Re: PATCH: subshell into a new process group

On Sun, May 10, 2026, at 1:43 PM, Bart Schaefer wrote:
> I don't think it needs a new ESUB flag.  I've tried several test cases
> and can't find any way in which the following results in materially
> different behavior (other than the intended).  Anyone?  With this, the
> process group has already been changed before "disown".

Is that not too big a change? Consider eg

setsid zsh -c '(sleep 1 && echo alive) & kill -- -$$'

I'd expect that to continue to kill everything, and was thinking
that the extra flag is needed if we want the subshell to survive
only if we use '&!' instead.

It's made confusing by the fact that ESUB_PGRP is a misnomer,
it's applied to every call path except one (very unusual) area
something to do with pipelines where the call omits the flag:
   entersubsh(ESUB_ASYNC, NULL);
but immediately (after a long comment) does this manually anyway:
  setpgrp(0L, mypgrp = getpid()); 

I think the idea is that ESUB_PGRP really indicates whether jobtable
bookkeeping is supposed to be reset, not whether we want a new
process group.  The flags could be reworked to be more explanatory
(current ESUB_PGRP could be inverted to be eg ESUB_CONTINUATION and
passed only for the one special case, etc), but it would be a lot of
mechanical code churn.