Re: Need help with serious memory gremlins

[Marcus G. Daniels]

>>>>> "MB" == Matthew M Burke <address@hidden> writes:

MB> P.S. Here's the stack trace.  I get very similar results from a
MB> number of different fiddling around attempts.  In particular, at
MB> one point I had merely added a variable of type id to an
MB> interface, set it to NULL in the class's createBegin method and
MB> was still getting core dumps!

Program received signal SIGSEGV, Segmentation fault.
0x4010b8dd in chunk_alloc () at malloc.c:2680
malloc.c:2680: No such file or directory.
(gdb) where
#0  0x4010b8dd in chunk_alloc () at malloc.c:2680
#1  0x4010b295 in malloc () at malloc.c:2349

Check for things like malloc's of size zero, `free's of already
freed things, array accesses and assignments that step over their 
allocated boundaries. 

This sort of investigation can be partially automated by using a
special malloc library.  One of my favorite packages for debugging
malloc problems is dmalloc:

  http://www.letters.com/dmalloc/

It takes some time to learn how to use, but usually will find misuses
of malloc.

Another package, which is more powerful in some ways is Checker.

  http://www.gnu.org/software/checker/checker.html

However, Checker only works on GNU/Linux and Solaris.  Checker is
integrated with gcc.


                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.
                  ==================================