On May 24, 2022, Bart Schaefer <schaefer@brasslantern.com> wrote:
> zargs was re-coded as of the 5.9 release so that it waits for
> individual jobs instead of using a single "wait" for all of them.
> This means it can detect nonzero exit in cases where earlier revisions
> missed it.

Got it. So we're now seeing a more precise return code than before 5.9.

> Something is causing the subshell that was running f to exit with
> status 19 even though f returned 0.

Any idea what could be returning 19? Looks like it's something outside
the scope of f. At least the code is consistently 19, when it happens.

Kind regards,
Eric