[python-win32] Re: pythonw.exe hardcrash on py2exe wmi-sample / Pythin 2.4

[Thomas Heller]

"Massa, Harald" <harald.massa at suedvers.de> writes:

> (xposting from py2exe users)
>  
>  
>  Hello,
>
> I just upgraded my system to Python 2.4. I tried to do an former 
> 2.3-setup.py script to repackage my application within Python 2.4
>
>
> This application uses wmi and therefor has an typelib-argumentlist.
>
> on **generating typelib stubs** the run fails. I went down to 
> py2exe/samples/advanced where the wmitest is, and used the setup.py there.
>
> same error. py2exe tries to generate typelib stubs, and crashes HARD. 
> Not a simple traceback, but
>
> """Fehler in Anwendung" ... "Die Anweisung in 0x1e03ab43´verweist auf 
> Speicher in "0x34202c39". Der Vorgang "read" konnte nicht auf dem 
> Speicher duchgeführt werden."""
>
>
> (that is the german error from windows, when Memory at not allowed areas 
> is being accessed)
>
> After that, only "ok" and "Abbrechen" (which do the same :( )
>
> I tracked that down to the line (in build_exe)
>
> makepy.GenerateFromTypeLibSpec(info, bForDemand = False)
>
> that is the one that crashes badly. It kills everything ... I put 
> debugging-prints around it, and even the debugging-prints immediately 
> before are not flushed to stdout anymor.
>
> What can I do? OS is Windows XP prof SP2, acual updates. Is there an old 
> pywin32 bug exposed just through sp2??

I seem to rememeber that this was a problem, reported somewhere, with
makepy generating source code with line length > 512 characters, and /
or had it to do something with the coding declaration at the top of the
file?  And could it be that the bForDemand=... value did also have
something to do with it?  And did it only occur with Python 2.4?

But I don't remember more details...

Thomas