Re: maintainer mode parallel make issue

2009-04-22 Thread Ralf Wildenhues
Hi Eric, * Ralf Wildenhues wrote on Mon, Apr 20, 2009 at 10:11:12PM CEST: > * Eric Blake wrote on Mon, Apr 20, 2009 at 06:44:17PM CEST: > > But the point remains that we DO want perl's flock to work for > > cross-process locking (automake and autoconf both need a lock on > > autom4te.cache/request

Re: maintainer mode parallel make issue

2009-04-20 Thread Ralf Wildenhues
Hi Eric, * Eric Blake wrote on Mon, Apr 20, 2009 at 06:44:17PM CEST: > Ralf Wildenhues gmx.de> writes: > > > You can try this; test against git next of Automake. > > However, it also causes a hang on Cygwin, so there may be an independent > > bug (in the test or in perl); the lock() called from

Re: maintainer mode parallel make issue

2009-04-20 Thread Eric Blake
Ralf Wildenhues gmx.de> writes: > You can try this; test against git next of Automake. > However, it also causes a hang on Cygwin, so there may be an independent > bug (in the test or in perl); the lock() called from the thread seems to > never complete. Hmm. On some systems flock(2) is documen

Re: maintainer mode parallel make issue

2009-04-18 Thread Eric Blake
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 According to Ralf Wildenhues on 4/18/2009 4:27 PM: > Tangential to this issue, I think XFile::lock should warn resp. fail > also when parallel non-GNU make is used and exposes a locking issue. I agree; and the patch looks good to me. At any rate, my

Re: maintainer mode parallel make issue

2009-04-18 Thread Ralf Wildenhues
Tangential to this issue, I think XFile::lock should warn resp. fail also when parallel non-GNU make is used and exposes a locking issue. I'm not quite sure whether this can actually happen in practice, as for example BSD make does not update the target `Makefile' before updating regular targets, b

Re: maintainer mode parallel make issue

2009-04-16 Thread Eric Blake
Eric Blake byu.net> writes: > > The above looks like XFile::lock fails to do its work. We should have a > > test that ensures that, in either Autoconf or Automake. > > Indeed it is possible that some of the cygwin 1.7 improvements might be the > cause of XFile::lock behaving differently; if th

Re: maintainer mode parallel make issue

2009-04-16 Thread Ralf Wildenhues
* Eric Blake wrote on Thu, Apr 16, 2009 at 09:31:29PM CEST: > At any rate, I'll repeat this information to the cygwin developers, along > with > any ideas we have of how to trim it into a smaller testcase implicating just > XFile::lock. You can try this; test against git next of Automake. Howev

Re: maintainer mode parallel make issue

2009-04-16 Thread Eric Blake
Eric Blake byu.net> writes: > > > > Do you see these failures with Automake 1.10.2 also? > > I'll start checking that, but thought I'd at least give a first round of > replies. > > > Which Autoconf versions are you using? > > This was with bleeding edge autoconf.git, but as part of downgradi

Re: maintainer mode parallel make issue

2009-04-16 Thread Eric Blake
Eric Blake byu.net> writes: > > Have you seen them on systems other than Cygwin? > > How old is this Cygwin version? > > Haven't tested other systems yet. This was on cygwin 1.7 (still in beta), so > it could very well be a cygwin problem, although I am pretty sure that I have > also seen it

Re: maintainer mode parallel make issue

2009-04-16 Thread Ralf Wildenhues
Hi Eric, thanks for the bug report. * Eric Blake wrote on Thu, Apr 16, 2009 at 04:17:32PM CEST: > I'm using origin/next at the moment, and seeing failures like this anytime I > touch configure.ac then attempt a parallel make on cygwin: Do you see these failures with Automake 1.10.2 also? Which

maintainer mode parallel make issue

2009-04-16 Thread Eric Blake
I'm using origin/next at the moment, and seeing failures like this anytime I touch configure.ac then attempt a parallel make on cygwin: $ make -j2 check CDPATH="${ZSH_VERSION+.}:" && cd . && /bin/sh /home/eblake/m4-patch/build- aux/missing --run aclocal-1.10c -I ltdl/m4 CDPATH="${ZSH_VERSION+.}:"