🚀 go-pugleaf

Under gcc 2.6.3, I'm reaching a condition where I'm getting segment 
faults on a malloc (no switches on gcc).  Compiled under level 3 debug 
(gdb) and executed, at the point of the segmentation fault, the back 
trace shows a call into __gcc_malloc, which calls _malloc.  When I try a 
"print malloc(7)", it faults, as it does with any value less than, say 
9.  Thus "print malloc(7)" yields a segmentation fault, while a "print 
malloc(9)" works.  I figure I've corrupted heap somewhere, so I slowly 
step through the code from the beginning.

At one point in the code (before the segmentation fault), I do two 
mallocs, one after the other, with valid positive arguments.  The 
returns are queer...they are identical and seem to be large values (e.g. 
0x1ff78).

The question is: Is there a location I can go to look up debug problems 
like this?  By the way, it does compile and execute clean under MS 
Visual C, although I am aware that there are some differences...I 
believe I've avoided the issues of constant strings (can be overwritten 
in MSC, not in gcc), and a check for NULL returns on malloc.

Thanks,

Carlos da Roza
daroza@cn.ca