tags 812626 unreproducible thanks On Wed, 27 Jan 2016 gregor herrmann wrote: > > On Tue, 26 Jan 2016 21:14:35 +0100, Andreas Tille wrote: > > > > I cloned the git repo (very handy :)) and had a look, but I can't > > > reproduce the test failure, not even when running them manually under > > > high load. > > Hmmmm, I should have mentioned this: I also can not reproduce the > > problem. > > Ok, maybe the submitter can shed some light on this issue.
Michael, could you please be more verbose about the build conditions since two people can not reproduce this independently. > > > > > SSE2 veczweight, wchunk = 32 > > I get > > SSE2 veczweight, wchunk = 64 > > here instead (when building on amd64). > > Me too. > > > > I noticed that the non-perl-test output comes before t/02test.t while > > > it comes later (after the "ok 3" of t/02test.t) for me. I thought this > > > might be a parallelization problem but the test was run with -j1, and > > > it also passes for me with -jN. > > From my (limited) experience this kind of non-reproducible FTBFS are > > often connected to parallelization. I'm tempted to drop the --parallel > > from d/rules. > > Doesn't hurt, but since since the logs have > make -j1 test TEST_VERBOSE=1 > this doesn't seem to have any effect anyway. Kind regards Andreas. -- http://fam-tille.de