Hi,
I'm not quite understanding what I'm doing (happens often). This pseudocode works:

#!/bin/pseudoksh

function kill_parent {
when i_want_to ; do
sleep 2
kill -TERM $PPID
exit
done
}

kill_parent &
ssh remote_host sh <<-EOF
notify_kill_parent # let the bg function know it's ok to kill parent
run_command # something x-based
EOF


Now, if I start my script in an xterm:
xterm -e run_my_script
It works as expected! The xterm pops up, ssh queries me for a password, the command runs and displays, and, seconds later, the xterm is killed. This is what I want to have happen. But I'm not sure why the remote shell and my application (run_command, above) still stick around. I've been pretty sloppy with stdin/stdout so I'm surprised they don't disappear when I kill the parent (xterm).

I'm happy it works, but I'm worried about a bug creeping in later- for example, since I'm not handling stdout, am I in danger of filling a buffer that is going to make the app hang? Or something like this.

I'm looking for insight into what happens when xterm calls ssh which calls a shell which calls a command. Xterm dies but the command does not. Why?
Thanks.
-Schwage