Coredump on SunOS

I am running an application on SunOS 5.7 against an Oracle 8.1.6 database. The problem is when my application terminates, a SEGV occurs and a core is dumped. The very last line of my code that is executed is a return statement from main().

Here is a look at the callstack:

>dbx $bin/testmsg core

...snip...
detected a multithreaded program
t_at_1 (l_at_1) terminated by signal SEGV (no mapping at the fault address) (/opt/SUNWspro/bin/../WS6U1/bin/sparcv9/dbx) where current thread: t_at_1
=>[1] 0xfd9f9b0c(0x3, 0x0, 0xff3a02fc, 0x11e00, 0xff3dc7ac, 0x1), at 0xfd9f9b0b
  [2] _exithandle(0xfeab6108, 0xfeab9e94, 0xfec7a054, 0xffbecd54, 0xffbecd5c, 0xffbecd64), at 0xfea20258
  [3] exit(0x0, 0xffbecfd4, 0xffbecfdc, 0x27000, 0x0, 0x0), at0xfea983e0

As you can see, there is not much to go on there. What we have determined is that if the database connection is not made, the program will end gracefully. It is only when the database connection is made that the coredumps occur. We are using a custom database driver built on top of Oracle's OCI.

How do I proceed diagnosing/fixing the problem?

Many thanks,

Matt Received on Tue Jan 30 2001 - 18:51:17 CET