[galib] GAList and GAGenome Destructors - Segmentation Fault

[Miriam.ORiordan]

To whom it concerns,
I have derived a class GEListGenome using the galib class GAListGenome.  The version of galib I'm using is 2.4.7. Up until recently my experiments were dealing with small population sizes (usually < 90) and they were running to completion with no errors. When I increased the population size to 110 I got a segmentation fault. Using GDB I was able to isolate the line of code which caused the fault. It is in GAList.C in the destructor function. The following is the line of code

while(hd) delete GAListBASE:: remove(DYN_CAST(GANode<T>*,hd));


Now another strange thing is happening also.
When I increased the popsize to 150 again my experiment crashed with a segmentation fault but this time the line of code is in GAGenome.C in the GAGenome destructor function.

delete evd;


I have gone through my own code with a fine toothcomb and I can't see what could be causing the segmentation faults. If my code is the issue why doesn't the problem occur at smaller population sizes.

Has anybody ever seen the two segmentation faults that I've described and if so how did they overcome them?

I would greatly appreciate any help or advice that would help me solve the problems.

Thanks,
Miriam
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/galib/attachments/20120509/efc7bb86/attachment.htm