Re: Different handling of "trap" in bash 4.4.20(1)-release (x86_64-pc-linux-gnu) and bash version 5.2.15(1)-release (x86_64-aeos5-linux-gnu)

[Chet Ramey]

On 9/27/23 5:08 AM, MESSNER Andreas via wrote:
> Classified as: {OPEN}
>
> Hello,
>
> I have a question concerning "trap"  handling in bash
> Script for testing:
> #/bin/bash
>
> function cleanup ()
> {
>      echo "Executing cleanup"
>
>      # trying to kill non-existing process
>      killall doesnotexist >/dev/null 2>&1
>      killall doesnotexist >/dev/null 2>&1
>
>      echo "Leaving cleanup"
> }
> trap cleanup INT CHLD
>
> echo "Calling cleanup"
> cleanup
> echo "Back from cleanup"
>
> sleep 1
>
> Executing the script in bash 4.4.20(1) shows the following result:
> Calling cleanup
> Executing cleanup
> Leaving cleanup
> Back from cleanup
>
> Executing the script in bash 5.2.15(1) shows the following result:

The change came in bash-5.0. From CHANGES:

n. The SIGCHLD trap is run once for each exiting child process even if job
   control is not enabled when the shell is in Posix mode.

(That isn't strictly correct; the SIGCHLD behavior is in effect even when
the shell is not in Posix mode.)

This made the behavior the same whether or not job control was enabled.
There was nothing in the documentation to imply it was limited. It also
allows recursive (nested) trap command execution if a SIGCHLD arrives
during the execution of a SIGCHLD trap.

There was a lively austin-group (POSIX) discussion about this back in
April, 2017.

Chet
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
                 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, UTech, CWRU    chet@case.edu    http://tiswww.cwru.edu/~chet/