Re: Issues with larger numbers with -j

["J. Lewis Muir" <jlmuir%imca-cat.org@localhost>]

On 06/13, Jason Bacon wrote:
> On 06/13/17 20:03, John Klos wrote:
> >Hi,
> >
> >I've noticed a number of instances when compiling pkgsrc packages
> >with a large value for -j failures, where compiling with a small
> >number or -j 1 compiles just fine. For instance, devel/readline
> >and lang/python36 have failed in the last few minutes with -j 40.

I assume the file system is not NFS or similar?

> I saw this with htcondor in FreeBSD ports a few years ago.  The
> build consistently failed on a 16 core machine and consistently
> worked on a 4 core machine.  Some Makefiles are not really job-safe
> because they don't accurately describe all dependencies, but a small
> number of make jobs won't trigger the problem because prerequisites
> happen to get built before they're needed.

I agree with Jason that it seems likely a make-jobs safe issue where
the devel/readline and lang/python36 distribution Makefiles are either
not make-jobs safe or have a bug.  If they're not make-jobs safe, they
should be marked as such with MAKE_JOBS_SAFE=no.  If they're supposed to
be make-jobs safe, then they likely contain a bug that makes them not
actually make-jobs safe.

I reported a similar issue with editors/vim-share a while back,
and, with the help of some people on the list, found the bug in the
distribution's Makefile:

  https://mail-index.netbsd.org/pkgsrc-users/2016/11/15/msg023989.html

Regards,

Lewis