python 2.3.2 hangs when returning an int

omission9 omission9 at invalid.email.info
Tue Feb 24 21:27:32 EST 2004


Thanks to all who responded.  This took me forever to replicate and is 
not that easy to demonstrate outside of my fairly complex application. 
Basically, I feel that this is the fault of the Python garbage 
collector. What was happening was that in one function I locally defined 
a class which extended wxGrid. In  another function, which was called by 
the first one mentioned, I  instantiated another instance, local to the 
function, of that class. It  seems that even though the wxGrid class 
used in the calling function was  only used after the function call 
some stray references or something to  the other instance were still 
drifting around in the interpreter which  were causing some bizarro 
problems like the one in my original post. By  eliminating the instance 
in the function that was being called I  eliminated the problem. 
BIZARRE!! What made this so awful was that it  took many many iterations 
of these functions to generate the error. I  have no idea why it didn't 
blow the first, say, hundred times through.
Maybe if I ever get the motivation I'll attach a C debugger to the whole 
thing and look deeper into the guts.


omission9 wrote:

> Any advice on the following would be much appreciated. I have thrown 
> everything I have at it and am completely stumped. I apologize for the 
> length, I have tried to be as succint as possible.
> 
> I have a pythoncard/wxPython application. The application works well 
> except that users were reporting occasional crashes. The application 
> would simply stop and just hang. I was able to reproduce the bug 
> faithfully and ran the application in the python debugger pdb.py.
> The trace is below and I simply cannot understand why it is hanging 
> where it is. The application gets stuck executing
> 
> return len(self.data[0])
> 
> which is in this function
> 
>     def GetNumberCols(self):
>         if len(self.data) <= 0:
>             return 0
>         else:    
>             return len(self.data[0])
> 
> This works many many times during the applications execution, indeed 
> several successful runs of this line are visible immediatley above the 
> hang. I have also printed out the value of self.data. This is a list 
> containing 12 inner lists each of length 3 at the point of the hang and 
> is never larger than that.
> What on earth could cause this to stop the interpreter cold in its 
> tracks without an exception being thrown? CPU usage on the machine is 
> high throughout the application's execution and stays as such after the 
> "hang" begins. I eventually kill the app although I have let it run for 
> a long time and it never leaves this line.
> For what its worth, I am using python 2.3.2 on windows2k.
> 
> Anyone have idea what could be happening?
> 
>   c:\python23\lib\site-packages\wxpython\grid.py(664)_setOORInfo()->None
>   -> return val
>   (Pdb) s
>   --Return--
>    c:\python23\lib\site-packages\wxpython\grid.py(850)__init__()->None
>   -> self._setOORInfo(self)
>   (Pdb) s
> 
>  
> c:\cygwin\home\ar881\development\bt_analysis\console\ndbbtable.py(17)__init
> 
>   __()
>   -> self.data = data
>   (Pdb) s
> 
>  
> c:\cygwin\home\ar881\development\bt_analysis\console\ndbbtable.py(18)__init
> 
>   __()
>   -> self.colLabels = headings
>   (Pdb) s




More information about the Python-list mailing list