There is a repeatable crash in rxapi at revision 4300. It's probably always been there, I just threw in the 4300 so the bug title would be unique.
I'm stumped with it, so it will probably need your touch. I've attached a zip file with the stuff to recreate it.
What makes it hard, for me, is that it only happens with a release build. I can run the program that causes the crash under a debug build a thousand times and it never crashes.
The zip file has source and a simple make file, also the compiled executable.
The executable uses RexxStart() in a loop to run the included simple Rexx program 5 times. Run the program from the command line several times in a row. On the third execution from the command line, at the first invocaton of RexxStart(), it always crashes in rxapi.
I threw in a copy of the stack at the time of the crash, I always have the exact same stack. I can trace the sequence using the rxapi map file. It always goes:
processMessages -> dispatch -> createSessionQueue ->
queue = new DataQueue(session);
then a malloc in rxapi and a crash down in the kernel.
But, from there I'm stuck.