cvs commit: ports/audio/lame Makefile

[Alexey Dokuchaev]

On Mon, Mar 30, 2009 at 07:40:44PM +0400, Max Brazhnikov wrote:
> On Mon, 30 Mar 2009 18:10:42 +0300, Ion-Mihai Tetcu wrote:
> > On Mon, 30 Mar 2009 14:26:22 +0000
> > Alexey Dokuchaev <danfe at FreeBSD.org> wrote:
> > > On Mon, Mar 30, 2009 at 04:00:43PM +0200, Emanuel Haupt wrote:
> > > > >
> > > > > lame fails on my tinderbox after enabling jobs. Removing
> > > > > MAKE_JOBS_SAFE helps.
> > > >
> > > > In that case it should be marked MAKE_JOBS_UNSAFE.
> > >
> > > Weird.  I could not reproduce it on i386 with make -DFORCE_MAKE_JOBS
> > > MAKE_JOBS_NUMBER=8.  Is there build log available?  I'd rather fix the
> > > port than blindly mark it as MAKE_JOBS_UNSAFE.
> >
> > It also builds OK with MAKE_JOBS_NUMBER=16, on amd64-{6,7,8}
> 
> Build log:
> http://people.freebsd.org/~makc/tmp/lame-3.98.2_1.log
> 
> ~> sysctl -n kern.smp.cpus
> 4

I could not reprocude it, maybe my box is too slow or something.
However, many Linux distros include this patch named ``lame-pmake.patch'',
and say something about fixing parallel builds.  It looks like it might
fix what you see (some race between test -d and mkdir).

Can you check please?  Patch (this one is from ASP Linux) is attached.

./danfe
-------------- next part --------------
diff -up lame-398-2/libmp3lame/i386/Makefile.in.pmake lame-398-2/libmp3lame/i386/Makefile.in
--- lame-398-2/libmp3lame/i386/Makefile.in.pmake	2008-10-20 23:14:56.000000000 +0200
+++ lame-398-2/libmp3lame/i386/Makefile.in	2008-10-20 23:15:18.000000000 +0200
@@ -482,7 +482,7 @@ uninstall-am:
 	$(NASM) $(NASMFLAGS) $< -o $@ -l $@.lst
 
 .nas.lo: $< nasm.h
-	test -d .libs || mkdir .libs
+	mkdir -p .libs
 	$(ECHO) '# Generated by ltmain.sh - GNU libtool 1.5.22 (1.1220.2.365 2005/12/18 22:14:06)' >$@
 	$(ECHO) "pic_object='.libs/$*.o'" >>$@
 	$(ECHO) "non_pic_object='$*.o'" >>$@