Need advice debugging SIGSEV in memcpy()

Hi,

First of all thanks to those who replied to my question on the prerelease
directory!

I'm making some improvements in the Rewriter; now runs one hour on
a lightly-loaded production machine without crashing. :)

When I run it under gdb (this is Red Hat Linux 4.0, kernel 2.0.29),
I get a SIGSEGV in memcpy(), and a stack backtrace ("bt") loses
the context. In contrast, this works for SIGPIPE, making it easy to
show what exactly triggered the signal.

Any debugging hints on how to find out where the problem is?
Memory problems seem to be bite rather nastily. :)

Thanks,

-- 
miguel a.l. paraz  <map@iphil.net>                              +63-2-893-0850
iphil communications, makati city, philippines          <http://www.iphil.net>