On Fri, 16 Apr 1999, Brian Buhrow wrote:

# they had going, stay running after window exits.  Sending a SIGHUP to these
# shells causes them to exit normally, killing all their children along the
# way.  It seems that any children of window aren't getting properly signaled
# when window exits.  Is this more fallout from that funny business from Grey
# Wolf about children not getting signals?  In any case, I would consider
# this a pretty big failure, as it means I need to manually kill all my
# shells each time I log out, otherwise, the next time I run window, csh
# complains that it cannot give me job control because it cannot access the
# tty.

I have been experiencing this myself as well, most especially if I neglect
to log out of all my xterms before terminating my session.  Something is
not getting properly cleaned up.

I have noticed that POSIX specified that only setpgid(0,0) is really per-
mitted; I don't know if setpgid(pid,pgrp) results in undefined behaviour
or what.

I *do* know that if I setpgid(pid,pgrp) under NetBSD, I get ESRCH, even
if pid exists.  Something Is Intrinsically Wrong, here.