I am running HP-Unix 10.2 and am getting a core dump in a
multithreaded application.  My real problem is that I am unable to
determine the location of the core dump.  I tried using the dde
debugger to debug the core file and got the following error msg
******************************************************************
Target process died because of: SIGSEGV - segmentation violation.
end.o has not been linked in.
Call to fopen(/var/opt/dde/dde_error_log, a) failed: No such file or
directory
Target process died in unknown location (possibly in dynamically
linked library).
Executing image in process 858: "/<location>/<filename>".
Stopped at: `va(7f76dd64) (7F76DD64)
******************************************************************

I know the program doesn't explicitly use any dynamically linked
library, and I have also tried stepping through each step in the
different threads but I still cant seem to find were the program is
getting the segmentation fault.

Is there anyway to see which thread caused the fault or the
instruction that was being executed when the fault occurred with the
DDE debugger?

jvadn0@uky.edu (Jon Vadney) writes:

> I tried using the dde
> debugger to debug the core file and got the following error msg

The DDE is the absolute worst debugger I have ever used.
Unfortunately, it is the only one that can debug MT apps on hpux 10.20
If at all possible, try moving to hpux 11.23 -- it has *real*
threads (not the DCE garbage), and gdb works quite well on it.

> Target process died because of: SIGSEGV - segmentation violation.
> end.o has not been linked in.

The missing 'end.o' may be the first problem: be sure to compile
*and* link with '-g'

The other thing you can do is run the app under gdb until it SIGSERVs
-- gdb will not be able to switch between threads, but it should
be able to tell you the stack trace for the thread that crashes.

Good luck.
-- 
In order to understand recursion you must first understand recursion.
Remove /-nsp/ for email.