trap .. EXIT doesn't execute when exit is due to a signal

trap .. EXIT doesn't execute when exit is due to a signal

[Matthias B.]

trap ... EXIT doesn't execute ... when the shell is terminated due to a
signal (e.g. Ctrl-C). I think this is a bug. It's certainly undesirable,
because it causes cleanup tasks to not be executed. Bash doesn't have this
problem.

MSB

-- 
Physics is about prediction, not truth.

Re: trap .. EXIT doesn't execute when exit is due to a signal

[Dan Nelson]

In the last episode (Feb 12), Matthias B. said:
> trap ... EXIT doesn't execute ... when the shell is terminated due to
> a signal (e.g. Ctrl-C). I think this is a bug. It's certainly
> undesirable, because it causes cleanup tasks to not be executed. Bash
> doesn't have this problem.

zsh-users 3513 describes what is happening here, I think.  Ash and
/Solaris bin/sh have the same behaviour.

-- 
	Dan Nelson
	dnelson@allantgroup.com

Re: trap .. EXIT doesn't execute when exit is due to a signal

[Bart Schaefer]

On Feb 12,  3:04pm, Dan Nelson wrote:
} Subject: Re: trap .. EXIT doesn't execute when exit is due to a signal
}
} In the last episode (Feb 12), Matthias B. said:
} > trap ... EXIT doesn't execute ... when the shell is terminated due to
} > a signal (e.g. Ctrl-C). I think this is a bug. It's certainly
} > undesirable, because it causes cleanup tasks to not be executed. Bash
} > doesn't have this problem.
} 
} zsh-users 3513 describes what is happening here, I think.  Ash and
} /Solaris bin/sh have the same behaviour.

Yep, that's exactly it.

It just occured to me that it may very well be the case that bash is
using the POSIX atexit() handler to execute its exit traps, which I
believe would handle the situation described in 3513 without having
to do any special handling of SIGINT.

However, I'm not positive that covers all the cases in which zsh runs
the exit trap, and it'd have to be used carefully to be sure the exit
trap wasn't run more than once.

Re: trap .. EXIT doesn't execute when exit is due to a signal

[Peter Stephenson]

Bart Schaefer wrote:
> It just occured to me that it may very well be the case that bash is
> using the POSIX atexit() handler to execute its exit traps, which I
> believe would handle the situation described in 3513 without having
> to do any special handling of SIGINT.
> 
> However, I'm not positive that covers all the cases in which zsh runs
> the exit trap, and it'd have to be used carefully to be sure the exit
> trap wasn't run more than once.

No, it's not good enough for traps exiting functions.  bash only has an
EXIT trap for the whole shell.

bash-2.05b$ fn() { trap 'echo exiting' EXIT; echo fn run; }
bash-2.05b$ fn
fn run
bash-2.05b$ exit
exit
exiting

You couldn't use atexit() to fix the problem in zsh.  You'd need to
return from the interrupt and unwind the call stack to the top, then
exit.

It is a nuisance that EXIT traps aren't always run, however.

-- 
Peter Stephenson <pws@csr.com>                  Software Engineer
CSR PLC, Churchill House, Cambridge Business Park, Cowley Road
Cambridge, CB4 0WZ, UK                          Tel: +44 (0)1223 692070


**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.

**********************************************************************