Memory leaks in win pidgin 2.5.1?

[dave1g]

found this GPLed sf project  winleak - it was able to read pidgins symbols.
http://sourceforge.net/projects/winleak/

attached is the output of running pidgin for a few minutes...doing almost
nothing so its a bit surprising it was limited to keeping track of 50
stacks.  http://www.nabble.com/file/p19387252/MemoryHooksReport.txt
MemoryHooksReport.txt 


dave1g wrote:
> 
> I ran a free trial of a program i found online called Memory Validator on
> Debug version of windows pidgin 2.5.1. [
> http://www.softwareverify.com/cpp/memory/index.html ]
> 
> It claimed to have found quite a few memory leaks. Unfortunately as is the
> case with most (all?) windows debugging type applications. It required a
> PDB file that MS VC++ produces in order to see the symbols even with the
> debug version of pidgin. And so the stack traces it gave were just a bunch
> of memory addresses but they were scattered around in various pidgin dlls. 
> After much searching there appears to be no way to convert form the
> symbols that GDB produces in the exe to a PDB file. Even saw some post by
> datallah involved in a related discussion.
> 
> Would it be possible to maybe link in the dmalloc library [
> http://dmalloc.com/ ] for the debug version of pidgin? (even better to
> actually #include<dmalloc.h> I think it would be more successful in
> identifying human readable traces for where memory has been leaked if that
> is in fact the case. The symbols must be there since the RPT files have
> fully readable stack traces.
> 
> DMalloc is licensed creative commons attribution share alike, sounds
> GPLish but not sure. There is atleast 1 gpled program that uses it,
> busybox.
> 
> Just an idea, thanks
> David Grohmann
> 
> 
> 

-- 
View this message in context: http://www.nabble.com/Memory-leaks-in-win-pidgin-2.5.1--tp19386704p19387252.html
Sent from the Pidgin Support List mailing list archive at Nabble.com.