Re: [PATCH 2/2] job.c: implementing child_execute_job() using posix_spawn(), and use it if present

[Paul Smith]

On Mon, 2018-07-23 at 07:10 +0200, Barath Aron wrote:
> On 07/22/2018 09:58 PM, Paul Smith wrote:
> > On Mon, 2018-07-09 at 09:05 +0200, Aron Barath wrote:
> > > ---
> > 
> > Thanks for the work you put into this!
> 
> You're welcome! :)
> > Unfortunately this change reveals some deeper problems that I will
> > need to address.  Basically, the old code never really expected
> > fork() to fail: if it did we intended to just give up completely. 
> > Thus, it expected child_execute_job() to always succeed except in
> > exceptional situations, where make itself was not able to continue.
> 
> That's good news and bad news.

This weekend I reworked the handling of failures in child_execute_job()
so that it behaves properly even when the child failure happens in the
current process (e.g., the fork() or posix_spawn()).

It needs a bit of cleanup--in particular writing a test is complicated
by the fact that errors currently have different formats when built
with fork() vs. posix_spawn().  I need to decide if I want to align
them or fix the test to understand the difference.

I also added a configure --disable-posix-spawn option for those who
want to continue to use fork()/vfork() even when posix_spawn() is
available.  Since there's one missing bit of functionality in the
posix_spawn() case (resetting the stack size in the child) and I don't
think that using posix_spawn() buys you much if anything over vfork()
in GNU/Linux, at least.