Re: maintainer mode parallel make issue

[Ralf Wildenhues]

* Eric Blake wrote on Thu, Apr 23, 2009 at 12:29:01AM CEST:
> Ralf Wildenhues <Ralf.Wildenhues <at> gmx.de> writes:
> 
> > > Yes, definitely.  And a testsuite addition to Autoconf or Automake to
> > > ensure this would be good, I guess.
> > 
> > Proposed patch against Autoconf.  Tested on GNU/Linux and Cygwin: as it
> 
> cygwin 1.5.x, I presume?

Yes, 1.5.25; sorry, I should have been clear here.

> > is, the patch hasn't caused a failure in several runs.  When I comment
> > out the
> >   $icache_file->lock (LOCK_EX);
> > 
> > line in bin/autom4te.in or go for a shared lock with LOCK_SH, rebuild,
> > then the test has a fairly high chance of failing.
> 
> Your methods of testing seem sane.  I agree that using no lock or just 
> LOCK_SH 
> are insufficient to protect the second process from seeing modifications made 
> while the first still owns the file.

But note that that's not what I'm testing.  Again, I should have
mentioned this right away, but: trying to find out for sure which
process finished last by looking at configure contents is much harder,
because then you need to make them use different input, take into
account time stamp issues (i.e., the second one needs to start >2s
later than the first) and so on.  I figured this was more work than
necessary if failures were easily seen otherwise, too.

> > OK to apply?

> Yes, looks good to me.

Done, thanks.

Cheers,
Ralf