System V Hangup Problems

[clapper at nadc.arpa]

Thanks to everyone who replied to my posting about System V hangup problems.
We had conjectured that somehow our software was receiving more than one
hangup signal; many of the replies confirmed this theory, providing convincing
background information to support it.  We changed our software so that the
hangup handlers ignore further hangup signals as their first actions; our
problem as effectively disappeared.  The processes can still be killed while
attempting to reset the signal; however, this hasn't happened yet, and
there's not much we can do about that problem anyway.

As a side note, I called the AT&T Software Support Hotline to see if they
could help.  I don't think the gentleman on the other end of the line
believed me when I explained the problem.  I had to let him log onto our
machine for a walk-through demonstration before he acknowledged that 
something was definitely wrong.  He said he'd get back to me; that was 
last week, and I haven't heard from him yet.  I'm not overly surprised, 
however; AT&T Software Support hasn't been that helpful to us in the past.

Thanks again for the prompt replies; the information saved us a lot of
valuable time.

Brian M. Clapper                        (clapper at NADC.ARPA)
Naval Air Development Center
Warminster, PA