gEDA-dev: PCB MyFree

[Dan McMahill]

DJ Delorie wrote:

>>Given the choice, I'd rather add this feature to dmalloc.
>>Maybe dmalloc already has it?


> Perhaps, but dmalloc isn't the only malloc debugger out there, and who
> am I do try to predict what kind of weird debugging printfs I or
> others might need in the future?  I've sometimes found myself adding
> printfs to EVERY malloc/free, dumping the pointer, as well as other
> printfs printing out pointers, then using grep or perl to figure out
> the patterns and "what happened?"s.

I'd like to underscore this point.  There was a long standing bug in PCB 
that was fixed fairly recently where it was a malloc problem.  dmalloc 
didn't find it at all for me and neither did electric fence.  It took 
DJ's magic to find where we were writing way past the end of a buffer.

If we can keep things done in a way where alternative malloc debuggers 
can work, that would be good.

-Dan