From matteo.boscolo at boscolini.eu  Wed Dec  7 08:08:43 2011
From: matteo.boscolo at boscolini.eu (Matteo Boscolo)
Date: Wed, 07 Dec 2011 08:08:43 +0100
Subject: [python-win32] Loose com pointer
Message-ID: <4EDF10FB.5080705@boscolini.eu>

Hi all,
I need some help or idea to a com problem..

I got this class:

class foo(object):
     def setComObject(comObject):
         self.comO=comObject #This is a com object from a cad application

     def showForm(self)
         # use the self.comO to read some information from the cad 
application
         # Show the pyqt form as child of cad application
         # do somthing with the form
         # do somthing with the self.comO <----- Here if pass some time 
I'm not able to call any method to the com object

a=foo()
o="get istance of a cad application via com"
a.setComObject(o)
a.showForm() #<---- here if pass some time I'm not able to call any 
method to the com object

but I I' re call the
a.setComObject(o)  #Faster and I take less the 30 seconds on the form 
object it works well

It seems a problem of the garbage collector .. but I'm not sure how to 
debug it ..

any help is really appreciated.. it's the last dangerous bug in our 
application ...

Regards,
Matteo







From rickmmmbra at gmail.com  Thu Dec  8 08:34:36 2011
From: rickmmmbra at gmail.com (Rikus de Milander)
Date: Thu, 8 Dec 2011 09:34:36 +0200
Subject: [python-win32] Excel graphs and saving excel workbook .xls
Message-ID: <CAOrqxnBF64Pt8Sq7=d6F3fvFiLsGTh_WeK4zb8N=nUphPL2EVw@mail.gmail.com>

Hey,
I am currently working on a python script to fetch data from a database and
populate a excel sheet and draw some graphs. Currently I have only been
able to generate a scatter graph with  chart.ChartType =
constants.xlXYScatter . Is there a way that one can set the position of the
graph on the excel sheet or the size of the graph ?
Is there maybe a option to draw a bar graph or line graph ?

Another problem that I have is when saving the .xls workbook. When I use
self.workbook.save(filename) [I created a object of type workbook] it does
save but it also seems to close the application and remove it from the
memory so then any actions after the save for eg, adding a worksheet does
not work. Is there maybe a way to save the workbook without closing it ?

Regards
Rikus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111208/4962745b/attachment.html>

From rodperson at rodperson.com  Thu Dec  8 13:30:22 2011
From: rodperson at rodperson.com (Rod Person)
Date: Thu, 8 Dec 2011 07:30:22 -0500
Subject: [python-win32] Excel graphs and saving excel workbook .xls
In-Reply-To: <CAOrqxnBF64Pt8Sq7=d6F3fvFiLsGTh_WeK4zb8N=nUphPL2EVw@mail.gmail.com>
References: <CAOrqxnBF64Pt8Sq7=d6F3fvFiLsGTh_WeK4zb8N=nUphPL2EVw@mail.gmail.com>
Message-ID: <20111208073022.00005015@unknown>

On Thu, 8 Dec 2011 09:34:36 +0200
Rikus de Milander <rickmmmbra at gmail.com> wrote:

> Hey,
> I am currently working on a python script to fetch data from a
> database and populate a excel sheet and draw some graphs. Currently I
> have only been able to generate a scatter graph with  chart.ChartType
> = constants.xlXYScatter . Is there a way that one can set the
> position of the graph on the excel sheet or the size of the graph ?
> Is there maybe a option to draw a bar graph or line graph ?
> 
> Another problem that I have is when saving the .xls workbook. When I
> use self.workbook.save(filename) [I created a object of type
> workbook] it does save but it also seems to close the application and
> remove it from the memory so then any actions after the save for eg,
> adding a worksheet does not work. Is there maybe a way to save the
> workbook without closing it ?
> 
> Regards
> Rikus

Can't help with the graph as I don't use them, sorry.

But you should be able to use 

              self.workbook.SaveAs(filename)

this does not close excel, at least for me.


-- 

Rod Person  http://www.rodperson.com  rodperson at rodperson.com

"Silence is a fence around wisdom"

From caesar at inmail24.com  Sat Dec 10 15:52:50 2011
From: caesar at inmail24.com (Cesar Romani)
Date: Sat, 10 Dec 2011 09:52:50 -0500
Subject: [python-win32] pywintypes.com_error
Message-ID: <jbvro5$fmu$1@dough.gmane.org>

What's wrong with the following code?

--------------------
from win32com.shell import shell
import pythoncom, os

class PyShortcut:
   def __init__(self):
     self._base = pythoncom.CoCreateInstance(shell.CLSID_ShellLink, None,
       pythoncom.CLSCTX_INPROC_SERVER,
       shell.IID_IShellLink)
   def load(self, filename):
     self._base.QueryInterface(pythoncom.IID_IPersistFile).Load(filename)
   def save(self, filename):
     self._base.QueryInterface(pythoncom.IID_IPersistFile).Save(filename, 0)
   def __getattr__(self, name):
     if name != "_base":
       return getattr(self._base, name)

if __name__=='__main__':
   import sys

   shortcut = PyShortcut()
   if len(sys.argv) > 1 and os.path.exists(sys.argv[1]):
     shortcut.load(sys.argv[1])
     print '''Shortcut in file %s to
file:\n\t%s\nArguments:\n\t%s\nDescription:\n\t%s\nWorking
Directory:\n\t%s'''%(
       file,
       shortcut.GetPath(shell.SLGP_UNCPRIORITY)[0],
       shortcut.GetArguments(),
       shortcut.GetDescription(),
       shortcut.GetWorkingDirectory()
     )
--------------------

If I try to use it, as in:
python shortcut.py test.c
I get:

--------------------
Traceback (most recent call last):
   File "shortcut.py", line 22, in <module>
     shortcut.load(sys.argv[1])
   File "shortcut.py", line 10, in load
     self._base.QueryInterface(pythoncom.IID_IPersistFile).Load(filename)
pywintypes.com_error: (-2147467259, 'Unspecified error', None, None)
--------------------

I'm using python 2.7 on Windows 7 Home Premium with
pywin32-216.win32-py2.7

Many thanks in advance,

-- 
Cesar


From boctom at hotmail.com  Mon Dec 12 19:42:19 2011
From: boctom at hotmail.com (Boc Tom)
Date: Mon, 12 Dec 2011 18:42:19 +0000
Subject: [python-win32]  makepy.py usage and COM
Message-ID: <DUB111-W325BA2BCDE4AA5EBA5F717D7BC0@phx.gbl>


Hi,

I have Python 3.2.2 and pyWin32 build 216 :

I have the same issue already mentionned here :
http://mail.python.org/pipermail/python-win32/2008-August/008093.html

I just can't get the code right,
Is there any update that would fix that bug ?
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111212/d7b37d24/attachment.html>

From boctom at hotmail.com  Mon Dec 12 19:56:26 2011
From: boctom at hotmail.com (Boc Tom)
Date: Mon, 12 Dec 2011 18:56:26 +0000
Subject: [python-win32] JavaScript - Python
Message-ID: <DUB111-W41CA00305C01212D2B7F02D7BC0@phx.gbl>


Hi,

I want to use Python to automate InDesign,
There is very little help on that subject !
I would like to use some of the sample scripts provided in the  InDesign Sripting documentation,
So I translate them from VBScript to Python which is ok except there are some parts written in JavaScript as well :

Quote from the tutorial "Importing a graphic" :
""" VBScript does not have the ability to do this, so we'll use a JavaScript to get a file name.
We'll run the JavaScript using InDesign's DoScript feature.
myJavaScriptString = "var myTextFile = File.openDialog(""Choose a graphicfile"");myTextFile.fsName;"
Run the JavaScript using DoScript.
myGraphicFileName = myInDesign.DoScript(myJavaScriptString, idScriptLanguage.idJavascript) """

So my question would be :
- how to import a graphic into Indesign using Python Win32 ?
- Should I try to make the doScript command work, the following command doesn't work :
myGraphicFileName = myInDesign.DoScript(myJavaScriptString, language = Javascript)


Thanks for your help,

Regards,

Thomas


 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111212/53e08503/attachment.html>

From pfraser at spatialmedia.com  Tue Dec 13 03:22:33 2011
From: pfraser at spatialmedia.com (Peter Fraser)
Date: Tue, 13 Dec 2011 15:22:33 +1300
Subject: [python-win32] win32com and the GIL
Message-ID: <4EE6B6E9.1040107@spatialmedia.com>

Hello

Can win32com be made to release the GIL when it calls into COM 
components?  Presumably it doesn't do this by default .

I'm calling an apartment threaded COM component (VB6) and NOT passing 
any Python objects to that component (I promise).


thankyou and have a splendid day

Peter






From boctom at hotmail.com  Wed Dec 14 18:42:50 2011
From: boctom at hotmail.com (Boc Tom)
Date: Wed, 14 Dec 2011 17:42:50 +0000
Subject: [python-win32] COM Browser Adobe InDesign Type Library not loading
Message-ID: <DUB111-W125B2FF7C5ADD51811D5B95D7A20@phx.gbl>


Hi,

Using PythonWin with InDesign,
"The type info can not be loaded " when browsing the Adobe Type Library,
Any idea how to fix this ?
Thanks  for your help,

Thomas
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111214/a6ea7303/attachment.html>

From admin at mbnoimi.net  Thu Dec 15 10:48:15 2011
From: admin at mbnoimi.net (Muhammad Bashir Al-Noimi)
Date: Thu, 15 Dec 2011 11:48:15 +0200
Subject: [python-win32] Unable to build .msi package for pywin32
Message-ID: <4EE9C25F.1070802@mbnoimi.net>

An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111215/e921ddd6/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: build.log.bz2
Type: application/octet-stream
Size: 19926 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111215/e921ddd6/attachment-0001.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: admin.vcf
Type: text/x-vcard
Size: 331 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111215/e921ddd6/attachment-0001.vcf>

From skippy.hammond at gmail.com  Fri Dec 16 04:42:43 2011
From: skippy.hammond at gmail.com (Mark Hammond)
Date: Fri, 16 Dec 2011 14:42:43 +1100
Subject: [python-win32] win32com and the GIL
In-Reply-To: <4EE6B6E9.1040107@spatialmedia.com>
References: <4EE6B6E9.1040107@spatialmedia.com>
Message-ID: <4EEABE33.70801@gmail.com>

On 13/12/2011 1:22 PM, Peter Fraser wrote:
> Hello
>
> Can win32com be made to release the GIL when it calls into COM
> components? Presumably it doesn't do this by default .

It does do it by default and there is no way to have it *not* do it.

> I'm calling an apartment threaded COM component (VB6) and NOT passing
> any Python objects to that component (I promise).

Assuming you mean you aren't passing a COM object that happens to be 
implemented in Python, then it doesn't matter if you do - the GIL will 
be re-grabbed when necessary.

Cheers,

Mark

>
>
> thankyou and have a splendid day
>
> Peter
>
>
>
>
>
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32


From skippy.hammond at gmail.com  Fri Dec 16 04:44:33 2011
From: skippy.hammond at gmail.com (Mark Hammond)
Date: Fri, 16 Dec 2011 14:44:33 +1100
Subject: [python-win32] manifest "Microsoft.Windows.Common-Controls"
 version="6.0.0.0" => strange problem with pywin32 build 216 ?
In-Reply-To: <ja60qp$95p$1@dough.gmane.org>
References: <ja60qp$95p$1@dough.gmane.org>
Message-ID: <4EEABEA1.1040705@gmail.com>

On 19/11/2011 3:23 AM, Robert wrote:
> In C:\Python23 and C:\Python26 installations there are
> python(s).exe.manifest files which contain
> '"Microsoft.Windows.Common-Controls" version="6.0.0.0"'
> (see attached/below)
>
> With pywin32 build 216 (was not in 212; and 214 I guess) win32ui apps,
> which are started by these python(w).exe's , a strange auto-magical
> button pressing is going on, immediately after application start. almost
> any button is pressed when the mouse moves.. By chance its possible to
> stop the show in the Task Manager.
>
> It seems (seen by a debug output), that notification messages (BN_..
> (Which could be routed via HookNotify?)) or so do trigger the normal
> HookCommand'ed handler.
> When that Common-Controls section (or the whole .manifest) is removed,
> the apps work normal in build 216.
>
> What could be the reason?

I've no idea - that sounds really strange.

> What does this Common-Control section actually do?

I believe it says a new version of the commctrl library which does "xp 
themes" should be used by the app.

Cheers,

Mark

From skippy.hammond at gmail.com  Fri Dec 16 04:59:40 2011
From: skippy.hammond at gmail.com (Mark Hammond)
Date: Fri, 16 Dec 2011 14:59:40 +1100
Subject: [python-win32] python-win32 Digest, Vol 104, Issue 14
In-Reply-To: <CABL-cAZyHB1eVrCsJH-Sg=cEFsK65tbkKNMf1yNxcYL4=+tExw@mail.gmail.com>
References: <mailman.37.1321873203.24986.python-win32@python.org>
	<CABL-cAZyHB1eVrCsJH-Sg=cEFsK65tbkKNMf1yNxcYL4=+tExw@mail.gmail.com>
Message-ID: <4EEAC22C.2030106@gmail.com>

On 22/11/2011 2:10 PM, ??? wrote:
> http://sourceforge.net/tracker/?func=detail&aid=2831327&group_id=78018&atid=551954
> <http://sourceforge.net/tracker/?func=detail&aid=2831327&group_id=78018&atid=551954>
>
> UTC / Local time break pythonwin for Python32.
>
> A lot of function does not work. Simple throws following messages:
>
> ValueError: astimezone() cannot be applied to a naive datetime

I'll try and look at this issue soon.

Cheers,

Mark

From skippy.hammond at gmail.com  Fri Dec 16 05:00:49 2011
From: skippy.hammond at gmail.com (Mark Hammond)
Date: Fri, 16 Dec 2011 15:00:49 +1100
Subject: [python-win32] addin fails to load the first time after login
In-Reply-To: <jajndu$dq$1@dough.gmane.org>
References: <jajndu$dq$1@dough.gmane.org>
Message-ID: <4EEAC271.4050806@gmail.com>

I'm afraid I can't guess without a traceback.  Try registering your 
addin with --debug and then run "python -m win32traceutil" - if you are 
luck you will also see the traceback there.

Mark

On 24/11/2011 8:09 AM, Christian K. wrote:
> Hi,
>
> I am observing a strange behaviour of two python com-servers/addins, one
> of them is called from Excel VBA the other one is an Outlook (both from
> Office 2003) addin.
>
> Both of them work if I register them and start using Outlook or Excel
> thereafter. However when I log out and login the next time, both of them
> fail to load. Excel is showing an error message when
> CreateObject("addin") is called, showing a python traceback from
> somewhere within win32com (only parts of it since the window cannot be
> resized). Outlook silently ignores the addin and in the list of
> com_addins it is marked as deactivated. Re-registering solves the
> problem until the next login.
>
> Interestingly I get them to work if I start e.g. spyder (a python IDE)
> before making calls to the com servers. It does not seem to help though,
> if I only open a python shell. This sounds like a timeout issue to me.
>
> Do you know what might be the cause and what I can do about it?
>
> Best regards, Christian
>
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32


From skippy.hammond at gmail.com  Fri Dec 16 04:58:52 2011
From: skippy.hammond at gmail.com (Mark Hammond)
Date: Fri, 16 Dec 2011 14:58:52 +1100
Subject: [python-win32] Setting a 64bit FILETIME PT_SYSTIME property
 using pywin32's SetProps()
In-Reply-To: <CAFkt3UNXCjwipCxt_Zk2RCSk+JjhVxEqTcXjr3B-XQ7rJu=wWQ@mail.gmail.com>
References: <CAFkt3UNXCjwipCxt_Zk2RCSk+JjhVxEqTcXjr3B-XQ7rJu=wWQ@mail.gmail.com>
Message-ID: <4EEAC1FC.1080203@gmail.com>

Attempting to pass an integer will assume the int is a regular "unix 
timestamp" and pywin32 will do the conversions as necessary.  So there 
should be no need to convert it to a 64bit FILETIME integer.  You should 
also find that passing a datetime object directly works, even in py2k 
builds - pywin32 will use the timetuple() method to perform the conversion.

Mark

On 21/11/2011 6:03 PM, Sriram ET. wrote:
> I am using pywin32 mapi routines to interact with my outlook contacts
> store. I am trying to set the PT_BIRTHDAY property and have the
> following code for doing that:
>
> The conversion from and to FILETIME values is from filetimes.py
> available at:http://reliablybroken.com/b/wp-content/filetimes.py
>
> |     dt = datetime.strptime('1980-01-01','%Y-%m-%d')
>      filetime = filetimes.dt_to_filetime(dt)
>      msg.SetProps([(mapitags.PR_BIRTHDAY, filetime)])
> |
>
> I get an OverflowError: Python int too large to convert to c long
>
> I am on a 32 bit machine, and I understand FILETME is a 64 bit value. I
> am at a loss. How can I set the underlying MAPI property using this
> library? Is there a workaround?
>
> (Setup: Outlook 2010, Window 7 running on a 32bit virtual machine via
> VMWare Fusion 4 on Mac OS X Lion)
>
> Thanks!
>
>
>
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32


From nawijn at gmail.com  Fri Dec 16 16:25:32 2011
From: nawijn at gmail.com (Marco Nawijn)
Date: Fri, 16 Dec 2011 16:25:32 +0100
Subject: [python-win32] Problem with Dispatch in normal Python shell
Message-ID: <CA+pnd0kg7VsaYPCz+W8-U-4DEhYVEtxiR9L26j30hJqeB=6M9w@mail.gmail.com>

Dear All,

I hope someone can assist me with a dispatch problem that I am facing.
Let me briefly explain my case. I am implementing a monitoring and
reporting application for one of our manufacturing machines. I use the
COM interface provided by the PLC manufacturer (Siemens). Since it is
a long running program and the PythonWin GUI becomes sort of
unresponsive I would like to run the data extraction operation through
a normal Python shell (DOS/PowerShell terminal). The problem that I
face is that I cannot dispatch the automation object through the
normal Python shell (see traceback below). It works fine in the
interactive PythonWin console (see at the bottom of the post). I would
be very gratefull if someone can provide me with some hints on how to
solve this problem. Maybe somebody knows the difference in running an
interactive console or a normal Python shell?

I have been struggling with the issue for a few days now. It is
becoming quite urgent now, because we are required to log the
manufacturing data on a higher rate than we do now. We have a tool
implemented in VB, but is is completely hardcoded to only run at a one
minute time interval. We would like to be more flexible, but use
Python to do so.

Thanks in advance,

Marco



Python 2.7.2 (default, Jun 12 2011, 15:08:59) [MSC v.1500 32 bit (Intel)] on win
32
Type "help", "copyright", "credits" or "license" for more information.
>>> from win32com.client import Dispatch
>>> server = Dispatch('OPC.Automation.1')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "c:\python27\lib\site-packages\win32com\client\__init__.py", line 95, in
Dispatch
    dispatch, userName = dynamic._GetGoodDispatchAndUserName(dispatch,userName,c
lsctx)
  File "c:\python27\lib\site-packages\win32com\client\dynamic.py", line 108, in
_GetGoodDispatchAndUserName
    return (_GetGoodDispatch(IDispatch, clsctx), userName)
  File "c:\python27\lib\site-packages\win32com\client\dynamic.py", line 85, in _
GetGoodDispatch
    IDispatch = pythoncom.CoCreateInstance(IDispatch, None, clsctx, pythoncom.II
D_IDispatch)
pywintypes.com_error: (-2147023898, 'Invalid access to memory location.', None,
None)



PythonWin 2.7.2 (default, Jun 12 2011, 15:08:59) [MSC v.1500 32 bit
(Intel)] on win32.
Portions Copyright 1994-2008 Mark Hammond - see 'Help/About PythonWin'
for further copyright
information.

>>> from win32com.client import Dispatch

>>> server = Dispatch('OPC.Automation.1')

>>> print server

<win32com.gen_py.OPC DA Automation Wrapper 2.02.IOPCGroups instance at
0x16331400>

>>>

From pacopyc at gmail.com  Fri Dec 16 22:42:17 2011
From: pacopyc at gmail.com (pacopyc pacopyc)
Date: Fri, 16 Dec 2011 22:42:17 +0100
Subject: [python-win32] send message remote host
Message-ID: <CAJDkAOHYt9uYnZcibjGdBdJgUFdRfpUVRMSj7tQ7FpXQgkyPhg@mail.gmail.com>

Hi, I'd like send a message to remote host. I want that remote host's
user read my short message immediately. I have remote host
administrator credentials. Can you help me?

From timr at probo.com  Sat Dec 17 02:47:06 2011
From: timr at probo.com (Tim Roberts)
Date: Fri, 16 Dec 2011 17:47:06 -0800
Subject: [python-win32] send message remote host
In-Reply-To: <CAJDkAOHYt9uYnZcibjGdBdJgUFdRfpUVRMSj7tQ7FpXQgkyPhg@mail.gmail.com>
References: <CAJDkAOHYt9uYnZcibjGdBdJgUFdRfpUVRMSj7tQ7FpXQgkyPhg@mail.gmail.com>
Message-ID: <4EEBF49A.8000103@probo.com>

pacopyc pacopyc wrote:
> Hi, I'd like send a message to remote host. I want that remote host's
> user read my short message immediately. I have remote host
> administrator credentials. Can you help me?

You'll have to be more specific with what you mean by "message".  Are
you talking about the same operation that the "net send" command performs?

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From sleerssen at gmail.com  Sat Dec 17 16:34:17 2011
From: sleerssen at gmail.com (Scott Leerssen)
Date: Sat, 17 Dec 2011 10:34:17 -0500
Subject: [python-win32] file descriptor issues using
	msvcrt.open_osfhandle()+os.fdopen()
Message-ID: <32FE529F-0D6B-4D8A-91C3-10809322966B@gmail.com>

Greetings!

Apologies in advance for the length of this post;  after many days of scouring Python and Windows news groups, I've seen reports of issues similar to mine, but no cause or resolution other than re-trying an operation.  I'm in no way a Windows programmer; I've spent my entire career on Unix/Linux, so the Windows ways are a bit of a mystery to me at times.  That being said, I'm hoping that someone might shed some light on something stupid I've done (limited to this topic, please) or that I might actually save someone else's time when they see this post.

I have been tracking down random descriptor issues with a multithreaded Windows service written in Python.  The issue presents itself as "bad file descriptor", "attempt to read from something that is not a socket", and various errors closing files and sockets.  

In my service, I needed to be able to open a file with FILE_FLAG_BACKUP_SEMANTICS, create a Python file object and pass it off to another method to read.  Since the Python posix file open() method doesn't provide a way to pass special Windozy flags, and I didn't see a way of setting that post-open, I followed a pattern that I found that looks like this:

       h = win32file.CreateFile(
           fname,
           win32file.GENERIC_READ,
           win32file.FILE_SHARE_READ,
           None,
           win32file.OPEN_EXISTING,
           win32file.FILE_FLAG_BACKUP_SEMANTICS,
           None
           )
       fd = msvcrt.open_osfhandle(h, os.O_RDONLY)
       f = os.fdopen(fd, 'rb')

That worked well and I was able to pass that 'f' off to be handled like any other file-like object.  However, my multi-threaded service would randomly raise file descriptor errors reading files, logging and writing data on the network.  What I found is that the file descriptors in the PyHANDLE and the file object seem to be somewhat out of sync during garbage collection.  Closing the PyHANDLE leaves the file object with an invalid descriptor, and closing the file object leaves the PyHANDLE with an invalid descriptor, but only sometimes.  If you put the code above into a loop, most of the time it will raise with a bad file descriptor error as the old objects go out of scope and are collected.  I tried to remedy that by first calling CloseHandle() on the PyHANDLE and then close() on the file object (wrapped in a a try:..except IOError to deal with the already closed descriptor), and that makes the loop run, but still causes problems with other threads that just happen to open/close a descriptor at the right/wrong time.

So, my completely inexperienced (in Windows) empirical diagnosis is that there is some sort of synchronization issue between the msvcrt library and the CPython runtime that results in descriptors being shared among file and socket objects in a multithreaded application.

At any rate, I replaced the CreateFile()/msvcrt.open_osfhandle()/os.fdopen() with a wrapper class that uses CreateFile()/ReadFile() and provides a minimal set of file-like object methods and all of my file descriptor and socket problems went away, so either I'm correct, or I just moved my problem.

As I said in the beginning of this epic, I'm no more sure that I've found a problem with the underlying classes that manage descriptors on Windows than I've just replaced my own bad code with other not-as-bad code, but I hope that this rings a bell for someone that can set me straight.

Thanks,
Scott Leerssen


From amauryfa at gmail.com  Sat Dec 17 17:36:06 2011
From: amauryfa at gmail.com (Amaury Forgeot d'Arc)
Date: Sat, 17 Dec 2011 17:36:06 +0100
Subject: [python-win32] file descriptor issues using
	msvcrt.open_osfhandle()+os.fdopen()
In-Reply-To: <32FE529F-0D6B-4D8A-91C3-10809322966B@gmail.com>
References: <32FE529F-0D6B-4D8A-91C3-10809322966B@gmail.com>
Message-ID: <CAGmFidaaB9oEGaex9xBHwPo3VUsfLhSmCLex80CtJ7dkhi2WRQ@mail.gmail.com>

2011/12/17 Scott Leerssen <sleerssen at gmail.com>

> Closing the PyHANDLE leaves the file object with an invalid descriptor,
> and closing the file object leaves the PyHANDLE with an invalid descriptor,
> but only sometimes


After open_osfhandle(), the handle is owned by the file object, and should
not be closed when the PyHANDLE object is deallocated.
Did you try to use the Detach() method? Something like
   fd = open_osfhandle(h.Detach(), os.O_RDONLY)
Documentation says: "You would call this function when you need the
underlying win32 handle to exist beyond the lifetime of the handle object"
which seems adequate here.

-- 
Amaury Forgeot d'Arc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111217/d0c24f5a/attachment.html>

From sleerssen at gmail.com  Sat Dec 17 18:08:53 2011
From: sleerssen at gmail.com (Scott Leerssen)
Date: Sat, 17 Dec 2011 12:08:53 -0500
Subject: [python-win32] file descriptor issues using
	msvcrt.open_osfhandle()+os.fdopen()
In-Reply-To: <CAGmFidaaB9oEGaex9xBHwPo3VUsfLhSmCLex80CtJ7dkhi2WRQ@mail.gmail.com>
References: <32FE529F-0D6B-4D8A-91C3-10809322966B@gmail.com>
	<CAGmFidaaB9oEGaex9xBHwPo3VUsfLhSmCLex80CtJ7dkhi2WRQ@mail.gmail.com>
Message-ID: <84739C66-AF7C-4D0E-AA4D-E4F1AEFDDD1D@gmail.com>


On Dec 17, 2011, at 11:36 AM, Amaury Forgeot d'Arc wrote:

> 2011/12/17 Scott Leerssen <sleerssen at gmail.com>
> Closing the PyHANDLE leaves the file object with an invalid descriptor, and closing the file object leaves the PyHANDLE with an invalid descriptor, but only sometimes
> 
> After open_osfhandle(), the handle is owned by the file object, and should not be closed when the PyHANDLE object is deallocated.
> Did you try to use the Detach() method? Something like
>    fd = open_osfhandle(h.Detach(), os.O_RDONLY)
> Documentation says: "You would call this function when you need the underlying win32 handle to exist beyond the lifetime of the handle object" which seems adequate here.

I did see that, but I interpreted that to mean that the PyHANDLE would be dereferenced from the underlying Windows file handle, and I'm not sure what the consequences of that would be.  That is, how would I free that resource if I lose it from the Python context?  Also, it didn't seem to fit since I didn't want a win32 handle to live past the close.  I just wanted all the associated handles and descriptors to close at the same time.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111217/08d032e7/attachment.html>

From amauryfa at gmail.com  Sat Dec 17 18:25:57 2011
From: amauryfa at gmail.com (Amaury Forgeot d'Arc)
Date: Sat, 17 Dec 2011 18:25:57 +0100
Subject: [python-win32] file descriptor issues using
	msvcrt.open_osfhandle()+os.fdopen()
In-Reply-To: <84739C66-AF7C-4D0E-AA4D-E4F1AEFDDD1D@gmail.com>
References: <32FE529F-0D6B-4D8A-91C3-10809322966B@gmail.com>
	<CAGmFidaaB9oEGaex9xBHwPo3VUsfLhSmCLex80CtJ7dkhi2WRQ@mail.gmail.com>
	<84739C66-AF7C-4D0E-AA4D-E4F1AEFDDD1D@gmail.com>
Message-ID: <CAGmFidZ_08_jKRE3b85VCMMJ7UYMFiuO0Pa=H-H3TxJ-korp7w@mail.gmail.com>

2011/12/17 Scott Leerssen <sleerssen at gmail.com>

> I did see that, but I interpreted that to mean that the PyHANDLE would be
> dereferenced from the underlying Windows file handle,
>

It's the other way round: the PyHandle just forgets the win32 handle value,
and will not close it in its __del__.

-- 
Amaury Forgeot d'Arc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111217/d6ee72f5/attachment.html>

From sleerssen at gmail.com  Sat Dec 17 18:34:37 2011
From: sleerssen at gmail.com (Scott Leerssen)
Date: Sat, 17 Dec 2011 12:34:37 -0500
Subject: [python-win32] file descriptor issues using
	msvcrt.open_osfhandle()+os.fdopen()
In-Reply-To: <CAGmFidZ_08_jKRE3b85VCMMJ7UYMFiuO0Pa=H-H3TxJ-korp7w@mail.gmail.com>
References: <32FE529F-0D6B-4D8A-91C3-10809322966B@gmail.com>
	<CAGmFidaaB9oEGaex9xBHwPo3VUsfLhSmCLex80CtJ7dkhi2WRQ@mail.gmail.com>
	<84739C66-AF7C-4D0E-AA4D-E4F1AEFDDD1D@gmail.com>
	<CAGmFidZ_08_jKRE3b85VCMMJ7UYMFiuO0Pa=H-H3TxJ-korp7w@mail.gmail.com>
Message-ID: <0DD1C7B3-CB99-44FE-AAEA-8E0F998B16C8@gmail.com>


On Dec 17, 2011, at 12:25 PM, Amaury Forgeot d'Arc wrote:

> 2011/12/17 Scott Leerssen <sleerssen at gmail.com>
> I did see that, but I interpreted that to mean that the PyHANDLE would be dereferenced from the underlying Windows file handle, 
> 
> It's the other way round: the PyHandle just forgets the win32 handle value, and will not close it in its __del__. 

Sorry, I did write that backward.  How is the win32 handle freed?

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111217/7d51b369/attachment-0001.html>

From pacopyc at gmail.com  Sat Dec 17 22:48:01 2011
From: pacopyc at gmail.com (pacopyc pacopyc)
Date: Sat, 17 Dec 2011 22:48:01 +0100
Subject: [python-win32] send message remote host
Message-ID: <CAJDkAOFvW7Lw01XBfZy__Pwavhfe_AaVVwEhfSnmovOxR2J7EA@mail.gmail.com>

Yes, I'm thinking about "net send" of Windows

From christophedeze at wanadoo.fr  Mon Dec 19 12:57:49 2011
From: christophedeze at wanadoo.fr (cd)
Date: Mon, 19 Dec 2011 12:57:49 +0100 (CET)
Subject: [python-win32] rename computer
Message-ID: <24572038.67645.1324295869582.JavaMail.www@wwinf2210>

hello
I didn't find any example of howto rename a windows computer with python ?

i've tried with wmi unsuccessfully ....

thanks for any halps 

Christophe

From mail at timgolden.me.uk  Mon Dec 19 14:47:16 2011
From: mail at timgolden.me.uk (Tim Golden)
Date: Mon, 19 Dec 2011 13:47:16 +0000
Subject: [python-win32] rename computer
In-Reply-To: <24572038.67645.1324295869582.JavaMail.www@wwinf2210>
References: <24572038.67645.1324295869582.JavaMail.www@wwinf2210>
Message-ID: <4EEF4064.6030700@timgolden.me.uk>

On 19/12/2011 11:57, cd wrote:
> hello
> I didn't find any example of howto rename a windows computer with python ?
>
> i've tried with wmi unsuccessfully ....

What did you try with WMI, and what happened?

TJG

From christophedeze at wanadoo.fr  Mon Dec 19 15:26:17 2011
From: christophedeze at wanadoo.fr (cd)
Date: Mon, 19 Dec 2011 15:26:17 +0100 (CET)
Subject: [python-win32] rename computer
In-Reply-To: <B0206DB0-9A3C-4103-AE88-B9CB6C8491B6@gmail.com>
References: <24572038.67645.1324295869582.JavaMail.www@wwinf2210>
	<B0206DB0-9A3C-4103-AE88-B9CB6C8491B6@gmail.com>
Message-ID: <1085245363.199514.1324304777960.JavaMail.www@wwinf1g27>




thanks
it works
but I prefere to use a builtin" system call or somethings like this, in order to avoid invalid hostname.
I hope wmi can valid hostname for me





> Message du 19/12/11 14:31
> De : "Scott Leerssen" 
> A : "cd" 
> Copie ? : 
> Objet : Re: [python-win32] rename computer
> 
> 
> On Dec 19, 2011, at 6:57 AM, cd wrote:
> 
> > hello
> > I didn't find any example of howto rename a windows computer with python ?
> > 
> > i've tried with wmi unsuccessfully ....
> > 
> > thanks for any halps 
> > 
> > Christophe
> 
> I do it by setting the following registry values:
> 
> 'HKLM\\System\\CurrentControlSet\\Control\\ComputerName\\ComputerName',
> 'ComputerName',
> 
> 'HKLM\\System\\CurrentControlSet\\Control\\ComputerName\\ActiveComputerName',
> 'ComputerName',
> 
> DHCP might set the following for you, but if you are using static addressing, you'll want to set them as well:
> 
> 'HKLM'\\System\\CurrentControlSet\\Services\\Tcpip\\Parameters',
> 'Hostname',
> 
> 'HKLM\\System\\CurrentControlSet\\Services\\Tcpip\\Parameters',
> 'NV Hostname',
> 
> 
> 

From christophedeze at wanadoo.fr  Mon Dec 19 15:20:38 2011
From: christophedeze at wanadoo.fr (cd)
Date: Mon, 19 Dec 2011 15:20:38 +0100 (CET)
Subject: [python-win32] rename computer
In-Reply-To: <4EEF4064.6030700@timgolden.me.uk>
References: <24572038.67645.1324295869582.JavaMail.www@wwinf2210>
	<4EEF4064.6030700@timgolden.me.uk>
Message-ID: <210200286.198995.1324304438070.JavaMail.www@wwinf1g27>

import wmi
c = wmi.WMI().Win32_ComputerSystem
computer = c()[0]
for propertyName in sorted( list( c.properties ) ):
if propertyName == 'Name':
print setattr( computer, propertyName, 'newname' )
...
Traceback (most recent call last):
File "", line 3, in 
File "C:\Python27\lib\site-packages\wmi.py", line 571, in __setattr__
handle_com_error ()
File "C:\Python27\lib\site-packages\wmi.py", line 241, in handle_com_error
raise klass (com_error=err)
wmi.x_wmi: 












> Message du 19/12/11 14:48
> De : "Tim Golden" 
> A : python-win32 at python.org
> Copie ? : 
> Objet : Re: [python-win32] rename computer
> 
> On 19/12/2011 11:57, cd wrote:
> > hello
> > I didn't find any example of howto rename a windows computer with python ?
> >
> > i've tried with wmi unsuccessfully ....
> 
> What did you try with WMI, and what happened?
> 
> TJG
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32
> 

From mail at timgolden.me.uk  Mon Dec 19 15:32:36 2011
From: mail at timgolden.me.uk (Tim Golden)
Date: Mon, 19 Dec 2011 14:32:36 +0000
Subject: [python-win32] rename computer
In-Reply-To: <210200286.198995.1324304438070.JavaMail.www@wwinf1g27>
References: <24572038.67645.1324295869582.JavaMail.www@wwinf2210>
	<4EEF4064.6030700@timgolden.me.uk>
	<210200286.198995.1324304438070.JavaMail.www@wwinf1g27>
Message-ID: <4EEF4B04.8050300@timgolden.me.uk>

On 19/12/2011 14:20, cd wrote:
> import wmi
> c = wmi.WMI().Win32_ComputerSystem
> computer = c()[0]
> for propertyName in sorted( list( c.properties ) ):
> if propertyName == 'Name':
> print setattr( computer, propertyName, 'newname' )

Ah.

I'm afraid, good as WMI is, it's not that good. You need to
invoke the right method on the Win32_ComputerSystem instance.

The appropriate method is referenced here:

http://msdn.microsoft.com/en-us/library/windows/desktop/aa393056%28v=vs.85%29.aspx


There are some examples here:

http://msdn.microsoft.com/en-us/library/windows/desktop/aa394586%28v=vs.85%29.aspx


and the (definitely untested) Python equivalent is something
like this:

<code>
import wmi

c = wmi.WMI ()
for system in c.Win32_ComputerSystem ():
   system.Rename ("NEW-NAME", "TIM", "PASSWORD")

</code>

TJG

From christophedeze at wanadoo.fr  Mon Dec 19 15:55:01 2011
From: christophedeze at wanadoo.fr (cd)
Date: Mon, 19 Dec 2011 15:55:01 +0100 (CET)
Subject: [python-win32] rename computer
In-Reply-To: <4EEF4B04.8050300@timgolden.me.uk>
References: <24572038.67645.1324295869582.JavaMail.www@wwinf2210>
	<4EEF4064.6030700@timgolden.me.uk>
	<210200286.198995.1324304438070.JavaMail.www@wwinf1g27>
	<4EEF4B04.8050300@timgolden.me.uk>
Message-ID: <1845251643.201646.1324306501654.JavaMail.www@wwinf1g27>


thanks genius it works !!



> Message du 19/12/11 15:34
> De : "Tim Golden" 
> A : python-win32 at python.org
> Copie ? : 
> Objet : Re: [python-win32] rename computer
> 
> On 19/12/2011 14:20, cd wrote:
> > import wmi
> > c = wmi.WMI().Win32_ComputerSystem
> > computer = c()[0]
> > for propertyName in sorted( list( c.properties ) ):
> > if propertyName == 'Name':
> > print setattr( computer, propertyName, 'newname' )
> 
> Ah.
> 
> I'm afraid, good as WMI is, it's not that good. You need to
> invoke the right method on the Win32_ComputerSystem instance.
> 
> The appropriate method is referenced here:
> 
> http://msdn.microsoft.com/en-us/library/windows/desktop/aa393056%28v=vs.85%29.aspx
> 
> 
> There are some examples here:
> 
> http://msdn.microsoft.com/en-us/library/windows/desktop/aa394586%28v=vs.85%29.aspx
> 
> 
> and the (definitely untested) Python equivalent is something
> like this:
> 
> 
> import wmi
> 
> c = wmi.WMI ()
> for system in c.Win32_ComputerSystem ():
> system.Rename ("NEW-NAME", "TIM", "PASSWORD")
> 
> 
> 
> TJG
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32
> 

From mail at timgolden.me.uk  Mon Dec 19 16:06:37 2011
From: mail at timgolden.me.uk (Tim Golden)
Date: Mon, 19 Dec 2011 15:06:37 +0000
Subject: [python-win32] rename computer
In-Reply-To: <4EEF4B04.8050300@timgolden.me.uk>
References: <24572038.67645.1324295869582.JavaMail.www@wwinf2210>
	<4EEF4064.6030700@timgolden.me.uk>
	<210200286.198995.1324304438070.JavaMail.www@wwinf1g27>
	<4EEF4B04.8050300@timgolden.me.uk>
Message-ID: <4EEF52FD.4010901@timgolden.me.uk>

On 19/12/2011 14:32, Tim Golden wrote:
> <code>
> import wmi
>
> c = wmi.WMI ()
> for system in c.Win32_ComputerSystem ():
> system.Rename ("NEW-NAME", "TIM", "PASSWORD")
>
> </code>

For the record, the OP reported in private email
that this solution works for him.

TJG

From theller at ctypes.org  Mon Dec 19 16:26:08 2011
From: theller at ctypes.org (Thomas Heller)
Date: Mon, 19 Dec 2011 16:26:08 +0100
Subject: [python-win32] Several,
	totally isolated Python interpreters in the same process
Message-ID: <jcnl2h$6rv$1@dough.gmane.org>

I made an experiment that I wanted to share because it is imo kind of cool:

I managed to embed a second Python interpreter in a Python process, with
help of WinSxS (WinSxS is a Microsoft technology that allows to isolate
plugins from other plugins, or from the calling process).

The second interpreter is totally isolated from the calling process.

The idea is to use this technology to decouple COM extensions 
implemented in Python from the caller and from other COM extensions.
But there may be other uses as well...

The code is here:

http://code.google.com/p/ctypes-stuff/source/browse/trunk/winsxs


Information I found useful about WinSxS and assemblies:

http://msdn.microsoft.com/en-us/library/dd408052%28v=VS.85%29.aspx

http://omnicognate.wordpress.com/2009/10/05/winsxs/


Thomas


From sdementen at gmail.com  Mon Dec 19 17:17:07 2011
From: sdementen at gmail.com (=?ISO-8859-1?Q?S=E9bastien_de_Menten?=)
Date: Mon, 19 Dec 2011 17:17:07 +0100
Subject: [python-win32] wrapping a IUnknown pointer into a Dispatch interface
Message-ID: <CAB2pxDuYZvzOzXAd-AQ0jFr0TS-QHZ=7KFBoB628M-7rTDFOjQ@mail.gmail.com>

Hello,

I am using a call to a windows function via ctypes to retrieve a pointer to
the IDispatch interface of an Excel application, something like:

oledll.oleacc.AccessibleObjectFromWindow(hnd, OBJID_NATIVEOM,
byref(automation.IDispatch._iid_), byref(pointer_idispatch))

where hnd is the window handle to the proper excel window.

In the variable result (with type POINTER(IUnknown)), I have the pointer to
the interface.
Currently, to be able to call methods/properties on this COM object, I call
the "wrap" function from the comtypes.client library and then use the COM
object, something like:

app = wrap( pointer_idispatch )
print app.Application.Name '==> "Microsoft Excel"
app.Application.ActiveWorkbook.ActiveSheet.Cells(1,1).Value[()] = 1


However, this is slow (a coupe of seconds to do the wrap) and the resulting
object is less friendly to use than the COM object I can get with win32.
For instance, i need to use

app.Application.ActiveWorkbook.ActiveSheet.Cells(1,1).Value[()] = 1

instead of

app.Application.ActiveWorkbook.ActiveSheet.Cells(1,1).Value = 1


So my question, how could I instantiate a COM IDispatch object from win32
(equivalent of 'app' hereabove) when I have the pointer to the IDispatch
interface ('pointer_idispatch' hereabove) ?

thank you

Sebastien
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111219/4e6a2832/attachment-0001.html>

From timr at probo.com  Mon Dec 19 19:19:32 2011
From: timr at probo.com (Tim Roberts)
Date: Mon, 19 Dec 2011 10:19:32 -0800
Subject: [python-win32] send message remote host
In-Reply-To: <CAJDkAOFvW7Lw01XBfZy__Pwavhfe_AaVVwEhfSnmovOxR2J7EA@mail.gmail.com>
References: <CAJDkAOFvW7Lw01XBfZy__Pwavhfe_AaVVwEhfSnmovOxR2J7EA@mail.gmail.com>
Message-ID: <4EEF8034.6050808@probo.com>

pacopyc pacopyc wrote:
> Yes, I'm thinking about "net send" of Windows

As Google would have told you, "net send" uses the API
NetMessageBufferSend, which is in the win32net module in PyWin32.  There
are samples on the Internet, although you might have to translate the
C++ to Python.

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From skippy.hammond at gmail.com  Mon Dec 19 22:29:41 2011
From: skippy.hammond at gmail.com (Mark Hammond)
Date: Tue, 20 Dec 2011 08:29:41 +1100
Subject: [python-win32] Problem with Dispatch in normal Python shell
In-Reply-To: <CA+pnd0kg7VsaYPCz+W8-U-4DEhYVEtxiR9L26j30hJqeB=6M9w@mail.gmail.com>
References: <CA+pnd0kg7VsaYPCz+W8-U-4DEhYVEtxiR9L26j30hJqeB=6M9w@mail.gmail.com>
Message-ID: <4EEFACC5.7010305@gmail.com>

I'm afraid it looks like the object itself has a problem when running 
outside a GUI environment.  You might like to see if you can use Windows 
Scripting Host (cscript.exe) to load the object but I suspect that might 
fail in the same way.  If so, your best bet might be to talk to the vendor.

Mark

On 17/12/2011 2:25 AM, Marco Nawijn wrote:
> Dear All,
>
> I hope someone can assist me with a dispatch problem that I am facing.
> Let me briefly explain my case. I am implementing a monitoring and
> reporting application for one of our manufacturing machines. I use the
> COM interface provided by the PLC manufacturer (Siemens). Since it is
> a long running program and the PythonWin GUI becomes sort of
> unresponsive I would like to run the data extraction operation through
> a normal Python shell (DOS/PowerShell terminal). The problem that I
> face is that I cannot dispatch the automation object through the
> normal Python shell (see traceback below). It works fine in the
> interactive PythonWin console (see at the bottom of the post). I would
> be very gratefull if someone can provide me with some hints on how to
> solve this problem. Maybe somebody knows the difference in running an
> interactive console or a normal Python shell?
>
> I have been struggling with the issue for a few days now. It is
> becoming quite urgent now, because we are required to log the
> manufacturing data on a higher rate than we do now. We have a tool
> implemented in VB, but is is completely hardcoded to only run at a one
> minute time interval. We would like to be more flexible, but use
> Python to do so.
>
> Thanks in advance,
>
> Marco
>
>
>
> Python 2.7.2 (default, Jun 12 2011, 15:08:59) [MSC v.1500 32 bit (Intel)] on win
> 32
> Type "help", "copyright", "credits" or "license" for more information.
>>>> from win32com.client import Dispatch
>>>> server = Dispatch('OPC.Automation.1')
> Traceback (most recent call last):
>    File "<stdin>", line 1, in<module>
>    File "c:\python27\lib\site-packages\win32com\client\__init__.py", line 95, in
> Dispatch
>      dispatch, userName = dynamic._GetGoodDispatchAndUserName(dispatch,userName,c
> lsctx)
>    File "c:\python27\lib\site-packages\win32com\client\dynamic.py", line 108, in
> _GetGoodDispatchAndUserName
>      return (_GetGoodDispatch(IDispatch, clsctx), userName)
>    File "c:\python27\lib\site-packages\win32com\client\dynamic.py", line 85, in _
> GetGoodDispatch
>      IDispatch = pythoncom.CoCreateInstance(IDispatch, None, clsctx, pythoncom.II
> D_IDispatch)
> pywintypes.com_error: (-2147023898, 'Invalid access to memory location.', None,
> None)
>
>
>
> PythonWin 2.7.2 (default, Jun 12 2011, 15:08:59) [MSC v.1500 32 bit
> (Intel)] on win32.
> Portions Copyright 1994-2008 Mark Hammond - see 'Help/About PythonWin'
> for further copyright
> information.
>
>>>> from win32com.client import Dispatch
>
>>>> server = Dispatch('OPC.Automation.1')
>
>>>> print server
>
> <win32com.gen_py.OPC DA Automation Wrapper 2.02.IOPCGroups instance at
> 0x16331400>
>
>>>>
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32


From skippy.hammond at gmail.com  Mon Dec 19 22:32:59 2011
From: skippy.hammond at gmail.com (Mark Hammond)
Date: Tue, 20 Dec 2011 08:32:59 +1100
Subject: [python-win32] file descriptor issues using
	msvcrt.open_osfhandle()+os.fdopen()
In-Reply-To: <0DD1C7B3-CB99-44FE-AAEA-8E0F998B16C8@gmail.com>
References: <32FE529F-0D6B-4D8A-91C3-10809322966B@gmail.com>
	<CAGmFidaaB9oEGaex9xBHwPo3VUsfLhSmCLex80CtJ7dkhi2WRQ@mail.gmail.com>
	<84739C66-AF7C-4D0E-AA4D-E4F1AEFDDD1D@gmail.com>
	<CAGmFidZ_08_jKRE3b85VCMMJ7UYMFiuO0Pa=H-H3TxJ-korp7w@mail.gmail.com>
	<0DD1C7B3-CB99-44FE-AAEA-8E0F998B16C8@gmail.com>
Message-ID: <4EEFAD8B.5050901@gmail.com>

On 18/12/2011 4:34 AM, Scott Leerssen wrote:
>
> On Dec 17, 2011, at 12:25 PM, Amaury Forgeot d'Arc wrote:
>
>> 2011/12/17 Scott Leerssen <sleerssen at gmail.com
>> <mailto:sleerssen at gmail.com>>
>>
>>     I did see that, but I interpreted that to mean that the PyHANDLE
>>     would be dereferenced from the underlying Windows file handle,
>>
>>
>> It's the other way round: the PyHandle just forgets the win32 handle
>> value, and will not close it in its __del__.
>
> Sorry, I did write that backward. How is the win32 handle freed?

The .Detach() method returns an integer handle value which can then be 
closed using win32api.CloseHandle().

Mark

From sleerssen at gmail.com  Mon Dec 19 22:34:42 2011
From: sleerssen at gmail.com (Scott Leerssen)
Date: Mon, 19 Dec 2011 16:34:42 -0500
Subject: [python-win32] file descriptor issues using
	msvcrt.open_osfhandle()+os.fdopen()
In-Reply-To: <4EEFAD8B.5050901@gmail.com>
References: <32FE529F-0D6B-4D8A-91C3-10809322966B@gmail.com>
	<CAGmFidaaB9oEGaex9xBHwPo3VUsfLhSmCLex80CtJ7dkhi2WRQ@mail.gmail.com>
	<84739C66-AF7C-4D0E-AA4D-E4F1AEFDDD1D@gmail.com>
	<CAGmFidZ_08_jKRE3b85VCMMJ7UYMFiuO0Pa=H-H3TxJ-korp7w@mail.gmail.com>
	<0DD1C7B3-CB99-44FE-AAEA-8E0F998B16C8@gmail.com>
	<4EEFAD8B.5050901@gmail.com>
Message-ID: <921F76B1-25C1-4E69-B0FE-F05B48796DEB@gmail.com>


On Dec 19, 2011, at 4:32 PM, Mark Hammond wrote:

> On 18/12/2011 4:34 AM, Scott Leerssen wrote:
>> 
>> On Dec 17, 2011, at 12:25 PM, Amaury Forgeot d'Arc wrote:
>> 
>>> 2011/12/17 Scott Leerssen <sleerssen at gmail.com
>>> <mailto:sleerssen at gmail.com>>
>>> 
>>>    I did see that, but I interpreted that to mean that the PyHANDLE
>>>    would be dereferenced from the underlying Windows file handle,
>>> 
>>> 
>>> It's the other way round: the PyHandle just forgets the win32 handle
>>> value, and will not close it in its __del__.
>> 
>> Sorry, I did write that backward. How is the win32 handle freed?
> 
> The .Detach() method returns an integer handle value which can then be closed using win32api.CloseHandle().
> 

Excellent!  Thanks!


From randfb at yahoo.com  Mon Dec 19 23:24:00 2011
From: randfb at yahoo.com (Rand Batchelder)
Date: Mon, 19 Dec 2011 14:24:00 -0800 (PST)
Subject: [python-win32] Problem : COM with Array Argument
Message-ID: <1324333440.84831.YahooMailNeo@web160901.mail.bf1.yahoo.com>

I ran makepy on the DesignCad 20 type library.

I get Type mismatch errors with the function GetPoints:

a subset of the class definition is:
"""
class IDcadEntity(DispatchBaseClass):
??? CLSID = IID('{B4D5DA6E-EDEB-4816-BD4C-88C04C119C24}')
??? coclass_clsid = IID('{4FCCD850-D360-458C-8FB1-61CB90A28137}')


??? def GetPoint(self, iIndex=defaultNamedNotOptArg, dX=pythoncom.Missing, dY=pythoncom.Missing, dZ=pythoncom.Missing):
??? ??? return self._ApplyTypes_(13, 1, (11, 0), ((2, 1), (16389, 2), (16389, 2), (16389, 2)), u'GetPoint', None,iIndex
??? ??? ??? , dX, dY, dZ)

??? def GetPoints(self, iStartIndex=defaultNamedNotOptArg, iCount=defaultNamedNotOptArg, pXYZArray=defaultNamedNotOptArg):
??? ??? return self._ApplyTypes_(15, 1, (2, 0), ((2, 1), (2, 1), (24581, 3)), u'GetPoints', None,iStartIndex
??? ??? ??? , iCount, pXYZArray)

"""
I tried Ent.GetPoints(1,1, d)
where d is a list of 3 floats,

I get the following errors:
Traceback (most recent call last):
? File "<interactive input>", line 1, in <module>
? File "C:\Python27\lib\site-packages\win32com\gen_py\DD0F69E1-1E0D-4064-94FA-2ED7ADC72C5Dx0x20x0.py", line 4631, in GetPoints
??? , iCount, pXYZArray)
? File "C:\Python27\lib\site-packages\win32com\client\__init__.py", line 456, in _ApplyTypes_
??? self._oleobj_.InvokeTypes(dispid, 0, wFlags, retType, argTypes, *args),
com_error: (-2147352571, 'Type mismatch.', None, 3)

Please advise
thanks
Rand Batchelder
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111219/8a40488e/attachment.html>

From sleerssen at gmail.com  Mon Dec 19 23:37:09 2011
From: sleerssen at gmail.com (Scott Leerssen)
Date: Mon, 19 Dec 2011 17:37:09 -0500
Subject: [python-win32] file descriptor issues using
	msvcrt.open_osfhandle()+os.fdopen()
In-Reply-To: <0FFA640B-7F3C-4B73-B32A-65ED196D6B1F@gmail.com>
References: <32FE529F-0D6B-4D8A-91C3-10809322966B@gmail.com>
	<CAGmFidaaB9oEGaex9xBHwPo3VUsfLhSmCLex80CtJ7dkhi2WRQ@mail.gmail.com>
	<84739C66-AF7C-4D0E-AA4D-E4F1AEFDDD1D@gmail.com>
	<CAGmFidZ_08_jKRE3b85VCMMJ7UYMFiuO0Pa=H-H3TxJ-korp7w@mail.gmail.com>
	<0DD1C7B3-CB99-44FE-AAEA-8E0F998B16C8@gmail.com>
	<4EEFAD8B.5050901@gmail.com>
	<919906B4-07CF-44E0-B332-2D1804A9737C@gmail.com>
	<4EEFB429.9050300@skippinet.com.au>
	<0FFA640B-7F3C-4B73-B32A-65ED196D6B1F@gmail.com>
Message-ID: <CE835F1A-2AB4-4FAD-8835-CEA57FF3A86F@gmail.com>

adding python-win32 to the rest of the thread...  I missed a reply-all in here somewhere...

On Dec 19, 2011, at 5:06 PM, Scott Leerssen wrote:

> 
> On Dec 19, 2011, at 5:01 PM, Mark Hammond wrote:
> 
>> Note you are also closing h.  But more generally, I wouldn't be at all surprised to find that closing fd also closes the win32 handle associated with it, otherwise every CRT handle ever opened would leak the OS handle.
> 
> It does close the descriptor.  However, I still see screwy behavior with other sockets and descriptors in other threads.  It's almost like there are two sides of the descriptor, and one gets reused before the other is released, resulting in unexpected closure of descriptors or "stealing" of one threads descriptor when another thread opens a file or socket.  It is truly odd.
> 
>> IOW, I think you can just detach it and ignore closing it.
>> 
>> Mark
>> 
>> On 20/12/2011 8:43 AM, Scott Leerssen wrote:
>>> 
>>> On Dec 19, 2011, at 4:32 PM, Mark Hammond wrote:
>>> 
>>>> On 18/12/2011 4:34 AM, Scott Leerssen wrote:
>>>>> 
>>>>> On Dec 17, 2011, at 12:25 PM, Amaury Forgeot d'Arc wrote:
>>>>> 
>>>>>> 2011/12/17 Scott Leerssen<sleerssen at gmail.com
>>>>>> <mailto:sleerssen at gmail.com>>
>>>>>> 
>>>>>>   I did see that, but I interpreted that to mean that the PyHANDLE
>>>>>>   would be dereferenced from the underlying Windows file handle,
>>>>>> 
>>>>>> 
>>>>>> It's the other way round: the PyHandle just forgets the win32 handle
>>>>>> value, and will not close it in its __del__.
>>>>> 
>>>>> Sorry, I did write that backward. How is the win32 handle freed?
>>>> 
>>>> The .Detach() method returns an integer handle value which can then be closed using win32api.CloseHandle().
>>> 
>>> I just tried that, and after closing the file object and attempting to call win32api.ClosHandle() on the value returned by detach, I get an exception of "error: (6, 'CloseHandle', 'The handle is invalid.')"  Here's what the code looks like:
>>> 
>>>    h = win32file.CreateFile(
>>>        fname,
>>>        win32file.GENERIC_READ,
>>>        win32file.FILE_SHARE_READ,
>>>        None,
>>>        win32file.OPEN_EXISTING,
>>>        win32file.FILE_FLAG_BACKUP_SEMANTICS,
>>>        None
>>>        )
>>>    hd = h.Detach()
>>>    fd = msvcrt.open_osfhandle(hd, os.O_RDONLY)
>>>    f = os.fdopen(fd, 'rb')
>>>    f.read()
>>>    f.close()
>>>    h.Close()
>>>    win32file.CloseHandle(hd)
>>> 
>>> 
>> 
> 


From timr at probo.com  Mon Dec 19 23:59:22 2011
From: timr at probo.com (Tim Roberts)
Date: Mon, 19 Dec 2011 14:59:22 -0800
Subject: [python-win32] Problem : COM with Array Argument
In-Reply-To: <1324333440.84831.YahooMailNeo@web160901.mail.bf1.yahoo.com>
References: <1324333440.84831.YahooMailNeo@web160901.mail.bf1.yahoo.com>
Message-ID: <4EEFC1CA.7080107@probo.com>

Rand Batchelder wrote:
> I ran makepy on the DesignCad 20 type library.
>
> I get Type mismatch errors with the function GetPoints:
> ...
>     def GetPoints(self, iStartIndex=defaultNamedNotOptArg,
> iCount=defaultNamedNotOptArg, pXYZArray=defaultNamedNotOptArg):
>         return self._ApplyTypes_(15, 1, (2, 0), ((2, 1), (2, 1),
> (24581, 3)), u'GetPoints', None,iStartIndex
>             , iCount, pXYZArray)
> """
> I tried Ent.GetPoints(1,1, d)
> where d is a list of 3 floats,
>
> I get the following errors:
> Traceback (most recent call last):
> ...
>   File "C:\Python27\lib\site-packages\win32com\client\__init__.py",
> line 456, in _ApplyTypes_
>     self._oleobj_.InvokeTypes(dispid, 0, wFlags, retType, argTypes,
> *args),
> com_error: (-2147352571, 'Type mismatch.', None, 3)

There's not an awful lot we can do here.  Type code 24581 is a COM safe
array of doubles, passed by reference.  That seems to be what the
documentation shows.  You'd sure your values were all floats?

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From amauryfa at gmail.com  Tue Dec 20 01:38:47 2011
From: amauryfa at gmail.com (Amaury Forgeot d'Arc)
Date: Tue, 20 Dec 2011 01:38:47 +0100
Subject: [python-win32] file descriptor issues using
	msvcrt.open_osfhandle()+os.fdopen()
In-Reply-To: <CE835F1A-2AB4-4FAD-8835-CEA57FF3A86F@gmail.com>
References: <32FE529F-0D6B-4D8A-91C3-10809322966B@gmail.com>
	<CAGmFidaaB9oEGaex9xBHwPo3VUsfLhSmCLex80CtJ7dkhi2WRQ@mail.gmail.com>
	<84739C66-AF7C-4D0E-AA4D-E4F1AEFDDD1D@gmail.com>
	<CAGmFidZ_08_jKRE3b85VCMMJ7UYMFiuO0Pa=H-H3TxJ-korp7w@mail.gmail.com>
	<0DD1C7B3-CB99-44FE-AAEA-8E0F998B16C8@gmail.com>
	<4EEFAD8B.5050901@gmail.com>
	<919906B4-07CF-44E0-B332-2D1804A9737C@gmail.com>
	<4EEFB429.9050300@skippinet.com.au>
	<0FFA640B-7F3C-4B73-B32A-65ED196D6B1F@gmail.com>
	<CE835F1A-2AB4-4FAD-8835-CEA57FF3A86F@gmail.com>
Message-ID: <CAGmFidZLxoH036scaZP1Q_44Pg5j5-zOPfjk=BYjwPt6GxFmYw@mail.gmail.com>

2011/12/19 Scott Leerssen <sleerssen at gmail.com>

> >>> I just tried that, and after closing the file object and attempting to
> call win32api.ClosHandle() on the value returned by detach, I get an
> exception of "error: (6, 'CloseHandle', 'The handle is invalid.')"  Here's
> what the code looks like:
> >>>
> >>>    h = win32file.CreateFile(
> >>>        fname,
> >>>        win32file.GENERIC_READ,
> >>>        win32file.FILE_SHARE_READ,
> >>>        None,
> >>>        win32file.OPEN_EXISTING,
> >>>        win32file.FILE_FLAG_BACKUP_SEMANTICS,
> >>>        None
> >>>        )
> >>>    hd = h.Detach()
> >>>    fd = msvcrt.open_osfhandle(hd, os.O_RDONLY)
> >>>    f = os.fdopen(fd, 'rb')
> >>>    f.read()
> >>>    f.close()
> >>>    h.Close()
> >>>    win32file.CloseHandle(hd)
>

- h was Detach()ed, so h.Close() is a no-op.
- you should not win32file.CloseHandle(hd), the handle now owned by the
file object:
f.close() called the C function fclose(), which called close(fd) which
called CloseHandle.

f.close() is enough here!

-- 
Amaury Forgeot d'Arc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111220/904c081e/attachment.html>

From sleerssen at gmail.com  Tue Dec 20 04:38:58 2011
From: sleerssen at gmail.com (Scott Leerssen)
Date: Mon, 19 Dec 2011 22:38:58 -0500
Subject: [python-win32] file descriptor issues using
	msvcrt.open_osfhandle()+os.fdopen()
In-Reply-To: <CAGmFidZLxoH036scaZP1Q_44Pg5j5-zOPfjk=BYjwPt6GxFmYw@mail.gmail.com>
References: <32FE529F-0D6B-4D8A-91C3-10809322966B@gmail.com>
	<CAGmFidaaB9oEGaex9xBHwPo3VUsfLhSmCLex80CtJ7dkhi2WRQ@mail.gmail.com>
	<84739C66-AF7C-4D0E-AA4D-E4F1AEFDDD1D@gmail.com>
	<CAGmFidZ_08_jKRE3b85VCMMJ7UYMFiuO0Pa=H-H3TxJ-korp7w@mail.gmail.com>
	<0DD1C7B3-CB99-44FE-AAEA-8E0F998B16C8@gmail.com>
	<4EEFAD8B.5050901@gmail.com>
	<919906B4-07CF-44E0-B332-2D1804A9737C@gmail.com>
	<4EEFB429.9050300@skippinet.com.au>
	<0FFA640B-7F3C-4B73-B32A-65ED196D6B1F@gmail.com>
	<CE835F1A-2AB4-4FAD-8835-CEA57FF3A86F@gmail.com>
	<CAGmFidZLxoH036scaZP1Q_44Pg5j5-zOPfjk=BYjwPt6GxFmYw@mail.gmail.com>
Message-ID: <FAE3D76C-0CF5-48B5-ACCE-256CC6793A90@gmail.com>


On Dec 19, 2011, at 7:38 PM, Amaury Forgeot d'Arc wrote:

> 2011/12/19 Scott Leerssen <sleerssen at gmail.com>
> >>> I just tried that, and after closing the file object and attempting to call win32api.ClosHandle() on the value returned by detach, I get an exception of "error: (6, 'CloseHandle', 'The handle is invalid.')"  Here's what the code looks like:
> >>>
> >>>    h = win32file.CreateFile(
> >>>        fname,
> >>>        win32file.GENERIC_READ,
> >>>        win32file.FILE_SHARE_READ,
> >>>        None,
> >>>        win32file.OPEN_EXISTING,
> >>>        win32file.FILE_FLAG_BACKUP_SEMANTICS,
> >>>        None
> >>>        )
> >>>    hd = h.Detach()
> >>>    fd = msvcrt.open_osfhandle(hd, os.O_RDONLY)
> >>>    f = os.fdopen(fd, 'rb')
> >>>    f.read()
> >>>    f.close()
> >>>    h.Close()
> >>>    win32file.CloseHandle(hd)
> 
> - h was Detach()ed, so h.Close() is a no-op.
> - you should not win32file.CloseHandle(hd), the handle now owned by the file object:
> f.close() called the C function fclose(), which called close(fd) which called CloseHandle.
> 
> f.close() is enough here!


Thanks, that's very helpful.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111219/102a0ad8/attachment.html>

From mail at timgolden.me.uk  Tue Dec 20 14:20:55 2011
From: mail at timgolden.me.uk (Tim Golden)
Date: Tue, 20 Dec 2011 13:20:55 +0000
Subject: [python-win32] Several,
 totally isolated Python interpreters in the same process
In-Reply-To: <jcnl2h$6rv$1@dough.gmane.org>
References: <jcnl2h$6rv$1@dough.gmane.org>
Message-ID: <4EF08BB7.1070102@timgolden.me.uk>

On 19/12/2011 15:26, Thomas Heller wrote:
> I made an experiment that I wanted to share because it is imo kind of cool:
>
> I managed to embed a second Python interpreter in a Python process, with
> help of WinSxS (WinSxS is a Microsoft technology that allows to isolate
> plugins from other plugins, or from the calling process).
>
> The second interpreter is totally isolated from the calling process.
>
> The idea is to use this technology to decouple COM extensions
> implemented in Python from the caller and from other COM extensions.

Looks really interesting. I haven't a clue how I might
use it, but I'd love to find a reason to ...

TJG

From randfb at yahoo.com  Wed Dec 21 15:18:28 2011
From: randfb at yahoo.com (randfb)
Date: Wed, 21 Dec 2011 06:18:28 -0800 (PST)
Subject: [python-win32] Problem : COM with Array Argument
In-Reply-To: <4EEFC1CA.7080107@probo.com>
References: <1324333440.84831.YahooMailNeo@web160901.mail.bf1.yahoo.com>
	<4EEFC1CA.7080107@probo.com>
Message-ID: <1324477108139-2080239.post@n6.nabble.com>

>There's not an awful lot we can do here.  Type code 24581 is a COM safe
>array of doubles, passed by reference.  That seems to be what the
>documentation shows.  You'd sure your values were all floats?

>> I tried Ent.GetPoints(1,1, d)
>> where d is a list of 3 floats, 

Yes I tried various arrangements like:
d=[0.0, 0.0, 0.0] 

Has anyone ever been able to read or write a SAFEARRY from python thru COM ?

--
View this message in context: http://python.6.n6.nabble.com/Problem-COM-with-Array-Argument-tp1964132p2080239.html
Sent from the Python - python-win32 mailing list archive at Nabble.com.

From kxroberto at googlemail.com  Tue Dec 27 17:58:16 2011
From: kxroberto at googlemail.com (Robert)
Date: Tue, 27 Dec 2011 17:58:16 +0100
Subject: [python-win32] SIGSEGV crash at the end of apps using win32uiole /
	OCX
Message-ID: <jdcteh$2nb$1@dough.gmane.org>

This is about a ugly old bug which I dislike for long, which I can 
only suppress by a os._exit() near Python end. Now I found some 
more info. Perhaps somebody sees what it is about?

When a GUI app has ever used a win32uiole OCX like shown in the 
attached minimal example (pythonwin\pywin\Demos\ocx\webbrowser.py, 
just changed to use a standalone frame window), then at the very 
end of the app after all windows have been destroyed and Pythons 
sort of main() with destruction of all Python objects is reached, 
there is a SIGSEGV like shown below in 
python26!PyInterpreterState_Clear ().
Reproduced with various Python 2 versions.


Notes:


* Obviously extra threads are created by the browser OCX if at 
least one .Navigate() occurred.

* if .Navigate() is commented out, and just the Window/OCX 
creations/destructions are done the problem is not. (But then gdb 
doesn't show creation of extra threads. Thus the problem seems to 
have to do with extra threads.)

* "app.Run()" is commented out for a mere standalone run, to 
reproduce the bug easily. Thus the windows are drawn down right 
after creation. If not, then at exit, this (simple) app somehow 
doesn't terminate but seems to wait eternally because of extra IE4 
threads (?). When run with gdb as shown, the hang/eternal wait yet 
is not and the SIGSEGV occurs after .Run() too (!?)

* the problem is not, when the example is run in Pythonwin.exe at 
the end of Pythonwin.exe. The problem is not when such script is 
run with py2exe v0.5.2 stub. (maybe that does rather quickly 
something like os._exit() at the end). But the problem occurs with 
current py2exe 0.6.9 stub, and with all normal python(w).exe's - 
direct start or via interactive sessions.
When Pythonwin is run via python(w).exe and  "import 
pywin.framework.startup ....  app.Run()", and when then the 
webbcrash.py is executed, then after exit of Pythonwin the crash 
is there too. Thus Pythonwin.exe unlike python(w).exe seems to do 
just a sloppy cleanup (like py2exe0.5.2) and thus avoid the problem.

* when os._exit() is done at the end of the script, the problem is 
not existing: the ugly "solution" ;-)
sys.exit() doesn't "solve"  - probably because it doesn't prevent 
against destruction of all Python objects, threads ...

* pywin32 build 216  (but same with 214)

* BTW: if the original pythonwin\pywin\Demos\ocx\webbrowser.py 
(using a MDIClient) is run standalone through python(w).exe, there 
is also a ugly SIGSEGV or even worse C-level crash - now at start. 
It should only raise a Python exception due to the non-working MDI 
setup I guess.




Robert




=========== SIGSEV of webbcrash.py watched with gdb :


SIGSEGV: code "mov 0x4(%ebx),%eax" at 0x1e02e449 points to memory 
0x00000004. cannot read.

C:\scratch>python26 webbcrash.py
Run() ...
Run() terminated.
==> SIGSEGV crash
[uncomment "##app.Run()"]
C:\scratch>gdb python26 webbcrash.py
GNU gdb (GDB) 7.2
Copyright (C) 2010 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later 
<http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show 
copying"
and "show warranty" for details.
This GDB was configured as "mingw32".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from C:\bin/python26.exe...(no debugging symbols 
found)...done.
"C:\scratch/webbcrash.py" is not a core dump: File format not 
recognized
(gdb) run
Starting program: C:\bin/python26.exe
          (NOTE: ITS THE SAME WITH C:\Python26\python.exe)
[New Thread 6740.0x10bc]
Python 2.6.6 (r266:84297, Aug 24 2010, 18:46:32) [MSC v.1500 32 
bit (Intel)] on2
win32
Type "help", "copyright", "credits" or "license" for more information.
 >>> execfile('webbcrash.py')
Web Browser - http://www.google.com/
[New Thread 6740.0x19d0]
[New Thread 6740.0xdec]
[New Thread 6740.0x1d90]
[New Thread 6740.0x1f00]
[New Thread 6740.0x19cc]
[New Thread 6740.0x26c]
[New Thread 6740.0x118c]
[New Thread 6740.0x153c]
[New Thread 6740.0x1da0]
 >>> exit()

Program received signal SIGSEGV, Segmentation fault.
0x1e02e449 in python26!PyInterpreterState_Clear ()
    from C:\WINDOWS\system32\python26.dll
(gdb) bt
#0  0x1e02e449 in python26!PyInterpreterState_Clear ()
    from C:\WINDOWS\system32\python26.dll
#1  0x00000000 in ?? ()
(gdb) disass
...
    0x1e02e408 <+246>:   call   0x1e02e464 
<python26!PyThread_get_thread_ident>
    0x1e02e40d <+251>:   mov    %eax,0x50(%esi)
    0x1e02e410 <+254>:   mov    %edi,0x40(%esi)
    0x1e02e413 <+257>:   mov    %edi,0x28(%esi)
    0x1e02e416 <+260>:   mov    %edi,0x2c(%esi)
    0x1e02e419 <+263>:   mov    %edi,0x30(%esi)
    0x1e02e41c <+266>:   mov    %edi,0x34(%esi)
    0x1e02e41f <+269>:   mov    %edi,0x38(%esi)
    0x1e02e422 <+272>:   mov    %edi,0x3c(%esi)
    0x1e02e425 <+275>:   mov    %edi,0x18(%esi)
    0x1e02e428 <+278>:   mov    %edi,0x1c(%esi)
    0x1e02e42b <+281>:   mov    %edi,0x20(%esi)
    0x1e02e42e <+284>:   mov    %edi,0x24(%esi)
    0x1e02e431 <+287>:   cmp    %edi,0xc(%esp)
    0x1e02e435 <+291>:   je     0x1e02e43c 
<python26!PyInterpreterState_Clear+298
 >
    0x1e02e437 <+293>:   call   0x1e03111b 
<python26!PyThread_create_key+31>
    0x1e02e43c <+298>:   push   $0x1
    0x1e02e43e <+300>:   pushl  0x1e204f74
    0x1e02e444 <+306>:   call   0x1e02e248 
<python26!PyThread_acquire_lock>
=> 0x1e02e449 <+311>:   mov    0x4(%ebx),%eax
    0x1e02e44c <+314>:   pushl  0x1e204f74
    0x1e02e452 <+320>:   mov    %eax,(%esi)
    0x1e02e454 <+322>:   mov    %esi,0x4(%ebx)
    0x1e02e457 <+325>:   call   0x1e02e2eb 
<python26!PyThread_release_lock>
    0x1e02e45c <+330>:   add    $0xc,%esp
    0x1e02e45f <+333>:   pop    %edi
    0x1e02e460 <+334>:   mov    %esi,%eax
    0x1e02e462 <+336>:   pop    %esi
    0x1e02e463 <+337>:   ret
End of assembler dump.
(gdb)




=========== script to reproduce the crash (attached too)
=== slightly modified pythonwin\pywin\Demos\ocx\webbrowser.py


# This demo uses the IE4 Web Browser control.

# It catches an "OnNavigate" event, and updates the frame title.
# (event stuff by Neil Hodgson)

import win32ui, win32con, win32api, regutil
from pywin.mfc import window, activex
from win32com.client import gencache
import sys, os

WebBrowserModule = 
gencache.EnsureModule("{EAB22AC0-30C1-11CF-A7EB-0000C05BAE0B}", 0, 
1, 1)
if WebBrowserModule is None:
     raise ImportError("IE4 does not appear to be installed.")

class MyWebBrowser(activex.Control, WebBrowserModule.WebBrowser):
     pass
##    def OnBeforeNavigate2(self, pDisp, URL, Flags, 
TargetFrameName, PostData, Headers, Cancel):
##        self.GetParent().OnNavigate(URL)
##        #print "BeforeNavigate2", pDisp, URL, Flags, 
TargetFrameName, PostData, Headers, Cancel

##class BrowserFrame(window.MDIChildWnd):
class BrowserFrame(window.FrameWnd):
     def __init__(self, url = None):
         if url is None:
             self.url = "http://www.google.com"
         else:
             self.url = url
         pass # Dont call base class doc/view version...
     def Create(self, title, rect = None):
         style = win32con.WS_VISIBLE | win32con.WS_OVERLAPPEDWINDOW
         ##style = win32con.WS_CHILD | win32con.WS_VISIBLE | 
win32con.WS_OVERLAPPEDWINDOW
         ##self._obj_ = win32ui.CreateMDIChild()
         self._obj_ = win32ui.CreateFrame()
         self._obj_.AttachObject(self)
         self._obj_.CreateWindow(None, title, style, rect, self)
         rect = self.GetClientRect()
         rect = (0,0,rect[2]-rect[0], rect[3]-rect[1])
         self.ocx = MyWebBrowser()
         self.ocx.CreateControl("Web Browser", win32con.WS_VISIBLE 
| win32con.WS_CHILD, rect, self, 1000)
         self.ocx.Navigate(self.url)
         self.HookMessage (self.OnSize, win32con.WM_SIZE)
     def OnSize (self, params):
         rect = self.GetClientRect()
         rect = (0,0,rect[2]-rect[0], rect[3]-rect[1])
         self.ocx.SetWindowPos(0, rect, 0)
     def OnNavigate(self, url):
         title = "Web Browser - %s" % (url,)
         print title
         self.SetWindowText(title)


import pywin.mfc.thread
class MyApp(pywin.mfc.thread.WinApp):
     frame = None
     def InitInstance(self): #$pycheck_no
         global f
         self.frame = f = BrowserFrame()
         self.frame.Create("Web Browser")
         if 'pywin.debugger' not in sys.modules:
             self.SetMainFrame(self.frame)

if __name__=='__main__':
     ##win32ui.Enable3dControls()
     win32ui.EnableControlContainer()
     ##app = win32ui.GetApp()
     app = MyApp()
     app.InitInstance()
     ##f = BrowserFrame()
     ##f.Create("Web Browser")
     if 'pywin.debugger' not in sys.modules:
         ##app.SetMainFrame(f)
         print "Run() ..."
         ##app.Run()
         print "Run() terminated."
         ##f.DestroyWindow()
         ##print "f.DestroyWindow() 'ed."
         ##os._exit(0)

-------------- next part --------------
A non-text attachment was scrubbed...
Name: webbcrash.py
Type: text/x-python
Size: 2860 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111227/42b575e3/attachment.py>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: startframework.pyw
URL: <http://mail.python.org/pipermail/python-win32/attachments/20111227/42b575e3/attachment.ksh>

From kxroberto at googlemail.com  Tue Dec 27 18:02:55 2011
From: kxroberto at googlemail.com (Robert)
Date: Tue, 27 Dec 2011 18:02:55 +0100
Subject: [python-win32] Several,
	totally isolated Python interpreters in the same process
In-Reply-To: <jcnl2h$6rv$1@dough.gmane.org>
References: <jcnl2h$6rv$1@dough.gmane.org>
Message-ID: <jdctn7$4ft$1@dough.gmane.org>

Thomas Heller wrote:
> I made an experiment that I wanted to share because it is imo kind
> of cool:
>
> I managed to embed a second Python interpreter in a Python
> process, with
> help of WinSxS (WinSxS is a Microsoft technology that allows to
> isolate
> plugins from other plugins, or from the calling process).
>
> The second interpreter is totally isolated from the calling process.
>
> The idea is to use this technology to decouple COM extensions
> implemented in Python from the caller and from other COM extensions.
> But there may be other uses as well...
>
> The code is here:
>
> http://code.google.com/p/ctypes-stuff/source/browse/trunk/winsxs
>


is it necessary to have extra python files for each instance? or 
could one existing python installation run in multiple instances 
too this way ?


Robert


>
> Information I found useful about WinSxS and assemblies:
>
> http://msdn.microsoft.com/en-us/library/dd408052%28v=VS.85%29.aspx
>
> http://omnicognate.wordpress.com/2009/10/05/winsxs/
>
>
> Thomas