with a "business module"
plug-in framework that encourages the development of clearly-distinct
(from the core components) managed plug-ins that extend the system and
adapt it to specific use cases. We would like to see an aftermarket of
plug-ins. Because of the way we've implemented the system, it provides a
variety of configurations, from low-cost, low-scale, essentially
all-freeware system to high-cost, high-scale configurations using
enterprise-hardened stuff. I think that if we can make the low-cost
versions of the components available for free, that we'll see people
doing some really cool stuff with it for projects that need this kind of
functionality but that don't have any money to spend (or that require
open source, such as Doug Englebart's Open Hypermedia System project).

Anyway, we're trying to do what we can to contribute back to the
community.
 
Cheers,

Eliot
-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m




From larsga@garshol.priv.no  Mon Apr 16 10:29:29 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 16 Apr 2001 11:29:29 +0200
Subject: [XML-SIG] One-Line Enhancement to catalog.py
In-Reply-To: <200104120920.f3C9KnJ06117@mira.informatik.hu-berlin.de>
References: <3AD510AC.7782ED7A@isogen.com> <200104120920.f3C9KnJ06117@mira.informatik.hu-berlin.de>
Message-ID: <m3puedi45i.fsf@lambda.garshol.priv.no>

* Martin v. Loewis
| 
| It appears that just adding this line is an incomplete change. At a
| minimum, the application should be informed using, say,
| handle_sgmldecl. Can you come up with a patch that does that?

Well, the question is whether that is appropriate or not. This
information is only of interest to an SGML system. On the other hand,
we probably need to support all of the SGML-specific entries in the
end, and then we might perhaps just as well make the catalog parser a
full SGML catalog parser, since someone may find that useful.

Hmmmmmm.

Yeah, let's do it. I'll make the changes and commit them.

--Lars M.



From larsga@garshol.priv.no  Mon Apr 16 11:47:22 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 16 Apr 2001 12:47:22 +0200
Subject: [XML-SIG] Using character entities in external DTD without validating.
In-Reply-To: <200104092155.f39Lt0u03361@mira.informatik.hu-berlin.de>
References: <LPBBKIJMCOEOJAFDNMFEOEILCEAA.pyxml@xhaus.com> <200104092155.f39Lt0u03361@mira.informatik.hu-berlin.de>
Message-ID: <m3oftxi0jp.fsf@lambda.garshol.priv.no>

* Martin v. Loewis
| 
| With some efforts, you might manage to talk xmlproc (the validating
| parser) into not producing validation errors.

You can tell xmlproc to read the external subset even when it does not
validate. Call parser.set_read_external_subset(1) to tell it to do so.

The SAX 2.0 property 

  'http://xml.org/sax/features/external-parameter-entities'

can be used to tell parsers to read the external subset. The only SAX
driver supporting this is that of xmlproc; support for it was checked
into CVS two minutes ago.

--Lars M.



From larsga@garshol.priv.no  Mon Apr 16 12:25:36 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 16 Apr 2001 13:25:36 +0200
Subject: [XML-SIG] PyTRaX?
Message-ID: <m3lmp1hyrz.fsf@lambda.garshol.priv.no>

We now have two XSLT processors usable in Python: 4XSLT and Sablotron,
and it seems that we soon will have another. Is it time to start
thinking about a standardized API to these processors, something along
the lines of TRaX?

I can think of the following areas that might be supported:

 - embedding processors; support for running transformations

 - support for providing structured input (streams, event streams,
   document trees)

 - support for information exchange (error handlers, URI resolvers,
   location information provision etc)

 - support for writing extension elements and functions (difficult)

--Lars M.



From fdrake@acm.org  Mon Apr 16 13:13:01 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Mon, 16 Apr 2001 08:13:01 -0400 (EDT)
Subject: [XML-SIG] PyTRaX?
In-Reply-To: <m3lmp1hyrz.fsf@lambda.garshol.priv.no>
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no>
Message-ID: <15066.57805.329884.23693@cj42289-a.reston1.va.home.com>

Lars Marius Garshol writes:
 > We now have two XSLT processors usable in Python: 4XSLT and Sablotron,
 > and it seems that we soon will have another. Is it time to start
 > thinking about a standardized API to these processors, something along
 > the lines of TRaX?

  Can you provide a pointer for TRaX?  Thanks!


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From larsga@garshol.priv.no  Mon Apr 16 13:47:02 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 16 Apr 2001 14:47:02 +0200
Subject: [XML-SIG] PyTRaX?
In-Reply-To: <15066.57805.329884.23693@cj42289-a.reston1.va.home.com>
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no> <15066.57805.329884.23693@cj42289-a.reston1.va.home.com>
Message-ID: <m3d7adhv09.fsf@lambda.garshol.priv.no>

* Fred L. Drake, Jr.
| 
| Can you provide a pointer for TRaX?

It's in the JAXP specification from Sun, at 

  <URL: http://java.sun.com/aboutJava/communityprocess/review/jsr063/ >

There used to also be a separate TRaX site, but as far as I can tell
that no longer exists.

--Lars M.



From larsga@garshol.priv.no  Mon Apr 16 13:55:22 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 16 Apr 2001 14:55:22 +0200
Subject: [XML-SIG] How to leave character entities alone
In-Reply-To: <200104060535.f365Ylg17064@borgia.local>
References: <200104060535.f365Ylg17064@borgia.local>
Message-ID: <m3ae5hhumd.fsf@lambda.garshol.priv.no>

* Uche Ogbuji
| 
| There was a bug in 4DOM where entities weren't being handled
| properly, but this should now be sorted out.

I'm not convinced that creating EntityReference nodes for entities is
the correct way to handle them. The presence of such nodes is likely
to be surprising to most newbies and to expose bugs in quite a few DOM
applications.

I think the default should be that these nodes are not produced, but
that there may be a flag to for turning them on.

A better fix would be to redesign the way the DOM represents these
nodes, but that is obviously too late now.

--Lars M.



From MichaelDyck@home.com  Tue Apr 17 06:10:56 2001
From: MichaelDyck@home.com (Michael Dyck)
Date: Mon, 16 Apr 2001 22:10:56 -0700
Subject: [XML-SIG] licencing of PyXML contributions
Message-ID: <3ADBD060.3004DD2B@home.com>

I'm working on a Python implementation of XQuery, and I'm thinking of
contributing it to PyXML (if you want it). Before I start showing it to
anyone, I want to put it under an open-source licence. Are there any
licences that would make it easier/harder to add to PyXML?

-Michael Dyck


From pyxml@xhaus.com  Tue Apr 17 12:21:50 2001
From: pyxml@xhaus.com (Alan Kennedy)
Date: Tue, 17 Apr 2001 12:21:50 +0100
Subject: [XML-SIG] "Selling" Open Source.
In-Reply-To: <3AD870E8.5EEAAE70@isogen.com>
Message-ID: <LPBBKIJMCOEOJAFDNMFEGEOPCEAA.pyxml@xhaus.com>

W. Eliot Kimber wrote:

> However I
> spoke with our CEO just yesterday about providing an open
> source version of what we've got, which is all entirely
> Python based. She said she would give it serious
> consideration.

> I'm trying to sell my
> employer, DataChannel, on the benefit of other integrators
> having easy access to our software so they can develop
> additional add ons.

For interest, The Economist has a series of articles this month about
Software and Web Services, including an article on Open Source.

It might be a useful for persuading CEO's, CIO's and CFO's to know that such
an economics/finance/free market oriented publication as the Economist is
writing positively about Open Source, in language that C?O's understand. You
can read the article at

http://www.economist.com/printedition/displayStory.cfm?Story_ID=568269

Contents for the "Survey of Software" can be read at

http://www.economist.com/printedition/index.cfm

Regards,

Alan.



From rsalz@zolera.com  Tue Apr 17 17:39:02 2001
From: rsalz@zolera.com (Rich Salz)
Date: Tue, 17 Apr 2001 12:39:02 -0400
Subject: [XML-SIG] bug in dom.ext.GetElementByID
Message-ID: <3ADC71A6.236051BF@zolera.com>

It looks like GetElementyID is pre-NS (at least).  Does the following
rewrite make sense?  At least it works ... 

_id_key = ('', 'ID')

def GetElementById(startNode, targetId):
    '''
    Return the element in the given tree with an ID attribute of the
given
    value
    '''
    snit = startNode.ownerDocument.createNodeIterator(startNode,
NodeFilter.SHOW_ELEMENT, None, 0)
    curr_node = snit.nextNode()
    while curr_node:
        attr = curr_node.attributes.get(_id_key, None)
        if attr and attr._get_nodeValue() == targetId:
            return curr_node
        curr_node = snit.nextNode()
    return None

Just double-checking before I check it in.
	/r$


From martin@loewis.home.cs.tu-berlin.de  Tue Apr 17 21:12:36 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 17 Apr 2001 22:12:36 +0200
Subject: [XML-SIG] licencing of PyXML contributions
In-Reply-To: <3ADBD060.3004DD2B@home.com> (message from Michael Dyck on Mon,
 16 Apr 2001 22:10:56 -0700)
References: <3ADBD060.3004DD2B@home.com>
Message-ID: <200104172012.f3HKCau01647@mira.informatik.hu-berlin.de>

> Before I start showing it to anyone, I want to put it under an
> open-source licence. Are there any licences that would make it
> easier/harder to add to PyXML?

I guess inclusion of GPL'ed software would be difficult unless we
could reasonably claim that it is "mere packaging", since all derived
code must then appear under GPL, also, and it might be hard to get all
contributors to agree to change there license to GPL.

We have the habit of copying all license text into the file LICENCE.
For that, it would simplify maintainance if the license text would not
read, say, "MICHAEL DYCK DISCLAIMS ALL WARRANTIES", since that could
easily change into "MICHALE DYCK, LARS MARIUS GARSHOL, FRED DRAKE, AND
MARTIN VON LOEWIS DISCLAIM ALL WARRANTIES". So you might chose a
licensing text that does not mention the author's name. In recent
expat (I believe), the text 

<license>
Permission is hereby granted, free of charge, to any person obtaining
a copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:
 
The above copyright notice and this permission notice shall be included
in all copies or substantial portions of the Software.
 
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
</license>

is used. In the copy in LICENSE, we don't copy "the above copyright",
but instead mention the files where it can be found. This license
essentially allows any use, as long as the copyright notice in the
source files is maintained.

Of course, this is a suggestion only - he who writes the code choses
the license.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Tue Apr 17 21:15:10 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 17 Apr 2001 22:15:10 +0200
Subject: [XML-SIG] Using character entities in external DTD without validating.
In-Reply-To: <m3oftxi0jp.fsf@lambda.garshol.priv.no> (message from Lars Marius
 Garshol on 16 Apr 2001 12:47:22 +0200)
References: <LPBBKIJMCOEOJAFDNMFEOEILCEAA.pyxml@xhaus.com> <200104092155.f39Lt0u03361@mira.informatik.hu-berlin.de> <m3oftxi0jp.fsf@lambda.garshol.priv.no>
Message-ID: <200104172015.f3HKFAO01672@mira.informatik.hu-berlin.de>

> The only SAX driver supporting this is that of xmlproc; support for
> it was checked into CVS two minutes ago.

:-)


From doc@sympatico.ca  Wed Apr 18 01:09:59 2001
From: doc@sympatico.ca (DOC)
Date: Tue, 17 Apr 2001 17:09:59 -0700
Subject: [XML-SIG] PyXML-0.6.5 win download
Message-ID: <000c01c0c79b$e945d4e0$13733dcf@b1wyft68>

This is a multi-part message in MIME format.

------=_NextPart_000_0009_01C0C761.3BF45F80
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable



Hi:

I did a download of this file but the self extracting process=20
stalls out after the second or third frame.

PyXML-0.6.5.win32-py2.0.exe

I tried it on two diff machines. One was running win 95 the other ME.

DOC



PyXML-0.6.5.win32-py2.0.exe

------=_NextPart_000_0009_01C0C761.3BF45F80
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Diso-8859-1">
<META content=3D"MSHTML 5.50.4134.600" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff><FONT face=3DArial size=3D2>
<DIV>&nbsp;</DIV>
<DIV>&nbsp;</DIV>
<DIV>Hi:</DIV>
<DIV>&nbsp;</DIV>
<DIV>I did a download of this file but the self extracting process =
</DIV>
<DIV>stalls out after the second or third frame.</DIV>
<DIV>&nbsp;</DIV>
<DIV>
<DIV><A=20
href=3D"http://prdownloads.sourceforge.net/pyxml/PyXML-0.6.5.win32-py2.0.=
exe">PyXML-0.6.5.win32-py2.0.exe</A></DIV></DIV>
<DIV>&nbsp;</DIV>
<DIV>I tried it on two diff machines. One was running win 95 the other =
ME.</DIV>
<DIV>&nbsp;</DIV>
<DIV>DOC</DIV>
<DIV>&nbsp;</DIV>
<DIV>&nbsp;</DIV>
<DIV>&nbsp;</DIV>
<DIV><A=20
href=3D"http://prdownloads.sourceforge.net/pyxml/PyXML-0.6.5.win32-py2.0.=
exe">PyXML-0.6.5.win32-py2.0.exe</A></DIV></FONT></BODY></HTML>

------=_NextPart_000_0009_01C0C761.3BF45F80--



From martin@loewis.home.cs.tu-berlin.de  Tue Apr 17 22:36:59 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 17 Apr 2001 23:36:59 +0200
Subject: [XML-SIG] PyXML-0.6.5 win download
In-Reply-To: <000c01c0c79b$e945d4e0$13733dcf@b1wyft68> (doc@sympatico.ca)
References: <000c01c0c79b$e945d4e0$13733dcf@b1wyft68>
Message-ID: <200104172136.f3HLax202355@mira.informatik.hu-berlin.de>

> I did a download of this file but the self extracting process
> stalls out after the second or third frame.
> 
> PyXML-0.6.5.win32-py2.0.exe

Can you please elaborate? Did you download the file completely, or did
the download stall after the "second or third frame"? If the latter,
what do you mean by "frame"?

If the download completed, can you please report the exact order of
commands executed, and buttons pressed and clicked?

Thanks,
Martin



From Mike.Olson@fourthought.com  Wed Apr 18 00:35:20 2001
From: Mike.Olson@fourthought.com (Mike Olson)
Date: Tue, 17 Apr 2001 17:35:20 -0600
Subject: [XML-SIG] PyTRaX?
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no>
Message-ID: <3ADCD338.ACCF9907@FourThought.com>

Lars Marius Garshol wrote:
> 
> We now have two XSLT processors usable in Python: 4XSLT and Sablotron,
> and it seems that we soon will have another. Is it time to start
> thinking about a standardized API to these processors, something along
> the lines of TRaX?
> 
> I can think of the following areas that might be supported:
> 
>  - embedding processors; support for running transformations
> 
>  - support for providing structured input (streams, event streams,
>    document trees)
> 
>  - support for information exchange (error handlers, URI resolvers,
>    location information provision etc)
> 
>  - support for writing extension elements and functions (difficult)


I was thinking along these lines as well.  Something else I thought of
as needed is a way to register and retrieve instances of processors. 
Something like xml.xslt.newProcessor('4XSLT') and
xml.xslt.RegisterProcessor('Sablotron',Sablotron.Processor)

Mike
> 
> --Lars M.
> 
> _______________________________________________
> XML-SIG maillist  -  XML-SIG@python.org
> http://mail.python.org/mailman/listinfo/xml-sig

-- 
Mike Olson				 Principal Consultant
mike.olson@fourthought.com               (303)583-9900 x 102
Fourthought, Inc.                         http://Fourthought.com 
Software-engineering, knowledge-management, XML, CORBA, Linux, Python


From jmurray@agyinc.com  Wed Apr 18 02:21:04 2001
From: jmurray@agyinc.com (Joe Murray)
Date: Tue, 17 Apr 2001 18:21:04 -0700
Subject: [XML-SIG] sax expatreader and unicode
Message-ID: <3ADCEC00.3796F640@agyinc.com>

What am I missing: the sax expatreader can't handle some unicode
characters?  I thought this was supported.  I believe the xml.dom
modules handle unicode characters just fine...

>From the text:

"...LEX. IN NA=EFVE H4 AND CHO CELLS, PS1 CO-IMM..."

Output:
=2E
=2E
=2E

  File "analyzexml.py", line 68, in analyze_sax
    parser.parse(stream)
  File
"/usr/local/lib/python2.0/site-packages/_xmlplus/sax/expatreader.py",
line 43, in parse
    xmlreader.IncrementalParser.parse(self, source)
  File
"/usr/local/lib/python2.0/site-packages/_xmlplus/sax/xmlreader.py", line
120, in parse
    self.feed(buffer)
  File
"/usr/local/lib/python2.0/site-packages/_xmlplus/sax/expatreader.py",
line 87, in feed
    self._parser.Parse(data, isFinal)
UnicodeError: UTF-8 decoding error: invalid data


jm

--=20
Joseph Murray
Bioinformatics Specialist, AGY Therapeutics
290 Utah Avenue, South San Francisco, CA 94080
(650) 228-1146


From martin@loewis.home.cs.tu-berlin.de  Wed Apr 18 06:25:56 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Wed, 18 Apr 2001 07:25:56 +0200
Subject: [XML-SIG] sax expatreader and unicode
In-Reply-To: <3ADCEC00.3796F640@agyinc.com> (message from Joe Murray on Tue,
 17 Apr 2001 18:21:04 -0700)
References: <3ADCEC00.3796F640@agyinc.com>
Message-ID: <200104180525.f3I5Puf03550@mira.informatik.hu-berlin.de>

> What am I missing: the sax expatreader can't handle some unicode
> characters?

Most likely, the error is in your data, not in Expat.

> >From the text:
> 
> "...LEX. IN NAïVE H4 AND CHO CELLS, PS1 CO-IMM..."

You did not give the complete document. Did it include a <?xml
declaration, with an encoding= attribute?

> UnicodeError: UTF-8 decoding error: invalid data

That error is properly reported: Your data, atleast as transmitted in
your message, is not valid UTF-8. In this message, the character in
question is primarily the byte \xef. If taken as Latin-1, it is the
character LATIN SMALL LETTER I WITH DIAERESIS. You have to declare
that the document is Latin-1, or else an XML processor will assume
UTF-8.

Regards,
Martin



From larsga@garshol.priv.no  Wed Apr 18 17:03:10 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 18 Apr 2001 18:03:10 +0200
Subject: [XML-SIG] PyTRaX?
In-Reply-To: <3ADCD338.ACCF9907@FourThought.com>
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no> <3ADCD338.ACCF9907@FourThought.com>
Message-ID: <m37l0i9ow1.fsf@lambda.garshol.priv.no>

* Mike Olson
| 
| I was thinking along these lines as well.  

Excellent! So perhaps we should go out and do it. :-)

| Something else I thought of as needed is a way to register and
| retrieve instances of processors.

Good point. There does need to be something along the lines of the
make_parser function in SAX.

--Lars M.



From mrklaw@mindless.com  Wed Apr 18 17:46:18 2001
From: mrklaw@mindless.com (Sloan Poe)
Date: Wed, 18 Apr 2001 12:46:18 -0400 (EDT)
Subject: [XML-SIG] question...
Message-ID: <200104181646.f3IGkJY04773@raven.warren-wilson.edu>

--684802-1804289383-987612378=:4739
Content-Type: TEXT/PLAIN; CHARSET=US-ASCII
Content-Disposition: INLINE

--Please respond to me directly as I am not on the list--

I'm having a small problem getting some of the demos included with the
python xml package to run.  I've installed the latest version from
sourceforge and confirmed that all the files are installed in
/usr/lib/python1.5/site-packages/
but when I try do use the statement from xml.sax.saxutils import escape I
get an error.

I can "import xml" or even "from xml.sax import saxutils" I just can't
import escape or any function for that matter.
I'm trying to run one of the scripts in the xbel dir.  I can get it to work
on one computer but not on a computer with an almost identical
installation.


I'm sure this is some ignorance on my part. 

thanks in advance for your help.

--684802-1804289383-987612378=:4739
Content-Type: TEXT/x-vcard; CHARSET=US-ASCII; NAME="vcard.vcf"
Content-Disposition: INLINE; FILENAME="vcard.vcf"

BEGIN:VCARD
FN:Mr. R. Sloan Poe, Jr.
TITLE:Web Programmer
ORG:Warren Wilson College;Web Crew
ADR;DOM;PARCEL;WORK:;6294 Warren Wilson College;701 Warren Wilson Road;Swannanoa;NC;28778;
LABEL;POSTAL;WORK;ENCODING=QUOTED-PRINTABLE:6294 Warren Wilson College=0D=0A=
701 Warren Wilson Road=0D=0A=
Swannanoa, NC 28778
ADR;DOM;POSTAL;WORK:;6294 Warren Wilson College;P.O. Box 9000;Asheville;NC;28815-9000;
LABEL;POSTAL;WORK;ENCODING=QUOTED-PRINTABLE:6294 Warren Wilson College=0D=0A=
P.0. Box 9000=0D=0A=
Asheville, NC 28815-9000
TEL;Home;VOICE;MESG;PREF:1-828-771-5906
EMAIL;Internet:mrklaw@mindless.com
URL:http://www.warren-wilson.edu/~rpoe/card.vcf
UID:http://www.warren-wilson.edu/~rpoe/card.vcf
TZ:-0500
BDAY:1978-03-16
REV:20010403T174502
VERSION:2.1
END:VCARD

--684802-1804289383-987612378=:4739
Content-Type: TEXT/PLAIN; CHARSET=US-ASCII
Content-Disposition: INLINE


--

Sloan Poe
mrklaw@mindless.com

If I'm insane, who are you in?

--684802-1804289383-987612378=:4739--


From martin@loewis.home.cs.tu-berlin.de  Wed Apr 18 20:42:09 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Wed, 18 Apr 2001 21:42:09 +0200
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <m37l0i9ow1.fsf@lambda.garshol.priv.no> (message from Lars Marius
 Garshol on 18 Apr 2001 18:03:10 +0200)
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no> <3ADCD338.ACCF9907@FourThought.com> <m37l0i9ow1.fsf@lambda.garshol.priv.no>
Message-ID: <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de>

> Good point. There does need to be something along the lines of the
> make_parser function in SAX.

My proposal would be to add two keyword arguments, properties= and
features=. Each is a list of binary tuples, each tuple has name and
value. Alternatively, dictionaries might be better.

make_parser will the iterate over all known parser factories, invoking
create_parser for each, then trying to set all the properties and
features. It will return the first parser that supports all of them,
and return a configured instance.

There should also be a function xml.sax.register_parser, which accepts
an object that has a create_parser function, or a string naming a
module that has a create_parser function.

What do you think?

Regards,
Martin



From fdrake@acm.org  Wed Apr 18 21:03:49 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Wed, 18 Apr 2001 16:03:49 -0400 (EDT)
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de>
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no>
 <3ADCD338.ACCF9907@FourThought.com>
 <m37l0i9ow1.fsf@lambda.garshol.priv.no>
 <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de>
Message-ID: <15069.62245.424602.782995@cj42289-a.reston1.va.home.com>

Martin v. Loewis writes:
 > make_parser will the iterate over all known parser factories, invoking
...
 > There should also be a function xml.sax.register_parser, which accepts

  Sounds pretty reasonable to me.  My biggest concern is the possible
cost of creating all those parsers, but I'm not sure how important
that is.  I'm thinking of how I set up my DOMBuilder object, and it's
actually pretty thin; the low-level parser doesn't get created until
it's actually needed.


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From martin@loewis.home.cs.tu-berlin.de  Wed Apr 18 20:57:40 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Wed, 18 Apr 2001 21:57:40 +0200
Subject: [XML-SIG] question...
In-Reply-To: <200104181646.f3IGkJY04773@raven.warren-wilson.edu> (message from
 Sloan Poe on Wed, 18 Apr 2001 12:46:18 -0400 (EDT))
References: <200104181646.f3IGkJY04773@raven.warren-wilson.edu>
Message-ID: <200104181957.f3IJveV00955@mira.informatik.hu-berlin.de>

> I'm sure this is some ignorance on my part. 

Hard to say, with as little information as you've provided.

> I'm having a small problem getting some of the demos included with the
> python xml package to run.  I've installed the latest version from
> sourceforge and confirmed that all the files are installed in
> /usr/lib/python1.5/site-packages/
> but when I try do use the statement from xml.sax.saxutils import escape I
> get an error.

It would be helpful if you could report exactly what error you get,
best with a full traceback as reported by python, copying that
literally from the terminal where you have tried executing the code in
question.

> I can "import xml" or even "from xml.sax import saxutils" I just can't
> import escape or any function for that matter.

When you do

>>> from xml.sax import saxutils
>>> saxutils
<module 'xml.sax.saxutils' from '/usr/lib/python1.5/site-packages/xml/sax/saxutils.pyc'> 

what do you get? If you look at the source file of the module
(/usr/lib/python1.5/site-packages/xml/sax/saxutils.py in my case),
can you spot the escape function?

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Wed Apr 18 21:26:43 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Wed, 18 Apr 2001 22:26:43 +0200
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <15069.62245.424602.782995@cj42289-a.reston1.va.home.com>
 (fdrake@acm.org)
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no>
 <3ADCD338.ACCF9907@FourThought.com>
 <m37l0i9ow1.fsf@lambda.garshol.priv.no>
 <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de> <15069.62245.424602.782995@cj42289-a.reston1.va.home.com>
Message-ID: <200104182026.f3IKQhm01211@mira.informatik.hu-berlin.de>

>   Sounds pretty reasonable to me.  My biggest concern is the possible
> cost of creating all those parsers, but I'm not sure how important
> that is.  I'm thinking of how I set up my DOMBuilder object, and it's
> actually pretty thin; the low-level parser doesn't get created until
> it's actually needed.

I'd be happy to use a different procedure, if somebody can propose on.
Providing lists of all supported properties and their possible values
in advance does not seem attractive: there could be an infinite number
of supported values.

Regards,
Martin


From fdrake@acm.org  Wed Apr 18 22:03:06 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Wed, 18 Apr 2001 17:03:06 -0400 (EDT)
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <200104182026.f3IKQhm01211@mira.informatik.hu-berlin.de>
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no>
 <3ADCD338.ACCF9907@FourThought.com>
 <m37l0i9ow1.fsf@lambda.garshol.priv.no>
 <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de>
 <15069.62245.424602.782995@cj42289-a.reston1.va.home.com>
 <200104182026.f3IKQhm01211@mira.informatik.hu-berlin.de>
Message-ID: <15070.266.458475.645549@cj42289-a.reston1.va.home.com>

Martin v. Loewis writes:
 > I'd be happy to use a different procedure, if somebody can propose on.
 > Providing lists of all supported properties and their possible values
 > in advance does not seem attractive: there could be an infinite number
 > of supported values.

  Agreed, and I'd expect maintainability would be seriously
compromised as well.
  If somethon has an implementation of your proposal, I'll certainly
not stand in the way.  ;-)  My misgivings about instantiating the
parsers are fairly minor, so please don't consider my comment as
reason to worry about it.  That (worrying about it) is a small enough
job that I can handle it alone.  ;-)


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From rsalz@zolera.com  Thu Apr 19 02:32:58 2001
From: rsalz@zolera.com (Rich Salz)
Date: Wed, 18 Apr 2001 21:32:58 -0400
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no> <3ADCD338.ACCF9907@FourThought.com> <m37l0i9ow1.fsf@lambda.garshol.priv.no> <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de>
Message-ID: <3ADE404A.76EFC23B@zolera.com>

I think this is pretty close. :)

I think the keyword arguments properties and features should take
dictionaries.

> There should also be a function xml.sax.register_parser, which accepts
> an object that has a create_parser function, or a string naming a
> module that has a create_parser function.

I think perhaps
	def can_create_parser(properties={}, features={}):
		'If you can create a parser with the desired properties and features,
		return non-None.  If you cannot, return None.'

	def create_parser(properties={}, features={}):
		'Create a parser with the specified prperties and features, return
		None (raise an exception?) if not possible.'

Is a little cleaner set of callbacks.


From uche.ogbuji@fourthought.com  Thu Apr 19 05:06:34 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Wed, 18 Apr 2001 22:06:34 -0600
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: Message from "Martin v. Loewis" <martin@loewis.home.cs.tu-berlin.de>
 of "Wed, 18 Apr 2001 21:42:09 +0200." <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de>
Message-ID: <200104190406.f3J46Ym01936@borgia.local>

> > Good point. There does need to be something along the lines of the
> > make_parser function in SAX.
> 
> My proposal would be to add two keyword arguments, properties= and
> features=. Each is a list of binary tuples, each tuple has name and
> value. Alternatively, dictionaries might be better.
> 
> make_parser will the iterate over all known parser factories, invoking
> create_parser for each, then trying to set all the properties and
> features. It will return the first parser that supports all of them,
> and return a configured instance.
> 
> There should also be a function xml.sax.register_parser, which accepts
> an object that has a create_parser function, or a string naming a
> module that has a create_parser function.
> 
> What do you think?

I like.  But I think we'll want more than just a parser factory:

properties = {"http://factory.pyxml.org/properties/encoding": "BIG5",
              "http://uche.ogbuji.net/my-extended-properties/spam": "eggs"}
features = ["http://factory.pyxml.org/properties/xpointer"]
xml.factory.getXsltProcessor(properties, features)


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From uche.ogbuji@fourthought.com  Thu Apr 19 05:46:09 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Wed, 18 Apr 2001 22:46:09 -0600
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: Message from Uche Ogbuji <uche.ogbuji@fourthought.com>
 of "Wed, 18 Apr 2001 22:06:34 MDT." <200104190406.f3J46Ym01936@borgia.local>
Message-ID: <200104190446.f3J4kAK02141@borgia.local>

> I like.  But I think we'll want more than just a parser factory:
> 
> properties = {"http://factory.pyxml.org/properties/encoding": "BIG5",
>               "http://uche.ogbuji.net/my-extended-properties/spam": "eggs"}
> features = ["http://factory.pyxml.org/properties/xpointer"]
> xml.factory.getXsltProcessor(properties, features)

Or should that be

urn:factory.pyxml.org:properties:encoding

etc.?

At any rate, looks as if the pyxml.org domain is just a-sitting there.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From fdrake@acm.org  Thu Apr 19 06:01:38 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Thu, 19 Apr 2001 01:01:38 -0400 (EDT)
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <200104190446.f3J4kAK02141@borgia.local>
References: <uche.ogbuji@fourthought.com>
 <200104190406.f3J46Ym01936@borgia.local>
 <200104190446.f3J4kAK02141@borgia.local>
Message-ID: <15070.28978.544846.154129@cj42289-a.reston1.va.home.com>

Uche Ogbuji writes:
 > Or should that be
 > 
 > urn:factory.pyxml.org:properties:encoding

  Gosh, I wish the W3C would be consistent!  They seem to like URIs
until the next time they write a spec, and then use short strings!
(Look at the DOM: feature names are short strings, DOMBuilder/
DOMWriter properties are short strings... except they want everyone
else to use Java-style reversed domain names.  Sheesh!  That's
pathetic!
  At least the SAX crew didn't start introducing other types of
names.

 > At any rate, looks as if the pyxml.org domain is just a-sitting there.

  Is there any reason we can't use the python.org domain for this?


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From uche.ogbuji@fourthought.com  Thu Apr 19 06:14:55 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Wed, 18 Apr 2001 23:14:55 -0600
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: Message from "Fred L. Drake, Jr." <fdrake@acm.org>
 of "Thu, 19 Apr 2001 01:01:38 EDT." <15070.28978.544846.154129@cj42289-a.reston1.va.home.com>
Message-ID: <200104190515.f3J5Ete02212@borgia.local>

> Uche Ogbuji writes:
>  > Or should that be
>  > 
>  > urn:factory.pyxml.org:properties:encoding
> 
>   Gosh, I wish the W3C would be consistent!

W3C?  Consistent?  In a pig's een.

Didntcha hear?  Now they produce "standards", not "recommendations".

>   At least the SAX crew didn't start introducing other types of
> names.

Yeah.  Of course SAX was my inspiration, until my mind wandered to URNs.

>  > At any rate, looks as if the pyxml.org domain is just a-sitting there.
> 
>   Is there any reason we can't use the python.org domain for this?

Not really.

http://xml.python.org/factory/property

works for me as well.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From martin@loewis.home.cs.tu-berlin.de  Thu Apr 19 07:06:19 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 19 Apr 2001 08:06:19 +0200
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <200104190446.f3J4kAK02141@borgia.local> (message from Uche
 Ogbuji on Wed, 18 Apr 2001 22:46:09 -0600)
References: <200104190446.f3J4kAK02141@borgia.local>
Message-ID: <200104190606.f3J66Jn01217@mira.informatik.hu-berlin.de>

> > I like.  But I think we'll want more than just a parser factory:
> > 
> > properties = {"http://factory.pyxml.org/properties/encoding": "BIG5",
> >               "http://uche.ogbuji.net/my-extended-properties/spam": "eggs"}
> > features = ["http://factory.pyxml.org/properties/xpointer"]
> > xml.factory.getXsltProcessor(properties, features)
> 
> Or should that be
> 
> urn:factory.pyxml.org:properties:encoding
> 
> etc.?

I'm not sure whether requiring URIs as configuration keys is such a
smart idea, anyway. SAX uses that approach, DOM uses simple strings
(like 'core' or 'mutationevents'). Works equally well, IMO. Even for
SAX, most Python users probably don't write 

  p.setFeature("http://xml.org/sax/features/namespaces", 1)

but

  from xml.sax.handler import feature_namespaces
  ...
  p.setFeature(feature_namespaces, 1)

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Thu Apr 19 07:01:33 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 19 Apr 2001 08:01:33 +0200
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <200104190406.f3J46Ym01936@borgia.local> (message from Uche
 Ogbuji on Wed, 18 Apr 2001 22:06:34 -0600)
References: <200104190406.f3J46Ym01936@borgia.local>
Message-ID: <200104190601.f3J61XW01112@mira.informatik.hu-berlin.de>

> I like.  But I think we'll want more than just a parser factory:

Certainly. This one was specifically about xml.sax.make_parser.

> properties = {"http://factory.pyxml.org/properties/encoding": "BIG5",
>               "http://uche.ogbuji.net/my-extended-properties/spam": "eggs"}
> features = ["http://factory.pyxml.org/properties/xpointer"]
> xml.factory.getXsltProcessor(properties, features)

I'm not sure the same factory mechanism would work for all
classes. E.g. to select and configure a SAX parser, you need to do
setFeature/setProperty calls. To select a DOM implementation, you need
to check hasFeature on it, which triggered the
xml.dom.getDOMImplementation API. I'd rather obtain an XSLT processor
by calling xml.xslt.getProcessor (or xml.xslt.make_processor).

Regards,
Martin



From larsga@garshol.priv.no  Thu Apr 19 08:35:01 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 19 Apr 2001 09:35:01 +0200
Subject: [XML-SIG] Re: SAX parser factories (Was: PyTRaX?)
In-Reply-To: <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de>
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no> <3ADCD338.ACCF9907@FourThought.com> <m37l0i9ow1.fsf@lambda.garshol.priv.no> <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de>
Message-ID: <m3lmox8hqy.fsf@lambda.garshol.priv.no>

* Martin v. Loewis
| 
| My proposal would be to add two keyword arguments, properties= and
| features=. Each is a list of binary tuples, each tuple has name and
| value. Alternatively, dictionaries might be better.

I like this, and have been thinking about making something similar for
SAX as an add-on outside the core API. I would much prefer
dictionaries, BTW.
 
| make_parser will the iterate over all known parser factories, invoking
| create_parser for each, then trying to set all the properties and
| features. It will return the first parser that supports all of them,
| and return a configured instance.

Yep. This was my idea as well.
 
| There should also be a function xml.sax.register_parser, which accepts
| an object that has a create_parser function, or a string naming a
| module that has a create_parser function.

Sounds good to me.

But is this a proposal for a PyTRAX processor factory or for a SAX
parser factory? If the latter, do you think this should go into core
Python?

--Lars M.



From tpassin@home.com  Thu Apr 19 14:02:37 2001
From: tpassin@home.com (Thomas B. Passin)
Date: Thu, 19 Apr 2001 09:02:37 -0400
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
References: <200104190446.f3J4kAK02141@borgia.local> <200104190606.f3J66Jn01217@mira.informatik.hu-berlin.de>
Message-ID: <004201c0c8d1$0294cc40$7cac1218@reston1.va.home.com>

Martin v. Loewis said -

>
> I'm not sure whether requiring URIs as configuration keys is such a
> smart idea, anyway. SAX uses that approach, DOM uses simple strings
> (like 'core' or 'mutationevents'). Works equally well, IMO. Even for
> SAX, most Python users probably don't write
>

I favor short (non-URI) feature names.  Using URIs to get unique strings may
work well when you have the possibility of many people working independently
on many projects accidentally producing name collisions.  I don't thnk that
applies here.  Even if, say, the RDF people duplicated some of our feature
names, it wouldn't matter since they would be used in a different context.

Let's keep them short.

Cheers,

Tom P



From tpassin@home.com  Thu Apr 19 14:10:43 2001
From: tpassin@home.com (Thomas B. Passin)
Date: Thu, 19 Apr 2001 09:10:43 -0400
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
References: <200104190406.f3J46Ym01936@borgia.local> <200104190601.f3J61XW01112@mira.informatik.hu-berlin.de>
Message-ID: <004a01c0c8d2$241f6ae0$7cac1218@reston1.va.home.com>

I don't like make_parser() and create_parser() approach because the names
are so similar as to lead to confusion.  I do like the proposal to ask for a
list of processors that support a feature list, which ability wuld be
separate from actual parser creation.  I say a list because if more than one
had my features, I might want to choose one rather than another, not just
take whichever one the system wanted to give me.

Also, with the has_features() approach, it would be possible to keep a
simple features catalog which would avoid the need to instantiate a
processor so it could be asked if it had some feature.  If each processor
(or wrapper) could respond to a request for its features, a script could
automatically query it when it was first registered and save the data in the
catalog.  Perhaps the catalog could be in xml, more likely a dictionary
format.

The next step in this evolution could be named feature sets!

Cheers,

Tom P



From fdrake@acm.org  Thu Apr 19 15:40:32 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Thu, 19 Apr 2001 10:40:32 -0400 (EDT)
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <004201c0c8d1$0294cc40$7cac1218@reston1.va.home.com>
References: <200104190406.f3J46Ym01936@borgia.local>
 <200104190601.f3J61XW01112@mira.informatik.hu-berlin.de>
 <004a01c0c8d2$241f6ae0$7cac1218@reston1.va.home.com>
 <200104190446.f3J4kAK02141@borgia.local>
 <200104190606.f3J66Jn01217@mira.informatik.hu-berlin.de>
 <004201c0c8d1$0294cc40$7cac1218@reston1.va.home.com>
Message-ID: <15070.63712.948366.946492@cj42289-a.reston1.va.home.com>

Thomas B. Passin writes:
 > I favor short (non-URI) feature names.  Using URIs to get unique strings may
 > work well when you have the possibility of many people working independently
 > on many projects accidentally producing name collisions.  I don't thnk that
 > applies here.  Even if, say, the RDF people duplicated some of our feature
 > names, it wouldn't matter since they would be used in a different context.

  Why would it not apply here?  I know the "Parsed XML" product for
Zope supports some additional features that can be checked with
hasFeature(); why would this not be the case for other APIs which can
support customizable features?
  If we do go with short names, we should at least strongly recommend
a way to formulate the names for additional features.  I'd stick with
what's recommended for the DOM in this case; for example:
"org.zope.dom.persistence".

Thomas B. Passin writes:
 > I don't like make_parser() and create_parser() approach because the names
 > are so similar as to lead to confusion.  I do like the proposal to ask for a
 > list of processors that support a feature list, which ability wuld be
 > separate from actual parser creation.  I say a list because if more than one
 > had my features, I might want to choose one rather than another, not just
 > take whichever one the system wanted to give me.

  So perhaps what we have is an interface like this:

	object SAXImplementation:
            def can_set_feature(feature, enabled):
                """Return true if the parsers returned by create()
                can support 'feature', and false if they don't."""

            def can_support_property(property, value):
                """Return true if the parsers returned by create()
                can support 'property' with the given value, and
                false if they can't."""

            def get_features_list():
                """Return a list of supported feature names.
                Inclusion of a name in the list does not imply that
                the feature is supported in both enabled and disabled
                forms."""

            def get_properties_list():
                """Return a list of supported property names.
                Inclusion of a name in the list does not imply that
                all values for that property are supported."""

            def create():
                """Return a new XMLReader instance."""

        def find_parsers(features={}, properties={}):
            """Return a list of SAXImplementation objects that support
            the given features and properties."""

        def create_parser(features={}, properties={}):
            """Return a configured parser object that supports the
            given feature and property settings.  If more than one
            SAXImplementation supports the given settings, one will be
            selected arbitrarily."""

        def register_parser(impl):
            """Add the SAXImplementation 'impl' to the set of parsers
            known to the factory."""

(I expect SAXImplementation objects will usually be modules.)

 > Also, with the has_features() approach, it would be possible to keep a
 > simple features catalog which would avoid the need to instantiate a
 > processor so it could be asked if it had some feature.  If each processor
 > (or wrapper) could respond to a request for its features, a script could
 > automatically query it when it was first registered and save the data in the

  This would be fine with me.  I think the SAXImplementation interface
I outlined above would support this implementation of the factory
functions.

 > catalog.  Perhaps the catalog could be in xml, more likely a dictionary
 > format.

  I expect such a catalog would have to be a volatile data structure
rather than something persistent -- the interaction with sys.path
would be a nightmare for a persistent catalog!

 > The next step in this evolution could be named feature sets!

  That would be nice to have!


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From ken@bitsko.slc.ut.us  Thu Apr 19 16:14:55 2001
From: ken@bitsko.slc.ut.us (Ken MacLeod)
Date: 19 Apr 2001 10:14:55 -0500
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: Uche Ogbuji's message of "Wed, 18 Apr 2001 22:06:34 -0600"
References: <200104190406.f3J46Ym01936@borgia.local>
Message-ID: <m04rvkq5u8.fsf@bitsko.slc.ut.us>

[please do not Cc me msgs posted to list, thx.]

Uche Ogbuji <uche.ogbuji@fourthought.com> writes:

> I like.  But I think we'll want more than just a parser factory:
> 
> properties = {"http://factory.pyxml.org/properties/encoding": "BIG5",
>               "http://uche.ogbuji.net/my-extended-properties/spam": "eggs"}
> features = ["http://factory.pyxml.org/properties/xpointer"]
> xml.factory.getXsltProcessor(properties, features)

Note when extending this beyond just SAX, "properties" were added to
SAX a short time after "features" were.  David Megginson later stated
that having both was probably redundant, that a property with a value
of true or false was effectively the same as a feature.

  -- Ken


From fdrake@acm.org  Thu Apr 19 16:26:12 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Thu, 19 Apr 2001 11:26:12 -0400 (EDT)
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <m04rvkq5u8.fsf@bitsko.slc.ut.us>
References: <200104190406.f3J46Ym01936@borgia.local>
 <m04rvkq5u8.fsf@bitsko.slc.ut.us>
Message-ID: <15071.916.14259.394048@cj42289-a.reston1.va.home.com>

Ken MacLeod writes:
 > Note when extending this beyond just SAX, "properties" were added to
 > SAX a short time after "features" were.  David Megginson later stated
 > that having both was probably redundant, that a property with a value
 > of true or false was effectively the same as a feature.

  I agree, but I also understand why Java programmers might like
having them as distinct things (the old boolean vs. Object static
typing thing).
  For the Python SAX binding, I think we should keep the same
distinction that exists in SAX, but that APIs originating in Python
should unify properties and features, and just call them properties.


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From uche.ogbuji@fourthought.com  Thu Apr 19 16:40:36 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Thu, 19 Apr 2001 09:40:36 -0600
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: Message from Ken MacLeod <ken@bitsko.slc.ut.us>
 of "19 Apr 2001 10:14:55 CDT." <m04rvkq5u8.fsf@bitsko.slc.ut.us>
Message-ID: <200104191540.f3JFeaf04858@borgia.local.dhcp.fourthought.com>

> [please do not Cc me msgs posted to list, thx.]

This is really something that should be fixed in the list config.  Mailman 
does cc by default.  Personally, I don't have a problem with it, but it's 
probably because of the way I filter my mail.

> Uche Ogbuji <uche.ogbuji@fourthought.com> writes:
> 
> > I like.  But I think we'll want more than just a parser factory:
> > 
> > properties = {"http://factory.pyxml.org/properties/encoding": "BIG5",
> >               "http://uche.ogbuji.net/my-extended-properties/spam": "eggs"}
> > features = ["http://factory.pyxml.org/properties/xpointer"]
> > xml.factory.getXsltProcessor(properties, features)
> 
> Note when extending this beyond just SAX, "properties" were added to
> SAX a short time after "features" were.  David Megginson later stated
> that having both was probably redundant, that a property with a value
> of true or false was effectively the same as a feature.

Agreed.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From uche.ogbuji@fourthought.com  Thu Apr 19 16:49:07 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Thu, 19 Apr 2001 09:49:07 -0600
Subject: [XML-SIG] RDF Parser -> PyXML
Message-ID: <200104191549.f3JFn7x04884@borgia.local.dhcp.fourthought.com>

James Tauber and I met at XML DevCon and had a good chat.  We agreed that it 
would be great to get Redfoot's RDF parser into PyXML as a basis for RDF 
support.  Then both RedFoot and 4RDF could use this as the parser, and build 
their unique additional functions on top of it.

I really liek this idea, especially since I'm quite dissatisfied with the RDF 
parser in 4RDF and would like to have an alternative.  the quickest way to get 
this would be if there were one in PyXML that we could use.

So I'd like to wind up the discussion here so that we can maybe check in the 
Redfoot parser as early as this week, if James and co. are still agreeable.

Any general comments before a discussion of technical issues?

James?


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From Mike.Olson@fourthought.com  Thu Apr 19 16:12:35 2001
From: Mike.Olson@fourthought.com (Mike Olson)
Date: Thu, 19 Apr 2001 09:12:35 -0600
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no> <3ADCD338.ACCF9907@FourThought.com> <m37l0i9ow1.fsf@lambda.garshol.priv.no> <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de>
Message-ID: <3ADF0063.4499C271@FourThought.com>

"Martin v. Loewis" wrote:
> 
> > Good point. There does need to be something along the lines of the
> > make_parser function in SAX.
> 
> My proposal would be to add two keyword arguments, properties= and
> features=. Each is a list of binary tuples, each tuple has name and
> value. Alternatively, dictionaries might be better.
> 
> make_parser will the iterate over all known parser factories, invoking
> create_parser for each, then trying to set all the properties and
> features. It will return the first parser that supports all of them,
> and return a configured instance.
> 
> There should also be a function xml.sax.register_parser, which accepts
> an object that has a create_parser function, or a string naming a
> module that has a create_parser function.
> 
> What do you think?


Can you still specify a parser to avoid the iteration?  What about
setting a default?

Mike
> 
> Regards,
> Martin
> 
> _______________________________________________
> XML-SIG maillist  -  XML-SIG@python.org
> http://mail.python.org/mailman/listinfo/xml-sig

-- 
Mike Olson				 Principal Consultant
mike.olson@fourthought.com               (303)583-9900 x 102
Fourthought, Inc.                         http://Fourthought.com 
Software-engineering, knowledge-management, XML, CORBA, Linux, Python


From Mike.Olson@fourthought.com  Thu Apr 19 16:19:35 2001
From: Mike.Olson@fourthought.com (Mike Olson)
Date: Thu, 19 Apr 2001 09:19:35 -0600
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
References: <200104190406.f3J46Ym01936@borgia.local>
Message-ID: <3ADF0207.FB270CF6@FourThought.com>

Uche Ogbuji wrote:
> 
> 
> I like.  But I think we'll want more than just a parser factory:
> 
> properties = {"http://factory.pyxml.org/properties/encoding": "BIG5",
>               "http://uche.ogbuji.net/my-extended-properties/spam": "eggs"}
> features = ["http://factory.pyxml.org/properties/xpointer"]
> xml.factory.getXsltProcessor(properties, features)


Would there be an idea of precedence, what if two parsers meet all of
the criteria which one would we pick?

Mike

> 
> --
> Uche Ogbuji                               Principal Consultant
> uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
> Fourthought, Inc.                         http://Fourthought.com
> 4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
> Software-engineering, knowledge-management, XML, CORBA, Linux, Python
> 
> _______________________________________________
> XML-SIG maillist  -  XML-SIG@python.org
> http://mail.python.org/mailman/listinfo/xml-sig

-- 
Mike Olson				 Principal Consultant
mike.olson@fourthought.com               (303)583-9900 x 102
Fourthought, Inc.                         http://Fourthought.com 
Software-engineering, knowledge-management, XML, CORBA, Linux, Python


From rsalz@zolera.com  Thu Apr 19 17:39:57 2001
From: rsalz@zolera.com (Rich Salz)
Date: Thu, 19 Apr 2001 12:39:57 -0400
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
References: <200104190406.f3J46Ym01936@borgia.local> <3ADF0207.FB270CF6@FourThought.com>
Message-ID: <3ADF14DD.28DC8F68@zolera.com>

> Would there be an idea of precedence, what if two parsers meet all of
> the criteria which one would we pick?

the one with the most lines of code?
As in, "my python's bigger than yours" :)


From Olivier.Cayrol@logilab.fr  Thu Apr 19 19:44:05 2001
From: Olivier.Cayrol@logilab.fr (Olivier CAYROL (Logilab))
Date: Thu, 19 Apr 2001 20:44:05 +0200 (CEST)
Subject: [XML-SIG] [4XSLT] bug report and patch for complex XML and XSL nesting
Message-ID: <Pine.LNX.4.21.0104192040590.3976-101000@leo.logilab.fr>

  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.
  Send mail to mime@docserver.cac.washington.edu for more info.

---1463794431-39797830-987705845=:3976
Content-Type: TEXT/PLAIN; charset=ISO-8859-1
Content-Transfer-Encoding: QUOTED-PRINTABLE


  Hello,

  I found a vicious bug in 4XSLT (hidden very deeply in the code).=20
Attached to this message you will find a tar.gz file containing a=20
directory tree that exhibits the bug. It is a little application for=20
managing Easter rabbits and eggs distribution (!). There is an XML file=20
that contains the data: easter_mng.xml, an XSL Transformation file:=20
xsl/transf.xsl and XML files containing data for localization:=20
lib/common.xml, lib/en.xml, lib/fr.xml.

  The lib/common.xml file is imported in the XSLT stylesheets with the=20
'document()' function and is used to insert language-dependant tags in=20
the output. This common.xml file imports other XML files (one per=20
language) with the classic external ENTITY mechanism of XML.

  When trying to transform the data file from the main directory with=20
the following line command:=20
    4xslt -Dlang=3Den  easter_mng.xml xsl/transf.xsl
, I got this exception:
    ...
      File "/usr/lib/python1.5/site-packages/xml/xslt/XsltFunctions.py",=20
    line 63, in Document
        doc =3D context.stylesheet._docReader.fromUri(uri, baseUri=3DbaseUr=
i)
      File "/usr/lib/python1.5/site-packages/Ft/Lib/ReaderBase.py", line=20
    67, in fromUri
        rt =3D self.fromStream(stream, baseUri, ownerDoc, stripElements)
      File "/usr/lib/python1.5/site-packages/Ft/Lib/pDomlette.py", line 5
    78, in fromStream
        raise FtException(Error.XML_PARSE_ERROR, p.ErrorLineNumber, p.Err
    orColumnNumber, expat.ErrorString(p.ErrorCode))
      Ft.Lib.FtException: ('XML parse error at line 16, column 2: error i
    n processing external entity reference', (16, 2, 'error in processing
     external entity reference'))

  In fact, there is a problem when 4XSLT reads the XML document=20
referenced in the 'document()' function: this XML file contains ENTITYs=20
that import XML tree parts by giving local paths from the current=20
document directory whereas in 4XSLT, the baseUri is always the URI of=20
the initial XSLT. The XML reader is unable to find the external entities
and the bug appears.

  Replacing line 67 of Ft.Lib.ReaderBase.py in DomletteReader.fromUri
function:
    rt =3D self.fromStream(stream, baseUri, ownerDoc, stripElements)
with :
    newBaseUri =3D urllib.basejoin(baseUri, uri)
    rt =3D self.fromStream(stream, newBaseUri, ownerDoc, stripElements)
fixes the bug.

  I initially found the bug while trying to process Norman Walsh's XSLT
stylesheets for turning docbook files in XSL formatting objects files (I
am unfortunately not working for the Easter Rabbit).

  Regards,

    O. CAYROL.
_________________________________________________________________________
Olivier CAYROL                                   LOGILAB - Paris (France)
                                                 http://www.logilab.com/
Change your millenium, try NARVAL the Intelligent Personal Assistant.
Changez de mill=E9naire, essayez NARVAL l'Assistant Personnel Intelligent.
_________________________________________________________________________

---1463794431-39797830-987705845=:3976
Content-Type: APPLICATION/x-gzip; name="easter_mng.tar.gz"
Content-Transfer-Encoding: BASE64
Content-ID: <Pine.LNX.4.21.0104192044050.3976@leo.logilab.fr>
Content-Description: bug exhibitor
Content-Disposition: attachment; filename="easter_mng.tar.gz"
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---1463794431-39797830-987705845=:3976--


From larsga@garshol.priv.no  Thu Apr 19 20:21:45 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 19 Apr 2001 21:21:45 +0200
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: <200104191549.f3JFn7x04884@borgia.local.dhcp.fourthought.com>
References: <200104191549.f3JFn7x04884@borgia.local.dhcp.fourthought.com>
Message-ID: <m3pue8pueu.fsf@lambda.garshol.priv.no>

* Uche Ogbuji
|
| James Tauber and I met at XML DevCon and had a good chat.  We agreed
| that it would be great to get Redfoot's RDF parser into PyXML as a
| basis for RDF support.  Then both RedFoot and 4RDF could use this as
| the parser, and build their unique additional functions on top of
| it.

I would like to see RDF and other things that are not core XML
functionality kept outside the PyXML package. A PyRDF package might be
a better solution for both PyXML and all those interested in RDF
support in Python. SourceForge is there, so it should be easy to set up.

--Lars M.



From martin@loewis.home.cs.tu-berlin.de  Thu Apr 19 22:06:58 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 19 Apr 2001 23:06:58 +0200
Subject: [XML-SIG] Re: SAX parser factories (Was: PyTRaX?)
In-Reply-To: <m3lmox8hqy.fsf@lambda.garshol.priv.no> (message from Lars Marius
 Garshol on 19 Apr 2001 09:35:01 +0200)
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no> <3ADCD338.ACCF9907@FourThought.com> <m37l0i9ow1.fsf@lambda.garshol.priv.no> <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de> <m3lmox8hqy.fsf@lambda.garshol.priv.no>
Message-ID: <200104192106.f3JL6ws00987@mira.informatik.hu-berlin.de>

> But is this a proposal for a PyTRAX processor factory or for a SAX
> parser factory? If the latter, do you think this should go into core
> Python?

The latter, so the API should be in core Python (even though it will
only appear in 2.2). I think Python will then still support the
expatreader only, unless additional parsers have registered
themselves.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Thu Apr 19 22:20:30 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 19 Apr 2001 23:20:30 +0200
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <004a01c0c8d2$241f6ae0$7cac1218@reston1.va.home.com>
 (tpassin@home.com)
References: <200104190406.f3J46Ym01936@borgia.local> <200104190601.f3J61XW01112@mira.informatik.hu-berlin.de> <004a01c0c8d2$241f6ae0$7cac1218@reston1.va.home.com>
Message-ID: <200104192120.f3JLKUo01190@mira.informatik.hu-berlin.de>

> I don't like make_parser() and create_parser() approach because the names
> are so similar as to lead to confusion.  I do like the proposal to ask for a
> list of processors that support a feature list, which ability wuld be
> separate from actual parser creation.  I say a list because if more than one
> had my features, I might want to choose one rather than another, not just
> take whichever one the system wanted to give me.

So what would be your proposed API, given that xml.sax.make_parser
already exists (but only allows to specify parser names)?

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Thu Apr 19 22:28:55 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 19 Apr 2001 23:28:55 +0200
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <15070.63712.948366.946492@cj42289-a.reston1.va.home.com>
 (fdrake@acm.org)
References: <200104190406.f3J46Ym01936@borgia.local>
 <200104190601.f3J61XW01112@mira.informatik.hu-berlin.de>
 <004a01c0c8d2$241f6ae0$7cac1218@reston1.va.home.com>
 <200104190446.f3J4kAK02141@borgia.local>
 <200104190606.f3J66Jn01217@mira.informatik.hu-berlin.de>
 <004201c0c8d1$0294cc40$7cac1218@reston1.va.home.com> <15070.63712.948366.946492@cj42289-a.reston1.va.home.com>
Message-ID: <200104192128.f3JLStl01204@mira.informatik.hu-berlin.de>

[...]
>             def create():
>                 """Return a new XMLReader instance."""
> 
>         def create_parser(features={}, properties={}):
>             """Return a configured parser object that supports the
>             given feature and property settings.  If more than one
>             SAXImplementation supports the given settings, one will be
>             selected arbitrarily."""

I'd prefer to extend on existing API, where the SAXImplementation
currently has a create_parser, and xml.sax currently has make_parser.

Apart from that, this proposal requires quite some API from an
individual parser module, so I'd rather see this implemented for, say,
expatreader, before changing the infrastructure to use the API. Even
writing a Python PEP might be appropriate.

Also, the question will be whether we provide that infrastructure only
for SAX2 parsers (of which we two at the moment), or also for the SAX1
drivers, which don't have the notion of features and properties at all
at the moment.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Thu Apr 19 22:49:12 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 19 Apr 2001 23:49:12 +0200
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <3ADF0063.4499C271@FourThought.com> (message from Mike Olson on
 Thu, 19 Apr 2001 09:12:35 -0600)
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no> <3ADCD338.ACCF9907@FourThought.com> <m37l0i9ow1.fsf@lambda.garshol.priv.no> <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de> <3ADF0063.4499C271@FourThought.com>
Message-ID: <200104192149.f3JLnCU01364@mira.informatik.hu-berlin.de>

> Can you still specify a parser to avoid the iteration?  

Certainly.

> What about setting a default?

I think PY_SAX_PARSER needs continued support, if for no other reason
than backwards compatibility. The question is how it would integrate
with properties requested by the application. There is no issue if the
application did not request any features, or if the PY_SAX_PARSER(s)
support the requested features.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Thu Apr 19 22:52:04 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 19 Apr 2001 23:52:04 +0200
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <3ADF0207.FB270CF6@FourThought.com> (message from Mike Olson on
 Thu, 19 Apr 2001 09:19:35 -0600)
References: <200104190406.f3J46Ym01936@borgia.local> <3ADF0207.FB270CF6@FourThought.com>
Message-ID: <200104192152.f3JLq4l01366@mira.informatik.hu-berlin.de>

> Would there be an idea of precedence, what if two parsers meet all of
> the criteria which one would we pick?

For SAX, I think the registry ought to have an ordered list. For the
parsers known to PyXML, this list should be sorted roughly by expected
resource consumption (both time and space).

One proposal is to return the list of all matching parsers, which
would off-load selection to the application (not that the application
would be in any better position to make a choice - unless it can
off-load the choice to an experienced user).

Regards,
Martin


From uche.ogbuji@fourthought.com  Fri Apr 20 03:59:39 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Thu, 19 Apr 2001 20:59:39 -0600
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: Message from Lars Marius Garshol <larsga@garshol.priv.no>
 of "19 Apr 2001 21:21:45 +0200." <m3pue8pueu.fsf@lambda.garshol.priv.no>
Message-ID: <200104200300.f3K2xdZ07783@borgia.local>

> 
> * Uche Ogbuji
> |
> | James Tauber and I met at XML DevCon and had a good chat.  We agreed
> | that it would be great to get Redfoot's RDF parser into PyXML as a
> | basis for RDF support.  Then both RedFoot and 4RDF could use this as
> | the parser, and build their unique additional functions on top of
> | it.
> 
> I would like to see RDF and other things that are not core XML
> functionality kept outside the PyXML package. A PyRDF package might be
> a better solution for both PyXML and all those interested in RDF
> support in Python. SourceForge is there, so it should be easy to set up.

Can you define "core XML"?

My guess is that this is one of those "I know it when I see it" things.  You 
don't use RDF, so it's probably harder for you to see it as core XML.  For me, 
it's hard to see RDF as anything but core XML.

I also think that your stated standard would tend to exclude large parts of 
the current contents of PyXML.

I might be persuaded of your viewpoint, but I'd need to see an argument for 
your viewpoint.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From jtauber@bowstreet.com  Fri Apr 20 05:28:32 2001
From: jtauber@bowstreet.com (James Tauber)
Date: Fri, 20 Apr 2001 00:28:32 -0400
Subject: [XML-SIG] RDF Parser -> PyXML
Message-ID: <C4EE90263CBFD411A0C800B0D0490095CCC681@bst-mail02>

I can certainly make it available. It was written from the start to be
usable apart from Redfoot so it won't be any problem.

There's no support for containers yet but that shouldn't hold things up.

James

> -----Original Message-----
> From: Uche Ogbuji [mailto:uche.ogbuji@fourthought.com]
> Sent: Thursday, April 19, 2001 11:49 AM
> To: xml-sig@python.org
> Subject: [XML-SIG] RDF Parser -> PyXML
> 
> 
> James Tauber and I met at XML DevCon and had a good chat.  We 
> agreed that it 
> would be great to get Redfoot's RDF parser into PyXML as a 
> basis for RDF 
> support.  Then both RedFoot and 4RDF could use this as the 
> parser, and build 
> their unique additional functions on top of it.
> 
> I really liek this idea, especially since I'm quite 
> dissatisfied with the RDF 
> parser in 4RDF and would like to have an alternative.  the 
> quickest way to get 
> this would be if there were one in PyXML that we could use.
> 
> So I'd like to wind up the discussion here so that we can 
> maybe check in the 
> Redfoot parser as early as this week, if James and co. are 
> still agreeable.
> 
> Any general comments before a discussion of technical issues?
> 
> James?
> 
> 
> -- 
> Uche Ogbuji                               Principal Consultant
> uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
> Fourthought, Inc.                         http://Fourthought.com 
> 4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
> Software-engineering, knowledge-management, XML, CORBA, Linux, Python
> 
> 
> 
> _______________________________________________
> XML-SIG maillist  -  XML-SIG@python.org
> http://mail.python.org/mailman/listinfo/xml-sig
> 


From Mike.Olson@fourthought.com  Fri Apr 20 04:52:42 2001
From: Mike.Olson@fourthought.com (Mike Olson)
Date: Thu, 19 Apr 2001 21:52:42 -0600
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no> <3ADCD338.ACCF9907@FourThought.com> <m37l0i9ow1.fsf@lambda.garshol.priv.no> <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de> <3ADF0063.4499C271@FourThought.com> <200104192149.f3JLnCU01364@mira.informatik.hu-berlin.de>
Message-ID: <3ADFB28A.B3337C33@FourThought.com>

"Martin v. Loewis" wrote:
> 
> > Can you still specify a parser to avoid the iteration?
> 
> Certainly.
> 
> > What about setting a default?
> 
> I think PY_SAX_PARSER needs continued support, if for no other reason
> than backwards compatibility. The question is how it would integrate
> with properties requested by the application. There is no issue if the
> application did not request any features, or if the PY_SAX_PARSER(s)
> support the requested features.


I guess what I am worried about is making it to complex/slow for an
application that knows what parser/processor
it wants to use.

Mike

> 
> Regards,
> Martin

-- 
Mike Olson				 Principal Consultant
mike.olson@fourthought.com               (303)583-9900 x 102
Fourthought, Inc.                         http://Fourthought.com 
Software-engineering, knowledge-management, XML, CORBA, Linux, Python


From tpassin@home.com  Fri Apr 20 06:06:17 2001
From: tpassin@home.com (Thomas B. Passin)
Date: Fri, 20 Apr 2001 01:06:17 -0400
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
References: <200104190406.f3J46Ym01936@borgia.local><200104190601.f3J61XW01112@mira.informatik.hu-berlin.de><004a01c0c8d2$241f6ae0$7cac1218@reston1.va.home.com><200104190446.f3J4kAK02141@borgia.local><200104190606.f3J66Jn01217@mira.informatik.hu-berlin.de><004201c0c8d1$0294cc40$7cac1218@reston1.va.home.com> <15070.63712.948366.946492@cj42289-a.reston1.va.home.com>
Message-ID: <003201c0c957$a2181980$7cac1218@reston1.va.home.com>

Fred L. Drake said -
>
> Thomas B. Passin writes:
>  > I favor short (non-URI) feature names.  Using URIs to get unique
strings may
>  > work well when you have the possibility of many people working
independently
>  > on many projects accidentally producing name collisions.  I don't thnk
that
>  > applies here.  Even if, say, the RDF people duplicated some of our
feature
>  > names, it wouldn't matter since they would be used in a different
context.
>
>   Why would it not apply here?  I know the "Parsed XML" product for
> Zope supports some additional features that can be checked with
> hasFeature(); why would this not be the case for other APIs which can
> support customizable features?

Well, XML namespaces envision a situation where different element of the
same name might get used in the same document.  With our features,  if I ask
for feature X on an RDF processor, it can't be confused with feature X on an
xslt processsor or feature X on a validating XML parser.  So there is no
need to disambiguate, I thought.  If there were likely to be a lot of
uncoordinated efforts all adding features to python processors, I'd see it
differently.  But I'm not hard over about it.

>   If we do go with short names, we should at least strongly recommend
> a way to formulate the names for additional features.  I'd stick with
> what's recommended for the DOM in this case; for example:
> "org.zope.dom.persistence".
>

Why, do you see a unification with java processor APIs in the future?  If
we're going to go to dotted names, let's just use URIs and be done with it.
If we use the dotted name method, would it be tied to the current package
structure?  I don;t favor that becaues what should be done if the packages
are refactored?  I favor making up a URN-like prefix and notation even if we
never register it anywhere.  Otherwise, a (possibly fake) url as others have
suggested.  If we use a url, we could consider pointing it to a RDDL-like
document that could contain machine and human readible information on the
features.

Cheers,

Tom P




From tpassin@home.com  Fri Apr 20 06:27:42 2001
From: tpassin@home.com (Thomas B. Passin)
Date: Fri, 20 Apr 2001 01:27:42 -0400
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
References: <200104190406.f3J46Ym01936@borgia.local><200104190601.f3J61XW01112@mira.informatik.hu-berlin.de><004a01c0c8d2$241f6ae0$7cac1218@reston1.va.home.com><200104190446.f3J4kAK02141@borgia.local><200104190606.f3J66Jn01217@mira.informatik.hu-berlin.de><004201c0c8d1$0294cc40$7cac1218@reston1.va.home.com> <15070.63712.948366.946492@cj42289-a.reston1.va.home.com>
Message-ID: <003701c0c95a$9fae1700$7cac1218@reston1.va.home.com>

I was thinking of something like this (ignoring any distinction between
properties and features):

def get_processor_list(featurelist):
    '''Return a list of parsers (by name or registered ID, these would not
be instances)
    that support the requested features.  Each item of the
    list can be instantiated by a call to create_named_parser().'''

def create_named_processor(parsername, featurelist):
    '''Instantiate a parser using a parser name as returned by
get_parser_list.'''

These would be in addition to Fred's suggestions (copied below), except that
neither of these methods should belong to a specific processor
implementation (since they look across all of them). I'm not yet quite clear
on what class they should be part of.   With this approach, you can take the
first processor on the list if you don't care which one you use:

processors=get_processor_list(featurelist)
if processors:
    parser=create_named_processor(processors[0],featurelist)
else:
   bye_bye_right_now()

Or if you know you a particular one, use it by name if it shows up in the
list.  This api should work nicely for other types of processors in the
future, so it would give us a uniform approach to wrapping processors that
may have various features.

Cheers,

Tom P

Fred L. Drake wrote -
>
>
>   So perhaps what we have is an interface like this:
>
> object SAXImplementation:
>             def can_set_feature(feature, enabled):
>                 """Return true if the parsers returned by create()
>                 can support 'feature', and false if they don't."""
>
>             def can_support_property(property, value):
>                 """Return true if the parsers returned by create()
>                 can support 'property' with the given value, and
>                 false if they can't."""
>
>             def get_features_list():
>                 """Return a list of supported feature names.
>                 Inclusion of a name in the list does not imply that
>                 the feature is supported in both enabled and disabled
>                 forms."""
>
>             def get_properties_list():
>                 """Return a list of supported property names.
>                 Inclusion of a name in the list does not imply that
>                 all values for that property are supported."""
>
>             def create():
>                 """Return a new XMLReader instance."""
>
>         def find_parsers(features={}, properties={}):
>             """Return a list of SAXImplementation objects that support
>             the given features and properties."""
>
>         def create_parser(features={}, properties={}):
>             """Return a configured parser object that supports the
>             given feature and property settings.  If more than one
>             SAXImplementation supports the given settings, one will be
>             selected arbitrarily."""
>
>         def register_parser(impl):
>             """Add the SAXImplementation 'impl' to the set of parsers
>             known to the factory."""
>
> (I expect SAXImplementation objects will usually be modules.)
>
>  > Also, with the has_features() approach, it would be possible to keep a
>  > simple features catalog which would avoid the need to instantiate a
>  > processor so it could be asked if it had some feature.  If each
processor
>  > (or wrapper) could respond to a request for its features, a script
could
>  > automatically query it when it was first registered and save the data
in the
>
>   This would be fine with me.  I think the SAXImplementation interface
> I outlined above would support this implementation of the factory
> functions.
>
>  > catalog.  Perhaps the catalog could be in xml, more likely a dictionary
>  > format.
>
>   I expect such a catalog would have to be a volatile data structure
> rather than something persistent -- the interaction with sys.path
> would be a nightmare for a persistent catalog!
>
>  > The next step in this evolution could be named feature sets!
>
>   That would be nice to have!
>
>
>   -Fred
>
> --
> Fred L. Drake, Jr.  <fdrake at acm.org>
> PythonLabs at Digital Creations
>
>



From larsga@garshol.priv.no  Fri Apr 20 09:47:27 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 20 Apr 2001 10:47:27 +0200
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <004a01c0c8d2$241f6ae0$7cac1218@reston1.va.home.com>
References: <200104190406.f3J46Ym01936@borgia.local> <200104190601.f3J61XW01112@mira.informatik.hu-berlin.de> <004a01c0c8d2$241f6ae0$7cac1218@reston1.va.home.com>
Message-ID: <m37l0got40.fsf@lambda.garshol.priv.no>

* Thomas B. Passin
|
| I don't like make_parser() and create_parser() approach because the names
| are so similar as to lead to confusion. 

That's a good point, but only make_parser() is externally visible. The
only people who need to care about create_parser() is those making SAX
drivers that they want to make available to the make_parser() factory.

So far that has only been the XML-SIG, and I think we could change
that without too much pain for anyone, if we find it worthwhile.

| I do like the proposal to ask for a list of processors that support
| a feature list, which ability wuld be separate from actual parser
| creation.  I say a list because if more than one had my features, I
| might want to choose one rather than another, not just take
| whichever one the system wanted to give me.
 
Also a good point. So perhaps this should be a different function,
then. It also seems that we should have a parser name property to make
it easier to see which parsers have been returned.

| Also, with the has_features() approach, it would be possible to keep
| a simple features catalog which would avoid the need to instantiate
| a processor so it could be asked if it had some feature.

I think it's worth the trouble to instantiate a processor, since that
is the only reliable way to check if the processor actually has all
the modules it needs to work.

--Lars M.



From larsga@garshol.priv.no  Fri Apr 20 09:48:43 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 20 Apr 2001 10:48:43 +0200
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <3ADFB28A.B3337C33@FourThought.com>
References: <m3lmp1hyrz.fsf@lambda.garshol.priv.no> <3ADCD338.ACCF9907@FourThought.com> <m37l0i9ow1.fsf@lambda.garshol.priv.no> <200104181942.f3IJg9d00877@mira.informatik.hu-berlin.de> <3ADF0063.4499C271@FourThought.com> <200104192149.f3JLnCU01364@mira.informatik.hu-berlin.de> <3ADFB28A.B3337C33@FourThought.com>
Message-ID: <m366g0ot1w.fsf@lambda.garshol.priv.no>

* Mike Olson
| 
| I guess what I am worried about is making it to complex/slow for an
| application that knows what parser/processor it wants to use.

That application will simply give make_parser() the name of the parser
and have it created and returned without further ado. That's how it
works now and that needs to be kept for backwards compatibility. It
also makes sense to keep it.

--Lars M.



From larsga@garshol.priv.no  Fri Apr 20 09:49:29 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 20 Apr 2001 10:49:29 +0200
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: <200104192152.f3JLq4l01366@mira.informatik.hu-berlin.de>
References: <200104190406.f3J46Ym01936@borgia.local> <3ADF0207.FB270CF6@FourThought.com> <200104192152.f3JLq4l01366@mira.informatik.hu-berlin.de>
Message-ID: <m34rvkot0m.fsf@lambda.garshol.priv.no>

* Martin v. Loewis
| 
| For SAX, I think the registry ought to have an ordered list. For the
| parsers known to PyXML, this list should be sorted roughly by expected
| resource consumption (both time and space).

Agreed.
 
| One proposal is to return the list of all matching parsers, which
| would off-load selection to the application (not that the application
| would be in any better position to make a choice - unless it can
| off-load the choice to an experienced user).

I think this is worth considering, at least. 

--Lars M.



From larsga@garshol.priv.no  Fri Apr 20 09:56:37 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 20 Apr 2001 10:56:37 +0200
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: <200104200300.f3K2xdZ07783@borgia.local>
References: <200104200300.f3K2xdZ07783@borgia.local>
Message-ID: <m33db4osoq.fsf@lambda.garshol.priv.no>

* Uche Ogbuji
| 
| Can you define "core XML"?

Well, it would have to be things that 

 - use or support the XML data model (RDF fails here)
 - are not XML applications themselves (whether RDF fails here can be argued)
 - help you build XML applications and software

Catalogs, TREX, XSLT, XPath and so on all meet these criteria while
RDF does not. RDF has its own parsers, its own data model, its own
schema language, its own databases and its own everything. 

To me these are strong arguments for why a separate package of RDF
functionality makes more sense than to provide parsers, object model
implementations and validators for two different data models in the
same package.

| You don't use RDF, so it's probably harder for you to see it as core
| XML.  For me, it's hard to see RDF as anything but core XML.

You are right that I don't use RDF and so may be looking at it the
wrong way, but to be honest I don't consider either RDF or topic maps
as core XML technologies or even XML technologies at all. Sure, they
use an XML serialization format, but that is the only connection apart
from that they all are used for information management.

What is, really, the connection between RDF and XML?  (I am asking for
enlightenment here. :)

| I also think that your stated standard would tend to exclude large
| parts of the current contents of PyXML.
 
Which ones?

--Lars M.



From Nicolas.Chauvat@logilab.fr  Fri Apr 20 11:19:04 2001
From: Nicolas.Chauvat@logilab.fr (Nicolas Chauvat)
Date: Fri, 20 Apr 2001 12:19:04 +0200 (CEST)
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: <m33db4osoq.fsf@lambda.garshol.priv.no>
Message-ID: <Pine.LNX.4.21.0104201206310.17297-100000@aries>

> You are right that I don't use RDF and so may be looking at it the
> wrong way, but to be honest I don't consider either RDF or topic maps
> as core XML technologies or even XML technologies at all. Sure, they
> use an XML serialization format, but that is the only connection apart
> from that they all are used for information management.
>=20
> What is, really, the connection between RDF and XML?  (I am asking for
> enlightenment here. :)

I used to think the same and the more I use RDF, the more I consider it as
XML's other half. Ever used references to identifiers (IDs) in XML to
materialize relations between objects? Where did you stuff them... in
attributes or in text nodes? How would you tie two different source
objects to the same target object using the same type of relation... with
two different attributes (same name?) or element nodes, one for each
source object?

I you've ever faced that kind of situation and considered it to be a
problem, then RDF will be your friend.=20

I don't pretend to perfectly answer your question, just to hint at key RDF
advantages. BTW, I don't use topic maps, but what I read about it does not
make me consider it a core XML technology... as opposed to RDF.

--=20
Nicolas Chauvat

http://www.logilab.com - "Mais o=F9 est donc Ornicar ?" - LOGILAB, Paris (F=
rance)



From tpassin@home.com  Fri Apr 20 13:52:49 2001
From: tpassin@home.com (Thomas B. Passin)
Date: Fri, 20 Apr 2001 08:52:49 -0400
Subject: [XML-SIG] RDF Parser -> PyXML
References: <200104200300.f3K2xdZ07783@borgia.local> <m33db4osoq.fsf@lambda.garshol.priv.no>
Message-ID: <000e01c0c998$ce9a3b00$7cac1218@reston1.va.home.com>

One thing to consider is that a separate effort for pyRDF might have trouble
gaining critical mass - enough developers to really keep it going - if it
were not part of a larger activity like PyXML.  Here's what I suggest.  By
all means let's welcome the RDF contributions into the PyXML fold and CVS.
Keep the RDF code in a separate "rdf" package, not the "xml" package.  The
RDF work should have its own co-captain, I would thnk.  I'm sure Andrew has
more than enough to keep him busy as it is!

Then in the future if the time should come to split off a separate PyRDF
SIG, it will be easy to do so and the XML SIG will have become its proud
parent.

In the meantime, thinking about RDF will be like getting a whole new group
of use cases, whiich could lead to improvements in the mainline xml work
too.

Cheers,

Tom P

Lars Marius Garshol wrote -

>  - use or support the XML data model (RDF fails here)
>  - are not XML applications themselves (whether RDF fails here can be
argued)
>  - help you build XML applications and software
>
> Catalogs, TREX, XSLT, XPath and so on all meet these criteria while
> RDF does not. RDF has its own parsers, its own data model, its own
> schema language, its own databases and its own everything.
>
> To me these are strong arguments for why a separate package of RDF
> functionality makes more sense than to provide parsers, object model
> implementations and validators for two different data models in the
> same package.
>




From jennifer@homedr.org  Fri Apr 20 10:11:38 2001
From: jennifer@homedr.org (jennifer@homedr.org)
Date: Fri, 20 Apr 2001 09:11:38
Subject: [XML-SIG] Excellent Real Estate Investment Opportunity!!!!
Message-ID: <731.16182.813432@unknown>

<html>

<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dwindows=
-1252">
<meta name=3D"GENERATOR" content=3D"Microsoft FrontPage 4.0">
<meta name=3D"ProgId" content=3D"FrontPage.Editor.Document">
<title>Excellent Real Estate Investment Opportunites.</title>
</head>

<body>

<p>Remove instructions found at the bottom of this message<br>
<br>
You are reading today's free Real Estate Investment and Financial Newslet=
ter !!!!!<br>
-------------------------------------------------------------------------=
--------------<br>
Today's Feature Investment Ad<br>
-------------------------------------------------------------------------=
-------------<br>
NORTH IDAHO PROPERTY!<br>
Great Price!    $84,000<br>
Post Falls, Idaho<br>
One year old - 960 square foot single family home<br>
2 bedroom - 2 bath<br>
kitchen, dining room, living room<br>
wood deck porch and single car garage<br>
One block from Middle School<br>
Presently rented at $600 per month<br>
Post Falls (pop. 12,000) is located 30 minutes east of Spokane, WA. (pop.=
 200,000) within the&nbsp;<br>
heart of the Inland Empire (pop 500,000).<br>
Pristine lakes and rivers surround Post Falls.<br>
For pictures or more information email inquires to <a href=3D"mailto:jenn=
ifer@homedr.org">Jennifer@homedr.org</a> with the word "NIProperty" in th=
e subject line.<br>
<br>
NOT AN AGENT<br>
<br>
Our Supporters<br>
Get paid to complete online Surveys!  Get up to $100 per survey!!!!!<br>
<a href=3D"http://www.cyberbounty.com/ad?a=3D71&amp;b=3D9999&amp;c=3D2334=
">http://www.cyberbounty.com/ad?a=3D71&amp;b=3D9999&amp;c=3D2334<br>
</a><br>
Free Debt Consolidation - Cut your monthly bills by %50 NOW!!!!<br>
<a href=3D"http://www.tshirtnews.com/x/mags.htm">http://www.tshirtnews.co=
m/x/mags.htm</a><br>
<br>
<br>
Need a loan?  Match your needs with Hundreds of Lenders.&nbsp;<br>
Get approved in minutes.  Apply Now!!!<br>
<a href=3D"http://www.onresponse.com/onR_Ads.asp?a=3D8663&amp;d=3D2006">h=
ttp://www.onresponse.com/onR_Ads.asp?a=3D8663&amp;d=3D2006<br>
</a><br>
<br>
-----Premier Bank Visa Card -----<br>
<br>
Bad credit, No credit, No Problem, Get your credit card?<br>
Apply Now for a First Premier Bank Visa<br>
<a href=3D"http://www.onresponse.com/onR_Ads.asp?a=3D8663&amp;d=3D2006">h=
ttp://www.tshirtnews.com/x/prem.htm<br>
</a><br>
<br>
---- Attention Homeowners!!!------<br>
Put some extra cash in your pocket, consolidate your bills, and lower you=
r bills.&nbsp;<br>
  Online Pre qualification is Free and there's no obligation.  Apply Now =
regardless of your past credit!<br>
<br>
<a href=3D"http://tshirtnews.com/x/spectrum.htm">http://tshirtnews.com/x/=
spectrum.htm<br>
</a><br>
-------------------------------------------------------------------------=
-----------------<br>
<br>
Unsubscribe information=85..<br>
<br>
Ok, ok, ok. You don't want another copy of this e-rag. So, we don't want =
to waste any bits sending it to anyone who does not want it. A friend gav=
e us your email address, thinking you might want to see a free copy. So d=
on't cuss us out. Please send us an email message with the single word UN=
SUBSCRIBE in the message subject line and we will definitely remove your =
from our list. No we don't sell our remove list.<br>
</p>

</body>

</html>


From uche.ogbuji@fourthought.com  Fri Apr 20 14:32:22 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Fri, 20 Apr 2001 07:32:22 -0600
Subject: [XML-SIG] SAX parser factories (Was: PyTRaX?)
In-Reply-To: Message from "Thomas B. Passin" <tpassin@home.com>
 of "Fri, 20 Apr 2001 01:06:17 EDT." <003201c0c957$a2181980$7cac1218@reston1.va.home.com>
Message-ID: <200104201332.f3KDWMZ11051@borgia.local>

> >   If we do go with short names, we should at least strongly recommend
> > a way to formulate the names for additional features.  I'd stick with
> > what's recommended for the DOM in this case; for example:
> > "org.zope.dom.persistence".
> >
> 
> Why, do you see a unification with java processor APIs in the future?  If
> we're going to go to dotted names, let's just use URIs and be done with it.

This is my inclination as well.

> If we use the dotted name method, would it be tied to the current package
> structure?  I don;t favor that becaues what should be done if the packages
> are refactored?  I favor making up a URN-like prefix and notation even if we
> never register it anywhere.  Otherwise, a (possibly fake) url as others have
> suggested.  If we use a url, we could consider pointing it to a RDDL-like
> document that could contain machine and human readible information on the
> features.

Possibly something like that, but RDDL itself would probably be a tad 
heavyweight, unless we come up with a scheme where the processor can look up a 
RDDL document at the base URI that covers a whole family of properties.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From uche.ogbuji@fourthought.com  Fri Apr 20 14:47:47 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Fri, 20 Apr 2001 07:47:47 -0600
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: Message from Lars Marius Garshol <larsga@garshol.priv.no>
 of "20 Apr 2001 10:56:37 +0200." <m33db4osoq.fsf@lambda.garshol.priv.no>
Message-ID: <200104201348.f3KDlme11097@borgia.local>

> 
> * Uche Ogbuji
> | 
> | Can you define "core XML"?
> 
> Well, it would have to be things that 
> 
>  - use or support the XML data model (RDF fails here)
>  - are not XML applications themselves (whether RDF fails here can be argued)
>  - help you build XML applications and software

I'm unclear on "are not XML applications themselves"

I certainly think that RDF falls well into your third category.

I think that the result of this discussion between you and me will be "agree 
to disagree", so I'd like to hear from others.

But let me touch on your other questions.

> What is, really, the connection between RDF and XML?  (I am asking for
> enlightenment here. :)

I think RDF is the most natural tool available for managing XML meta-data.  
Indeed I don't often use XML without using RDF as well.  In this regard RDF is 
as important to me as XSLT.

* Rather than process large XML documents, I tend to break things down to 
small, cohesive mini-docs, and use RDF-encoded relationships to "glue" them 
together.

* When managing large quantities of XML documents, even with varying 
structure, I use RDF for indexing and aggregation.

* RDF itself is my most common XML vocabulary for documents that do a lot of 
description, and represent object relationships.  See

http://www-106.ibm.com/developerworks/library/ws-rdf/index.html?dwzone=ws

For an example of this line of thought as I apply it to the normally XML-only 
WSDL.

> | I also think that your stated standard would tend to exclude large
> | parts of the current contents of PyXML.
>  
> Which ones?

Well, what I had in mind was formed without the benefit of your clarification 
above.

I was thinking things like

unicode
xpath
xslt
marshal
iso8601

By your criteria above, most of these do indeed belong in the package, but so 
then does RDF, IMO.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From akuchlin@mems-exchange.org  Fri Apr 20 17:28:33 2001
From: akuchlin@mems-exchange.org (Andrew Kuchling)
Date: Fri, 20 Apr 2001 12:28:33 -0400
Subject: [XML-SIG] Canonicalizing XML
Message-ID: <E14qdlp-0007P4-00@ute.cnri.reston.va.us>

Has anyone written code for producing XML in Canonical XML format?
(http://www.w3.org/TR/xml-c14n)

--amk





From rsalz@zolera.com  Fri Apr 20 20:19:39 2001
From: rsalz@zolera.com (Rich Salz)
Date: Fri, 20 Apr 2001 15:19:39 -0400
Subject: [XML-SIG] Canonicalizing XML
References: <E14qdlp-0007P4-00@ute.cnri.reston.va.us>
Message-ID: <3AE08BCB.5B8A5045@zolera.com>

I've done some of it, but the code is horrible, and I want to re-do it. 
It uses xpath and 4dom; if someone wants the code to use as a source of
ideas for a real version, let me know.
	/r$


From martin@loewis.home.cs.tu-berlin.de  Fri Apr 20 21:10:07 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Fri, 20 Apr 2001 22:10:07 +0200
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: <200104201348.f3KDlme11097@borgia.local> (message from Uche
 Ogbuji on Fri, 20 Apr 2001 07:47:47 -0600)
References: <200104201348.f3KDlme11097@borgia.local>
Message-ID: <200104202010.f3KKA7E01355@mira.informatik.hu-berlin.de>

> I think that the result of this discussion between you and me will be "agree 
> to disagree", so I'd like to hear from others.

I know nothing about RDF, so I can't comment from an architectural
point of view. However, I think I can represent or understand the
following positions.

1. PyXML user who also uses RDF. I'd certainly appreciate if I had to
   install only a single package, not two - unless the release
   schedule for PyXML would compromise timely delivery of RDF code
   updates.

2. PyXML user who does not use RDF. I would not care if RDF code was
   included in PyXML, as long that does not consume an intolerable
   percentage of the entire package (which would be bad for download
   times and disk consumption).

3. PyXML maintainer and packager. I appreciate contributions from
   whoever is willing to contribute, as long as it is open source, and
   as long there is willingness to maintain the contribution for the
   months (and years) to come. My biggest concern is that the original
   author runs away and leaves me with all the bug reports.

Regards,
Martin


From fdrake@acm.org  Fri Apr 20 21:29:39 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Fri, 20 Apr 2001 16:29:39 -0400 (EDT)
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: <200104202010.f3KKA7E01355@mira.informatik.hu-berlin.de>
References: <200104201348.f3KDlme11097@borgia.local>
 <200104202010.f3KKA7E01355@mira.informatik.hu-berlin.de>
Message-ID: <15072.39987.36359.665251@cj42289-a.reston1.va.home.com>

Martin v. Loewis writes:
 > 1. PyXML user who also uses RDF. I'd certainly appreciate if I had to
 >    install only a single package, not two - unless the release
 >    schedule for PyXML would compromise timely delivery of RDF code
 >    updates.

  Which indicates that the RDF code should be separate, at least until
it has stabilized and has good reception from users.  I don't know
where Redfoot is on the development curve.

 > 2. PyXML user who does not use RDF. I would not care if RDF code was
 >    included in PyXML, as long that does not consume an intolerable
 >    percentage of the entire package (which would be bad for download
 >    times and disk consumption).

  This is Python code; it should never be too long!  ;-)

 > 3. PyXML maintainer and packager. I appreciate contributions from
 >    whoever is willing to contribute, as long as it is open source, and
 >    as long there is willingness to maintain the contribution for the
 >    months (and years) to come. My biggest concern is that the original
 >    author runs away and leaves me with all the bug reports.

  Reinforces my comment on #1 -- don't include until it's stable.
  Perhaps the right thing to do would be to have a separate package,
with an option to consider merging it in later.  I think it would be
good to keep development discussion here -- it *is* a general XML
thing, whether or not all of us use it (we don't all use validation
either, but it's in the package).  The cross-pollination and expansion
of the set of use cases is good.  I don't mind sharing the CVS
repository, either -- CVS supports multiple modules quite well.
  Why don't we set Redfoot up as the "redfoot" CVS module and reserve
the package name xml.rdf for it?  (As in, let it require PyXML and
then install itself to ../site-package/_xmlplus/rdf/.)  If a decision
is made to keep it separate or merge it in later, there are no real
disruptions for user code.  Is this reasonable, or am I off my rocker
again?


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From martin@loewis.home.cs.tu-berlin.de  Fri Apr 20 22:01:56 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Fri, 20 Apr 2001 23:01:56 +0200
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: <m33db4osoq.fsf@lambda.garshol.priv.no> (message from Lars Marius
 Garshol on 20 Apr 2001 10:56:37 +0200)
References: <200104200300.f3K2xdZ07783@borgia.local> <m33db4osoq.fsf@lambda.garshol.priv.no>
Message-ID: <200104202101.f3KL1uY01684@mira.informatik.hu-berlin.de>

> Catalogs, TREX, XSLT, XPath and so on all meet these criteria while
> RDF does not. RDF has its own parsers, its own data model, its own
> schema language, its own databases and its own everything. 

Maybe I'm missing something here, but ... Isn't any RDF description a
well-formed XML document? and don't RDF 'parsers' typically operate on
top of XML parsers? atleast 4RDF seems to operate that way.

That you then perform some processing, further correctness checks and
so on is expected for a specific XML application.

Given that, I'm not so sure I understand what an RDF parser would do
that a plain XML parser with a SAX interface wouldn't.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Fri Apr 20 22:04:57 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Fri, 20 Apr 2001 23:04:57 +0200
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: <000e01c0c998$ce9a3b00$7cac1218@reston1.va.home.com>
 (tpassin@home.com)
References: <200104200300.f3K2xdZ07783@borgia.local> <m33db4osoq.fsf@lambda.garshol.priv.no> <000e01c0c998$ce9a3b00$7cac1218@reston1.va.home.com>
Message-ID: <200104202104.f3KL4vG01741@mira.informatik.hu-berlin.de>

> Then in the future if the time should come to split off a separate PyRDF
> SIG, it will be easy to do so and the XML SIG will have become its proud
> parent.

I think nobody has questioned whether RDF discussion should take place
on xml-sig@python.org; this is certainly the right mailing list (IMO).
Discussion seems to be more about packaging. 

After all, PyXML <> xml-sig.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Fri Apr 20 22:14:09 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Fri, 20 Apr 2001 23:14:09 +0200
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: <15072.39987.36359.665251@cj42289-a.reston1.va.home.com>
 (fdrake@acm.org)
References: <200104201348.f3KDlme11097@borgia.local>
 <200104202010.f3KKA7E01355@mira.informatik.hu-berlin.de> <15072.39987.36359.665251@cj42289-a.reston1.va.home.com>
Message-ID: <200104202114.f3KLE9m01831@mira.informatik.hu-berlin.de>

>   Why don't we set Redfoot up as the "redfoot" CVS module and reserve
> the package name xml.rdf for it?  (As in, let it require PyXML and
> then install itself to ../site-package/_xmlplus/rdf/.)  If a decision
> is made to keep it separate or merge it in later, there are no real
> disruptions for user code.  Is this reasonable, or am I off my rocker
> again?

Sounds good to me. There is plenty of experience of installing things
into the _xmlplus tree (xml for you Python 1.5 users :-), as well as
with multiple modules in CVS (the SF operation with typically a single
module in each repository only is actually a strange setup).

Regards,
Martin



From chris.arndt@web.de  Fri Apr 20 22:25:00 2001
From: chris.arndt@web.de (Christopher Arndt)
Date: Fri, 20 Apr 2001 23:25:00 +0200
Subject: [XML-SIG] xbel demo patch
Message-ID: <3AE0A92C.B2F4B5F1@web.de>

Hi xml'ers!

I have recently toyed with the XBEL demos included in the PyXML
distribution and found they did not work for be because of problems with
german umlauts (and iso-latin characters in general).

I fixed the parsing scripts to convert non-ascii into xml charrefs but
that posed problems in the xbel parsing script, because the CDATA was
not accumulated in the right way accross event handler calls.

So I fixed that to and now the scripts are working for me.

I made a patch against the most recent (0.6.5) distribution. Should I
post it on thsi list (7k bzipped) or send it to someone else?

--

Christopher Arndt                                   [t] +49 6221-303918
Hildastr. 2                                         [c] +49 173-9542751
69115 Heidelberg                                    [e]
chris.arndt@web.de



From fdrake@acm.org  Fri Apr 20 22:31:00 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Fri, 20 Apr 2001 17:31:00 -0400 (EDT)
Subject: [XML-SIG] xbel demo patch
In-Reply-To: <3AE0A92C.B2F4B5F1@web.de>
References: <3AE0A92C.B2F4B5F1@web.de>
Message-ID: <15072.43668.443715.641248@cj42289-a.reston1.va.home.com>

Christopher Arndt writes:
 > I made a patch against the most recent (0.6.5) distribution. Should I
 > post it on thsi list (7k bzipped) or send it to someone else?

  I'd have to dig around to see who wrote the modules in that
directory, but I seem to be taking the lead on further development of
XBEL (whcih I enjoy), so you can send them to me and I'll take a look
at them.
  I wouldn't be surprised if the issue is that xmllib is being used
for parsing rather than Expat.  I'll try and poke around it this
weekend.


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From jtauber@jtauber.com  Sat Apr 21 00:15:59 2001
From: jtauber@jtauber.com (James Tauber)
Date: Fri, 20 Apr 2001 19:15:59 -0400
Subject: [XML-SIG] RDF Parser -> PyXML
Message-ID: <008501c0c9ef$df0dbc80$bd020a0a@EHUD>

> Which indicates that the RDF code should be separate, at least until
> it has stabilized and has good reception from users. I don't know
> where Redfoot is on the development curve.

Redfoot should be considered 1.0 beta at the moment but note: the RDF
support we are talking about is only 5% of what Redfoot is.

> Why don't we set Redfoot up as the "redfoot" CVS module and reserve
> the package name xml.rdf for it? (As in, let it require PyXML and
> then install itself to ../site-package/_xmlplus/rdf/.) If a decision
> is made to keep it separate or merge it in later, there are no real
> disruptions for user code. Is this reasonable, or am I off my rocker
> again?

Using the "redfoot" name for this would not be appropriate for the reason I
give above: RDF parsing / serialization is only a tiny part of Redfoot.
Other parts of Redfoot may be appropriate to donate later on but what we're
talking about now is small.

Personally, I don't care whether it is part of PyXML or not. I just want to:
    - make it easier for Python people to use RDF
    - to encourage RDF people to use Python

James



From jtauber@bowstreet.com  Sat Apr 21 00:14:58 2001
From: jtauber@bowstreet.com (James Tauber)
Date: Fri, 20 Apr 2001 19:14:58 -0400
Subject: [XML-SIG] RDF Parser -> PyXML
Message-ID: <C4EE90263CBFD411A0C800B0D0490095CCC69C@bst-mail02>

> Maybe I'm missing something here, but ... Isn't any RDF description a
> well-formed XML document?

Yes

> and don't RDF 'parsers' typically operate on
> top of XML parsers? atleast 4RDF seems to operate that way.

As does Redfoot's RDF parser.

> That you then perform some processing, further correctness checks and
> so on is expected for a specific XML application.
> 
> Given that, I'm not so sure I understand what an RDF parser would do
> that a plain XML parser with a SAX interface wouldn't.

An RDF parser sits on top of an XML parser and converts RDF's various
serializations into triples. 

James


From uche.ogbuji@fourthought.com  Sat Apr 21 00:18:17 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Fri, 20 Apr 2001 17:18:17 -0600
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: Message from "Fred L. Drake, Jr." <fdrake@acm.org>
 of "Fri, 20 Apr 2001 16:29:39 EDT." <15072.39987.36359.665251@cj42289-a.reston1.va.home.com>
Message-ID: <200104202318.f3KNIHJ14021@borgia.local.dhcp.fourthought.com>

> Martin v. Loewis writes:
>  > 1. PyXML user who also uses RDF. I'd certainly appreciate if I had to
>  >    install only a single package, not two - unless the release
>  >    schedule for PyXML would compromise timely delivery of RDF code
>  >    updates.
> 
>   Which indicates that the RDF code should be separate, at least until
> it has stabilized and has good reception from users.  I don't know
> where Redfoot is on the development curve.

Hmm.  I don't know that we can claim much history of waiting until code is 
mature before we merge it in.

>  > 3. PyXML maintainer and packager. I appreciate contributions from
>  >    whoever is willing to contribute, as long as it is open source, and
>  >    as long there is willingness to maintain the contribution for the
>  >    months (and years) to come. My biggest concern is that the original
>  >    author runs away and leaves me with all the bug reports.
> 
>   Reinforces my comment on #1 -- don't include until it's stable.
>   Perhaps the right thing to do would be to have a separate package,
> with an option to consider merging it in later.  I think it would be
> good to keep development discussion here -- it *is* a general XML
> thing, whether or not all of us use it (we don't all use validation
> either, but it's in the package).  The cross-pollination and expansion
> of the set of use cases is good.  I don't mind sharing the CVS
> repository, either -- CVS supports multiple modules quite well.
>   Why don't we set Redfoot up as the "redfoot" CVS module and reserve
> the package name xml.rdf for it?  (As in, let it require PyXML and
> then install itself to ../site-package/_xmlplus/rdf/.)  If a decision
> is made to keep it separate or merge it in later, there are no real
> disruptions for user code.  Is this reasonable, or am I off my rocker
> again?

Well this is all fair enough, but the only problem is that there is little 
value to it if it doesn't get packaged in.  If it doesn't, there aren't people 
working with it, finding bugs, fixing them, etc., and it won't magically 
become mature sitting in a dungeon.

I think it might help this discussion to note that James's code is really just 
a SAX handler that emits RDF triples in simple Python form.  It's by no means 
a monster.  And it's by no means a full-blown RDF system.  4RDF consists of 
almost 100 Python files, only *one* of which implements a parser.  Checking in 
Redfoot's parser would provide a simple, common root from which the (hopefully 
many) Python RDF implementations can flourish, packaged separately.  Sort of 
like an expat for which many different DOMs have grown.

Of course, my opinion is to go ahead and check it right into PyXML as an 
"xml.rdf" package.  After all, it only deals with the XML serialization of RDF.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From jtauber@bowstreet.com  Sat Apr 21 00:16:24 2001
From: jtauber@bowstreet.com (James Tauber)
Date: Fri, 20 Apr 2001 19:16:24 -0400
Subject: [XML-SIG] RDF Parser -> PyXML
Message-ID: <C4EE90263CBFD411A0C800B0D0490095CCC69D@bst-mail02>

> I think nobody has questioned whether RDF discussion should take place
> on xml-sig@python.org; this is certainly the right mailing list (IMO).
> Discussion seems to be more about packaging. 

I raised the RDF question here because I thought this mailing list was the
right place, regardless of whether the RDF code is part of PyXML or not. So
I'm glad someone agrees that we can at least talk about RDF here.

James


From jtauber@jtauber.com  Sat Apr 21 00:46:13 2001
From: jtauber@jtauber.com (James Tauber)
Date: Fri, 20 Apr 2001 19:46:13 -0400
Subject: [XML-SIG] RDF Parser -> PyXML
References: <200104202318.f3KNIHJ14021@borgia.local.dhcp.fourthought.com>
Message-ID: <00a801c0c9f4$1875cb30$bd020a0a@EHUD>

> I think it might help this discussion to note that James's code is really
just
> a SAX handler that emits RDF triples in simple Python form.

Almost right. It actually uses expat. Which raises the question: should I
change it to use SAX? Any advantage?

I plan on a refactor tonight so the candidate code will be available this
weekend.

James




From uche.ogbuji@fourthought.com  Sat Apr 21 02:46:38 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Fri, 20 Apr 2001 19:46:38 -0600
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: Message from "James Tauber" <jtauber@jtauber.com>
 of "Fri, 20 Apr 2001 19:15:59 EDT." <008501c0c9ef$df0dbc80$bd020a0a@EHUD>
Message-ID: <200104210146.f3L1kca15963@borgia.local>

> > Why don't we set Redfoot up as the "redfoot" CVS module and reserve
> > the package name xml.rdf for it? (As in, let it require PyXML and
> > then install itself to ../site-package/_xmlplus/rdf/.) If a decision
> > is made to keep it separate or merge it in later, there are no real
> > disruptions for user code. Is this reasonable, or am I off my rocker
> > again?
> 
> Using the "redfoot" name for this would not be appropriate for the reason I
> give above: RDF parsing / serialization is only a tiny part of Redfoot.
> Other parts of Redfoot may be appropriate to donate later on but what we're
> talking about now is small.

Agreed.  That's why I had in mind xml.rdf.

> Personally, I don't care whether it is part of PyXML or not. I just want to:
>     - make it easier for Python people to use RDF
>     - to encourage RDF people to use Python

Yes.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From martin@loewis.home.cs.tu-berlin.de  Sat Apr 21 06:42:16 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Sat, 21 Apr 2001 07:42:16 +0200
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: <200104202318.f3KNIHJ14021@borgia.local.dhcp.fourthought.com>
 (message from Uche Ogbuji on Fri, 20 Apr 2001 17:18:17 -0600)
References: <200104202318.f3KNIHJ14021@borgia.local.dhcp.fourthought.com>
Message-ID: <200104210542.f3L5gG501042@mira.informatik.hu-berlin.de>

> Of course, my opinion is to go ahead and check it right into PyXML
> as an "xml.rdf" package.  After all, it only deals with the XML
> serialization of RDF.

We didn't never have such a long discussion whether to check in a
single file before... For a single file, I'm not even sure a package
is appropriate, could be a module xml.rdf as well. But I guess you'd
like to call it xml.rdf.parser, or some such.

So I'd be now in favour of this addition, provided there would be some
indication (e.g. in xml.rdf.__init__) that this is *not* a complete
RDF library, just the parser.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Sat Apr 21 06:46:31 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Sat, 21 Apr 2001 07:46:31 +0200
Subject: [XML-SIG] RDF Parser -> PyXML
In-Reply-To: <00a801c0c9f4$1875cb30$bd020a0a@EHUD> (jtauber@jtauber.com)
References: <200104202318.f3KNIHJ14021@borgia.local.dhcp.fourthought.com> <00a801c0c9f4$1875cb30$bd020a0a@EHUD>
Message-ID: <200104210546.f3L5kVv01066@mira.informatik.hu-berlin.de>

> Almost right. It actually uses expat. Which raises the question: should I
> change it to use SAX? Any advantage?

I don't know whether RDF has the notion of validation, in which case
SAX would be an advantage: you could parse it through xmlproc. Also,
if speed matters, parsing through sgmlop might be a gain over using
expat - although this gain could be easily eaten by the SAX overhead.

IOW, if all validation/error checking happens in the RDF parser, then
using xml.parser.expat is the right choice.

Regards,
Martin


From jtauber@jtauber.com  Sat Apr 21 07:05:50 2001
From: jtauber@jtauber.com (James Tauber)
Date: Sat, 21 Apr 2001 02:05:50 -0400
Subject: [XML-SIG] RDF Parser -> PyXML
References: <200104202318.f3KNIHJ14021@borgia.local.dhcp.fourthought.com> <00a801c0c9f4$1875cb30$bd020a0a@EHUD> <200104210546.f3L5kVv01066@mira.informatik.hu-berlin.de>
Message-ID: <002001c0ca29$20ea4db0$c8020a0a@EHUD>

> IOW, if all validation/error checking happens in the RDF parser, then
> using xml.parser.expat is the right choice.

This is true, so I'll stick with expat for now.

James



From mbennett@ideaeng.com  Sun Apr 22 21:56:32 2001
From: mbennett@ideaeng.com (Mark Bennett)
Date: Sun, 22 Apr 2001 13:56:32 -0700
Subject: [XML-SIG] Can't install PyXML w/ Python 2.1 on NT
Message-ID: <jUsT.aNoTheR.mEsSaGe.iD.9879728568955@ideaeng.com>

Sorry to bother you all.  I did check the archives.

I've done a fresh install of PYthon 2.1
Then I got the .exe for PyXML (I've tried 6.2 and 6.5) - I got the
one for Python 2.0

When the install runs it asks me which installed Python to add it
to, but has blanks and won't let me spec a path, no Browse button.

I see in the March archives that somebody else asked this.  The advice
back then was to move up to 2.0 from 1.x.  I'd prefer not to move
backwards to 2.0 if I can avoid it.

Any ideas?

Thanks,
Mark










From martin@loewis.home.cs.tu-berlin.de  Mon Apr 23 04:03:30 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Mon, 23 Apr 2001 05:03:30 +0200
Subject: [XML-SIG] Can't install PyXML w/ Python 2.1 on NT
In-Reply-To: <jUsT.aNoTheR.mEsSaGe.iD.9879728568955@ideaeng.com> (message from
 Mark Bennett on Sun, 22 Apr 2001 13:56:32 -0700)
References: <jUsT.aNoTheR.mEsSaGe.iD.9879728568955@ideaeng.com>
Message-ID: <200104230303.f3N33U901641@mira.informatik.hu-berlin.de>

> I've done a fresh install of PYthon 2.1
> Then I got the .exe for PyXML (I've tried 6.2 and 6.5) - I got the
> one for Python 2.0
[...]
> Any ideas?

The binary releases are tied through a specific release of Python,
through associating the binary modules with a specific version of
pythonxy.dll. Therefore, you cannot use a binary release created for
Python 2.0 with Python 2.1.

Instead, you need to compile PyXML yourself, which requires Visual
C++.

Regards,
Martin


From sjoerd.mullender@oratrix.com  Mon Apr 23 16:52:35 2001
From: sjoerd.mullender@oratrix.com (Sjoerd Mullender)
Date: Mon, 23 Apr 2001 17:52:35 +0200
Subject: [XML-SIG] Canonicalizing XML
In-Reply-To: Your message of Fri, 20 Apr 2001 12:28:33 -0400.
 <E14qdlp-0007P4-00@ute.cnri.reston.va.us>
References: <E14qdlp-0007P4-00@ute.cnri.reston.va.us>
Message-ID: <20010423155236.71C71301CF7@bireme.oratrix.nl>

------- =_aaaaaaaaaa0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <2357.988041096.1@bireme.oratrix.nl>

I've written a validating XML parser in Python that can produce
Canonical XML.

I'll attach it.  Usage (for getting Canonical XML):

	python fxmllib.py -c file.xml

On Fri, Apr 20 2001 Andrew Kuchling wrote:

> Has anyone written code for producing XML in Canonical XML format?
> (http://www.w3.org/TR/xml-c14n)
> 
> --amk
> 
> 
> 
> 

-- Sjoerd Mullender <sjoerd.mullender@oratrix.com>


------- =_aaaaaaaaaa0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <2357.988041096.2@bireme.oratrix.nl>
Content-Description: Validating XML Parser
Content-Disposition: attachment; filename="fxmllib.py"
Content-Transfer-Encoding: quoted-printable

__version__ =3D "$Id: fxmllib.py,v 1.2 2001/04/20 15:12:49 sjoerd Exp $"

import re, string
import sys                              # need for CanonXMLParser

class Error(Exception):
    """Error class; raised when a syntax error is encountered.
       Instance variables are:
       lineno: line at which error was found;
       offset: offset into data where error was found;
       text: data in which error was found.
       If these values are unknown, they are set to None."""
    lineno =3D offset =3D text =3D filename =3D None
    def __init__(self, *args):
        self.args =3D args
        if len(args) > 1:
            self.lineno =3D args[1]
            if len(args) > 2:
                self.text =3D args[2]
                if len(args) > 3:
                    self.offset =3D args[3]
                    if len(args) > 4:
                        self.filename =3D args[4]

    def __str__(self):
        if self.filename:
            if self.lineno:
                msg =3D '"%s", line %d: ' % (self.filename, self.lineno)
            else:
                msg =3D '"%s": ' % self.filename
        elif self.lineno:
            msg =3D 'line %d: ' % self.lineno
        else:
            msg =3D ''
        return '%sSyntax error: %s' % (msg, self.args[0])

# The character sets below are taken directly from the XML spec.
_BaseChar =3D u'\u0041-\u005A\u0061-\u007A\u00C0-\u00D6\u00D8-\u00F6\u00F8=
-\u00FF' \
            u'\u0100-\u0131\u0134-\u013E\u0141-\u0148\u014A-\u017E' \
            u'\u0180-\u01C3\u01CD-\u01F0\u01F4-\u01F5\u01FA-\u0217' \
            u'\u0250-\u02A8\u02BB-\u02C1\u0386\u0388-\u038A\u038C' \
            u'\u038E-\u03A1\u03A3-\u03CE\u03D0-\u03D6\u03DA\u03DC\u03DE' \=

            u'\u03E0\u03E2-\u03F3\u0401-\u040C\u040E-\u044F\u0451-\u045C' =
\
            u'\u045E-\u0481\u0490-\u04C4\u04C7-\u04C8\u04CB-\u04CC' \
            u'\u04D0-\u04EB\u04EE-\u04F5\u04F8-\u04F9\u0531-\u0556\u0559' =
\
            u'\u0561-\u0586\u05D0-\u05EA\u05F0-\u05F2\u0621-\u063A' \
            u'\u0641-\u064A\u0671-\u06B7\u06BA-\u06BE\u06C0-\u06CE' \
            u'\u06D0-\u06D3\u06D5\u06E5-\u06E6\u0905-\u0939\u093D' \
            u'\u0958-\u0961\u0985-\u098C\u098F-\u0990\u0993-\u09A8' \
            u'\u09AA-\u09B0\u09B2\u09B6-\u09B9\u09DC-\u09DD\u09DF-\u09E1' =
\
            u'\u09F0-\u09F1\u0A05-\u0A0A\u0A0F-\u0A10\u0A13-\u0A28' \
            u'\u0A2A-\u0A30\u0A32-\u0A33\u0A35-\u0A36\u0A38-\u0A39' \
            u'\u0A59-\u0A5C\u0A5E\u0A72-\u0A74\u0A85-\u0A8B\u0A8D' \
            u'\u0A8F-\u0A91\u0A93-\u0AA8\u0AAA-\u0AB0\u0AB2-\u0AB3' \
            u'\u0AB5-\u0AB9\u0ABD\u0AE0\u0B05-\u0B0C\u0B0F-\u0B10' \
            u'\u0B13-\u0B28\u0B2A-\u0B30\u0B32-\u0B33\u0B36-\u0B39\u0B3D' =
\
            u'\u0B5C-\u0B5D\u0B5F-\u0B61\u0B85-\u0B8A\u0B8E-\u0B90' \
            u'\u0B92-\u0B95\u0B99-\u0B9A\u0B9C\u0B9E-\u0B9F\u0BA3-\u0BA4' =
\
            u'\u0BA8-\u0BAA\u0BAE-\u0BB5\u0BB7-\u0BB9\u0C05-\u0C0C' \
            u'\u0C0E-\u0C10\u0C12-\u0C28\u0C2A-\u0C33\u0C35-\u0C39' \
            u'\u0C60-\u0C61\u0C85-\u0C8C\u0C8E-\u0C90\u0C92-\u0CA8' \
            u'\u0CAA-\u0CB3\u0CB5-\u0CB9\u0CDE\u0CE0-\u0CE1\u0D05-\u0D0C' =
\
            u'\u0D0E-\u0D10\u0D12-\u0D28\u0D2A-\u0D39\u0D60-\u0D61' \
            u'\u0E01-\u0E2E\u0E30\u0E32-\u0E33\u0E40-\u0E45\u0E81-\u0E82' =
\
            u'\u0E84\u0E87-\u0E88\u0E8A\u0E8D\u0E94-\u0E97\u0E99-\u0E9F' \=

            u'\u0EA1-\u0EA3\u0EA5\u0EA7\u0EAA-\u0EAB\u0EAD-\u0EAE\u0EB0' \=

            u'\u0EB2-\u0EB3\u0EBD\u0EC0-\u0EC4\u0F40-\u0F47\u0F49-\u0F69' =
\
            u'\u10A0-\u10C5\u10D0-\u10F6\u1100\u1102-\u1103\u1105-\u1107' =
\
            u'\u1109\u110B-\u110C\u110E-\u1112\u113C\u113E\u1140\u114C' \
            u'\u114E\u1150\u1154-\u1155\u1159\u115F-\u1161\u1163\u1165' \
            u'\u1167\u1169\u116D-\u116E\u1172-\u1173\u1175\u119E\u11A8' \
            u'\u11AB\u11AE-\u11AF\u11B7-\u11B8\u11BA\u11BC-\u11C2\u11EB' \=

            u'\u11F0\u11F9\u1E00-\u1E9B\u1EA0-\u1EF9\u1F00-\u1F15' \
            u'\u1F18-\u1F1D\u1F20-\u1F45\u1F48-\u1F4D\u1F50-\u1F57\u1F59' =
\
            u'\u1F5B\u1F5D\u1F5F-\u1F7D\u1F80-\u1FB4\u1FB6-\u1FBC\u1FBE' \=

            u'\u1FC2-\u1FC4\u1FC6-\u1FCC\u1FD0-\u1FD3\u1FD6-\u1FDB' \
            u'\u1FE0-\u1FEC\u1FF2-\u1FF4\u1FF6-\u1FFC\u2126\u212A-\u212B' =
\
            u'\u212E\u2180-\u2182\u3041-\u3094\u30A1-\u30FA\u3105-\u312C' =
\
            u'\uAC00-\uD7A3'
_Ideographic =3D u'\u4E00-\u9FA5\u3007\u3021-\u3029'
_CombiningChar =3D u'\u0300-\u0345\u0360-\u0361\u0483-\u0486\u0591-\u05A1\=
u05A3-\u05B9' \
                 u'\u05BB-\u05BD\u05BF\u05C1-\u05C2\u05C4\u064B-\u0652\u06=
70' \
                 u'\u06D6-\u06DC\u06DD-\u06DF\u06E0-\u06E4\u06E7-\u06E8' \=

                 u'\u06EA-\u06ED\u0901-\u0903\u093C\u093E-\u094C\u094D' \
                 u'\u0951-\u0954\u0962-\u0963\u0981-\u0983\u09BC\u09BE\u09=
BF' \
                 u'\u09C0-\u09C4\u09C7-\u09C8\u09CB-\u09CD\u09D7\u09E2-\u0=
9E3' \
                 u'\u0A02\u0A3C\u0A3E\u0A3F\u0A40-\u0A42\u0A47-\u0A48' \
                 u'\u0A4B-\u0A4D\u0A70-\u0A71\u0A81-\u0A83\u0ABC\u0ABE-\u0=
AC5' \
                 u'\u0AC7-\u0AC9\u0ACB-\u0ACD\u0B01-\u0B03\u0B3C\u0B3E-\u0=
B43' \
                 u'\u0B47-\u0B48\u0B4B-\u0B4D\u0B56-\u0B57\u0B82-\u0B83' \=

                 u'\u0BBE-\u0BC2\u0BC6-\u0BC8\u0BCA-\u0BCD\u0BD7\u0C01-\u0=
C03' \
                 u'\u0C3E-\u0C44\u0C46-\u0C48\u0C4A-\u0C4D\u0C55-\u0C56' \=

                 u'\u0C82-\u0C83\u0CBE-\u0CC4\u0CC6-\u0CC8\u0CCA-\u0CCD' \=

                 u'\u0CD5-\u0CD6\u0D02-\u0D03\u0D3E-\u0D43\u0D46-\u0D48' \=

                 u'\u0D4A-\u0D4D\u0D57\u0E31\u0E34-\u0E3A\u0E47-\u0E4E\u0E=
B1' \
                 u'\u0EB4-\u0EB9\u0EBB-\u0EBC\u0EC8-\u0ECD\u0F18-\u0F19\u0=
F35' \
                 u'\u0F37\u0F39\u0F3E\u0F3F\u0F71-\u0F84\u0F86-\u0F8B' \
                 u'\u0F90-\u0F95\u0F97\u0F99-\u0FAD\u0FB1-\u0FB7\u0FB9' \
                 u'\u20D0-\u20DC\u20E1\u302A-\u302F\u3099\u309A'
_Digit =3D u'\u0030-\u0039\u0660-\u0669\u06F0-\u06F9\u0966-\u096F\u09E6-\u=
09EF' \
         u'\u0A66-\u0A6F\u0AE6-\u0AEF\u0B66-\u0B6F\u0BE7-\u0BEF' \
         u'\u0C66-\u0C6F\u0CE6-\u0CEF\u0D66-\u0D6F\u0E50-\u0E59' \
         u'\u0ED0-\u0ED9\u0F20-\u0F29'
_Extender =3D u'\u00B7\u02D0\u02D1\u0387\u0640\u0E46\u0EC6\u3005\u3031-\u3=
035' \
            u'\u309D-\u309E\u30FC-\u30FE'
_Letter =3D _BaseChar + _Ideographic
_NameChar =3D '-' + _Letter + _Digit + '._:' + _CombiningChar + _Extender

_S =3D '[ \t\r\n]+'                       # white space
_opS =3D '[ \t\r\n]*'                     # optional white space
_Name =3D '['+_Letter+'_:]['+_NameChar+']*' # XML Name
_QStr =3D "(?:'[^']*'|\"[^\"]*\")"        # quoted XML string
_Char =3D u'\x09\x0A\x0D\x20-\uD7FF\uE000-\uFFFD' # legal characters

comment =3D re.compile('<!--(?P<comment>(?:[^-]|-[^-])*)-->')
space =3D re.compile(_S)
interesting =3D re.compile('[&<]')
amp =3D re.compile('&')
name =3D re.compile('^'+_Name+'$')
names =3D re.compile('^'+_Name+'(?:'+_S+_Name+')*$')

ref =3D re.compile('&(?:(?P<name>'+_Name+')|#(?P<char>(?:[0-9]+|x[0-9a-fA-=
F]+)));')
entref =3D re.compile('(?:&#(?P<char>(?:[0-9]+|x[0-9a-fA-F]+))|%(?P<pname>=
'+_Name+'));')
_attrre =3D _S+'(?P<attrname>'+_Name+')'+_opS+'=3D'+_opS+'(?P<attrvalue>'+=
_QStr+')'
attrfind =3D re.compile(_attrre)
starttag =3D re.compile('<(?P<tagname>'+_Name+')(?P<attrs>(?:'+_attrre+')*=
)'+_opS+'(?P<slash>/?)>')
endtag =3D re.compile('</(?P<tagname>'+_Name+')'+_opS+'>')

illegal =3D re.compile(r'\]\]>')
illegal1 =3D re.compile('[^'+_Char+']')

cdata =3D re.compile('<!\\[CDATA\\[(?P<cdata>(?:[^]]|\\](?!\\]>)|\\]\\](?!=
>))*)\\]\\]>')

_SystemLiteral =3D '(?P<syslit>'+_QStr+')'
_PublicLiteral =3D '(?P<publit>"[-\'()+,./:=3D?;!*#@$_%% \n\ra-zA-Z0-9]*"|=
' \
                            "'[-()+,./:=3D?;!*#@$_%% \n\ra-zA-Z0-9]*')"
_ExternalId =3D '(?:SYSTEM|PUBLIC'+_S+_PublicLiteral+')'+_S+_SystemLiteral=

externalid =3D re.compile(_ExternalId)
ndata =3D re.compile(_S+'NDATA'+_S+'(?P<name>'+_Name+')')
doctype =3D re.compile('<!DOCTYPE'+_S+'(?P<docname>'+_Name+')(?:'+_S+_Exte=
rnalId+')?'+_opS+'(?:\\[(?P<data>(?:'+_S+'|%'+_Name+';|'+comment.pattern+'=
|<(?:![^-]|[^!])(?:[^\'">]|\'[^\']*\'|"[^"]*")*>)*)\\]'+_opS+')?>')

xmldecl =3D re.compile('<\?xml'+
                     _S+'version'+_opS+'=3D'+_opS+'(?P<version>'+_QStr+')'=
+
                     '(?:'+_S+'encoding'+_opS+'=3D'+_opS+
                        "(?P<encoding>'[A-Za-z][-A-Za-z0-9._]*'|"
                        '"[A-Za-z][-A-Za-z0-9._]*"))?'
                     '(?:'+_S+'standalone'+_opS+'=3D'+_opS+
                        '(?P<standalone>\'(?:yes|no)\'|"(?:yes|no)"))?'+
                     _opS+'\?>')
textdecl =3D re.compile('<\?xml'
                      '(?:'+_S+'version'+_opS+'=3D'+_opS+'(?P<version>'+_Q=
Str+'))?'+
                      '(?:'+_S+'encoding'+_opS+'=3D'+_opS+
                      "(?P<encoding>'[A-Za-z][-A-Za-z0-9._]*'|"
                      '"[A-Za-z][-A-Za-z0-9._]*"))?'+
                      _opS+'\?>')
pidecl =3D re.compile('<\\?(?![xX][mM][lL][ \t\r\n?])(?P<name>'+_Name+')(?=
:'+_S+'(?P<data>(?:[^?]|\\?(?!>))*))?\\?>')

# XML NAMESPACES
_NCName =3D '['+_Letter+'_]['+'-' + _Letter + _Digit + '._' + _CombiningCh=
ar + _Extender+']*'    # XML Name, minus the ":"
ncname =3D re.compile(_NCName + '$')
qname =3D re.compile('(?:(?P<prefix>' + _NCName + '):)?' # optional prefix=

                   '(?P<local>' + _NCName + ')$')
xmlns =3D re.compile('xmlns(?::(?P<ncname>' + _NCName + '))?$')

# DOCTYPE
_Nmtoken =3D '['+_NameChar+']+'
nmtoken =3D re.compile('^'+_Nmtoken+'$')
nmtokens =3D re.compile('^'+_Nmtoken+'(?:'+_S+_Nmtoken+')*$')
element =3D re.compile('<!ELEMENT'+_S+'(?P<name>'+_Name+')'+_S+r'(?P<conte=
nt>EMPTY|ANY|\()')
dfaelem0 =3D re.compile(_opS+r'(?P<token>\(|'+_Name+')')
dfaelem1 =3D re.compile(_opS+r'(?P<token>[)|,])')
dfaelem2 =3D re.compile(r'(?P<token>[+*?])')
mixedre =3D re.compile(r'\('+_opS+'#PCDATA'+'(('+_opS+r'\|'+_opS+_Name+')*=
'+_opS+r'\)\*|'+_opS+r'\))')
paren =3D re.compile('[()]')
attdef =3D re.compile(_S+'(?P<atname>'+_Name+')'+_S+'(?P<attype>CDATA|ID(?=
:REFS?)?|ENTIT(?:Y|IES)|NMTOKENS?|NOTATION'+_S+r'\((?P<notation>'+_opS+_Na=
me+'(?:'+_opS+r'\|'+_opS+_Name+')*'+_opS+r')\)|\('+_opS+_Nmtoken+'(?:'+_op=
S+r'\|'+_opS+_Nmtoken+')*'+_opS+r'\))'+_S+'(?P<atvalue>#REQUIRED|#IMPLIED|=
(?:#FIXED'+_S+')?(?P<atstring>'+_QStr+'))')
attlist =3D re.compile('<!ATTLIST'+_S+'(?P<elname>'+_Name+')(?P<atdef>(?:'=
+attdef.pattern+')*)'+_opS+'>')
_EntityVal =3D '"(?:[^"&%]|'+ref.pattern+'|%'+_Name+';)*"|' \
             "'(?:[^'&%]|"+ref.pattern+"|%"+_Name+";)*'"
entity =3D re.compile('<!ENTITY'+_S+'(?:%'+_S+'(?P<pname>'+_Name+')'+_S+'(=
?P<pvalue>'+_EntityVal+'|'+_ExternalId+')|(?P<ename>'+_Name+')'+_S+'(?P<va=
lue>'+_EntityVal+'|'+_ExternalId+'(?:'+_S+'NDATA'+_S+_Name+')?))'+_opS+'>'=
)
notation =3D re.compile('<!NOTATION'+_S+'(?P<name>'+_Name+')'+_S+'(?P<valu=
e>SYSTEM'+_S+_SystemLiteral+'|PUBLIC'+_S+_PublicLiteral+'(?:'+_S+_SystemLi=
teral+')?)'+_opS+'>')
peref =3D re.compile('%(?P<name>'+_Name+');')
ignore =3D re.compile(r'<!\[|\]\]>')
bracket =3D re.compile('[<>\'"%]')
conditional =3D re.compile(r'<!\['+_opS+'(?:(?P<inc>INCLUDE)|(?P<ign>IGNOR=
E))'+_opS+r'\[')

class XMLParser:
    """XMLParser([ xmlns ]) -> instance

       XML document parser.
       There is one optional argument:
       xmlns: understand XML Namespaces (default is 1)."""

    def __init__(self, xmlns =3D 1):
        self.__xmlns =3D xmlns            # whether or not to parse namesp=
aces
        self.reset()

    def reset(self):
        """reset()

           Reset parser to pristine state."""
        self.docname =3D None             # The outermost element in the d=
ocument (according to the DTD)
        self.rawdata =3D []
        self.entitydefs =3D {             # & entities defined in DTD (plu=
s the default ones)
            'lt': '&#60;',              # <
            'gt': '&#62;',              # >
            'amp': '&#38;',             # &
            'apos': '&#39;',            # '
            'quot': '&#34;',            # "
            }
        self.pentitydefs =3D {}           # % entities defined in DTD
        self.elems =3D {}                 # elements and their content/att=
rs
        self.baseurl =3D '.'              # base URL for external DTD
        self.ids =3D {}                   # IDs encountered in document
        self.notation =3D {}              # NOTATIONs
        self.doctype =3D None

    def feed(self, data):
        """feed(data)

           Feed data to parser."""
        self.rawdata.append(data)

    def close(self):
        """close()

           End of data, finish up parsing."""
        # Actually, this is where we start parsing.
        data =3D string.join(self.rawdata, '')
        self.rawdata =3D []
        self.parse(data)

    def __parse_textdecl(self, data, document =3D 0):
        # Figure out the encoding of a file by looking at the first
        # few bytes and the <?xml?> tag that may come at the very
        # beginning of the file.
        # This will convert the data to unicode from whatever format
        # it was originally.
        i =3D 0
        if data[:2] =3D=3D '\376\377':
            # UTF-16, big-endian
            enc =3D 'utf-16-be'
            i =3D 2
        elif data[:2] =3D=3D '\377\376':
            # UTF-16, little-endian
            enc =3D 'utf-16-le'
            i =3D 2
        elif data[:4] =3D=3D '\x00\x3C\x00\x3F':
            # UTF-16, big-endian
            enc =3D 'utf-16-be'
        elif data[:4] =3D=3D '\x3C\x00\x3F\x00':
            # UTF-16, little-endian
            enc =3D 'utf-16-le'
        else:
            enc =3D None                  # unknowns as yet
        if enc:
            try:
                data =3D unicode(data[i:], enc)
            except UnicodeError:
                self.__error("data cannot be converted to Unicode", data, =
i, self.baseurl, fatal =3D 1)
            i =3D 0
        # optional XMLDecl
        if document:
            res =3D xmldecl.match(data, i)
        else:
            res =3D textdecl.match(data, i)
        if res is not None:
            if document:
                version, encoding, standalone =3D res.group('version',
                                                          'encoding',
                                                          'standalone')
            else:
                version, encoding =3D res.group('version', 'encoding')
                standalone =3D None
            if version is not None and version[1:-1] !=3D '1.0':
                self.__error('only XML version 1.0 supported', data, res.s=
tart('version'), self.baseurl, fatal =3D 1)
            if encoding:
                encoding =3D encoding[1:-1]
                if enc and enc !=3D encoding.lower() and \
                   enc[:6] !=3D encoding.lower():
                    self.__error("declared encoding doesn't match actual e=
ncoding", data, res.start('encoding'), self.baseurl, fatal =3D 1)
                enc =3D encoding.lower()
            if standalone:
                standalone =3D standalone[1:-1]
##            self.handle_xml(encoding, standalone)
            i =3D res.end(0)
        if enc is None:
            # default is UTF 8
            enc =3D 'utf-8'
        if type(data) is not type(u'a'):
            try:
                data =3D unicode(data[i:], enc)
            except UnicodeError:
                self.__error("data cannot be converted to Unicode", data, =
i, self.baseurl, fatal =3D 1)
        else:
            data =3D data[i:]
        return data

    def __normalize_linefeed(self, data):
        # normalize line endings: first \r\n -> \n, then \r -> \n
        return u'\n'.join(u'\n'.join(data.split(u'\r\n')).split(u'\r'))

    def __normalize_space(self, data):
        # normalize white space: tab, linefeed and carriage return -> spac=
e
        data =3D ' '.join(data.split('\t'))
        data =3D ' '.join(data.split('\n'))
        data =3D u' '.join(data.split('\r'))
        return data

    def parse(self, data):
        """parse(data)

           Parse the data as an XML document."""
        from time import time
        t0 =3D time()
        data =3D self.__parse_textdecl(data, 1)
        data =3D self.__normalize_linefeed(data)
        # (Comment | PI | S)*
        i =3D self.__parse_misc(data, 0)
        # doctypedecl?
        res =3D doctype.match(data, i)
        if res is not None and self.doctype is None:
            docname, publit, syslit, docdata =3D res.group('docname', 'pub=
lit',
                                                        'syslit', 'data')
            self.docname =3D docname
            if publit: publit =3D string.join(string.split(publit[1:-1]))
            if syslit: syslit =3D syslit[1:-1]
            self.handle_doctype(docname, publit, syslit, docdata)
            i =3D res.end(0)
        elif self.doctype:
            # do as if there was a <!DOCTYPE> declaration
            self.handle_doctype(None, '', self.doctype, '')
        else:
            # self.doctype =3D=3D '' or no DOCTYPE
            # ignore DOCTYPE
            self.doctype =3D None
        t1 =3D time()
        # (Comment | PI | S)*
        i =3D self.__parse_misc(data, i)
        # the document itself
        res =3D starttag.match(data, i)
        if res is None:
            self.__error('no elements in document', data, i, self.baseurl,=
 fatal =3D 1)
        i =3D res.end(0)
        tagname, slash =3D res.group('tagname', 'slash')
        if self.docname and tagname !=3D self.docname:
            self.__error('starttag does not match DOCTYPE', data, res.star=
t('tagname'), self.baseurl, fatal =3D 0)
        val =3D self.__parse_attrs(tagname, data, res.start('tagname'), re=
s.span('attrs'), None)
        if val is None:
            return
        nstag, attrs, namespaces =3D val
        self.finish_starttag(nstag, attrs)
        if not slash:
            i =3D self.__parse_content(data, i, tagname, namespaces)
            if i is None:
                return
            if type(i) is type(res):
                res =3D i
            else:
                res =3D endtag.match(data, i)
            if res is None:
                self.__error('end tag missing', data, i, self.baseurl, fat=
al =3D 0)
            elif res.group('tagname') !=3D tagname:
                self.__error("end tag doesn't match start tag", data, res.=
start('tagname'), self.baseurl, fatal =3D 0)
            i =3D res.end(0)
        self.finish_endtag(nstag)
        i =3D self.__parse_misc(data, i)
        if i !=3D len(data):
            self.__error('garbage at end of document', data, i, self.baseu=
rl, fatal =3D 0)
        t2 =3D time()
        return t0, t1, t2

    def __parse_misc(self, data, i):
        # match any number of whitespace, processing instructions and comm=
ents
        matched =3D 1
        while matched:
            matched =3D 0
            res =3D comment.match(data, i)
            if res is not None:
                matched =3D 1
                c0, c1 =3D res.span('comment')
                ires =3D illegal1.search(data, c0, c1)
                if ires is not None:
                    self.__error('illegal characters in comment', data, ir=
es.start(0), self.baseurl, fatal =3D 0)
                self.handle_comment(data[c0:c1])
                i =3D res.end(0)
            res =3D pidecl.match(data, i)
            if res is not None:
                matched =3D 1
                c0, c1 =3D res.span('data')
                ires =3D illegal1.search(data, c0, c1)
                if ires is not None:
                    self.__error('illegal characters in Processing Instruc=
tion', data, ires.start(0), self.baseurl, fatal =3D 0)
                self.handle_proc(res.group('name'), res.group('data') or '=
')
                i =3D res.end(0)
            res =3D space.match(data, i)
            if res is not None:
                matched =3D 1
                i =3D res.end(0)
        return i

    def __update_state(self, dfa, states, tagname):
        # update the list of states in the dfa.  If tagname is None,
        # we're looking for the final state, so return a list of all
        # states reachable using epsilon transitions
        nstates =3D []
        seenstates =3D {}
        while states:
            s =3D states[0]
            seenstates[s] =3D 1
            del states[0]
            if tagname is not None and dfa[s].has_key(tagname):
                nstates =3D dfa[s][tagname][:]
            else:
                for s in dfa[s].get('', []):
                    if not seenstates.has_key(s):
                        states.append(s)
        if tagname is None:
            nstates =3D seenstates.keys()
        states[:] =3D nstates # change in-line

    def __check_dfa(self, dfa, initstate, tagname, data, i):
        states =3D [initstate]
        possibles =3D {}
        seenstates =3D {}
        while states:
            s =3D states[0]
            seenstates[s] =3D 1
            del states[0]
            for tag in dfa[s].keys():
                if tag and possibles.has_key(tag):
                    self.__error("non-deterministic content model for `%s'=
" % tagname, data, i, self.baseurl, fatal =3D 0)
                possibles[tag] =3D 1
            for s in dfa[s].get('', []):
                if not seenstates.has_key(s):
                    states.append(s)

    def __parse_content(self, data, i, ptagname, namespaces, states =3D No=
ne):
        # parse the content of an element (i.e. the string between
        # start tag and end tag)
        datalen =3D len(data)
        if self.elems.has_key(ptagname):
            content, attributes, start, end =3D self.elems[ptagname][:4] #=
 content model
            if states =3D=3D None:
                states =3D [start]
        else:
            content =3D None              # unknown content model
        while i < datalen:
            matched =3D 0
            res =3D interesting.search(data, i)
            if res is None:
                j =3D datalen
            else:
                j =3D res.start(0)
            if j > i:
                res =3D illegal.search(data, i, j)
                if res is not None:
                    self.__error("illegal data content in element `%s'" % =
ptagname, data, i, self.baseurl, fatal =3D 0)
                skip =3D 0
                complain =3D 0
                if content is not None:
                    res =3D space.match(data, i, j)
                    isspace =3D res is not None and res.span(0) =3D=3D (i,=
j)
                    if content =3D=3D 'EMPTY':
                        complain =3D 1
                        skip =3D 1
                    elif not isspace and  type(content) is type([]) and co=
ntent and type(content[0]) is type({}):
                        complain =3D 1
                    if complain:
                        self.__error("no character data allowed in element=
 `%s'" % ptagname, data, i, self.baseurl, fatal =3D 0)
                matched =3D 1
                if not skip:
                    self.handle_data(data[i:j])
                i =3D j
            res =3D starttag.match(data, i)
            if res is not None:
                tagname, slash =3D res.group('tagname', 'slash')
                if content =3D=3D 'EMPTY' or content =3D=3D '#PCDATA':
                    self.__error("empty element `%s' has content" % ptagna=
me, data, res.start(0), self.baseurl, fatal =3D 0)
                elif content =3D=3D 'ANY':
                    # always OK
                    pass
                elif type(content) is type([]) and content and type(conten=
t[0]) is not type({}):
                    # mixed
                    if tagname not in content:
                        self.__error("illegal content in element `%s'" % p=
tagname, data, res.start(0), self.baseurl, fatal =3D 0)
                elif content is not None:
                    self.__update_state(content, states, tagname)
                    if not states:
                        self.__error("illegal content for element `%s'" % =
ptagname, data, i, self.baseurl)
                val =3D self.__parse_attrs(tagname, data, res.start('tagna=
me'), res.span('attrs'), namespaces)
                if val is None:
                    return
                i =3D res.end(0)
                nstag, attrs, subnamespaces =3D val
                self.finish_starttag(nstag, attrs)
                if not slash:
                    i =3D self.__parse_content(data, i, tagname, subnamesp=
aces)
                    if i is None:
                        return
                    if type(i) is type(res):
                        res =3D i
                    else:
                        res =3D endtag.match(data, i)
                    if res is None:
                        self.__error('end tag missing', data, i, self.base=
url, fatal =3D 0)
                    elif res.group('tagname') !=3D tagname:
                        self.__error("end tag doesn't match start tag", da=
ta, res.start('tagname'), self.baseurl, fatal =3D 0)
                    i =3D res.end(0)
                self.finish_endtag(nstag)
                matched =3D 1
            res =3D endtag.match(data, i)
            if res is not None:
                if type(content) is type([]) and content and type(content[=
0]) is type({}):
                    self.__update_state(content, states, None)
                    if end not in states:
                        self.__error("content of element `%s' doesn't matc=
h content model" % ptagname, data, i, self.baseurl, fatal =3D 0)
                return res
            res =3D comment.match(data, i)
            if res is not None:
                c0, c1 =3D res.span('comment')
                ires =3D illegal1.search(data, c0, c1)
                if ires is not None:
                    self.__error('illegal characters in comment', data, ir=
es.start(0), self.baseurl, fatal =3D 0)
                self.handle_comment(data[c0:c1])
                i =3D res.end(0)
                matched =3D 1
            res =3D ref.match(data, i)
            if res is not None:
                name =3D res.group('name')
                if name:
                    if self.entitydefs.has_key(name):
                        sval =3D val =3D self.entitydefs[name]
                        baseurl =3D self.baseurl
                        if type(val) is type(()):
                            if val[2] is not None:
                                apply(self.handle_ndata, val)
                                val =3D None
                            else:
                                val =3D self.__read_pentity(val[0], val[1]=
)
                        if val is not None:
                            del self.entitydefs[name] # to break recursion=

                            n =3D self.__parse_content(val, 0, ptagname, n=
amespaces, states)
                            self.entitydefs[name] =3D sval # restore value=

                        if val is not None:
                            if n is None:
                                self.baseurl =3D baseurl
                                return
                            if type(n) is type(res) or n !=3D len(val):
                                if type(n) is type(res):
                                    n =3D res.start(0)
                                self.__error('misformed entity value', dat=
a, n, self.baseurl, fatal =3D 0)
                        self.baseurl =3D baseurl
                    else:
                        if self.docname:
                            self.__error("unknown entity reference `&%s;' =
in element `%s'" % (name, ptagname), data, i, self.baseurl, fatal =3D 0)
                        self.data =3D data
                        self.offset =3D res.start('name')
                        self.lineno =3D string.count(data, '\n', 0, self.o=
ffset)
                        self.unknown_entityref(name)
                else:
                    str =3D self.__parse_charref(res.group('char'), data, =
res.start(0))
                    if str is None:
                        return
                    self.handle_data(str)
                i =3D res.end(0)
                matched =3D 1
            res =3D pidecl.match(data, i)
            if res is not None:
                matched =3D 1
                c0, c1 =3D res.span('data')
                ires =3D illegal1.search(data, c0, c1)
                if ires is not None:
                    self.__error('illegal characters in Processing Instruc=
tion', data, ires.start(0), self.baseurl, fatal =3D 0)
                self.handle_proc(res.group('name'), res.group('data') or '=
')
                i =3D res.end(0)
            res =3D cdata.match(data, i)
            if res is not None:
                matched =3D 1
                c0, c1 =3D res.span('cdata')
                ires =3D illegal1.search(data, c0, c1)
                if ires is not None:
                    self.__error('illegal characters in CDATA section', da=
ta, ires.start(0), self.baseurl, fatal =3D 0)
                self.handle_cdata(res.group('cdata'))
                i =3D res.end(0)
            if not matched:
                self.__error("no valid content in element `%s'" % ptagname=
, data, i, self.baseurl)
                return
        return i

    def __check_attr(self, tagname, attrname, value, attributes, data, att=
rstart):
        # check that the attribute attrname on element tagname is of
        # the correct type with a legal value
        # return the normalized value (i.e. white space collapsed if
        # appropriate)
        # XXX this method needs work to be complete
        attype, atvalue, atstring =3D attributes[attrname]
        if atvalue[:6] =3D=3D '#FIXED':
            if value !=3D atstring:
                self.__error("attribute `%s' in element `%s' does not have=
 correct value" % (attrname, tagname), data, attrstart, self.baseurl, fata=
l =3D 0)
        if attype =3D=3D 'CDATA':
            return value                # always OK and don't change value=

        if type(attype) is type([]):    # enumeration
            if value not in attype:
                self.__error("attribute `%s' in element `%s' not valid" % =
(attrname, tagname), data, attrstart, self.baseurl, fatal =3D 0)
            return value
        if type(attype) is type(()):
            if value not in attype[1]:
                self.__error("attribute `%s' in element `%s' not valid" % =
(attrname, tagname), data, attrstart, self.baseurl, fatal =3D 0)
            return value
        if attype =3D=3D 'ID':
            if name.match(value) is None:
                self.__error("attribute `%s' in element `%s' is not an ID"=
 % (attrname, tagname), data, attrstart, self.baseurl, fatal =3D 0)
            if self.ids.has_key(value):
                self.__error("attrbute `%s' in element `%s' is not unique"=
 %  (attrname, tagname), data, attrstart, self.baseurl, fatal =3D 0)
            self.ids[value] =3D 1
            return value
        if attype =3D=3D 'IDREF':
            if name.match(value) is None:
                self.__error("attrbute `%s' in element `%s' is not an IDRE=
F" % (attrname, tagname), data, attrstart, self.baseurl, fatal =3D 0)
            # XXX should check ID exists
            return value
        if attype =3D=3D 'IDREFS':
            if names.match(value) is None:
                self.__error("attrbute `%s' in element `%s' is not an IDRE=
FS" % (attrname, tagname), data, attrstart, self.baseurl, fatal =3D 0)
            # XXX should check IDs exist
            return value
        if attype =3D=3D 'NMTOKEN':
            if nmtoken.match(value) is None:
                self.__error("attrbute `%s' in element `%s' is not a NMTOK=
EN" % (attrname, tagname), data, attrstart, self.baseurl, fatal =3D 0)
            return value
        if attype =3D=3D 'NMTOKENS':
            if nmtokens.match(value) is None:
                self.__error("attrbute `%s' in element `%s' is not a NMTOK=
ENS" % (attrname, tagname), data, attrstart, self.baseurl, fatal =3D 0)
            return value
        if attype =3D=3D 'ENTITY':
            if name.match(value) is None:
                self.__error("attrbute `%s' in element `%s' is not an ENTI=
TY" % (attrname, tagname), data, attrstart, self.baseurl, fatal =3D 0)
            # XXX should check ENTITY exists
            return value
        if attype =3D=3D 'ENTITIES':
            if names.match(value) is None:
                self.__error("attrbute `%s' in element `%s' is not an ENTI=
TIES" % (attrname, tagname), data, attrstart, self.baseurl, fatal =3D 0)
            # XXX should check ENTITIES exist
            return value
        # XXX other types?
        return value

    def __parse_attrs(self, tagname, data, tagstart, span, namespaces):
        # parse the string between the tag name and closing bracket
        # for attribute=3Dvalue pairs
        i, dataend =3D span
        attrlist =3D []
        namespace =3D None
        reqattrs =3D {}                   # attributes that are #REQUIRED
        if self.elems.has_key(tagname):
            attributes =3D self.elems[tagname][1]
            for key, (attype, atvalue, atstring) in attributes.items():
                if atvalue =3D=3D '#REQUIRED':
                    reqattrs[key] =3D 1
            attrseen =3D {}               # attributes that we've seen
        else:
            attributes =3D None
        while i < dataend:
            res =3D attrfind.match(data, i, dataend)
            if res is None:
                # couldn't match any attributes, but there is more
                # string to parse: complain and ignore rest of string
                self.__error('bad attributes', data, i, self.baseurl, fata=
l =3D 0)
                return
            name =3D res.group('attrname')
            if reqattrs.has_key(name):
                del reqattrs[name]      # seen this #REQUIRED attribute
            if attributes is not None and attributes.has_key(name):
                attype =3D attributes[name][0]
            else:
                attype =3D None
            start, end =3D res.span('attrvalue')
            value =3D self.__parse_attrval(data, attype, span =3D (start+1=
, end-1))
            if value is None:
                # bad attribute value: ignore, but continue parsing
                i =3D res.end(0)
                continue
            attrstart =3D res.start('attrname')
            if attributes is not None:
                if attributes.has_key(name):
                    attrseen[name] =3D 1
                    value =3D self.__check_attr(tagname, name, value, attr=
ibutes, data, attrstart)
                else:
                    self.__error("unknown attribute `%s' on element `%s'" =
% (name, tagname), data, attrstart, self.baseurl, fatal =3D 0)
            i =3D res.end(0)
            if self.__xmlns:
                res =3D xmlns.match(name)
                if res is not None:
                    # namespace declaration
                    ncname =3D res.group('ncname')
                    if namespace is None:
                        namespace =3D {}
                    namespace[ncname or ''] =3D value or None
                    continue
            attrlist.append((name, value, attrstart))
        if reqattrs:
            # there are #REQUIRED attributes that we haven't seen
            reqattrs =3D reqattrs.keys()
            reqattrs.sort()
            if len(reqattrs) > 1:
                s =3D 's'
            else:
                s =3D ''
            reqattrs =3D string.join(reqattrs, "', `")
            self.__error("required attribute%s `%s' of element `%s' missin=
g" % (s, reqattrs, tagname), data, dataend, self.baseurl, fatal =3D  0)
        if attributes is not None:
            # fill in missing attributes that have a default value
            for key, (attype, atvalue, atstring) in attributes.items():
                if atstring is not None and not attrseen.has_key(key):
                    attrlist.append((key, atstring, dataend))
        if namespace is not None:
            namespaces =3D (namespace, namespaces)
        if namespaces is not None:
            res =3D qname.match(tagname)
            if res is not None:
                prefix, nstag =3D res.group('prefix', 'local')
                if prefix is None: prefix =3D ''
                ns =3D None
                n =3D namespaces
                while n is not None:
                    d, n =3D n
                    if d.has_key(prefix):
                        ns =3D d[prefix]
                        break
                if ns is not None:
                    tagname =3D ns + ' ' + nstag
                elif prefix !=3D '':
                    self.__error("unknown namespace prefix `%s'" % prefix,=
 data, tagstart, self.baseurl, fatal =3D 0)
            else:
                self.__error("badly formed tag name `%s'" % tagname, data,=
 tagstart, self.baseurl, fatal =3D 0)
        attrdict =3D {}                   # collect attributes/values
        for attr, value, attrstart in attrlist:
            if namespaces is not None:
                res =3D qname.match(attr)
                if res is not None:
                    prefix, nsattr =3D res.group('prefix', 'local')
                    if prefix:
                        ans =3D None
                        n =3D namespaces
                        while n is not None:
                            d, n =3D n
                            if d.has_key(prefix):
                                ans =3D d[prefix]
                                break
                        if ans is not None:
                            attr =3D ans + ' ' + nsattr
                        elif prefix !=3D '':
                            self.__error("unknown namespace prefix `%s'" %=
 prefix, data, attrstart, self.baseurl, fatal =3D 0)
                else:
                    self.__error("badly formed attribute name `%s'" % attr=
, data, attrstart, self.baseurl, fatal =3D 0)
            if attrdict.has_key(attr):
                self.__error("duplicate attribute name `%s'" % attr, data,=
 attrstart, self.baseurl, fatal =3D 0)
            attrdict[attr] =3D value
        return tagname, attrdict, namespaces

    def __parse_attrval(self, data, attype, span =3D None):
        # parse an attribute value, replacing entity and character
        # references with their values
        if span is None:
            i =3D 0
            dataend =3D len(data)
        else:
            i, dataend =3D span
        res =3D illegal1.search(data, i, dataend)
        if res is not None:
            self.__error("illegal characters in attribute value", data, re=
s.start(0), self.baseurl, fatal =3D 0)
        newval =3D []
        while i < dataend:
            res =3D interesting.search(data, i, dataend)
            if res is None:
                str =3D data[i:dataend]
                if attype is None or attype =3D=3D 'CDATA':
                    str =3D self.__normalize_space(str)
                newval.append(str)
                break
            j =3D res.start(0)
            if data[j] =3D=3D '<':
                self.__error("no `<' allowed in attribute value", data, j,=
 self.baseurl, fatal =3D 0)
            if j > i:
                str =3D data[i:j]
                if attype is None or attype =3D=3D 'CDATA':
                    str =3D self.__normalize_space(str)
                newval.append(str)
            res =3D ref.match(data, j, dataend)
            if res is None:
                self.__error('illegal attribute value', data, j, self.base=
url, fatal =3D 0)
                newval.append(data[j])  # the &
                i =3D j + 1               # continue searching after the &=

                continue
            i =3D res.end(0)
            name =3D res.group('name')
            if name:
                # entity reference (e.g. "&lt;")
                if self.entitydefs.has_key(name):
                    val =3D self.entitydefs[name]
                    if type(val) is type(()):
                        self.__error("no external parsed entity allowed in=
 attribute value", data, res.start(0), self.baseurl, fatal =3D 1)
                    del self.entitydefs[name]
                    nval =3D self.__parse_attrval(val, attype)
                    self.entitydefs[name] =3D val
                    if nval is None:
                        return
                    newval.append(nval)
                else:
                    self.__error("reference to unknown entity `%s'" % name=
, data, res.start(0), self.baseurl, fatal =3D 0)
                    newval.append('&%s;' % name)
            else:
                val =3D self.__parse_charref(res.group('char'), data, res.=
start(0))
                if val is None:
                    newval.append('&#%s;' % res.group('char'))
                    continue
                newval.append(val)
        str =3D string.join(newval, '')
        if attype is not None and attype !=3D 'CDATA':
            str =3D string.join(string.split(str))
        return str

    def __parse_charref(self, name, data, i):
        # parse a character reference (e.g. "%#38;")
        # the "name" arg is just part between # and ;
        if name[0] =3D=3D 'x':
            # e.g. &#x26;
            n =3D int(name[1:], 16)
        else:
            # e.g. &#38;
            n =3D int(name)
        try:
            c =3D unichr(n)
        except ValueError:
            self.__error('bad character reference', data, i, self.baseurl,=
 fatal =3D 0)
            return
        if illegal1.search(c):
            self.__error('bad character reference', data, i, self.baseurl,=
 fatal =3D 0)
        return c

    def __read_pentity(self, publit, syslit):
        import urllib
        syslit =3D urllib.basejoin(self.baseurl, syslit)
        baseurl =3D self.baseurl
        self.baseurl =3D syslit
        val =3D self.read_external(publit, syslit)
        val =3D self.__parse_textdecl(val)
        return self.__normalize_linefeed(val)

    def parse_dtd(self, data, internal =3D 1):
        """parse_dtd(data[, internal ])

           Parse the DTD.
           This method is called by the parse_doctype method and is
           provided so that parse_doctype can be overridden.
           Argument is a string containing the full DTD.
           Optional argument internal is true (default) if the DTD is
           internal."""
        i =3D 0
        matched =3D 1
        ilevel =3D 0                      # nesting level of ignored secti=
ons
        while i < len(data) and matched:
            matched =3D 0
            res =3D peref.match(data, i)
            if res is not None:
                matched =3D 1
                name =3D res.group('name')
                if self.pentitydefs.has_key(name):
                    val =3D self.pentitydefs[name]
                    baseurl =3D self.baseurl
                    if type(val) is type(()):
                        val =3D self.__read_pentity(val[0], val[1])
                    self.parse_dtd(val, internal)
                    self.baseurl =3D baseurl
                else:
                    self.__error("unknown entity `%%%s;'" % name, data, i,=
 self.baseurl, fatal =3D 0)
                i =3D res.end(0)
            res =3D element.match(data, i)
            if res is not None:
                matched =3D 1
                name, content =3D res.group('name', 'content')
                i =3D res.end(0)
                elemval =3D (None, {}, None, None, None)
                if self.elems.has_key(name):
                    elemval =3D self.elems[name]
                    if elemval[0] is not None:
                        # XXX is this an error?
                        self.__error('non-unique element name declaration'=
, data, i, self.baseurl, fatal =3D 0)
                    elif content =3D=3D 'EMPTY':
                        # check for NOTATION on EMPTY element
                        for atname, (attype, atvalue, atstring) in elemval=
[1].items():
                            if type(attype) is type(()) and attype[0] =3D=
=3D 'NOTATION':
                                self.__error("NOTATION not allowed on EMPT=
Y element", data, i, self.baseurl)
                if content[0] =3D=3D '(':
                    i =3D res.start('content')
                    j, content, start, end =3D self.__dfa(data, i)
                    if type(content) is type([]) and content and type(cont=
ent[0]) is type({}):
                        self.__check_dfa(content, start, name, data, i)
                    contentstr =3D data[i:j]
                    i =3D j
                else:
                    contentstr =3D content
                    start =3D end =3D 0
                self.elems[name] =3D (content, elemval[1], start, end, con=
tentstr)
                res =3D space.match(data, i)
                if res is not None:
                    i =3D res.end(0)
                if data[i:i+1] !=3D '>':
                    self.__error('bad DOCTYPE', data, i, self.baseurl)
                    return
                i =3D i+1
            res =3D attlist.match(data, i)
            if res is not None:
                matched =3D 1
                elname, atdef =3D res.group('elname', 'atdef')
                if not self.elems.has_key(elname):
                    self.elems[elname] =3D (None, {}, None, None, None)
                ares =3D attdef.match(atdef)
                while ares is not None:
                    atname, attype, atvalue, atstring =3D ares.group('atna=
me', 'attype', 'atvalue', 'atstring')
                    if attype[0] =3D=3D '(':
                        attype =3D map(string.strip, string.split(attype[1=
:-1], '|'))
                    elif attype[:8] =3D=3D 'NOTATION':
                        if self.elems[elname][0] =3D=3D 'EMPTY':
                            self.__error("NOTATION not allowed on EMPTY el=
ement", data, ares.start('attype'), self.baseurl)
                        atnot =3D map(string.strip, string.split(ares.grou=
p('notation'), '|'))
                        attype =3D ('NOTATION', atnot)
                    if atstring:
                        atstring =3D atstring[1:-1] # remove quotes
                        atstring =3D self.__parse_attrval(atstring, attype=
)
                        if attype !=3D 'CDATA':
                            atstring =3D string.join(string.split(atstring=
))
                        else:
                            atstring =3D string.join(string.split(atstring=
, '\t'), ' ')
                    if type(attype) is type([]):
                        if atstring is not None and atstring not in attype=
:
                            self.__error("default value for attribute `%s'=
 on element `%s' not listed as possible value" % (atname, elname), data, i=
, self.baseurl)
                    elif type(attype) is type(()):
                        if atstring is not None and atstring not in attype=
[1]:
                            self.__error("default value for attribute `%s'=
 on element `%s' not listed as possible value" % (atname, elname), data, i=
, self.baseurl)
                    if not self.elems[elname][1].has_key(atname):
                        # first definition counts
                        self.elems[elname][1][atname] =3D attype, atvalue,=
 atstring
                    ares =3D attdef.match(atdef, ares.end(0))
                i =3D res.end(0)
            res =3D entity.match(data, i)
            if res is not None:
                matched =3D 1
                pname, name =3D res.group('pname', 'ename')
                if pname:
                    pvalue =3D res.group('pvalue')
                    if pvalue[0] in ('"',"'"):
                        c0, c1 =3D res.span('pvalue')
                        ires =3D illegal1.search(data, c0+1, c1-1)
                        if ires is not None:
                            self.__error("illegal characters in entity val=
ue", data, ires.start(0), self.baseurl, fatal =3D 0)
                    if self.pentitydefs.has_key(pname):
                        # first definition counts
                        pass
                    elif pvalue[0] in ('"',"'"):
                        pvalue =3D pvalue[1:-1]
                        pvalue =3D self.__normalize_space(pvalue)
                        cres =3D entref.search(pvalue)
                        while cres is not None:
                            chr, nm =3D cres.group('char', 'pname')
                            if chr:
                                repl =3D self.__parse_charref(cres.group('=
char'), data, i)
                            elif self.pentitydefs.has_key(nm):
                                repl =3D self.pentitydefs[nm]
                            else:
                                self.__error("unknown entity `%s' referenc=
ed" % nm, data, i, self.baseurl)
                                repl =3D '%%%s;' % nm
                            if type(repl) is type(()):
                                baseurl =3D self.baseurl
                                repl =3D self.__read_pentity(repl[0], repl=
[1])
                                self.baseurl =3D baseurl
                            pvalue =3D pvalue[:cres.start(0)] + repl + pva=
lue[cres.end(0):]
                            cres =3D entref.search(pvalue, cres.start(0)+l=
en(repl))
                        self.pentitydefs[pname] =3D pvalue
                    else:
                        r =3D externalid.match(pvalue)
                        publit, syslit =3D r.group('publit', 'syslit')
                        if publit: publit =3D string.join(string.split(pub=
lit[1:-1]))
                        if syslit: syslit =3D syslit[1:-1]
                        self.pentitydefs[pname] =3D publit, syslit
                else:
                    value =3D res.group('value')
                    if value[0] in ('"',"'"):
                        c0, c1 =3D res.span('value')
                        ires =3D illegal1.search(data, c0+1, c1-1)
                        if ires is not None:
                            self.__error("illegal characters in entity val=
ue", data, ires.start(0), self.baseurl, fatal =3D 0)
                    if self.entitydefs.has_key(name):
                        # use first definition
                        pass
                    elif value[0] in ('"',"'"):
                        value =3D value[1:-1]
                        value =3D self.__normalize_space(value)
                        cres =3D entref.search(value)
                        while cres is not None:
                            chr, nm =3D cres.group('char', 'pname')
                            if chr:
                                repl =3D self.__parse_charref(cres.group('=
char'), data, i)
                            elif self.pentitydefs.has_key(nm):
                                repl =3D self.pentitydefs[nm]
                                if type(repl) is type(()):
                                    baseurl =3D self.baseurl
                                    repl =3D self.__read_pentity(repl[0], =
repl[1])
                                    self.baseurl =3D baseurl
                            else:
                                self.__error("unknown entity `%s' referenc=
ed" % nm, data, i, self.baseurl)
                                repl =3D '%%%s;' % nm
                            value =3D value[:cres.start(0)] + repl + value=
[cres.end(0):]
                            cres =3D entref.search(value, cres.start(0)+le=
n(repl))
                        self.entitydefs[name] =3D value
                    else:
                        r =3D externalid.match(value)
                        publit, syslit =3D r.group('publit', 'syslit')
                        if publit: publit =3D string.join(string.split(pub=
lit[1:-1]))
                        if syslit: syslit =3D syslit[1:-1]
                        r1 =3D ndata.match(value, r.end(0))
                        if r1 is not None:
                            ndataname =3D r1.group('name')
                        else:
                            ndataname =3D None
                        self.entitydefs[name] =3D publit, syslit, ndatanam=
e
                i =3D res.end(0)
            res =3D notation.match(data, i)
            if res is not None:
                matched =3D 1
                name, value =3D res.group('name', 'value')
                if not self.notation.has_key(name):
                    self.notation[name] =3D value
                i =3D res.end(0)
            j =3D i                       # remember where we were
            i =3D self.__parse_misc(data, i)
            matched =3D matched or i > j  # matched anything?
            if not internal:
                if data[i:i+1] =3D=3D '<':
                    hlevel =3D 1
                    quote =3D None
                    j =3D i+1
                    while hlevel > 0:
                        res =3D bracket.search(data, j)
                        if res is None:
                            self.__error("unexpected EOF", data, i, self.b=
aseurl, fatal =3D 1)
                        j =3D res.end(0)
                        c =3D data[res.start(0)]
                        if c =3D=3D '<':
                            hlevel =3D hlevel + 1
                        elif quote and c =3D=3D quote:
                            quote =3D None
                        elif c in ('"', "'"):
                            quote =3D c
                        elif c =3D=3D '>':
                            hlevel =3D hlevel - 1
                        elif hlevel =3D=3D 1 and not quote:
                            # only expand parsed entities at lowest level
                            res =3D peref.match(data, res.start(0))
                            if res is not None:
                                pname =3D res.group('name')
                                if self.pentitydefs.has_key(pname):
                                    repl =3D self.pentitydefs[pname]
                                    if type(repl) is type(()):
                                        baseurl =3D self.baseurl
                                        repl =3D self.__read_pentity(repl[=
0], repl[1])
                                        self.baseurl =3D baseurl
                                    data =3D data[:res.start(0)] + ' ' + r=
epl + ' ' + data[res.end(0):]
                                    j =3D res.start(0) + len(repl) + 2
                                else:
                                    j =3D res.end(0)
                res =3D conditional.match(data, i)
                if res is not None:
                    inc, ign =3D res.group('inc', 'ign')
                    i =3D res.end(0)
                    if ign:
                        level =3D 1
                        while level > 0:
                            res =3D ignore.search(data, i)
                            if res.start(0) =3D=3D '<':
                                level =3D level + 1
                            else:
                                level =3D level - 1
                            i =3D res.end(0)
                    elif inc:
                        ilevel =3D ilevel + 1
                if ilevel and data[i:i+3] =3D=3D ']]>':
                    i =3D i+3
                    ilevel =3D ilevel - 1
        if i < len(data):
            self.__error('error while parsing DOCTYPE', data, i, self.base=
url)

    def __dfa(self, data, i):
        res =3D mixedre.match(data, i)
        if res is not None:
            mixed =3D res.group(0)
            if mixed[-1] =3D=3D '*':
                mixed =3D map(string.strip, string.split(mixed[1:-2], '|')=
)
            else:
                mixed =3D '#PCDATA'
            return res.end(0), mixed, 0, 0
        dfa =3D []
        i, start, end =3D self.__dfa1(data, i, dfa)
        return i, dfa, start, end

    def __dfa1(self, data, i, dfa):
        res =3D dfaelem0.match(data, i)
        if res is None:
            self.__error("syntax error in element content: `(' or Name exp=
ecter", data, i, self.baseurl, fatal =3D 1)
        token =3D res.group('token')
        if token =3D=3D '(':
            i, start, end =3D self.__dfa1(data, res.end(0), dfa)
            res =3D dfaelem1.match(data, i)
            if res is None:
                self.__error("syntax error in element content: `)', `|', o=
r `,' expected", data, i, self.baseurl, fatal =3D 1)
            token =3D res.group('token')
            sep =3D token
            while token in (',','|'):
                if sep !=3D token:
                    self.__error("syntax error in element content: `%s' or=
 `)' expected" % sep, data, i, self.baseurl, fatal =3D 1)
                i, nstart, nend =3D self.__dfa1(data, res.end(0), dfa)
                res =3D dfaelem1.match(data, i)
                if res is None:
                    self.__error("syntax error in element content: `%s' or=
 `)' expected" % sep, data, i, self.baseurl, fatal =3D 1)
                token =3D res.group('token')
                if sep =3D=3D ',':
                    # concatenate DFAs
                    e =3D dfa[end].get('', [])
                    e.append(nstart)
                    dfa[end][''] =3D e
                    end =3D nend
                else:
                    # make parallel
                    s =3D len(dfa)
                    dfa.append({'': [start, nstart]})
                    e =3D dfa[end].get('', [])
                    e.append(len(dfa))
                    dfa[end][''] =3D e
                    e =3D dfa[nend].get('', [])
                    e.append(len(dfa))
                    dfa[nend][''] =3D e
                    start =3D s
                    end =3D len(dfa)
                    dfa.append({})
            # token =3D=3D ')'
            i =3D res.end(0)
        else:
            # it's a Name
            start =3D len(dfa)
            dfa.append({token: [start+1]})
            end =3D len(dfa)
            dfa.append({})
            i =3D res.end(0)
        res =3D dfaelem2.match(data, i)
        if res is not None:
            token =3D res.group('token')
            s =3D len(dfa)
            e =3D s+1
            if token =3D=3D '+':
                dfa.append({'': [start]})
            else:
                dfa.append({'': [start, e]})
            dfa.append({})
            l =3D dfa[end].get('', [])
            dfa[end][''] =3D l
            if token !=3D '?':
                l.append(start)
            l.append(e)
            start =3D s
            end =3D e
            i =3D res.end(0)
        return i, start, end

    def parse_doctype(self, tag, publit, syslit, data):
        """parse_doctype(tag, publit, syslit, data)

           Parse the DOCTYPE.

           This method is called by the handle_doctype callback method
           and is provided so that handle_doctype can be overridden.
           The arguments are:
           tag: the name of the outermost element of the document;
           publit: the Public Identifier of the DTD (or None);
           syslit: the System Literal of the DTD (or None);
           data: the internal subset of the DTD (or None)."""
        if data:
            self.parse_dtd(data)
        if syslit:
            import urllib
            syslit =3D urllib.basejoin(self.baseurl, syslit)
            baseurl =3D self.baseurl
            self.baseurl =3D syslit
            external =3D self.read_external(publit, syslit)
            external =3D self.__parse_textdecl(external)
            external =3D self.__normalize_linefeed(external)
            self.parse_dtd(external, 0)
            self.baseurl =3D baseurl

    def __error(self, message, data =3D None, i =3D None, filename =3D Non=
e, fatal =3D 1):
        # called for all syntax errors
        # this either raises an exception (Error) or calls
        # self.syntax_error which may be overridden
        if data is not None and i is not None:
            self.lineno =3D lineno =3D string.count(data, '\n', 0, i) + 1
        else:
            self.lineno =3D None
        self.data =3D data
        self.offset =3D i
        self.filename =3D filename
        if fatal:
            raise Error(message, lineno, data, i, filename)
        self.syntax_error(message)

    # Overridable -- handle xml processing instruction
    def handle_xml(self, encoding, standalone):
        pass

    # Overridable -- handle DOCTYPE
    def handle_doctype(self, tag, publit, syslit, data):
        if self.doctype is not None:
            syslit =3D self.doctype
        self.parse_doctype(tag, publit, syslit, data)

    # Example -- read external file referenced from DTD with a SystemLiter=
al
    def read_external(self, publit, syslit):
        return ''

    # Example -- handle comment, could be overridden
    def handle_comment(self, data):
        pass

    # Example -- handle processing instructions, could be overridden
    def handle_proc(self, name, data):
        pass

    # Example -- handle data, should be overridden
    def handle_data(self, data):
        pass

    # Example -- handle cdata, should be overridden
    def handle_cdata(self, data):
        pass

    elements =3D {}                       # dict: tagname -> (startfunc, e=
ndfunc)
    def finish_starttag(self, tagname, attrs):
        method =3D self.elements.get(tagname, (None, None))[0]
        if method is None:
            self.unknown_starttag(tagname, attrs)
        else:
            self.handle_starttag(tagname, method, attrs)

    def finish_endtag(self, tagname):
        method =3D self.elements.get(tagname, (None, None))[1]
        if method is None:
            self.unknown_endtag(tagname)
        else:
            self.handle_endtag(tagname, method)

    # Overridable -- handle start tag
    def handle_starttag(self, tagname, method, attrs):
        method(tagname, attrs)

    # Overridable -- handle end tag
    def handle_endtag(self, tagname, method):
        method(tagname)

    # To be overridden -- handlers for unknown objects
    def unknown_starttag(self, tagname, attrs):
        pass

    def unknown_endtag(self, tagname):
        pass

    def unknown_entityref(self, name):
        self.__error('reference to unknown entity', self.data, self.offset=
, self.baseurl)

    # Example -- handle relatively harmless syntax errors, could be overri=
dden
    def syntax_error(self, message):
        raise Error(message, self.lineno, self.data, self.offset, self.fil=
ename)

class TestXMLParser(XMLParser):

    def __init__(self, xmlns =3D 1):
        self.testdata =3D ""
        XMLParser.__init__(self, xmlns)

    def handle_xml(self, encoding, standalone):
        self.flush()
        print 'xml: encoding =3D %s standalone =3D %s' % (encoding, standa=
lone)

    def read_external(self, publit, syslit):
        print 'reading %s' % name
        try:
            import urllib
            u =3D urllib.urlopen(syslit)
            data =3D u.read()
            u.close()
        except 'x':
            return ''
        return data

    def handle_doctype(self, tag, publit, syslit, data):
        self.flush()
        print 'DOCTYPE: %s %s' % (tag, `data`)
        XMLParser.handle_doctype(self, tag, publit, syslit, data)

    def handle_comment(self, data):
        self.flush()
        r =3D `data`
        if len(r) > 68:
            r =3D r[:32] + '...' + r[-32:]
        print 'comment: %s' % r

    def handle_proc(self, name, data):
        self.flush()
        print 'processing: %s %s' % (name,`data`)

    def handle_data(self, data):
        self.testdata =3D self.testdata + data
        if len(`self.testdata`) >=3D 70:
            self.flush()

    def handle_cdata(self, data):
        self.flush()
        print 'cdata: %s' % `data`

    def flush(self):
        data =3D self.testdata
        if data:
            self.testdata =3D ""
            print 'data: %s ' % `data`

##    def syntax_error(self, message):
##        if self.lineno is not None:
##            print 'Syntax error at line %d: %s' % (self.lineno, message)=

##        else:
##            print 'Syntax error: %s' % message

    def unknown_starttag(self, tag, attrs):
        self.flush()
        if not attrs:
            print 'start tag: <%s>' % tag
        else:
            print 'start tag: <%s' % tag,
            for name, value in attrs.items():
                print '%s =3D "%s"' % (name.encode('latin-1'), `value`),
            print '>'

    def unknown_endtag(self, tag):
        self.flush()
        print 'end tag: </%s>' % tag

    def unknown_entityref(self, name):
        self.flush()
        print '&%s;' % name

class CanonXMLParser(XMLParser):
    __cache =3D {}

    def read_external(self, publit, syslit):
        if publit and self.__cache.has_key(publit):
            return self.__cache[publit]
        try:
            import urllib
            u =3D urllib.urlopen(syslit)
            data =3D u.read()
            u.close()
        except 'x':
            return ''
        if publit:
            self.__cache[publit] =3D data
        return data

    def handle_data(self, data):
        sys.stdout.write(self.encode(data))

    def handle_cdata(self, data):
        sys.stdout.write(self.encode(data))

    def handle_proc(self, name, data):
        sys.stdout.write('<?%s %s?>' % (name.encode('utf-8'), data.strip()=
.encode('utf-8')))

    def unknown_starttag(self, tag, attrs):
        sys.stdout.write('<%s' % tag.encode('utf-8'))
        attrlist =3D attrs.items()
        attrlist.sort()
        for name, value in attrlist:
            sys.stdout.write(' %s=3D"%s"' % (name.encode('utf-8'), self.en=
code(value)))
        sys.stdout.write('>')

    def unknown_endtag(self, tag):
        sys.stdout.write('</%s>' % tag.encode('utf-8'))

    def unknown_entityref(self, name):
        print '&%s;' % name.encode('utf-8')

    def encode(self, data):
        for c, tr in [('&', '&amp;'),
                      ('>', '&gt;'),
                      ('<', '&lt;'),
                      ('"', '&quot;'),
                      ('\t', '&#9;'),
                      ('\n', '&#10;'),
                      ('\r', '&#13;')]:
            data =3D tr.join(data.split(c))
        return data.encode('utf-8')

class CheckXMLParser(XMLParser):
    __cache =3D {}

    def read_external(self, publit, syslit):
        if publit and self.__cache.has_key(publit):
            return self.__cache[publit]
        try:
            import urllib
            u =3D urllib.urlopen(syslit)
            data =3D u.read()
            u.close()
        except 'x':
            return ''
        if publit:
            self.__cache[publit] =3D data
        return data

def test(args =3D None):
    import sys, getopt

    if not args:
        args =3D sys.argv[1:]

    opts, args =3D getopt.getopt(args, 'cstnvCd:')
    klass =3D TestXMLParser
    do_time =3D 0
    namespace =3D 1
    verbose =3D 0
    doctype =3D None
    for o, a in opts:
        if o =3D=3D '-c':
            klass =3D CanonXMLParser
        elif o =3D=3D '-C':
            klass =3D CheckXMLParser
        elif o =3D=3D '-s':
            klass =3D XMLParser
        elif o =3D=3D '-t':
            do_time =3D 1
        elif o =3D=3D '-n':
            namespace =3D 0
        elif o =3D=3D '-v':
            verbose =3D 1
        elif o =3D=3D '-d':
            doctype =3D a

    if not args:
        args =3D ['test.xml']

    for file in args:
        if file =3D=3D '-':
            f =3D sys.stdin
            url =3D '.'
        else:
            try:
                f =3D open(file, 'r')
            except IOError, msg:
                print file, ":", msg
                sys.exit(1)
            import urllib
            url =3D urllib.pathname2url(file)

        data =3D f.read()
        if f is not sys.stdin:
            f.close()

        x =3D klass(xmlns =3D namespace)
        x.baseurl =3D url
        x.doctype =3D doctype
        if verbose:
            print '=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D',file
        try:
            t0, t1, t2 =3D x.parse(data)
        except Error, info:
            do_time =3D 0                 # can't print times now
            print str(info)
            if info.text is not None and info.offset is not None:
                i =3D string.rfind(info.text, '\n', 0, info.offset) + 1
                j =3D string.find(info.text, '\n', info.offset)
                if j =3D=3D -1: j =3D len(info.text)
                try:
                    print info.text[i:j]
                except UnicodeError:
                    print `info.text[i:j]`
                else:
                    print ' '*(info.offset-i)+'^'
        if klass is CanonXMLParser and (verbose or len(args) > 1):
            sys.stdout.write('\n')
        if do_time:
            print 'total time: %g' % (t2-t0)
            print 'parse DTD: %g' % (t1-t0)
            print 'parse body: %g' %(t2-t1)

if __name__ =3D=3D '__main__':
    test()

------- =_aaaaaaaaaa0--


From rsalz@zolera.com  Mon Apr 23 17:17:48 2001
From: rsalz@zolera.com (Rich Salz)
Date: Mon, 23 Apr 2001 12:17:48 -0400
Subject: [XML-SIG] Canonicalizing XML
References: <E14qdlp-0007P4-00@ute.cnri.reston.va.us> <20010423155236.71C71301CF7@bireme.oratrix.nl>
Message-ID: <3AE455AC.2A829E01@zolera.com>

> I've written a validating XML parser in Python that can produce
> Canonical XML.

As in the XML C14N recommendation?  It seems to do something else.  For
example, all namespace attributes are removed.

> I'll attach it.  Usage (for getting Canonical XML):
> 
>         python fxmllib.py -c file.xml

Do I need a DTD?
	/r$


From sjoerd.mullender@oratrix.com  Mon Apr 23 19:02:16 2001
From: sjoerd.mullender@oratrix.com (Sjoerd Mullender)
Date: Mon, 23 Apr 2001 20:02:16 +0200
Subject: [XML-SIG] Canonicalizing XML
References: <E14qdlp-0007P4-00@ute.cnri.reston.va.us> <20010423155236.71C71301CF7@bireme.oratrix.nl> <3AE455AC.2A829E01@zolera.com>
Message-ID: <3AE46E28.D062FA30@oratrix.com>

Ah, there is a recommendation.  In other words, no, I didn't follow the
C14N recommendation.  I emulated the output of the XML test suite.  I
forget where that came from, but I'm sure I can find out.
The test suite doesn't deal with namespaces since it's not for testing
the XML Namespace rec.

No, you don't need a DTD.

Rich Salz wrote:
> 
> > I've written a validating XML parser in Python that can produce
> > Canonical XML.
> 
> As in the XML C14N recommendation?  It seems to do something else.  For
> example, all namespace attributes are removed.
> 
> > I'll attach it.  Usage (for getting Canonical XML):
> >
> >         python fxmllib.py -c file.xml
> 
> Do I need a DTD?
>         /r$


From noreply@sourceforge.net  Mon Apr 23 19:36:25 2001
From: noreply@sourceforge.net (noreply@sourceforge.net)
Date: Mon, 23 Apr 2001 11:36:25 -0700
Subject: [XML-SIG] [ pyxml-Bugs-418315 ] ODS problems in Python 2.0
Message-ID: <E14rlCD-0004vi-00@usw-sf-web3.sourceforge.net>

Bugs item #418315, was updated on 2001-04-23 11:36
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418315&group_id=6473

Category: 4Suite
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Mike Olson (mikeolson)
Assigned to: Nobody/Anonymous (nobody)
Summary: ODS problems in Python 2.0

Initial Comment:
Email from alexander smishlajev


hello all!

i have run into some problems with 4Suite-0.10.2:

- the long long datatype formatter in Oracle and
Postgres
adapters:

Constants.Types.SIGNED_LONG_LONG:lambda
x:str(long(x))[:-1]

does not work on python 2.0: str() in v2.0 does not
append 'L'
suffix, so the last digit gets trimmed.  replacing
str() with
repr() makes types_test.py pass ok.

- i think that there is a bug in
Lib/DbmDatabase.CreateDatabase():
the DATABASE_DIR is joined twice.  the following patch
fixes the
problem:

--- DbmDatabase.py.orig Mon Feb 19 02:31:18 2001
+++ DbmDatabase.py      Tue Feb 20 21:15:46 2001
@@ -179,7 +179,7 @@
 
     dir_util._path_created = {}
     dir_util.mkpath(dbpath)
-    return Database(dbpath)
+    return Database(dbName)
 
 def DropDatabase(dbName):
     CheckVersion()

- DbmDatabase assumes that each table is stored in file
with the
same name.  in most cases, this is true because default
anydbm
backend module is dbhash, which works both on posix and
windows
platforms.  but dbm and dumbdbm databases create a pair
of files
having additional extensions.  IMHO whichdb module may
be used
instead of os.path.exists() to see if the database
exists, but
there still remain problems with GetAllDatabaseNames()
method.

- there is a problem that seems to be a bug in win32
dbhash: if
key+value length exceeds 400 characters, the database
may become
broken.  posix system (FreeBSD 4.2) is not affected. 
the
following interactive session demonstrates this:

Python 2.0 (#8, Oct 16 2000, 17:27:58) [MSC 32 bit
(Intel)] on
win32
Type "copyright", "credits" or "license" for more
information.
>>> import anydbm
>>> db =anydbm.open('test.db', 'c')
>>> db['2'] ='*' *1000
>>> db['2'] ='*'
>>> db.keys()
['22']
>>>

best wishes,
alex.

----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418315&group_id=6473


From noreply@sourceforge.net  Mon Apr 23 19:42:05 2001
From: noreply@sourceforge.net (noreply@sourceforge.net)
Date: Mon, 23 Apr 2001 11:42:05 -0700
Subject: [XML-SIG] [ pyxml-Bugs-418317 ] CDATA in XPath
Message-ID: <E14rlHh-00050Z-00@usw-sf-web3.sourceforge.net>

Bugs item #418317, was updated on 2001-04-23 11:42
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418317&group_id=6473

Category: 4Suite
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Mike Olson (mikeolson)
Assigned to: Nobody/Anonymous (nobody)
Summary: CDATA in XPath

Initial Comment:
Email from Kemani Driss

I have some elements whose data is enclosed in CDATA
tags and I am finding that the Evaluate method is
returning None for elements whose data is enclosed in 
CDATA tags. Works fine for other elements without
CDATA tags.

My XPath statement given to the Evaluate method is
"child::Options/child::text()".

Am I doing something wrong or is this a known
feature/bug of 4XPath?

thanks in advance.

----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418317&group_id=6473


From noreply@sourceforge.net  Mon Apr 23 19:44:31 2001
From: noreply@sourceforge.net (noreply@sourceforge.net)
Date: Mon, 23 Apr 2001 11:44:31 -0700
Subject: [XML-SIG] [ pyxml-Bugs-418319 ] more ODS issues
Message-ID: <E14rlK3-00052d-00@usw-sf-web3.sourceforge.net>

Bugs item #418319, was updated on 2001-04-23 11:44
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418319&group_id=6473

Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Mike Olson (mikeolson)
Assigned to: Nobody/Anonymous (nobody)
Summary: more ODS issues

Initial Comment:
Email from alexander smishlajev <alex@ank-sia.com>

hello!

_4odb_odmsdump.py and _4odb_odmsload.py use sys.argv
instead of
argv argument in Run()

DbAdapter.getAllObjectIds() in Dbm adapter seems to be
broken. 
following patch was locally applied:

--- Adapter.py.orig     Mon Feb 19 02:39:42 2001
+++ Adapter.py  Fri Mar 09 22:17:39 2001
@@ -208,8 +208,8 @@
             del objs[str(oid)]
 
 
-    def getAllObjectIds(self):
-        return map(lambda
x:x[0],self.statements[CompiliedStatement.GET_OBJECT_IDS].query(db))
+    def getAllObjectIds(self,db):
+        return map(int, db['objects'].keys())
 
 

_4odb_dig.py and _4odb_metadig.py both exit with usage
info if
len(argv) != 2 while corresponding .bat files always
pass a list
with 10 elements.

OifParser fails with long integers.  following patch
helps (i am
not sure if we can always return long integer):

--- OifParser.py.orig   Fri Feb 09 00:07:40 2001
+++ OifParser.py        Fri Mar 09 22:51:05 2001
@@ -104,7 +104,10 @@
             return float(literal[0][0])
         elif literal[0]['TYPE'] == 'INTEGER_LITERAL':
             # could be decimal, octal or hexadecimal
-            return eval(literal[0][0])
+            try:
+                return eval(literal[0][0])
+            except OverflowError:
+                return eval(literal[0][0] +'L')
         elif literal[0]['TYPE'] == 'OBJECT_TAG':
             return literal[0][0]
         else:


best wishes,
alex.

----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418319&group_id=6473


From noreply@sourceforge.net  Mon Apr 23 19:52:38 2001
From: noreply@sourceforge.net (noreply@sourceforge.net)
Date: Mon, 23 Apr 2001 11:52:38 -0700
Subject: [XML-SIG] [ pyxml-Bugs-418323 ] RDF problem
Message-ID: <E14rlRu-00059M-00@usw-sf-web3.sourceforge.net>

Bugs item #418323, was updated on 2001-04-23 11:52
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418323&group_id=6473

Category: 4Suite
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Mike Olson (mikeolson)
Assigned to: Nobody/Anonymous (nobody)
Summary: RDF problem

Initial Comment:
Email from "Joern v. Kattchee" <kattchee@ikv.de>

Hello,

there seems to be a bug in 4RDF (4Suite-0.10.2).
Parsing the following RDF

<rdf:RDF
xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
         xmlns:ns="http://my/ns#">
  <ns:item rdf:ID="this">
    <ns:prop>
      <ns:item>
        <ns:prop rdf:resource="http://foo/"/>
      </ns:item>
    </ns:prop>
  </ns:item>
</rdf:RDF>

results in these statements (p,s,o):

rdf:type, http://test#this, ns:item
rdf:type, http://test#anon-1, ns:prop
rdf:type, http://test#anon-2, ns:item
ns:prop,  http://test#anon-2, http://foo/
ns:item,  http://test#anon-1, http://test#anon-2
ns:prop,  http://test#this, http://test#anon-1

but correct would IMHO be something like this:

rdf:type, http://test#this, ns:item
rdf:type, http://test#anon-2, ns:item
ns:prop,  http://test#anon-2, http://foo/
ns:prop,  http://test#this, http://test#anon-2

4RDF somehow interprets the second 'item' as
propertyElt, but it is a typedNode.

Regards,

Joern v. Kattchee

----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418323&group_id=6473


From noreply@sourceforge.net  Mon Apr 23 19:56:44 2001
From: noreply@sourceforge.net (noreply@sourceforge.net)
Date: Mon, 23 Apr 2001 11:56:44 -0700
Subject: [XML-SIG] [ pyxml-Bugs-418324 ] [4XSLT] bug report and patch for complex
Message-ID: <E14rlVs-0005DO-00@usw-sf-web3.sourceforge.net>

Bugs item #418324, was updated on 2001-04-23 11:56
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418324&group_id=6473

Category: 4Suite
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Mike Olson (mikeolson)
Assigned to: Nobody/Anonymous (nobody)
Summary: [4XSLT] bug report and patch for complex

Initial Comment:
Email from "Olivier CAYROL (Logilab)"
<Olivier.Cayrol@logilab.fr>

  Hello,

  I found a vicious bug in 4XSLT (hidden very deeply in
the code). 
Attached to this message you will find a tar.gz file
containing a 
directory tree that exhibits the bug. It is a little
application for 
managing Easter rabbits and eggs distribution (!).
There is an XML file 
that contains the data: easter_mng.xml, an XSL
Transformation file: 
xsl/transf.xsl and XML files containing data for
localization: 
lib/common.xml, lib/en.xml, lib/fr.xml.

  The lib/common.xml file is imported in the XSLT
stylesheets with the 
'document()' function and is used to insert
language-dependant tags in 
the output. This common.xml file imports other XML
files (one per 
language) with the classic external ENTITY mechanism of
XML.

  When trying to transform the data file from the main
directory with 
the following line command: 
    4xslt -Dlang=en  easter_mng.xml xsl/transf.xsl
, I got this exception:
    ...
      File
"/usr/lib/python1.5/site-packages/xml/xslt/XsltFunctions.py", 
    line 63, in Document
        doc =
context.stylesheet._docReader.fromUri(uri,
baseUri=baseUri)
      File
"/usr/lib/python1.5/site-packages/Ft/Lib/ReaderBase.py",
line 
    67, in fromUri
        rt = self.fromStream(stream, baseUri, ownerDoc,
stripElements)
      File
"/usr/lib/python1.5/site-packages/Ft/Lib/pDomlette.py",
line 5
    78, in fromStream
        raise FtException(Error.XML_PARSE_ERROR,
p.ErrorLineNumber, p.Err
    orColumnNumber, expat.ErrorString(p.ErrorCode))
      Ft.Lib.FtException: ('XML parse error at line 16,
column 2: error i
    n processing external entity reference', (16, 2,
'error in processing
     external entity reference'))

  In fact, there is a problem when 4XSLT reads the XML
document 
referenced in the 'document()' function: this XML file
contains ENTITYs 
that import XML tree parts by giving local paths from
the current 
document directory whereas in 4XSLT, the baseUri is
always the URI of 
the initial XSLT. The XML reader is unable to find the
external entities
and the bug appears.

  Replacing line 67 of Ft.Lib.ReaderBase.py in
DomletteReader.fromUri
function:
    rt = self.fromStream(stream, baseUri, ownerDoc,
stripElements)
with :
    newBaseUri = urllib.basejoin(baseUri, uri)
    rt = self.fromStream(stream, newBaseUri, ownerDoc,
stripElements)
fixes the bug.

  I initially found the bug while trying to process
Norman Walsh's XSLT
stylesheets for turning docbook files in XSL formatting
objects files (I
am unfortunately not working for the Easter Rabbit).

  Regards,

    O. CAYROL.

----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418324&group_id=6473


From sskau@pchome.com.tw  Tue Apr 24 01:51:03 2001
From: sskau@pchome.com.tw (sskau)
Date: Tue, 24 Apr 2001 08:51:03 +0800
Subject: [XML-SIG] Persistence of HTMLDocument
Message-ID: <20010424083936.C5B4.SSKAU@pchome.com.tw>

Dear All:
	I used the PyXML 0.6.5 to parse some html files and then
try to save this html document object to disk (using shelve). (The
reason is that we
could speed up the loading process without further parsing)
I got the following exception:


Traceback (innermost last):
  File "docper.py", line 18, in ?
    newdoc=3Ddocobj["doc"]
  File "e:\Python\Lib\shelve.py", line 65, in __getitem__
    return Unpickler(f).load()
TypeError: ('too many arguments; expected 1, got 2', <class xml.dom.html.HT=
MLDo
ument.HTMLDocument at 988cb0>, (None,))


The Source code is:

=66rom xml.dom.ext.reader import HtmlLib
=66rom xml.dom import ext
import shelve

stream =3D open("c:\\temp\\new.html")
doc=3DHtmlLib.Reader().fromStream(stream)

# persistence of HtmlDocument
docobj =3D shelve.open("c:\\temp\\doc.obj")
docobj["doc"]=3Ddoc
docobj.close()

stream.close()

# reopen, load persistence document
docobj =3D shelve.open("c:\\temp\\doc.obj")
newdoc=3Ddocobj["doc"]
ext.PrettyPrint(newdoc)



Have anyone concern the persistence of paresed HtmlDocument DOM tree ?


Thanks a lot



Shi-Shiuan Kao



==========================================================
 PC home §K¶O¹q¤l«H½c¡A¥Ó½Ð½Ð¦Ü: http://www.pchome.com.tw 
 PC home Online ºô¸ô®a®x¡@¡@ ·|­û²Ä¤@¡A¥xÆW³Ì¤jªº¤J¤fºô¯¸ 
==========================================================


From jtauber@jtauber.com  Tue Apr 24 03:45:00 2001
From: jtauber@jtauber.com (James Tauber)
Date: Mon, 23 Apr 2001 22:45:00 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
Message-ID: <002601c0cc68$915a1940$20020a0a@EHUD>

We just discovered that Redfoot (and its RDF parser) breaks in Python 2.1
because expat is now more restrictive in what can be a namespace_separator.
Does anyone know why this change was made?

James



From martin@loewis.home.cs.tu-berlin.de  Tue Apr 24 07:07:55 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 24 Apr 2001 08:07:55 +0200
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <002601c0cc68$915a1940$20020a0a@EHUD> (jtauber@jtauber.com)
References: <002601c0cc68$915a1940$20020a0a@EHUD>
Message-ID: <200104240607.f3O67tR00949@mira.informatik.hu-berlin.de>

> We just discovered that Redfoot (and its RDF parser) breaks in Python 2.1
> because expat is now more restrictive in what can be a namespace_separator.
> Does anyone know why this change was made?

I'm not aware of any change to that respect. pyexpat passes the
namespace_separator as-is to Expat. What version of expat are you
using?

Regards,
Martin


From ssaky@kldp.org  Tue Apr 24 07:17:28 2001
From: ssaky@kldp.org (ssaky@kldp.org)
Date: Tue, 24 Apr 2001 15:17:28 +0900
Subject: [XML-SIG] about Python/XML HOWTO...
Message-ID: <20010424151728.A21831@ssaky.jerimo.org>

Hi.
Sorry for my rough english.

Could you mind if i translate your "Python/XML HOWTO" into korean,
and post it to http://kldp.org (Korean Linux Document Project) and
Python web site in Korean (http://www.python.or.kr) ?

Thanks you.

-- 
God Bless You~ =)

http://ssaky.jerimo.org
http://glory.python.or.kr



From eikeon@eikeon.com  Tue Apr 24 08:08:30 2001
From: eikeon@eikeon.com (Daniel "eikeon" Krech)
Date: Tue, 24 Apr 2001 03:08:30 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <002601c0cc68$915a1940$20020a0a@EHUD> <200104240607.f3O67tR00949@mira.informatik.hu-berlin.de>
Message-ID: <02c501c0cc8d$5e0a6b10$0401a8c0@VAIO>

> > We just discovered that Redfoot (and its RDF parser) breaks in Python
2.1
> > because expat is now more restrictive in what can be a
namespace_separator.
> > Does anyone know why this change was made?
>
> I'm not aware of any change to that respect. pyexpat passes the
> namespace_separator as-is to Expat. What version of expat are you
> using?
>

We are using the one that came with the Python2.1 final (source) release.
There is a test in Python-2.1/Lib/test/test_pyexpat.py at the end that is
testing for 'too short'. We have been using a namespace separator of an
empty string with no problems using Python2.0 and earlier. It does appear to
be a restriction imposed by the underlying pyexpat library... I am curious
what motivated this restriction as it seems like it should be a valid thing
for an application to want to do.

-eikeon



From martin@loewis.home.cs.tu-berlin.de  Tue Apr 24 13:23:41 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 24 Apr 2001 14:23:41 +0200
Subject: [XML-SIG] about Python/XML HOWTO...
In-Reply-To: <20010424151728.A21831@ssaky.jerimo.org>
References: <20010424151728.A21831@ssaky.jerimo.org>
Message-ID: <200104241223.f3OCNfv00849@mira.informatik.hu-berlin.de>

> Could you mind if i translate your "Python/XML HOWTO" into korean,
> and post it to http://kldp.org (Korean Linux Document Project) and
> Python web site in Korean (http://www.python.or.kr) ?

This would be very good. When you are done, please let us know so we
can make appropriate links. Also, I recommend that you remember the
CVS revision of the original document in your translation, so you can
find out better what has been changed later. As you probably know, the
original document is doc/xml-howto.tex.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Tue Apr 24 13:26:11 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 24 Apr 2001 14:26:11 +0200
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <02c501c0cc8d$5e0a6b10$0401a8c0@VAIO> (eikeon@eikeon.com)
References: <002601c0cc68$915a1940$20020a0a@EHUD> <200104240607.f3O67tR00949@mira.informatik.hu-berlin.de> <02c501c0cc8d$5e0a6b10$0401a8c0@VAIO>
Message-ID: <200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de>

> > I'm not aware of any change to that respect. pyexpat passes the
> > namespace_separator as-is to Expat. What version of expat are you
> > using?
> 
> We are using the one that came with the Python2.1 final (source) release.

There is no copy of Expat included in the source release of Python
2.1, so you must have got it from somewhere else...

> There is a test in Python-2.1/Lib/test/test_pyexpat.py at the end that is
> testing for 'too short'. We have been using a namespace separator of an
> empty string with no problems using Python2.0 and earlier. It does appear to
> be a restriction imposed by the underlying pyexpat library... I am curious
> what motivated this restriction as it seems like it should be a valid thing
> for an application to want to do.

So am I, but that is hard to answer until we find out what the exact
version of Expat is that you are using (expat is the underlying
library, not pyexpat), and until you can provide a test case that
demonstrates the problem...

Regards,
Martin



From fdrake@acm.org  Tue Apr 24 15:01:24 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Tue, 24 Apr 2001 10:01:24 -0400 (EDT)
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de>
References: <002601c0cc68$915a1940$20020a0a@EHUD>
 <200104240607.f3O67tR00949@mira.informatik.hu-berlin.de>
 <02c501c0cc8d$5e0a6b10$0401a8c0@VAIO>
 <200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de>
Message-ID: <15077.34612.333942.402686@cj42289-a.reston1.va.home.com>

Martin v. Loewis writes:
 > There is no copy of Expat included in the source release of Python
 > 2.1, so you must have got it from somewhere else...

  This is an issue with the bindings; the Expat version doesn't enter
into this.

 > So am I, but that is hard to answer until we find out what the exact
 > version of Expat is that you are using (expat is the underlying
 > library, not pyexpat), and until you can provide a test case that
 > demonstrates the problem...

  I added the check that a string passed as the namespace_separator
would be either omitted, None, or of length 1.  I don't understand why
you would want it to be of length 0 -- were you expecting parsing
without namespaces?  Is there a reason to pass namespace_separator as
an empty string rather than None, or just omit it?


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From eikeon@eikeon.com  Tue Apr 24 15:17:44 2001
From: eikeon@eikeon.com (Daniel "eikeon" Krech)
Date: Tue, 24 Apr 2001 10:17:44 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <002601c0cc68$915a1940$20020a0a@EHUD><200104240607.f3O67tR00949@mira.informatik.hu-berlin.de><02c501c0cc8d$5e0a6b10$0401a8c0@VAIO><200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> <15077.34612.333942.402686@cj42289-a.reston1.va.home.com>
Message-ID: <002001c0ccc9$54e28cc0$0401a8c0@VAIO>

>   I added the check that a string passed as the namespace_separator
> would be either omitted, None, or of length 1.  I don't understand why
> you would want it to be of length 0 -- were you expecting parsing
> without namespaces?  Is there a reason to pass namespace_separator as
> an empty string rather than None, or just omit it?

If we pass in None or omit it ends up parsing without namespaces, or at
least the names we are getting are of the form prefix:localname. We do want
namespaces, just with no separator between the namespace and the local name.
     [For <p:bar xmlns="foo"/> in our application (of the parser) we would
like to receive names in the form foobar. Passing in None yields p:bar]

-eikeon



From fdrake@acm.org  Tue Apr 24 15:31:05 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Tue, 24 Apr 2001 10:31:05 -0400 (EDT)
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <002001c0ccc9$54e28cc0$0401a8c0@VAIO>
References: <002601c0cc68$915a1940$20020a0a@EHUD>
 <200104240607.f3O67tR00949@mira.informatik.hu-berlin.de>
 <02c501c0cc8d$5e0a6b10$0401a8c0@VAIO>
 <200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de>
 <15077.34612.333942.402686@cj42289-a.reston1.va.home.com>
 <002001c0ccc9$54e28cc0$0401a8c0@VAIO>
Message-ID: <15077.36393.731033.157812@cj42289-a.reston1.va.home.com>

"Daniel \"eikeon\" Krech" writes:
 > If we pass in None or omit it ends up parsing without namespaces, or at
 > least the names we are getting are of the form prefix:localname. We do want
 > namespaces, just with no separator between the namespace and the local name.
 >      [For <p:bar xmlns="foo"/> in our application (of the parser) we would
 > like to receive names in the form foobar. Passing in None yields p:bar]

  This sounds like it would be ambiguous -- consider:

       <foo xmlns="bat" xmlns:baz="ba">
         <baz:tfoo/>
       </foo>

would give you the elements batfoo and batfoo.
  Admittedly, this is a contrived example, and probably pretty
unusually in practice.  But it definately shows that using '' as the
separator is not workable in the general case -- you have to know that
your input is reasonable.  I can easily imagine somethng that
generated artificial namespaces coming up with XML that will be
problematic for this.
  I'm not particularly opposed to opening up the bindings to support
using '', though.  If that's the SIG's concensus, I'll be glad to do
it.  Backward compatibility is a nice quality!


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From eikeon@eikeon.com  Tue Apr 24 16:41:53 2001
From: eikeon@eikeon.com (Daniel "eikeon" Krech)
Date: Tue, 24 Apr 2001 11:41:53 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <002601c0cc68$915a1940$20020a0a@EHUD><200104240607.f3O67tR00949@mira.informatik.hu-berlin.de><02c501c0cc8d$5e0a6b10$0401a8c0@VAIO><200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de><15077.34612.333942.402686@cj42289-a.reston1.va.home.com><002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com>
Message-ID: <001a01c0ccd5$f2f385c0$f3fc0a0a@bowstreet.com>

>   This sounds like it would be ambiguous -- consider:
>
>        <foo xmlns="bat" xmlns:baz="ba">
>          <baz:tfoo/>
>        </foo>
>
> would give you the elements batfoo and batfoo.

With no separator (aka '') there is some information loss, as you no longer
know what part of the name was contributed by the namespace and what part by
the local name. For our application of the parser we currently do not use
this piece of information and so the information only becomes a burden on
us.

Even with the ambiguity its seems fairly heavy handed to enforce at such a
low level. Often when the degenerate boundary cases are handled in a
consistent manor... it is a good thing. What are the benifits of being heavy
handed at the lower level?

IMHO, it would be really nice if batfoo and batfoo where the same. Does
anyone know if this is or is not the case? Or is it application specific as
to wether or not they are?

-eikeon



From loewis@informatik.hu-berlin.de  Tue Apr 24 17:13:06 2001
From: loewis@informatik.hu-berlin.de (Martin von Loewis)
Date: Tue, 24 Apr 2001 18:13:06 +0200 (MEST)
Subject: [XML-SIG] PyXML 0.6.5 for Python 2.1
Message-ID: <200104241613.SAA01273@pandora.informatik.hu-berlin.de>

I have rebuilt PyXML 0.6.5 with the Windows installer for Python 2.1,
and put that package on SourceForge for download at

http://prdownloads.sourceforge.net/pyxml/PyXML-0.6.5.win32-py2.1.exe

Regards,
Martin

P.S. It seems SF is constantly changing hostnames...


From rsalz@zolera.com  Tue Apr 24 17:38:08 2001
From: rsalz@zolera.com (Rich Salz)
Date: Tue, 24 Apr 2001 12:38:08 -0400
Subject: [XML-SIG] XML Schema
Message-ID: <200104241638.MAA19329@os390.zolera.com>

Is anyone doing anything with XML Schema?  My interest is in WSDL and SOAP
("web services"), so full schema -- in particular the subtyping,
restrictions, et al. -- isn't totally necessary.  If anyone's given
thought to what an internal parsed representation might look like (and
no, a set of domNodes doesn't count), I'd like to chat.
	/r$


From martin@loewis.home.cs.tu-berlin.de  Tue Apr 24 17:41:16 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 24 Apr 2001 18:41:16 +0200
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <15077.34612.333942.402686@cj42289-a.reston1.va.home.com>
 (fdrake@acm.org)
References: <002601c0cc68$915a1940$20020a0a@EHUD>
 <200104240607.f3O67tR00949@mira.informatik.hu-berlin.de>
 <02c501c0cc8d$5e0a6b10$0401a8c0@VAIO>
 <200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> <15077.34612.333942.402686@cj42289-a.reston1.va.home.com>
Message-ID: <200104241641.f3OGfGW01442@mira.informatik.hu-berlin.de>

>   This is an issue with the bindings; the Expat version doesn't enter
> into this.

Sorry for the confusion. I just missed the check you put in somehow.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Tue Apr 24 17:45:36 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 24 Apr 2001 18:45:36 +0200
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <15077.36393.731033.157812@cj42289-a.reston1.va.home.com>
 (fdrake@acm.org)
References: <002601c0cc68$915a1940$20020a0a@EHUD>
 <200104240607.f3O67tR00949@mira.informatik.hu-berlin.de>
 <02c501c0cc8d$5e0a6b10$0401a8c0@VAIO>
 <200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de>
 <15077.34612.333942.402686@cj42289-a.reston1.va.home.com>
 <002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com>
Message-ID: <200104241645.f3OGjaS01465@mira.informatik.hu-berlin.de>

>   I'm not particularly opposed to opening up the bindings to support
> using '', though.  If that's the SIG's concensus, I'll be glad to do
> it.  Backward compatibility is a nice quality!

Well, I'm in favour of backwards-compatibility, then.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Tue Apr 24 17:48:38 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 24 Apr 2001 18:48:38 +0200
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <001a01c0ccd5$f2f385c0$f3fc0a0a@bowstreet.com>
 (eikeon@eikeon.com)
References: <002601c0cc68$915a1940$20020a0a@EHUD><200104240607.f3O67tR00949@mira.informatik.hu-berlin.de><02c501c0cc8d$5e0a6b10$0401a8c0@VAIO><200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de><15077.34612.333942.402686@cj42289-a.reston1.va.home.com><002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <001a01c0ccd5$f2f385c0$f3fc0a0a@bowstreet.com>
Message-ID: <200104241648.f3OGmck01467@mira.informatik.hu-berlin.de>

> >        <foo xmlns="bat" xmlns:baz="ba">
> >          <baz:tfoo/>
> >        </foo>
> >
> > would give you the elements batfoo and batfoo.
[...]
> IMHO, it would be really nice if batfoo and batfoo where the same. Does
> anyone know if this is or is not the case? Or is it application specific as
> to wether or not they are?

They are certainly not, in XML! One is the foo element in the bat
namespace, the other is the tfoo element in the ba namespace.

Of course, on the application level, you could always treat apples and
oranges as the same elements, if you want to.

Regards,
Martin


From larsga@garshol.priv.no  Tue Apr 24 19:01:02 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 24 Apr 2001 20:01:02 +0200
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <15077.36393.731033.157812@cj42289-a.reston1.va.home.com>
References: <002601c0cc68$915a1940$20020a0a@EHUD> 	<200104240607.f3O67tR00949@mira.informatik.hu-berlin.de> 	<02c501c0cc8d$5e0a6b10$0401a8c0@VAIO> 	<200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> 	<15077.34612.333942.402686@cj42289-a.reston1.va.home.com> 	<002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com>
Message-ID: <m3u23e2n4x.fsf@lambda.garshol.priv.no>

* Fred L. Drake, Jr.
|
| I'm not particularly opposed to opening up the bindings to support
| using '', though.

I'm not very happy with the idea. This sounds very much like selling a
gun with a bent muzzle. It may be nice for firing around corners, but
it does seem to me to increase the chances of accidentally removing
one or more feet.

I see the potential problems with this. What I do not see is the
potential benefits. Can someone suggest any?

--Lars M.



From martin@loewis.home.cs.tu-berlin.de  Tue Apr 24 22:22:49 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Tue, 24 Apr 2001 23:22:49 +0200
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <m3u23e2n4x.fsf@lambda.garshol.priv.no> (message from Lars Marius
 Garshol on 24 Apr 2001 20:01:02 +0200)
References: <002601c0cc68$915a1940$20020a0a@EHUD> 	<200104240607.f3O67tR00949@mira.informatik.hu-berlin.de> 	<02c501c0cc8d$5e0a6b10$0401a8c0@VAIO> 	<200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> 	<15077.34612.333942.402686@cj42289-a.reston1.va.home.com> 	<002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <m3u23e2n4x.fsf@lambda.garshol.priv.no>
Message-ID: <200104242122.f3OLMnK11789@mira.informatik.hu-berlin.de>

> I see the potential problems with this. What I do not see is the
> potential benefits. Can someone suggest any?

Backwards compatibility. Code that used to work still does work - even
if you think it is stupid code. Happy developers, happy users.

Regards,
Martin



From tpassin@home.com  Tue Apr 24 23:07:11 2001
From: tpassin@home.com (Thomas B. Passin)
Date: Tue, 24 Apr 2001 18:07:11 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <002601c0cc68$915a1940$20020a0a@EHUD><200104240607.f3O67tR00949@mira.informatik.hu-berlin.de><02c501c0cc8d$5e0a6b10$0401a8c0@VAIO><200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de><15077.34612.333942.402686@cj42289-a.reston1.va.home.com><002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <001a01c0ccd5$f2f385c0$f3fc0a0a@bowstreet.com>
Message-ID: <001d01c0cd0a$e9b823a0$7cac1218@reston1.va.home.com>

Daniel "eikeon" Krech"  asked -

>
> IMHO, it would be really nice if batfoo and batfoo where the same. Does
> anyone know if this is or is not the case? Or is it application specific
as
> to wether or not they are?
>
Definitely completely different from an xml-namespaces point of view.  I'm
not sure how or why your code uses the concatenated names, but I'd suggest
they might lead to problems down the road, if yu want to start using other
tools or interfacing with other systems.  You couldn't count on any random
parser supporting your way, if you had an urge to start using some other
parser.

Cheers,

Tom P



From jtauber@jtauber.com  Tue Apr 24 23:45:57 2001
From: jtauber@jtauber.com (James Tauber)
Date: Tue, 24 Apr 2001 18:45:57 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <002601c0cc68$915a1940$20020a0a@EHUD> 	<200104240607.f3O67tR00949@mira.informatik.hu-berlin.de> 	<02c501c0cc8d$5e0a6b10$0401a8c0@VAIO> 	<200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> 	<15077.34612.333942.402686@cj42289-a.reston1.va.home.com> 	<002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <m3u23e2n4x.fsf@lambda.garshol.priv.no>
Message-ID: <003601c0cd10$57870090$ed020a0a@bowstreet.com>

The key issue that eikeon forgot to mention is that this is for RDF where
namespaceURI+localName are concatenated to form property URIs. Without the
ability to use '' we have to do a split+join on every element coming back
from expat.

James

> | I'm not particularly opposed to opening up the bindings to support
> | using '', though.
>
> I'm not very happy with the idea. This sounds very much like selling a
> gun with a bent muzzle. It may be nice for firing around corners, but
> it does seem to me to increase the chances of accidentally removing
> one or more feet.
>
> I see the potential problems with this. What I do not see is the
> potential benefits. Can someone suggest any?




From tpassin@home.com  Wed Apr 25 00:04:52 2001
From: tpassin@home.com (Thomas B. Passin)
Date: Tue, 24 Apr 2001 19:04:52 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <002601c0cc68$915a1940$20020a0a@EHUD> 	<200104240607.f3O67tR00949@mira.informatik.hu-berlin.de> 	<02c501c0cc8d$5e0a6b10$0401a8c0@VAIO> 	<200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> 	<15077.34612.333942.402686@cj42289-a.reston1.va.home.com> 	<002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <m3u23e2n4x.fsf@lambda.garshol.priv.no> <003601c0cd10$57870090$ed020a0a@bowstreet.com>
Message-ID: <004801c0cd12$f87cae80$7cac1218@reston1.va.home.com>

James Tauber said -

> The key issue that eikeon forgot to mention is that this is for RDF where
> namespaceURI+localName are concatenated to form property URIs. Without the
> ability to use '' we have to do a split+join on every element coming back
> from expat.
>
Aren't those two parts usually joined with a "#" symbol?  Ayway, if you did
go to a split/join, I bet it wouldn't slow you down hardly at all, since the
rest of the RDF parsing must take a lot more time.

Cheers,

Tom P



From jtauber@jtauber.com  Wed Apr 25 00:58:51 2001
From: jtauber@jtauber.com (James Tauber)
Date: Tue, 24 Apr 2001 19:58:51 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <002601c0cc68$915a1940$20020a0a@EHUD> 	<200104240607.f3O67tR00949@mira.informatik.hu-berlin.de> 	<02c501c0cc8d$5e0a6b10$0401a8c0@VAIO> 	<200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> 	<15077.34612.333942.402686@cj42289-a.reston1.va.home.com> 	<002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <m3u23e2n4x.fsf@lambda.garshol.priv.no> <003601c0cd10$57870090$ed020a0a@bowstreet.com> <004801c0cd12$f87cae80$7cac1218@reston1.va.home.com>
Message-ID: <001201c0cd1a$86429c50$6401a8c0@EHUD>

> > The key issue that eikeon forgot to mention is that this is for RDF
where
> > namespaceURI+localName are concatenated to form property URIs. Without
the
> > ability to use '' we have to do a split+join on every element coming
back
> > from expat.
> >
> Aren't those two parts usually joined with a "#" symbol?

Nope. The # ends up being at the end of the namespace, it isn't a separator.
That is done sometimes so that the URI can refer to an actual element in an
XML document (via its ID) but this don't not have to be the case. You can
have a property http://foo.com/bar which could be serialized any of

   <bar xmlns="http://foo.com/">...</bar>
   <ar xmlns="http://foo.com/b">...</ar>
   <r xmlns="http://foo.com/ba">...</r>

Now, I don't like this but it's the way it works.

James



From rsalz@zolera.com  Wed Apr 25 01:56:27 2001
From: rsalz@zolera.com (Rich Salz)
Date: Tue, 24 Apr 2001 20:56:27 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <002601c0cc68$915a1940$20020a0a@EHUD> 	<200104240607.f3O67tR00949@mira.informatik.hu-berlin.de> 	<02c501c0cc8d$5e0a6b10$0401a8c0@VAIO> 	<200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> 	<15077.34612.333942.402686@cj42289-a.reston1.va.home.com> 	<002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <m3u23e2n4x.fsf@lambda.garshol.priv.no> <003601c0cd10$57870090$ed020a0a@bowstreet.com> <004801c0cd12$f87cae80$7cac1218@reston1.va.home.com> <001201c0cd1a$86429c50$6401a8c0@EHUD>
Message-ID: <3AE620BB.9FB13F18@zolera.com>

>    <bar xmlns="http://foo.com/">...</bar>
>    <ar xmlns="http://foo.com/b">...</ar>
>    <r xmlns="http://foo.com/ba">...</r>

Hunh?  Can you point to a spec that shows this?
	/r$


From jtauber@jtauber.com  Wed Apr 25 02:48:13 2001
From: jtauber@jtauber.com (James Tauber)
Date: Tue, 24 Apr 2001 21:48:13 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <002601c0cc68$915a1940$20020a0a@EHUD> 	<200104240607.f3O67tR00949@mira.informatik.hu-berlin.de> 	<02c501c0cc8d$5e0a6b10$0401a8c0@VAIO> 	<200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> 	<15077.34612.333942.402686@cj42289-a.reston1.va.home.com> 	<002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <m3u23e2n4x.fsf@lambda.garshol.priv.no> <003601c0cd10$57870090$ed020a0a@bowstreet.com> <004801c0cd12$f87cae80$7cac1218@reston1.va.home.com> <001201c0cd1a$86429c50$6401a8c0@EHUD> <3AE620BB.9FB13F18@zolera.com>
Message-ID: <002301c0cd29$caa4b900$6401a8c0@EHUD>

You cut that part out of my email. It's RDF.

In particular see http://www.w3.org/TR/REC-rdf-syntax/ in section 6:

"This expansion is generated by concatenating the namespace name given in
the namespace declaration with the LocalPart of the qualified name"

----- Original Message -----
From: "Rich Salz" <rsalz@zolera.com>
To: "James Tauber" <jtauber@jtauber.com>
Cc: <xml-sig@python.org>
Sent: Tuesday, April 24, 2001 8:56 PM
Subject: Re: [XML-SIG] expat namespace_separator in Python 2.1


> >    <bar xmlns="http://foo.com/">...</bar>
> >    <ar xmlns="http://foo.com/b">...</ar>
> >    <r xmlns="http://foo.com/ba">...</r>
>
> Hunh?  Can you point to a spec that shows this?
> /r$



From rsalz@zolera.com  Wed Apr 25 03:27:48 2001
From: rsalz@zolera.com (Rich Salz)
Date: Tue, 24 Apr 2001 22:27:48 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <002601c0cc68$915a1940$20020a0a@EHUD> 	<200104240607.f3O67tR00949@mira.informatik.hu-berlin.de> 	<02c501c0cc8d$5e0a6b10$0401a8c0@VAIO> 	<200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> 	<15077.34612.333942.402686@cj42289-a.reston1.va.home.com> 	<002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <m3u23e2n4x.fsf@lambda.garshol.priv.no> <003601c0cd10$57870090$ed020a0a@bowstreet.com> <004801c0cd12$f87cae80$7cac1218@reston1.va.home.com> <001201c0cd1a$86429c50$6401a8c0@EHUD> <3AE620BB.9FB13F18@zolera.com> <002301c0cd29$caa4b900$6401a8c0@EHUD>
Message-ID: <3AE63624.9C763305@zolera.com>

> You cut that part out of my email. It's RDF.

Oh, it wasn't clear to me from your note.

Thanks.


From jtauber@jtauber.com  Wed Apr 25 03:29:00 2001
From: jtauber@jtauber.com (James Tauber)
Date: Tue, 24 Apr 2001 22:29:00 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <002601c0cc68$915a1940$20020a0a@EHUD> 	<200104240607.f3O67tR00949@mira.informatik.hu-berlin.de> 	<02c501c0cc8d$5e0a6b10$0401a8c0@VAIO> 	<200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> 	<15077.34612.333942.402686@cj42289-a.reston1.va.home.com> 	<002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <m3u23e2n4x.fsf@lambda.garshol.priv.no> <003601c0cd10$57870090$ed020a0a@bowstreet.com> <004801c0cd12$f87cae80$7cac1218@reston1.va.home.com> <001201c0cd1a$86429c50$6401a8c0@EHUD> <3AE620BB.9FB13F18@zolera.com> <002301c0cd29$caa4b900$6401a8c0@EHUD> <3AE63624.9C763305@zolera.com>
Message-ID: <003001c0cd2f$80210ae0$6401a8c0@EHUD>

Sorry I wasn't more clear. It is only RDF that works this way. It is unusual
but internally consistent.

James

> > You cut that part out of my email. It's RDF.
>
> Oh, it wasn't clear to me from your note.
>
> Thanks.



From uche.ogbuji@fourthought.com  Wed Apr 25 03:57:56 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Tue, 24 Apr 2001 20:57:56 -0600
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: Message from Rich Salz <rsalz@zolera.com>
 of "Tue, 24 Apr 2001 20:56:27 EDT." <3AE620BB.9FB13F18@zolera.com>
Message-ID: <200104250258.f3P2vu814912@borgia.local>

> >    <bar xmlns="http://foo.com/">...</bar>
> >    <ar xmlns="http://foo.com/b">...</ar>
> >    <r xmlns="http://foo.com/ba">...</r>
> 
> Hunh?  Can you point to a spec that shows this?

James is quite right.  The offending spec is RDF.  Dr. Jonathan Borden has 
been carrying on a spirited campaign to get this fixed, but as of now, that's 
the way it is, and the former permissiveness of expat NS separators will need 
to be restored for proper RDF support, which is important enough, IMO, to 
warrant the  sale of guns with bent muzzles.

;-)


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From jtauber@jtauber.com  Wed Apr 25 04:10:37 2001
From: jtauber@jtauber.com (James Tauber)
Date: Tue, 24 Apr 2001 23:10:37 -0400
Subject: [XML-SIG] expat namespace_separator in Python 2.1
References: <200104250258.f3P2vu814912@borgia.local>
Message-ID: <000d01c0cd35$5078f900$ed020a0a@bowstreet.com>

Actually, the whole RDF serialization needs overhauling. Yet another reason
for a standard RDF parser / serializer to hide people from this stuff :-)

James

> James is quite right.  The offending spec is RDF.  Dr. Jonathan Borden has
> been carrying on a spirited campaign to get this fixed, but as of now,
that's
> the way it is, and the former permissiveness of expat NS separators will
need
> to be restored for proper RDF support, which is important enough, IMO, to
> warrant the  sale of guns with bent muzzles.




From noreply@sourceforge.net  Wed Apr 25 13:26:27 2001
From: noreply@sourceforge.net (noreply@sourceforge.net)
Date: Wed, 25 Apr 2001 05:26:27 -0700
Subject: [XML-SIG] [ pyxml-Bugs-418808 ] [4XSLT] Bug in precise/general match
Message-ID: <E14sONH-0004Ls-00@usw-sf-web2.sourceforge.net>

Bugs item #418808, was updated on 2001-04-25 05:26
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418808&group_id=6473

Category: 4Suite
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Logilab (ornicar)
Assigned to: Nobody/Anonymous (nobody)
Summary: [4XSLT] Bug in precise/general match

Initial Comment:
Hello,

I was writing an XSL stylesheet for processing docbook
XML files when I discovered a very strange behaviour of
4XSLT that looks like a bug.

Consider the following XML file (testbug.xml):
  <?xml version="1.0" encoding="ISO-8859-1"?>

  <section>
   <para>This is a paragraph.</para>
   <formalpara>
    <title>Title</title>
    <para>This is a formal paragraph.</para>
   </formalpara>
  </section>

And the following XSLT file (testbug1.xslt):
  <?xml version="1.0" encoding="ISO-8859-1"?>

  <xsl:stylesheet version="1.0" 
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">

   <xsl:template match="formalpara/para">
    <xsl:apply-templates/>
   </xsl:template>

   <xsl:template match="para">
    <p><xsl:apply-templates/></p>
   </xsl:template>

   <xsl:template match="formalpara">
    <p>
     <xsl:apply-templates select="title"/>
     <xsl:apply-templates select="para"/>
    </p>
   </xsl:template>

   <xsl:template match="formalpara/title">
    <b><xsl:apply-templates/>: </b>
   </xsl:template>

  </xsl:stylesheet>

  Processing the XML file with the XSLT file gives the
following result:
  <p>This is a paragraph.</p>
  <p><b>Title: </b>This is a formal paragraph</p>

You have noticed that the template for processing a
<para> node is different if the <para> node is child of
<formalpara> node or not. 4XSLT chooses the template
with the most precise matching condition that matches
the current node (and this is what is written in the
XSLT norm).

  Now, consider the following XSLT file
(testbug2.xslt):
  <?xml version="1.0" encoding="ISO-8859-1"?>

  <xsl:stylesheet version="1.0" 
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">

   <xsl:template match="para">
    <p><xsl:apply-templates/></p>
   </xsl:template>

   <xsl:template match="formalpara/para">
    <xsl:apply-templates/>
   </xsl:template>

   <xsl:template match="formalpara">
    <p>
     <xsl:apply-templates select="title"/>
     <xsl:apply-templates select="para"/>
    </p>
   </xsl:template>

   <xsl:template match="formalpara/title">
    <b><xsl:apply-templates/>: </b>
   </xsl:template>

  </xsl:stylesheet>

You have noticed that the only modification I did was
to permute the "para" template and the
"formalpara/para" template.

  Processing the same XML file with this new XSLT file
outputs the following result:
  <p>This is a paragraph.</p>
  <p><b>Title: </b><p>This is a formal
paragraph.</p></p>

Now, 4XSLT doesn't use the "formalpara" template for
the <para> node child of <formalpara> but the "para"
template (thus we have nested <p> nodes in the output). 

  I feel that instead of choosing the template with the
most precise matching condition that matches the
current node, 4XSLT chooses the first one. Therefore,
my two XSLT that should have the same behaviour, output
two different results.

  Regards,

    O. CAYROL.

PS: I tried xalan on these examples and I got the same
result for each XSLT (just as I expected).


----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418808&group_id=6473


From larsga@garshol.priv.no  Wed Apr 25 07:23:52 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 25 Apr 2001 08:23:52 +0200
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <003601c0cd10$57870090$ed020a0a@bowstreet.com>
References: <002601c0cc68$915a1940$20020a0a@EHUD> 	<200104240607.f3O67tR00949@mira.informatik.hu-berlin.de> 	<02c501c0cc8d$5e0a6b10$0401a8c0@VAIO> 	<200104241226.f3OCQBw00851@mira.informatik.hu-berlin.de> 	<15077.34612.333942.402686@cj42289-a.reston1.va.home.com> 	<002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <m3u23e2n4x.fsf@lambda.garshol.priv.no> <003601c0cd10$57870090$ed020a0a@bowstreet.com>
Message-ID: <m3r8yhts3r.fsf@lambda.garshol.priv.no>

* James Tauber
|
| The key issue that eikeon forgot to mention is that this is for RDF
| where namespaceURI+localName are concatenated to form property
| URIs. Without the ability to use '' we have to do a split+join on
| every element coming back from expat.

Well, if expat used to support this and it's useful for RDF I guess we
should do it. Sorry I was so slow in the uptake here.

--Lars M.



From uche.ogbuji@fourthought.com  Wed Apr 25 20:52:05 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Wed, 25 Apr 2001 13:52:05 -0600
Subject: [XML-SIG] Checking to see whether the list is down again
Message-ID: <200104251952.f3PJq5m20145@borgia.local.dhcp.fourthought.com>

If everyone gets it, I guess it ain't.

--Uche




From fdrake@acm.org  Wed Apr 25 20:59:28 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Wed, 25 Apr 2001 15:59:28 -0400 (EDT)
Subject: [XML-SIG] Checking to see whether the list is down again
In-Reply-To: <200104251952.f3PJq5m20145@borgia.local.dhcp.fourthought.com>
References: <200104251952.f3PJq5m20145@borgia.local.dhcp.fourthought.com>
Message-ID: <15079.11424.15199.976834@cj42289-a.reston1.va.home.com>

Uche Ogbuji writes:
 > If everyone gets it, I guess it ain't.

  There was an Exim configuration bug on python.org & zope.org; mail
should be flooding through there about now...  ;-)


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From fdrake@acm.org  Wed Apr 25 14:50:40 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Wed, 25 Apr 2001 09:50:40 -0400 (EDT)
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <200104250258.f3P2vu814912@borgia.local>
References: <rsalz@zolera.com>
 <3AE620BB.9FB13F18@zolera.com>
 <200104250258.f3P2vu814912@borgia.local>
Message-ID: <15078.54832.157396.84051@cj42289-a.reston1.va.home.com>

Uche Ogbuji writes:
 > James is quite right.  The offending spec is RDF.  Dr. Jonathan
 > Borden has been carrying on a spirited campaign to get this fixed,
 > but as of now, that's the way it is, and the former permissiveness
 > of expat NS separators will need to be restored for proper RDF
 > support, which is important enough, IMO, to warrant the sale of
 > guns with bent muzzles.

  OK, I'll make the adjustment to pyexpat.c, but I think it's bad to
perpetuate such vile brokeness.  Whether or not its bad enough to use
guns with straight muzzles, well, I guess we'll all have to decide for
ourselves.  ;-)


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From noreply@sourceforge.net  Wed Apr 25 23:52:40 2001
From: noreply@sourceforge.net (noreply@sourceforge.net)
Date: Wed, 25 Apr 2001 15:52:40 -0700
Subject: [XML-SIG] [ pyxml-Bugs-418986 ] Cant find /usr/local/bin/python  RedHat7
Message-ID: <E14sY9I-0003AM-00@usw-sf-web3.sourceforge.net>

Bugs item #418986, was updated on 2001-04-25 15:52
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418986&group_id=6473

Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Richard Hebert (hebertrich)
Assigned to: Nobody/Anonymous (nobody)
Summary: Cant find /usr/local/bin/python  RedHat7

Initial Comment:
Well to sum up i have a recent official RH 7 disk Linux
install.
The packages are all where RH installs them.
Python has been installed at the same time RH Linux was
It has been updated through the update agent and showed
no bugs.

I tried to install PyXML with GnomeRPM and it plain
dont work
It says (retriple checking )
PyXML requires /usr/local/bin/python
of course  ...python is in /usr/bin 

This has been reproduced on two machines other than
mine.

Is there a workaround or patch for this ? 

Any help will be highly appreciated.


Richard Hebert
hebertrich@operamail

you can e mail me directly if you need more details.



----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=106473&aid=418986&group_id=6473


From mbennett@ideaeng.com  Thu Apr 26 00:43:06 2001
From: mbennett@ideaeng.com (Mark Bennett)
Date: Wed, 25 Apr 2001 16:43:06 -0700
Subject: [XML-SIG] Exception handling with xml.dom.ext.reader.HtmlLib
Message-ID: <jUsT.aNoTheR.mEsSaGe.iD.98824278431156@ideaeng.com>

I was thrilled to see sample python scripts for parsing HTML.  The 
library seems to handle lots of common mistakes like unbalanced tags,
etc, things that most XML parsers will reject by design.  By it's 
nature, HTML is rarely in proper XML format.

But I've hit a couple snags with the library and I was wondering 
if anybody had any ideas?

* There are some classes of common HTML mistakes that it doesn't
  handle, like unbalanced quotes.  As in <font color="red> or
  <font color=red">, the second form gives a stack dump.
* When it does crash it doesn't give you any information about
  the source file, like what line it was looking at.  Such info would
  be helpful.
* Though I don't know the exact cause, it doesn't handle pages
  like http://www.cnn.com

I'm not a parsing expert, but I'd be happy to contribute to any efforts 
to make the parser more robust.  Processing existing (poorly formed) 
HTML is the 800 pound gorilla for lots of XML applications.  This 
library does go a long way.

Thanks,
Mark









From Mike.Olson@fourthought.com  Thu Apr 26 03:52:06 2001
From: Mike.Olson@fourthought.com (Mike Olson)
Date: Wed, 25 Apr 2001 20:52:06 -0600
Subject: [XML-SIG] 4Suite and 4SuiteServer 0.11 alpha announcement
Message-ID: <3AE78D56.869210D@FourThought.com>

Well, friends, thanks for all your patience.  Finally, we are ready to
release 4Suite and 4Suite Server 0.11, and this is to announce an alpha
package of
the product.

The alpha test is available from 4Suite.org: no registration required. 
See

http://4suite.org/download.html

So why the looooong delay for this release?

We have taken all the lessons learned from developing, deploying and
using 4SS so far, huddled up for an intensive few weeks, and emerged
with a complete re-factoring of the architecture.

The most important change is that now the heart of 4SS, which is simply
a set of Python APIs for managing and processing XML in a
server-oriented framework, is directly exposed, rather than hidden
through CORBA.

This means that all the heavy requirements and complex installation are
a thing of the past.  You can still use CORBA and all that, but it is
just another integration protocol that sits beside HTTP, SOAP, WebDAV,
etc.  It is not required.  PostgresQL or other RDBMS is not required. 
In fact, all that is required now is Python and 4Suite.  It's quick to
install, and there is a great boost in performance.

The architecture of the repository is now much more "Pythonic" and easy
to grasp.

And we have also spent a good deal of time on docs.  Starting at

http://4Suite.org/topics/rdf/Top/Documentation/GeneralProgrammersGuide

you will find a comprehensive, step-by-step introduction to 4SS 0.11.

For the now very simple installation, see

http://stage.4Suite.org/documents/guides/4SuiteServer/UNIX_Installation

(should be applicable to Windows users as well but we'll get the Windows
installation
docs back on line before final release).

Note that the 4Suite 0.11 alpha is a prerequisite.

Pretty much all the features of 4SS 0.10.2 will be in the 0.11 final
release, and all but a few are fully implemented in this alpha.

See

http://4Suite.org/topics/rdf/Top/Documentation/features.html

For a listing of features in this release.

Please help us shake out this Alpha for bugs, or give 4SS a try if
you've been shy of doing so because of its complexity.  We have set a
pretty firm final release date of 30 April (next Monday), so quick bug
reports and suggestions would be ideal.  Note that we have been testing
it a great deal internally, and all our own 4SS apps have already been
ported to and tested on the new architecture, so you needn't feel like a
guinea pig.

Thanks.

The Fourthought guys

-- 
Mike Olson				 Principal Consultant
mike.olson@fourthought.com               (303)583-9900 x 102
Fourthought, Inc.                         http://Fourthought.com 
Software-engineering, knowledge-management, XML, CORBA, Linux, Python


From Mike.Olson@fourthought.com  Thu Apr 26 03:58:25 2001
From: Mike.Olson@fourthought.com (Mike Olson)
Date: Wed, 25 Apr 2001 20:58:25 -0600
Subject: [XML-SIG] Re: [4suite] 4Suite and 4SuiteServer 0.11 alpha announcement
References: <3AE78D56.869210D@FourThought.com>
Message-ID: <3AE78ED1.CE25170A@FourThought.com>

Mike Olson wrote:
> 
> http://4Suite.org/topics/rdf/Top/Documentation/GeneralProgrammersGuide

Sorry everyone, I sent a bad link in the first message, here is the
correct one.

http://4suite.org/documents/guides/4SuiteServer/GeneralProgrammersGuide


Mike

-- 
Mike Olson				 Principal Consultant
mike.olson@fourthought.com               (303)583-9900 x 102
Fourthought, Inc.                         http://Fourthought.com 
Software-engineering, knowledge-management, XML, CORBA, Linux, Python


From Juergen Hermann" <jh@web.de  Wed Apr 25 06:29:15 2001
From: Juergen Hermann" <jh@web.de (Juergen Hermann)
Date: Wed, 25 Apr 2001 07:29:15 +0200
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <200104250258.f3P2vu814912@borgia.local>
Message-ID: <m14sHq9-007M9DC@smtp.web.de>

On Tue, 24 Apr 2001 20:57:56 -0600, Uche Ogbuji wrote:
>the way it is, and the former permissiveness of expat NS separators wil=
l need 
>to be restored for proper RDF support, which is important enough, IMO, =
to 
>warrant the  sale of guns with bent muzzles.

Maybe the bend-o-matic could get a safety switch?

Ciao, J=FCrgen




From martin@loewis.home.cs.tu-berlin.de  Thu Apr 26 09:42:48 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Thu, 26 Apr 2001 10:42:48 +0200
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <m14sHq9-007M9DC@smtp.web.de> (jh@web.de)
References: <m14sHq9-007M9DC@smtp.web.de>
Message-ID: <200104260842.f3Q8gmv01500@mira.informatik.hu-berlin.de>

> Maybe the bend-o-matic could get a safety switch?

Very difficult. We could add a warning, using the Python 2.1 warning
framework, but that would probably not satisfy the users relying on
the feature. So the best we can do is to document that an empty
separator is allowed but not recommended.

Regards,
Martin



From pyxml@xhaus.com  Thu Apr 26 10:04:45 2001
From: pyxml@xhaus.com (Alan Kennedy)
Date: Thu, 26 Apr 2001 10:04:45 +0100
Subject: [XML-SIG] Exception handling with xml.dom.ext.reader.HtmlLib
In-Reply-To: <jUsT.aNoTheR.mEsSaGe.iD.98824278431156@ideaeng.com>
Message-ID: <LPBBKIJMCOEOJAFDNMFEOEGACFAA.pyxml@xhaus.com>

Mark,

I'm currently working on an application that needs to parse HTML, and have
looked at htmllib as a way to do it.

However, htmllib seems to only parse HTML 2.0, and as you have pointed out,
is not very tolerant of the structural errors that typify a lot of HTML
pages.

One avenue I'm currently investigating is to use Dave Raggetts TIDY program,
which takes a messy HTML file and outputs a cleaned up version, i.e. tags
rebalanced, attributes quoted, etc, etc. It also has some support for XML
and XHTML. While this support is not complete, it is very good.

You can find tidy at

http://www.w3.org/People/Raggett/tidy/

This program is written in C, so it should be possible to use it directly
from Python. The documentation for Tidy mentions that someone has done a
SWIG interface for it.

There is also a Java version, which could be used from Jython fairly easily.

I had a look at the code to see if it might be feasible to insert some hooks
into it to turn it into a generator of SAX events, but the code is quite
messy, and the printing/output works at a character and buffer level, not an
element and attribute level.

Alan.



From eliot@isogen.com  Thu Apr 26 21:39:28 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Thu, 26 Apr 2001 15:39:28 -0500
Subject: [XML-SIG] Processing of External Unparsed Entities
Message-ID: <3AE88780.8A08A119@isogen.com>

The code for processing external unparsed entities appears to be broken
in the current code, both in DOM1 and DOM2 (I am currently using DOM1 to
do development). Has anyone exercised this or tested it? Should it be
working? I'm looking at the latest code off the SourceForge CVS tree.

I've tried to trace through and get it to work, but I got as far as the
unparsedEntityDecl() method of XmlDomGenerator (in Sax.py). It was being
called with too many arguments, and the code as written didn't have a
"name" attribute, which an entity declaration obviously must have.

Fixing that, I then discovered that the DOM implementation object was an
HTMLDOMImplementation, which doesn't seem right (but maybe it's my
error)? Anyway, at that point, the call to createEntity() failed because
HTMLDOMImplementation has no such method.

I have only a dim understanding of how the Python DOM code works, but
I'm happy to take a stab at fixing it as I must be able to processing
external data entities in order to implement HyTime address resolution
in a DOM processing context.

Thanks,

Eliot
-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From eliot@isogen.com  Thu Apr 26 23:06:40 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Thu, 26 Apr 2001 17:06:40 -0500
Subject: [XML-SIG] Processing of External Unparsed Entities
References: <3AE88780.8A08A119@isogen.com>
Message-ID: <3AE89BF0.E4B1D86F@isogen.com>

"W. Eliot Kimber" wrote:
> 
> The code for processing external unparsed entities appears to be broken
> in the current code, both in DOM1 and DOM2 (I am currently using DOM1 to
> do development). Has anyone exercised this or tested it? Should it be
> working? I'm looking at the latest code off the SourceForge CVS tree.

Ok, I got the entity creation to work--there were several things that
clearly had never been tested because they couldn't have possibly ever
worked--for shame.

Can someone point me to info on how to run the regression tests?

Thanks,

Eliot
-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From eliot@isogen.com  Fri Apr 27 02:16:18 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Thu, 26 Apr 2001 20:16:18 -0500
Subject: [XML-SIG] Processing of External Unparsed Entities
References: <3AE88780.8A08A119@isogen.com>
Message-ID: <3AE8C862.41F12F26@isogen.com>

"W. Eliot Kimber" wrote:
> 
> The code for processing external unparsed entities appears to be broken
> in the current code, both in DOM1 and DOM2 (I am currently using DOM1 to
> do development). Has anyone exercised this or tested it? Should it be
> working? I'm looking at the latest code off the SourceForge CVS tree.

I had to do more work to get all the entity and notation processing for
unparsed entities working correctly with the Sax2 parser and drivers. (I
haven't tried to fix the Sax 1 stuff because it doesn't correctly handle
resolving relative system ID paths, so I can't use it anyway.) I also
had to fix attribute lookup using getNamedItem on Element.

I've touched the following files:

dom/DOMImplementation.py
dom/Document.py
dom/DocumentType.py
dom/Element.py
dom/Entity.py
dom/ext/reader/Sax2.py

The problems I found were:

- Orphaned unparsed entity and notation declarations were not getting
properly set with the owning document.
- Added createEntity to XmlDomGenerator (it was being called from Sax
driver but hadn't been implemented)
- Unparsed entities were not be constructed with a useful nodeName
value.
- lookup of attributes on NamedNodeList failed when using getAttribute()
because in Sax2 the keys are (Uri, name) but the code was only passing
in name. I hacked it by passing in (None, name) from getAttribute on
Element.

I have no idea if any of these fixes are appropriate, I just hacked it
until my document parsed and my particular application got the data it
wanted.

Please let me know how I should proceed with testing these changes and
contributing them back.

Thanks,

Eliot
-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From eliot@isogen.com  Fri Apr 27 16:16:46 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Fri, 27 Apr 2001 10:16:46 -0500
Subject: [XML-SIG] Another Sax2 Enhancement: dataSource
Message-ID: <3AE98D5E.477205D1@isogen.com>

I need to be able to get from a DOM the filename or URI of the file it
was constructed from (if it was constructed from a file). I didn't see
anything in the code that preserved this data, so I've hacked my local
copy of the code to add an optional "dataSource" parameter to the
FromFile/URI/Stream methods that propogates to the Sax reader. This is
set as the friend property _dataSource on the resulting DOM node.

Is this an appropriate solution?

I need this because uparsed entity resolution does not preserve the
absolute path of the entities, only the system ID as specified in the
declaration. As I am using relative paths, I must have the path of the
declaring document in order to be able to construct new DOMs from
referenced XML document entities.

Unfortunately, the DOM spec does not provide a property on Entity to
hold the fully-resolved system ID so I think my approach is the only
solution.

My business problem is processing a hyperdocument consisting of many
subordinate documents, where each document is declared as an unparsed
entity with a relative path (this hyperdocument is generated by an
automatic process and the intent is for the resulting package of
documents to be self contained so that it can be packaged and moved
without the need to rework any external identifiers or catalog files).

Thanks,

Eliot
-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From eliot@isogen.com  Fri Apr 27 16:35:58 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Fri, 27 Apr 2001 10:35:58 -0500
Subject: [XML-SIG] Question About PCDATA Content
Message-ID: <3AE991DE.6C2B9F87@isogen.com>

I have a question which may be a basic XML question, but I think is an
implementation question:

I have this data in my document:

<name>myname</name>

However, the text node that is the first child of the Element node
constructed from the name element returns 

'
myname'

(note the leading newline) as the value of the data property of the text
node.

Why is there that leading newline? It's not in the data anywhere and
there are certainly no rules for adding newlines to element content or
moving them around as there are in SGML.

Is this an implementation bug or have I forgotten something essential
about XML?

Thanks,

Eliot
-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From Alexandre.Fayolle@logilab.fr  Fri Apr 27 16:49:25 2001
From: Alexandre.Fayolle@logilab.fr (Alexandre Fayolle)
Date: Fri, 27 Apr 2001 17:49:25 +0200 (CEST)
Subject: [XML-SIG] Question About PCDATA Content
In-Reply-To: <3AE991DE.6C2B9F87@isogen.com>
Message-ID: <Pine.LNX.4.21.0104271747320.7169-100000@orion.logilab.fr>

On Fri, 27 Apr 2001, W. Eliot Kimber wrote:

> I have a question which may be a basic XML question, but I think is an
> implementation question:

I cannot reproduce your problem using 4Suite 0.11 (and the version of
4DOM that goes with it) with python 1.52 on a linux machine:

>>> s= '<name>myname</name>'
>>> from xml.dom.ext.reader import Sax2
>>> d = Sax2.FromXml(s)
>>> d.documentElement.firstChild.data
'myname'

Could you send a code sample and tell us what your platform is?

Alexandre Fayolle
-- 
http://www.logilab.com 
Narval is the first software agent available as free software (GPL).
LOGILAB, Paris (France).



From larsga@garshol.priv.no  Fri Apr 27 16:50:29 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 27 Apr 2001 17:50:29 +0200
Subject: [XML-SIG] Question About PCDATA Content
In-Reply-To: <3AE991DE.6C2B9F87@isogen.com>
References: <3AE991DE.6C2B9F87@isogen.com>
Message-ID: <m3n192qr3u.fsf@lambda.garshol.priv.no>

* W. Eliot Kimber
|
| I have a question which may be a basic XML question, but I think is an
| implementation question:
| 
| I have this data in my document:
| 
| <name>myname</name>
| 
| However, the text node that is the first child of the Element node
| constructed from the name element returns 
| 
| '
| myname'

This is a bug. There's nothing in XML that allows this to happen.
 
--Lars M.



From eliot@isogen.com  Fri Apr 27 17:04:49 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Fri, 27 Apr 2001 11:04:49 -0500
Subject: [XML-SIG] Question About PCDATA Content
References: <Pine.LNX.4.21.0104271747320.7169-100000@orion.logilab.fr>
Message-ID: <3AE998A1.2B9941C5@isogen.com>

Alexandre Fayolle wrote:
> 
> On Fri, 27 Apr 2001, W. Eliot Kimber wrote:
> 
> > I have a question which may be a basic XML question, but I think is an
> > implementation question:
> 
> I cannot reproduce your problem using 4Suite 0.11 (and the version of
> 4DOM that goes with it) with python 1.52 on a linux machine:

I'm running under Win2K, using the latest code from CVS as of yesterday
noon CDT with Linux 1.5.2

Thanks,

E.
-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From Alexandre.Fayolle@logilab.fr  Fri Apr 27 17:25:00 2001
From: Alexandre.Fayolle@logilab.fr (Alexandre Fayolle)
Date: Fri, 27 Apr 2001 18:25:00 +0200 (CEST)
Subject: [XML-SIG] Question About PCDATA Content (fwd)
Message-ID: <Pine.LNX.4.21.0104271822080.7169-100000@orion.logilab.fr>

I'm forwarding this to the list, since there may be people there who are
interested. 

Alexandre Fayolle
-- 
http://www.logilab.com 
Narval is the first software agent available as free software (GPL).
LOGILAB, Paris (France).

---------- Forwarded message ----------
Date: Fri, 27 Apr 2001 11:08:11 -0500
From: W. Eliot Kimber <eliot@isogen.com>
To: Alexandre Fayolle <alf@logilab.fr>
Subject: Re: [XML-SIG] Question About PCDATA Content

Alexandre Fayolle wrote:

> 
> Could you send a code sample and tell us what your platform is?

Here is the document I'm processing:

<?xml version="1.0"?>
<!DOCTYPE repository
  SYSTEM "repexport.dtd"
[
 <!ENTITY ThorDependencies
        SYSTEM "ThorDependencies\ThorDependencies.xml"
        NDATA sgml
>
<!ENTITY base
        SYSTEM "base\base.xml"
        NDATA sgml
>
<!ENTITY eliot.s_branch
        SYSTEM "eliot.s_branch\eliot.s_branch.xml"
        NDATA sgml
>
]>
<repository>
<objid>{Not Available in CORBA API}</objid>
<name>#DEFAULT</name>
<date>{Not Available in CORBA API}</date>
</repository>


I get '
#DEFAULT' from the name element.

Thanks,

Eliot
-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m



From martin@loewis.home.cs.tu-berlin.de  Fri Apr 27 18:10:31 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Fri, 27 Apr 2001 19:10:31 +0200
Subject: [XML-SIG] Processing of External Unparsed Entities
In-Reply-To: <3AE8C862.41F12F26@isogen.com> (eliot@isogen.com)
References: <3AE88780.8A08A119@isogen.com> <3AE8C862.41F12F26@isogen.com>
Message-ID: <200104271710.f3RHAV801062@mira.informatik.hu-berlin.de>

> I have no idea if any of these fixes are appropriate, I just hacked it
> until my document parsed and my particular application got the data it
> wanted.
> 
> Please let me know how I should proceed with testing these changes and
> contributing them back.

Hi Eliot,

To run the test suite, run regrtest.py in PyXML/test, and test.py in
test/dom. Ideally, you'd run them with and without your patches, since
not all tests may pass on your system (due to bugs in the test suite,
mostly).

Please submit your patches to sourceforge.net/projects/pyxml. Use
unified (-u) or context (-c) diffs, and copy your rationale for these
changes into the initial comment.

Thanks for contributing,
Martin



From martin@loewis.home.cs.tu-berlin.de  Fri Apr 27 18:18:40 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Fri, 27 Apr 2001 19:18:40 +0200
Subject: [XML-SIG] Processing of External Unparsed Entities
In-Reply-To: <3AE88780.8A08A119@isogen.com> (eliot@isogen.com)
References: <3AE88780.8A08A119@isogen.com>
Message-ID: <200104271718.f3RHIeG01213@mira.informatik.hu-berlin.de>

> The code for processing external unparsed entities appears to be broken
> in the current code, both in DOM1 and DOM2 (I am currently using DOM1 to
> do development). Has anyone exercised this or tested it? Should it be
> working?

It somewhat depends on the parser that you use, whether the SAX
readers report the proper events. I think it is fair to say that this
has seen little or no testing. So, even if you cannot contribute code
that fixes all the problems, contributing test cases is worthwhile.

> Fixing that, I then discovered that the DOM implementation object was an
> HTMLDOMImplementation, which doesn't seem right (but maybe it's my
> error)?

That is OK. In the DOM, an HTMLDOMImplementation is a
DOMImplementation that also supports HTML.

> Anyway, at that point, the call to createEntity() failed because
> HTMLDOMImplementation has no such method.

This is not surprising: There is no operation "createEntity" in the
DOM...

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Fri Apr 27 18:34:40 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Fri, 27 Apr 2001 19:34:40 +0200
Subject: [XML-SIG] Another Sax2 Enhancement: dataSource
In-Reply-To: <3AE98D5E.477205D1@isogen.com> (eliot@isogen.com)
References: <3AE98D5E.477205D1@isogen.com>
Message-ID: <200104271734.f3RHYei01291@mira.informatik.hu-berlin.de>

> I need to be able to get from a DOM the filename or URI of the file it
> was constructed from (if it was constructed from a file). I didn't see
> anything in the code that preserved this data, so I've hacked my local
> copy of the code to add an optional "dataSource" parameter to the
> FromFile/URI/Stream methods that propogates to the Sax reader. This is
> set as the friend property _dataSource on the resulting DOM node.
> 
> Is this an appropriate solution?

I would not think so. Are you sure you need this on every node?
That seems to be quite expensive for a rarely-used extension.

If you really only put the parameter to FromFile into the tree, isn't
putting it into the document sufficient?

But then, aren't you interested in the data sources of the elements
originating from an external parsed entity?

> I need this because uparsed entity resolution does not preserve the
> absolute path of the entities, only the system ID as specified in the
> declaration. As I am using relative paths, I must have the path of the
> declaring document in order to be able to construct new DOMs from
> referenced XML document entities.

I'm not sure I understand. Are you saying that the DOM requires to
store the relative path in Entity::systemId? Where does the spec say
so?

> My business problem is processing a hyperdocument consisting of many
> subordinate documents, where each document is declared as an unparsed
> entity with a relative path (this hyperdocument is generated by an
> automatic process and the intent is for the resulting package of
> documents to be self contained so that it can be packaged and moved
> without the need to rework any external identifiers or catalog files).

To solve this problem, isn't it sufficient to carry the document's
system ID along with the document, instead of putting it *into* the
document?

Regards,
Martin



From eliot@isogen.com  Fri Apr 27 19:28:01 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Fri, 27 Apr 2001 13:28:01 -0500
Subject: [XML-SIG] Another Sax2 Enhancement: dataSource
References: <3AE98D5E.477205D1@isogen.com> <200104271734.f3RHYei01291@mira.informatik.hu-berlin.de>
Message-ID: <3AE9BA31.201BF14A@isogen.com>

"Martin v. Loewis" wrote:
> 
> > I need to be able to get from a DOM the filename or URI of the file it
> > was constructed from (if it was constructed from a file). I didn't see
> > anything in the code that preserved this data, so I've hacked my local
> > copy of the code to add an optional "dataSource" parameter to the
> > FromFile/URI/Stream methods that propogates to the Sax reader. This is
> > set as the friend property _dataSource on the resulting DOM node.
> >
> > Is this an appropriate solution?
> 
> I would not think so. Are you sure you need this on every node?
> That seems to be quite expensive for a rarely-used extension.

It's only set on the document node, not every node in the DOM. I think
my comment above was misleading.
 
> If you really only put the parameter to FromFile into the tree, isn't
> putting it into the document sufficient?
> 
> But then, aren't you interested in the data sources of the elements
> originating from an external parsed entity?

Since I don't recognize the useful existence of external parsed
entities, I would never need to know anything about them :-)

Possibly, but since an external parsed entity would be relative to the
document that declares it, having the document's path and the entity's
relative path would be sufficient (as it is for unparsed entities).
 
> > I need this because uparsed entity resolution does not preserve the
> > absolute path of the entities, only the system ID as specified in the
> > declaration. As I am using relative paths, I must have the path of the
> > declaring document in order to be able to construct new DOMs from
> > referenced XML document entities.
> 
> I'm not sure I understand. Are you saying that the DOM requires to
> store the relative path in Entity::systemId? Where does the spec say
> so?

I don't think it explicitly requires it, but I would be very upset if a
DOM parser changed the value of the original system ID with no way to
get it back. I would consider that to be agregious destruction of
data--what if the system ID is a URN that I need to be able to
interrogate after the entity is resolved or what if I want to rewrite
the entity declarations as originally specified?

Thus I would not, personally suggest that the system ID value be the
fully-resolved location of the file.
 
> > My business problem is processing a hyperdocument consisting of many
> > subordinate documents, where each document is declared as an unparsed
> > entity with a relative path (this hyperdocument is generated by an
> > automatic process and the intent is for the resulting package of
> > documents to be self contained so that it can be packaged and moved
> > without the need to rework any external identifiers or catalog files).
> 
> To solve this problem, isn't it sufficient to carry the document's
> system ID along with the document, instead of putting it *into* the
> document?

I could, but then I'm requiring processing applications to do this, at
some significant cost in complexity. For example, in my code, I may be
receiving a node from any document in the hyperdocument. Thus I would
have to maintain a global mapping from document nodes to system IDs.
Doable, but why force all applications to do this when I should be able
to just ask the DOM "where did you come from?"

Given that the info is only stored on the document node, I think it's a
relatively small cost. It was certainly easier for me to patch the DOM
implementation then to implement my own dictionary (although not that
much easier). But now nobody else has to think about it.

Cheers,

E.

-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From martin@loewis.home.cs.tu-berlin.de  Fri Apr 27 19:38:38 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Fri, 27 Apr 2001 20:38:38 +0200
Subject: [XML-SIG] Another Sax2 Enhancement: dataSource
In-Reply-To: <3AE9BA31.201BF14A@isogen.com> (eliot@isogen.com)
References: <3AE98D5E.477205D1@isogen.com> <200104271734.f3RHYei01291@mira.informatik.hu-berlin.de> <3AE9BA31.201BF14A@isogen.com>
Message-ID: <200104271838.f3RIccF01642@mira.informatik.hu-berlin.de>

> Given that the info is only stored on the document node, I think it's a
> relatively small cost. It was certainly easier for me to patch the DOM
> implementation then to implement my own dictionary (although not that
> much easier). But now nobody else has to think about it.

Even easier would be to put the attribute into the Document object
without patching the DOM implementation:

root = FromXmlStream(foo)
root.dataSource = bar

This is Python; you can modify any instance's attributes (unless the
instance takes countermeasures).

IMO this is as clean, but simpler, than extending the DOM
implementation.

Regards,
Martin



From eliot@isogen.com  Fri Apr 27 19:53:26 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Fri, 27 Apr 2001 13:53:26 -0500
Subject: [XML-SIG] Another Sax2 Enhancement: dataSource
References: <3AE98D5E.477205D1@isogen.com> <200104271734.f3RHYei01291@mira.informatik.hu-berlin.de> <3AE9BA31.201BF14A@isogen.com> <200104271838.f3RIccF01642@mira.informatik.hu-berlin.de>
Message-ID: <3AE9C026.B7286C09@isogen.com>

"Martin v. Loewis" wrote:
> 
> > Given that the info is only stored on the document node, I think it's a
> > relatively small cost. It was certainly easier for me to patch the DOM
> > implementation then to implement my own dictionary (although not that
> > much easier). But now nobody else has to think about it.
> 
> Even easier would be to put the attribute into the Document object
> without patching the DOM implementation:
> 
> root = FromXmlStream(foo)
> root.dataSource = bar
> 
> This is Python; you can modify any instance's attributes (unless the
> instance takes countermeasures).

You're right. I've been so brainwashed by our use of CORBA on our
project that I forget that you can do that kind of thing. That would
certainly satisfy the requirement I have.

But I think that this is probably a requirement against the DOM spec
itself--that is, the notion of preserving the data source for DOMs is, I
think, a general requirement that all DOM implementations should
satisfy.

Cheers,

E.

-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From linudom@hotmail.com  Fri Apr 27 21:20:34 2001
From: linudom@hotmail.com (Dom Linu)
Date: Fri, 27 Apr 2001 20:20:34 -0000
Subject: [XML-SIG] Validating a DOM w/xmlproc?
Message-ID: <F558zDv60gxapzhad2V00004a50@hotmail.com>

Hi,

Quick question: I have been using xmlproc for validation with its 
XMLValidator parse_resource method and a filename to validate XML.  Now I 
find myself holding a DOM, and need to validate a DOM I have in memory (as 
opposed to a file).  What is the easiest way to have the DOM validated?

Pointers to past postings, TFM that covers this, or really obnoxious flames 
are all welcome!

Thanks,
dl
_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com



From gstein@lyra.org  Fri Apr 27 21:38:58 2001
From: gstein@lyra.org (Greg Stein)
Date: Fri, 27 Apr 2001 13:38:58 -0700
Subject: [XML-SIG] expat namespace_separator in Python 2.1
In-Reply-To: <003001c0cd2f$80210ae0$6401a8c0@EHUD>; from jtauber@jtauber.com on Tue, Apr 24, 2001 at 10:29:00PM -0400
References: <002001c0ccc9$54e28cc0$0401a8c0@VAIO> <15077.36393.731033.157812@cj42289-a.reston1.va.home.com> <m3u23e2n4x.fsf@lambda.garshol.priv.no> <003601c0cd10$57870090$ed020a0a@bowstreet.com> <004801c0cd12$f87cae80$7cac1218@reston1.va.home.com> <001201c0cd1a$86429c50$6401a8c0@EHUD> <3AE620BB.9FB13F18@zolera.com> <002301c0cd29$caa4b900$6401a8c0@EHUD> <3AE63624.9C763305@zolera.com> <003001c0cd2f$80210ae0$6401a8c0@EHUD>
Message-ID: <20010427133858.T1374@lyra.org>

Not just RDF, but WebDAV (RFC 2518) also works that way.

I *very* much agree with the concatenation mechanism. We also use it for
constructing property URIs in WebDAV.

So... we have two specs (and RFC and a REC) that use concatenation. IMO,
that means that pyexpat should support it. Plus the backwards compat thing.

Cheers,
-g

On Tue, Apr 24, 2001 at 10:29:00PM -0400, James Tauber wrote:
> 
> Sorry I wasn't more clear. It is only RDF that works this way. It is unusual
> but internally consistent.
> 
> James
> 
> > > You cut that part out of my email. It's RDF.
> >
> > Oh, it wasn't clear to me from your note.
> >
> > Thanks.
> 
> 
> _______________________________________________
> XML-SIG maillist  -  XML-SIG@python.org
> http://mail.python.org/mailman/listinfo/xml-sig

-- 
Greg Stein, http://www.lyra.org/


From eliot@isogen.com  Fri Apr 27 21:37:58 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Fri, 27 Apr 2001 15:37:58 -0500
Subject: [XML-SIG] Optimizing getElementsByTagName()
Message-ID: <3AE9D8A6.4635B154@isogen.com>

In my processing I'm using getElementsByTagName() a lot. I notice that
in the current implementation is recalculates the subelement list every
time you call it.

An obvious optimization would be to cache lists by tagName once
calculated.

Is this an appropriate optimization? It would have some memory cost.

Using getElementsByTagName() is very convenient in my business logic,
but I could replace it with my own iteration over child nodes. But
normally I would presume that these sorts of queries would be optimized
by the underlying implementation. Here they are not.

Any other thoughts on optimization for this method? I'll implement the
cache approach and see how it works.

Thanks,

E.
-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From eliot@isogen.com  Fri Apr 27 21:51:55 2001
From: eliot@isogen.com (W. Eliot Kimber)
Date: Fri, 27 Apr 2001 15:51:55 -0500
Subject: [XML-SIG] Optimizing getElementsByTagName()
References: <3AE9D8A6.4635B154@isogen.com>
Message-ID: <3AE9DBEB.E3467B0D@isogen.com>

"W. Eliot Kimber" wrote:
> 
> In my processing I'm using getElementsByTagName() a lot. I notice that
> in the current implementation is recalculates the subelement list every
> time you call it.
> 
> An obvious optimization would be to cache lists by tagName once
> calculated.
> 
> Is this an appropriate optimization? It would have some memory cost.

I thought about it some more and realized that it's probably not worth
the cost, especially when you factor in tracking mutations of the DOM,
which would invalidate the caches. Even in my own code, I was
constructing new lists much more than I was using cached lists.

Sorry for the bother.

Cheers,

E.

-- 
. . . . . . . . . . . . . . . . . . . . . . . .

W. Eliot Kimber | Lead Brain

1016 La Posada Dr. | Suite 240 | Austin TX  78752
    T 512.656.4139 |  F 512.419.1860 | eliot@isogen.com

w w w . d a t a c h a n n e l . c o m


From martin@loewis.home.cs.tu-berlin.de  Fri Apr 27 22:22:03 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Fri, 27 Apr 2001 23:22:03 +0200
Subject: [XML-SIG] Another Sax2 Enhancement: dataSource
In-Reply-To: <3AE9C026.B7286C09@isogen.com> (eliot@isogen.com)
References: <3AE98D5E.477205D1@isogen.com> <200104271734.f3RHYei01291@mira.informatik.hu-berlin.de> <3AE9BA31.201BF14A@isogen.com> <200104271838.f3RIccF01642@mira.informatik.hu-berlin.de> <3AE9C026.B7286C09@isogen.com>
Message-ID: <200104272122.f3RLM3602240@mira.informatik.hu-berlin.de>

> But I think that this is probably a requirement against the DOM spec
> itself--that is, the notion of preserving the data source for DOMs is, I
> think, a general requirement that all DOM implementations should
> satisfy.

That may well be the case. So you might report an issue to
www-dom@w3.org (I hope I got the address right). If such an attribute
becomes official, a patch adding it to the Python DOM implementations
would be certainly appreciated (somebody will probably implement it
when nobody else does).

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Fri Apr 27 22:33:16 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Fri, 27 Apr 2001 23:33:16 +0200
Subject: [XML-SIG] Validating a DOM w/xmlproc?
In-Reply-To: <F558zDv60gxapzhad2V00004a50@hotmail.com> (linudom@hotmail.com)
References: <F558zDv60gxapzhad2V00004a50@hotmail.com>
Message-ID: <200104272133.f3RLXGC02360@mira.informatik.hu-berlin.de>

> Quick question: I have been using xmlproc for validation with its
> XMLValidator parse_resource method and a filename to validate XML.
> Now I find myself holding a DOM, and need to validate a DOM I have
> in memory (as opposed to a file).  What is the easiest way to have
> the DOM validated?

I guess the easiest way is to linearise it, then parse it again. If it
is 4DOM, you can use xml.dom.ext.Printer for the linearisation.

Regards,
Martin


From uche.ogbuji@fourthought.com  Sat Apr 28 00:57:34 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Fri, 27 Apr 2001 17:57:34 -0600
Subject: [XML-SIG] Possible bug in sgmlop (charcter entityrefs in attributes)
Message-ID: <200104272357.f3RNvYv31305@borgia.local.dhcp.fourthought.com>

Seems that sgmlop.c does not handle character entities refs in attributes.  If 
the following HTML is parsed using 4DOM's HtmlLib reader, which uses 
Sgmlop.py, which uses sgmlop.c from PyXML, the handle_entiryref call-back is 
not being invoked for the "&Auml;" in the attribute, although it is working 
for the one in regular CDATA.

<HTML>
  <HEAD>
    <TITLE>Character entities in attributes</TITLE>
  </HEAD>
  <BODY>
    <FORM>
      &Auml;<INPUT VALUE='&Auml;RGERLICH!' NAME='murx' TYPE='text'>
    </FORM>
  </BODY>
</HTML>

I don't know enough about sgmlop to easily sort this out myself.

Any ideas?

Thanks


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From stuff4gary@hotmail.com  Sat Apr 28 01:20:18 2001
From: stuff4gary@hotmail.com (gary cor)
Date: Sat, 28 Apr 2001 00:20:18
Subject: [XML-SIG] Fool-proof XML examples, additional tutorials?
Message-ID: <F151CIqtQvAauNjVuMh0000d017@hotmail.com>

Hello all,

I am very impressed with the SAX2 search utility which I found on the latest 
python howto, I am searching and building XML again and again!!!  However, I 
seem to have hit a wall doing anything else but searching for XML bits in 
python and the DOM howto part is far to unambitious as well!!!

Listening to you guys go on about projects sounds very promissing but is 
also difficult to follow at times.  Sure, if you make python XML foolproof 
then only fools would use it , but I think there is the danger of some 
intellectual snobery as well!!  A few more people need to contribute to some 
good tutorials?.. Play safe as there will always the need for a wide variety 
of simple scripting methods for XML and python could rule there!   Possibly, 
full XML applications at there limits are better supported in Java anyway?

Gary
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.



From rsalz@zolera.com  Sat Apr 28 02:51:51 2001
From: rsalz@zolera.com (Rich Salz)
Date: Fri, 27 Apr 2001 21:51:51 -0400
Subject: [XML-SIG] Validating a DOM w/xmlproc?
References: <F558zDv60gxapzhad2V00004a50@hotmail.com> <200104272133.f3RLXGC02360@mira.informatik.hu-berlin.de>
Message-ID: <3AEA2237.DA5F01BF@zolera.com>

> I guess the easiest way is to linearise it, then parse it again. If it
> is 4DOM, you can use xml.dom.ext.Printer for the linearisation.

interesting question:  can you invoke DOM methods such that the
resultant structure is invalid?  Off the top of my head, I don't know.


From rsalz@zolera.com  Sat Apr 28 03:32:45 2001
From: rsalz@zolera.com (Rich Salz)
Date: Fri, 27 Apr 2001 22:32:45 -0400
Subject: [XML-SIG] Proposal for "xml.namespaces"
Message-ID: <3AEA2BCD.D2742E18@zolera.com>

Mike Olson suggested I bring this up.

I'd like to create "namespaces.py" that has the URI strings for various
namespaces.  Something like
	class SOAP:
		ENV = "http://schemas.xmlsoap.org/soap/envelope/"
		ENC = "http://schemas.xmlsoap.org/soap/encoding/"

	class XSD:
		v1999 = "http://www.w3.org/1999/XMLSchema"
		v2000 = "http://www.w3.org/2000/10/XMLSchema"
		v2001 = "http://www.w3.org/2001/XMLSchema"
		ALL = [ v1999, v2000, v2001 ]

etc.  Should it be "ns.py"?  That filename is somewhat terse, but being
able to write things like xml.ns.XSD.v2001, xml.ns.DSIG.RSA, etc., seems
like a good thing.
	/r$


From uche.ogbuji@fourthought.com  Sat Apr 28 04:20:01 2001
From: uche.ogbuji@fourthought.com (Uche Ogbuji)
Date: Fri, 27 Apr 2001 21:20:01 -0600
Subject: [XML-SIG] Optimizing getElementsByTagName()
In-Reply-To: Message from "W. Eliot Kimber" <eliot@isogen.com>
 of "Fri, 27 Apr 2001 15:37:58 CDT." <3AE9D8A6.4635B154@isogen.com>
Message-ID: <200104280320.f3S3K1F32244@borgia.local>

> In my processing I'm using getElementsByTagName() a lot. I notice that
> in the current implementation is recalculates the subelement list every
> time you call it.

This is just a naive way to satisfy at least a portion of the (silly, IMHO) 
liveness requirements of the DOM.

> An obvious optimization would be to cache lists by tagName once
> calculated.
> 
> Is this an appropriate optimization? It would have some memory cost.

I pretty much don't know how to *spell* getElementsbyTagName anymore.

My optimization for this is to use XPath.

I know it might not be suitable for your needs, but just a thought...


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python




From martin@loewis.home.cs.tu-berlin.de  Sat Apr 28 07:48:52 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Sat, 28 Apr 2001 08:48:52 +0200
Subject: [XML-SIG] Removing the Wise directory
Message-ID: <200104280648.f3S6mqx00941@mira.informatik.hu-berlin.de>

Is anybody still interested in the Wise directory of PyXML? If not,
I'd like to remove it from the package.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Sat Apr 28 07:33:15 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Sat, 28 Apr 2001 08:33:15 +0200
Subject: [XML-SIG] Fool-proof XML examples, additional tutorials?
In-Reply-To: <F151CIqtQvAauNjVuMh0000d017@hotmail.com>
 (stuff4gary@hotmail.com)
References: <F151CIqtQvAauNjVuMh0000d017@hotmail.com>
Message-ID: <200104280633.f3S6XF500890@mira.informatik.hu-berlin.de>

> A few more people need to contribute to some good tutorials?

Contributions of documentation is certainly welcome!

Please be aware that this is a volunteer's project, though: People
contribute what they have fun contributing; which is not necessarily
what users want most. It turns out that only few people have fun
writing documentation; this is why getting contributions of
documentation is so hard, and happens rarely.

Regards,
Martin


From martin@loewis.home.cs.tu-berlin.de  Sat Apr 28 07:47:13 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Sat, 28 Apr 2001 08:47:13 +0200
Subject: [XML-SIG] Proposal for "xml.namespaces"
In-Reply-To: <3AEA2BCD.D2742E18@zolera.com> (message from Rich Salz on Fri, 27
 Apr 2001 22:32:45 -0400)
References: <3AEA2BCD.D2742E18@zolera.com>
Message-ID: <200104280647.f3S6lDt00938@mira.informatik.hu-berlin.de>

> Should it be "ns.py"?  That filename is somewhat terse, but being
> able to write things like xml.ns.XSD.v2001, xml.ns.DSIG.RSA, etc., seems
> like a good thing.

I don't care much about that. What is wrong with defining symbolic
names for these in your application?

However, *if* these is an addition of such a module, I would require
that a module documentation goes with it - or else nobody but the
author of the module will ever make use of it.

Regards,
Martin



From martin@loewis.home.cs.tu-berlin.de  Sat Apr 28 07:45:16 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Sat, 28 Apr 2001 08:45:16 +0200
Subject: [XML-SIG] Validating a DOM w/xmlproc?
In-Reply-To: <3AEA2237.DA5F01BF@zolera.com> (message from Rich Salz on Fri, 27
 Apr 2001 21:51:51 -0400)
References: <F558zDv60gxapzhad2V00004a50@hotmail.com> <200104272133.f3RLXGC02360@mira.informatik.hu-berlin.de> <3AEA2237.DA5F01BF@zolera.com>
Message-ID: <200104280645.f3S6jGf00936@mira.informatik.hu-berlin.de>

> interesting question: can you invoke DOM methods such that the
> resultant structure is invalid?  Off the top of my head, I don't
> know.

Certainly. Suppose you have the DTD

<!ELEMENT foo EMPTY>

Still, it is certainly possible to do

>>> d=imp.createDocument(None,"foo",None)
>>> e=d.createElement("bar")
>>> d.documentElement.appendChild(e)
>>> d.toxml()
'<?xml version="1.0" ?>\n<foo><bar/></foo>'

How could the DOM implementation possibly know what DTD I had in mind?
I could have provided a DocumentType node also, but which of these
operations should then have failed, and how? Answer: the DOM does not
care about validity. 

During modification operations, it might well happen that the document
becomes invalid, eg. when removing one node and inserting a different
one elsewhere. If the DOM would reject such modifications, it would be
useless.

Regards,
Martin



From martin@loewis.home.cs.tu-berlin.de  Sat Apr 28 08:02:42 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Sat, 28 Apr 2001 09:02:42 +0200
Subject: [XML-SIG] Unicode literals in PyXML
Message-ID: <200104280702.f3S72g401019@mira.informatik.hu-berlin.de>

There are currently two files that use Unicode literals in PyXML;
xml.utils.characters, and xml.schema.trex. Unfortunately, this means
that distutils installation of PyXML fails under Python 1.5 - even
though the rest of the package works fine with 1.5.

There may be distutils tricks to avoid these problems, but I'm also
looking for a way to change the sources so that they will install with
Python 1.5 - getting an error if anybody uses them is acceptable.

For xml.utils.characters, wrapping each literal with a unicode() call
is probably safe; I'll best use UTF-16BE for these strings.

As for xml.schema.trex, I'm not sure what the possibilities are. It
*seems* like one could replace all Unicode literals with plain string
literals, but I may be wrong.

Regards,
Martin




From linudom@hotmail.com  Sat Apr 28 08:32:55 2001
From: linudom@hotmail.com (Dom Linu)
Date: Sat, 28 Apr 2001 07:32:55 -0000
Subject: [XML-SIG] Validating a DOM w/xmlproc?
Message-ID: <F173FRaqbC8wSA1XvpF00004723@hotmail.com>

<html><DIV>
<P>Yep, in this case the DOM is assembled by parsing a flat file and creating Nodes using 4DOM.&nbsp; It is then handed off to another process that wishes to validate it with xmlproc-- I guess it will have to "reconstitute" it to a form xmlproc can deal with-- I've played with using the InputSourceFactory interface&nbsp;and the parser's set_inputsource_factory in xmlproc to return a custom file-like object-- but then I'd have to wrap the DOM up like a file-- probably about as much work as turning it back into text.</P>
<P>Thanks for the input!!<BR><BR></P></DIV>
<DIV></DIV>
<DIV></DIV>&gt;From: "Martin v. Loewis" <MARTIN@LOEWIS.HOME.CS.TU-BERLIN.DE>
<DIV></DIV>&gt;To: rsalz@zolera.com 
<DIV></DIV>&gt;CC: linudom@hotmail.com, xml-sig@python.org 
<DIV></DIV>&gt;Subject: Re: [XML-SIG] Validating a DOM w/xmlproc? 
<DIV></DIV>&gt;Date: Sat, 28 Apr 2001 08:45:16 +0200 
<DIV></DIV>&gt; 
<DIV></DIV>&gt; &gt; interesting question: can you invoke DOM methods such that the 
<DIV></DIV>&gt; &gt; resultant structure is invalid? Off the top of my head, I don't 
<DIV></DIV>&gt; &gt; know. 
<DIV></DIV>&gt; 
<DIV></DIV>&gt;Certainly. Suppose you have the DTD 
<DIV></DIV>&gt; 
<DIV></DIV>&gt;<!ELEMENT foo EMPTY> 
<DIV></DIV>&gt; 
<DIV></DIV>&gt;Still, it is certainly possible to do 
<DIV></DIV>&gt; 
<DIV></DIV>&gt; &gt;&gt;&gt; d=imp.createDocument(None,"foo",None) 
<DIV></DIV>&gt; &gt;&gt;&gt; e=d.createElement("bar") 
<DIV></DIV>&gt; &gt;&gt;&gt; d.documentElement.appendChild(e) 
<DIV></DIV>&gt; &gt;&gt;&gt; d.toxml() 
<DIV></DIV>&gt;'<?xml version="1.0" ?>\n<FOO><BAR /></FOO>' 
<DIV></DIV>&gt; 
<DIV></DIV>&gt;How could the DOM implementation possibly know what DTD I had in mind? 
<DIV></DIV>&gt;I could have provided a DocumentType node also, but which of these 
<DIV></DIV>&gt;operations should then have failed, and how? Answer: the DOM does not 
<DIV></DIV>&gt;care about validity. 
<DIV></DIV>&gt; 
<DIV></DIV>&gt;During modification operations, it might well happen that the document 
<DIV></DIV>&gt;becomes invalid, eg. when removing one node and inserting a different 
<DIV></DIV>&gt;one elsewhere. If the DOM would reject such modifications, it would be 
<DIV></DIV>&gt;useless. 
<DIV></DIV>&gt; 
<DIV></DIV>&gt;Regards, 
<DIV></DIV>&gt;Martin 
<DIV></DIV>&gt; 
<DIV></DIV>&gt; 
<DIV></DIV>&gt;_______________________________________________ 
<DIV></DIV>&gt;XML-SIG maillist - XML-SIG@python.org 
<DIV></DIV>&gt;http://mail.python.org/mailman/listinfo/xml-sig 
<DIV></DIV><br clear=all><hr>Get your FREE download of MSN Explorer at <a href="http://explorer.msn.com">http://explorer.msn.com</a><br></p></html>


From martin@loewis.home.cs.tu-berlin.de  Sat Apr 28 08:43:25 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Sat, 28 Apr 2001 09:43:25 +0200
Subject: [XML-SIG] Yet another 0.6.x release
Message-ID: <200104280743.f3S7hPS01431@mira.informatik.hu-berlin.de>

It appears that we have to make another release of PyXML 0.6, to
provide patches that are requires by 4Suite 0.11. To prepare this
release, I have created a CVS branch "o6maint", starting from v065,
which should include bug fixes only. IOW, all the files that have been
added to PyXML since 0.6.5 will *not* appear in 0.6.6.

Regards,
Martin


From rnd@onego.ru  Sat Apr 28 10:52:37 2001
From: rnd@onego.ru (Roman Suzi)
Date: Sat, 28 Apr 2001 13:52:37 +0400 (MSD)
Subject: [XML-SIG] Fool-proof XML examples, additional tutorials?
Message-ID: <Pine.LNX.4.30.0104281352210.17566-100000@rnd.onego.ru>

>Hello all,


>I am very impressed with the SAX2 search utility which I found on the latest
>python howto, I am searching and building XML again and again!!!  However, I
>seem to have hit a wall doing anything else but searching for XML bits in
>python and the DOM howto part is far to unambitious as well!!!
>
>Listening to you guys go on about projects sounds very promissing but is
>also difficult to follow at times.  Sure, if you make python XML foolproof
>then only fools would use it

I am not experienced in XML and have not used it for more than play,
but I feel I need it in some places.

XML is a data model, as strings are for unmarked text. But XML
has more dimensions than string and more complex structure,
allowing it to model much more complex things. (while XML
code could could be represented by strings).

And look, string type has a
- built in syntactic support
- after decades of development, string operations are more or less
determined (find, sub, strip, +, ...).

After awhile, XML will no more complex to handle than strings or lists or
arrays, when it will be clear what is the right way. (Look at Numeric
Python for examples that multidimensional arrays could be very versatile).

For now, operations on XML are organized in several ways and that is
fixed with all that thingies: DOM, SAX, XSLT, XPath, XQL, ...

And XML-work is (IMHO) more complex than SQL one (due to less
regular (richer) internal structure than that of relational tables.)

So, for now we must experiment, and wait when X* will become as foolproof,
as relational databases are today ;-) Do you remember pre-SQL DBs?
FoxBase, for example?


Roman.

P.S. I am no specialist in XML, so probably I have some points wrong. But
its my opinion for now.




From rsalz@zolera.com  Sat Apr 28 13:15:09 2001
From: rsalz@zolera.com (Rich Salz)
Date: Sat, 28 Apr 2001 08:15:09 -0400
Subject: [XML-SIG] Proposal for "xml.namespaces"
References: <3AEA2BCD.D2742E18@zolera.com> <200104280647.f3S6lDt00938@mira.informatik.hu-berlin.de>
Message-ID: <3AEAB44D.B84CD471@zolera.com>

> I don't care much about that. What is wrong with defining symbolic
> names for these in your application?

For the same reason people put #define's in header files.
 
> However, *if* these is an addition of such a module, I would require
> that a module documentation goes with it

Sure.  What format would you like?
	/r$


From fdrake@acm.org  Sat Apr 28 13:54:34 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Sat, 28 Apr 2001 08:54:34 -0400 (EDT)
Subject: [XML-SIG] Proposal for "xml.namespaces"
In-Reply-To: <3AEAB44D.B84CD471@zolera.com>
References: <3AEA2BCD.D2742E18@zolera.com>
 <200104280647.f3S6lDt00938@mira.informatik.hu-berlin.de>
 <3AEAB44D.B84CD471@zolera.com>
Message-ID: <15082.48522.843043.135280@cj42289-a.reston1.va.home.com>

Rich Salz writes:
 > For the same reason people put #define's in header files.

  I think the idea is pretty good, but am unsure if its better to put
the namespace constants in a separate module of it the constants
should be defined along with the API for the relevant classes.

 > Sure.  What format would you like?

  If you were to provide docstrings that include links to the
specifications that define the namespaces, I'll be glad to create the
formal documentation in LaTeX.


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From martin@loewis.home.cs.tu-berlin.de  Sat Apr 28 15:38:31 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Sat, 28 Apr 2001 16:38:31 +0200
Subject: [XML-SIG] Proposal for "xml.namespaces"
In-Reply-To: <3AEAB44D.B84CD471@zolera.com> (message from Rich Salz on Sat, 28
 Apr 2001 08:15:09 -0400)
References: <3AEA2BCD.D2742E18@zolera.com> <200104280647.f3S6lDt00938@mira.informatik.hu-berlin.de> <3AEAB44D.B84CD471@zolera.com>
Message-ID: <200104281438.f3SEcVG03306@mira.informatik.hu-berlin.de>

> > However, *if* these is an addition of such a module, I would require
> > that a module documentation goes with it
> 
> Sure.  What format would you like?

A Python module documentation file, in the format of the Python
library documentation (ie TeX):

http://python.sourceforge.net/maint-docs/doc/doc.html

Regards,
Martin


From Mike.Olson@fourthought.com  Sat Apr 28 20:18:31 2001
From: Mike.Olson@fourthought.com (Mike Olson)
Date: Sat, 28 Apr 2001 13:18:31 -0600
Subject: [XML-SIG] 4Suite and 4Suite Server second alpha release
Message-ID: <3AEB1787.4A21CB5A@FourThought.com>

All,

  One more alpha release before we call it 0.11 final.

  Thank you all of the bug reports and comments.

  One note, we have removed PyXML from the 4Suite package.  PyXML 0.6.5
is now required.

  Improvements since the last alpha:

  * Tested 4ODS with Dbm Driver
  * General Windows fixes for 4SuiteServer
  * Documentation improvements
  * Nasty bug in 4ss agent
  * More through test harnesses of 4Suite Server
  * Removed PyXML from 4Suite package
  * Small bug fixes here and there

You can download the new packages at:

http://4Suite.org/download.html


Mike


-- 
Mike Olson				 Principal Consultant
mike.olson@fourthought.com               (303)583-9900 x 102
Fourthought, Inc.                         http://Fourthought.com 
Software-engineering, knowledge-management, XML, CORBA, Linux, Python


From fdrake@acm.org  Sun Apr 29 02:45:20 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Sat, 28 Apr 2001 21:45:20 -0400 (EDT)
Subject: [XML-SIG] Validating a DOM w/xmlproc?
In-Reply-To: <3AEA2237.DA5F01BF@zolera.com>
References: <F558zDv60gxapzhad2V00004a50@hotmail.com>
 <200104272133.f3RLXGC02360@mira.informatik.hu-berlin.de>
 <3AEA2237.DA5F01BF@zolera.com>
Message-ID: <15083.29232.579127.635176@cj42289-a.reston1.va.home.com>

Rich Salz writes:
 > interesting question:  can you invoke DOM methods such that the
 > resultant structure is invalid?  Off the top of my head, I don't know.

  Absolutely!  Whether or not any given change will do so depends on
the change and the schema.
  Now, if we're talking about well-formedness, I think that's
guaranteed to hold if the DOM implementation isn't seriously buggy.
But a non-well-formed DOM tree just doesn't make conceptual sense --
that the document is a tree is just about all well-formedness
requires!


  -Fred

-- 
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From larsga@garshol.priv.no  Sun Apr 29 11:47:51 2001
From: larsga@garshol.priv.no (Lars Marius Garshol)
Date: 29 Apr 2001 12:47:51 +0200
Subject: [XML-SIG] Validating a DOM w/xmlproc?
In-Reply-To: <F558zDv60gxapzhad2V00004a50@hotmail.com>
References: <F558zDv60gxapzhad2V00004a50@hotmail.com>
Message-ID: <m3oftgvv6w.fsf@lambda.garshol.priv.no>

* Dom Linu
| 
| Quick question: I have been using xmlproc for validation with its
| XMLValidator parse_resource method and a filename to validate XML.
| Now I find myself holding a DOM, and need to validate a DOM I have
| in memory (as opposed to a file).  What is the easiest way to have
| the DOM validated?

I've been wanting to modify xmlproc so that the validator can operate
without the XML parser itself. They are already separated to some
extent, but the validator is not yet up to only validating an event
stream. 

Once that is done I would like to make a validating SAX filter. What
you could then do would be to use a DOM2SAX event generator and pipe
the events through the validator.

This is not yet done, but it's something I hope that either I or
someone else can do in the not too distant future.

--Lars M.
 



From rsalz@zolera.com  Sun Apr 29 17:11:31 2001
From: rsalz@zolera.com (Rich Salz)
Date: Sun, 29 Apr 2001 12:11:31 -0400
Subject: [XML-SIG] Validating a DOM w/xmlproc?
References: <F558zDv60gxapzhad2V00004a50@hotmail.com>
 <200104272133.f3RLXGC02360@mira.informatik.hu-berlin.de>
 <3AEA2237.DA5F01BF@zolera.com> <15083.29232.579127.635176@cj42289-a.reston1.va.home.com>
Message-ID: <3AEC3D33.113AAA7D@zolera.com>

>   Absolutely!  Whether or not any given change will do so depends on
> the change and the schema.
>   Now, if we're talking about well-formedness

Gaargh, that's what I meant.  Sorry for the confusion.
	/r$


From donaldallingham@home.com  Mon Apr 30 02:47:26 2001
From: donaldallingham@home.com (Don Allingham)
Date: Sun, 29 Apr 2001 19:47:26 -0600
Subject: [XML-SIG] problems reading iso-8859-1 data
Message-ID: <3AECC42E.1040802@home.com>

I hope this is the correct spot to post this question.

I have an XML file with iso-8859-1 encoding. The sax parser (expat) 
seems to translating charaters above 128 to to separate characters. 
For example "é" in the xml file is being interpreted as "é" by the 
parser. (I'm running python 1.5.2 with PyXML 0.6.5)

Am I missing something obvious?

-- Don Allingham
  donaldallingham@home.com 



From fdrake@acm.org  Mon Apr 30 03:05:59 2001
From: fdrake@acm.org (Fred L. Drake, Jr.)
Date: Sun, 29 Apr 2001 22:05:59 -0400 (EDT)
Subject: [XML-SIG] problems reading iso-8859-1 data
In-Reply-To: <3AECC42E.1040802@home.com>
References: <3AECC42E.1040802@home.com>
Message-ID: <15084.51335.966370.432959@cj42289-a.reston1.va.home.com>

Don Allingham writes:
 > I have an XML file with iso-8859-1 encoding. The sax parser (expat)=20=

 > seems to translating charaters above 128 to to separate characters.=20=

 > For example "=E9" in the xml file is being interpreted as "=C3=A9" b=
y the=20
 > parser. (I'm running python 1.5.2 with PyXML 0.6.5)
 >=20
 > Am I missing something obvious?

  I'm not sure how obvious it is, but what you are missing is that the
expat output under Python 1.5.2 will always be UTF-8 encoded.  (Under
more recent versions of Python, Unicode strings are provided by
default, but UTF-8 can be requested if desired.)


  -Fred

--=20
Fred L. Drake, Jr.  <fdrake at acm.org>
PythonLabs at Digital Creations



From pyxml@xhaus.com  Mon Apr 30 09:24:12 2001
From: pyxml@xhaus.com (Alan Kennedy)
Date: Mon, 30 Apr 2001 09:24:12 +0100
Subject: [XML-SIG] problems reading iso-8859-1 data
In-Reply-To: <3AECC42E.1040802@home.com>
Message-ID: <LPBBKIJMCOEOJAFDNMFEEEJCCFAA.pyxml@xhaus.com>

Don,

Just a quick suggestion.

> I have an XML file with iso-8859-1 encoding. The sax
> parser (expat) seems to translating charaters above 128
> to to separate characters.
> For example "=E9" in the xml file is being interpreted as
> "=C3=A9" by the parser.
> (I'm running python 1.5.2 with PyXML 0.6.5)
>
> Am I missing something obvious?

Have you placed an encoding declaration at the top of your XML file, i.=
e.
something along the lines of

<?xml version=3D"1.0" encoding=3D"iso-8859-1"?>

I'm parsing iso-8859-1 files containing such characters without problem=
=2E

Just a suggestion.

Alan.



From jfk@informaticon.dk  Mon Apr 30 10:41:50 2001
From: jfk@informaticon.dk (=?iso-8859-1?Q?J=F8rgen=20Fr=F8jk=20Kj=E6rsgaard?=)
Date: Mon, 30 Apr 2001 11:41:50 +0200
Subject: [XML-SIG] problems reading iso-8859-1 data
References: <LPBBKIJMCOEOJAFDNMFEEEJCCFAA.pyxml@xhaus.com>
Message-ID: <3AED335E.AE1EBE13@informaticon.dk>

Alan Kennedy wrote:
> =

> Don,
> =

> Just a quick suggestion.
> =

> > I have an XML file with iso-8859-1 encoding. The sax
> > parser (expat) seems to translating charaters above 128
> > to to separate characters.
> > For example "=E9" in the xml file is being interpreted as
> > "=C3=A9" by the parser.
> > (I'm running python 1.5.2 with PyXML 0.6.5)
> >
> > Am I missing something obvious?

Expat always translates the parsed input to UTF-8 encoding. Python 2.0
handles this correctly but I'm not sure about Python 1.5.x as I've never
used it for XML processing.

> Have you placed an encoding declaration at the top of your XML file, i.=
e.
> something along the lines of
> =

> <?xml version=3D"1.0" encoding=3D"iso-8859-1"?>

This does not change the fact that Expat outputs UTF-8. However, if the
Expat parser hasn't been told to use iso-8859-1 as default encoding, it
will assume UTF-8 input unless you state the encoding in the input as
above.

/jfk

-- =

J=F8rgen Fr=F8jk Kj=E6rsgaard, Systemkonsulent (Systems Consultant)
Informaticon ApS * Web: www.informaticon.dk * Tlf: +45 8672 0093
Internet programmering * Systemudvikling p=E5 Linux, FreeBSD og PalmOS


From donaldallingham@home.com  Mon Apr 30 14:16:34 2001
From: donaldallingham@home.com (Don Allingham)
Date: 30 Apr 2001 07:16:34 -0600
Subject: [XML-SIG] problems reading iso-8859-1 data
In-Reply-To: <3AED335E.AE1EBE13@informaticon.dk>
References: <LPBBKIJMCOEOJAFDNMFEEEJCCFAA.pyxml@xhaus.com>
 <3AED335E.AE1EBE13@informaticon.dk>
Message-ID: <988636595.8711.5.camel@wallace>

--=-AotdqlOxl1oqQ3Q0ScLH
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

On 30 Apr 2001 11:41:50 +0200, J=F8rgen Fr=F8jk Kj=E6rsgaard wrote:
> Alan Kennedy wrote:
> >=20
> > <?xml version=3D"1.0" encoding=3D"iso-8859-1"?>
>=20
> This does not change the fact that Expat outputs UTF-8. However, if the
> Expat parser hasn't been told to use iso-8859-1 as default encoding, it
> will assume UTF-8 input unless you state the encoding in the input as
> above.
>=20

Yes, I have the encoding=3D"iso-8859-1" set the file.  It sounds as if
expat is converting the iso-8859-1 to UTF-8.  Is there a way to convert
back to iso-8859-1?  Unfortunately, I'm stuck with 1.5.2.

--=20
Don Allingham
donaldallingham@home.com=20

--=-AotdqlOxl1oqQ3Q0ScLH
Content-Type: text/html; charset=utf-8

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
  <META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
  <META NAME="GENERATOR" CONTENT="GtkHTML/0.9.3">
</HEAD>
<BODY><pre>On 30 Apr 2001 11:41:50 +0200, J&#248;rgen Fr&#248;jk Kj&#230;rsgaard wrote:
<FONT COLOR="#737373">&gt; Alan Kennedy wrote:</FONT>
<FONT COLOR="#737373">&gt; &gt; </FONT>
<FONT COLOR="#737373">&gt; &gt; &lt;?xml version=&quot;1.0&quot; encoding=&quot;iso-8859-1&quot;?&gt;</FONT>
<FONT COLOR="#737373">&gt; </FONT>
<FONT COLOR="#737373">&gt; This does not change the fact that Expat outputs UTF-8. However, if the</FONT>
<FONT COLOR="#737373">&gt; Expat parser hasn't been told to use iso-8859-1 as default encoding, it</FONT>
<FONT COLOR="#737373">&gt; will assume UTF-8 input unless you state the encoding in the input as</FONT>
<FONT COLOR="#737373">&gt; above.</FONT>
<FONT COLOR="#737373">&gt; </FONT></pre>Yes, I have the encoding=&quot;iso-8859-1&quot; set the file. &nbsp;It sounds as if expat is converting the iso-8859-1 to UTF-8. &nbsp;Is there a way to convert back to iso-8859-1? &nbsp;Unfortunately, I'm stuck with 1.5.2.<br>
<br>
-- <br>
Don Allingham<br>
donaldallingham@home.com </BODY>
</HTML>

--=-AotdqlOxl1oqQ3Q0ScLH--



From Alexandre.Fayolle@logilab.fr  Mon Apr 30 14:52:38 2001
From: Alexandre.Fayolle@logilab.fr (Alexandre Fayolle)
Date: Mon, 30 Apr 2001 15:52:38 +0200 (CEST)
Subject: [XML-SIG] problems reading iso-8859-1 data
In-Reply-To: <988636595.8711.5.camel@wallace>
Message-ID: <Pine.LNX.4.21.0104301551210.16081-100000@orion.logilab.fr>

On 30 Apr 2001, Don Allingham wrote:

> Yes, I have the encoding="iso-8859-1" set the file.  It sounds as if
> expat is converting the iso-8859-1 to UTF-8.  Is there a way to convert
> back to iso-8859-1?  Unfortunately, I'm stuck with 1.5.2.


Try this :

---------------------------8<---------------------------------
from xml.unicode.utf8_iso import utf8_to_code, code_to_utf8
import cStringIO
def utf8_to_latin(s):
    buff = cStringIO.StringIO()
    while s:
        try:
            head,s = utf8_to_code(1,s)
        except Exception,e:
            from traceback import print_exc
            print_exc()
            head = ''
            s = s[1:]
        buff.write(head)
    ans = buff.getvalue()
    buff.close()
    return ans

def latin_to_utf8(s):
    buff = cStringIO.StringIO()
    for c in s:
        try:
            cv = code_to_utf8(1,c)
        except Exception,e:
            from traceback import print_exc
            print_exc()
            cv = ''
        buff.write(cv)
    ans = buff.getvalue()
    buff.close()
    return ans
---------------------------------8<--------------------------

Alexandre Fayolle
-- 
http://www.logilab.com 
Narval is the first software agent available as free software (GPL).
LOGILAB, Paris (France).



From Alexandre.Fayolle@logilab.fr  Mon Apr 30 15:04:56 2001
From: Alexandre.Fayolle@logilab.fr (Alexandre Fayolle)
Date: Mon, 30 Apr 2001 16:04:56 +0200 (CEST)
Subject: [XML-SIG] One more remark on latin/UTF-8 and XBEL
Message-ID: <Pine.LNX.4.21.0104301553070.16081-100000@orion.logilab.fr>

I think there's a problem in the XBEL demos. With Python 1.52, the sgmlop
parser which is used for parsing netscape bookmarks will output latin1
strings which are passed 'as is' into 

Alexandre Fayolle
-- 
http://www.logilab.com 
Narval is the first software agent available as free software (GPL).
LOGILAB, Paris (France).



From Alexandre.Fayolle@logilab.fr  Mon Apr 30 15:12:06 2001
From: Alexandre.Fayolle@logilab.fr (Alexandre Fayolle)
Date: Mon, 30 Apr 2001 16:12:06 +0200 (CEST)
Subject: [XML-SIG] Wooops... (One more remark on latin/UTF-8 and XBEL)
In-Reply-To: <Pine.LNX.4.21.0104301553070.16081-100000@orion.logilab.fr>
Message-ID: <Pine.LNX.4.21.0104301608590.16081-100000@orion.logilab.fr>

Sorry, I hit the wrong key, sending the mail instead of canceling
it (^C and ^X are pretty close on my keyboard). Forget you even saw this
mail and delete it straight away...

Alexandre 'yes, I'm using Pine' Fayolle
-- 
http://www.logilab.com 
Narval is the first software agent available as free software (GPL).
LOGILAB, Paris (France).



From martin@loewis.home.cs.tu-berlin.de  Mon Apr 30 19:39:55 2001
From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis)
Date: Mon, 30 Apr 2001 20:39:55 +0200
Subject: [XML-SIG] problems reading iso-8859-1 data
In-Reply-To: <Pine.LNX.4.21.0104301551210.16081-100000@orion.logilab.fr>
 (message from Alexandre Fayolle on Mon, 30 Apr 2001 15:52:38 +0200
 (CEST))
References: <Pine.LNX.4.21.0104301551210.16081-100000@orion.logilab.fr>
Message-ID: <200104301839.f3UIdtL01407@mira.informatik.hu-berlin.de>

> > Yes, I have the encoding="iso-8859-1" set the file.  It sounds as if
> > expat is converting the iso-8859-1 to UTF-8.  Is there a way to convert
> > back to iso-8859-1?  Unfortunately, I'm stuck with 1.5.2.
> 
> 
> Try this :
> 
> ---------------------------8<---------------------------------
> from xml.unicode.utf8_iso import utf8_to_code, code_to_utf8
> import cStringIO

Or, shorter yet:

from xml.unicode.iso8859 import wstring
def utf8_to_latin1(s):
  wstring.decode("utf-8",s).encode("iso-8859-1")

def latin1_to_utf8(s):
  wstring.decode("iso-8859-1",s).encode("utf-8")

Regards,
Martin