Re: [PATCH] Fix foreground dead jobs in trap handlers reported like background ones in `jobs'

[Chet Ramey]

On 7/17/22 11:52 PM, Koichi Murase wrote:

> Bash Version: 5.1.4, also in the current devel 87a6e89e
>
> Description:
>
>    In interactive shells, the foreground dead jobs (i.e., foreground
>    jobs that have already terminated) in trap handlers, commands called
>    through `bind -x', and `PROMPT_COMMAND' (let us call these three
>    contexts as the affected context. Note: There might be even other
>    similar contexts) are treated as if they are the background dead
>    jobs.  This causes several unintended behaviors: (1) When `jobs' is
>    called in the affected contexts, all the foreground jobs that have
>    already terminated in the same call of the affected contexts are
>    printed.  (2) When `wait -n' is called in the affected contexts, it
>    immediately returns with the exit status of the previous foreground
>    dead job.
>
>    These behaviors are different from those in the normal case where
>    `jobs' and `wait -n' are executed outside the trap handler, and I do
>    not see the reasoning for this behavior difference between inside
>    and outside of the affected contexts.  I hope the behavior in the
>    affected contexts would be similar to that in the normal contexts
>    where the foreground dead job information is not printed.

Thanks for the detailed report and suggested patches. I applied a variant
of your patch 3; the change to print_job appears not to be needed (at least
all of your test cases work without it).

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/