Hello,
It's possible to have the following script never terminate:
#!/bin/bash
subfun() {
trap "echo received term for $*;exit 1" SIGTERM
echo my pid $BASHPID
i=0
while true; do
echo $BASHPID subfun $* $i
i=$((i+1))
sleep 5
done
}
trap 'echo main; exit 1' 0 1 2 3 15
subfun xxx &
funpid=$!
subfun yyy &
twopid=$!
echo funpid=$funpid twopid=$twopid mypid=$$
echo killing $funpid $twopid
kill $funpid $twopid
echo waiting on everything
wait
echo done
The output is:
my pid 12186
12186 subfun xxx 0
funpid=12186 twopid=12187 mypid=12185
killing 12186 12187
waiting on everything
my pid 12187
12187 subfun yyy 0
received term for xxx
12187 subfun yyy 1
12187 subfun yyy 2
12187 subfun yyy 3
and 12187 keeps printing. It would seem there $! can return a pid of the
subshell before it's in a state that it can be killed. This has to do
with setting of the trap handler, because if the trap handler of the
main shell is changed to : trap 'echo main; exit 1' EXIT - Then it never
hangs. I.e the more trap handlers that subshells have to reset to their
defaults the higher the chances this race condition is hit. Also if I
change the sleep in the subshell function to anything below 5 I also
seem to be unable to hit it. Also sometimes when it doesn't hang I get
the following output:
my pid 12283
12283 subfun xxx 0
funpid=12283 twopid=12284 mypid=12282
killing 12283 12284
waiting on everything
received term for xxx
done
main
It misses a "Received term for yyy" line.
I reproduced this behavior both with version 4.4.20(1)-release (that
comes with ubuntu 18.04) as well as 5.0.18(1)-release built from source.
Any ideas what might be causing it?
