Bug#549760: libipc-shareable-perl: FTBFS: tests failed

[Lucas Nussbaum]

On 19/10/09 at 22:34 +0200, gregor herrmann wrote:
> tag 549760 + unreproducible
> thanks
> 
> On Mon, 05 Oct 2009 20:44:48 +0200, Lucas Nussbaum wrote:
> 
> > During a rebuild of all packages in sid, your package failed to build on
> > amd64.
> > 
> > > t/38ipchv.t ... 
> > > Failed 2/5 subtests 
> > > t/40ipcref.t .. ok
> > > t/45obj.t ..... ok
> > > t/50ipcobj.t .. ok
> > > t/55lsync.t ... ok
> > > 
> > > Test Summary Report
> > > -------------------
> > > t/38ipchv.t (Wstat: 0 Tests: 3 Failed: 0)
> > >   Parse errors: Tests out of sequence.  Found (4) but expected (2)
> > >                 Tests out of sequence.  Found (5) but expected (3)
> > >                 Bad plan.  You planned 5 tests but ran 3.
> > > Files=14, Tests=91, 1916 wallclock secs ( 0.04 usr  0.04 sys +  0.53 cusr  0.11 csys =  0.72 CPU)
> 
> Hm, seems I haven't looked at that bug yet, and again it seems to be
> one of these heisenbugs; i.e. it builds fine in an i386 and a amd64
> cowbuilder sid chroot.
> 
> "Tests out of sequence" combined with the forking and sleeping in the
> test and the dynamic numbering for the test cases (!) might lead to
> some timing/load issues.
> 
> From looking at the code and adding a bit of debug output, the child
> process starts, sleeps until awakened by the parent, runs test (1),
> sleeps, gets awakened by the parent -- which has a waitpid($pid, 0)
> at that point, runs tests (2) and (3), then the parent should run
> tests (4) and (5).
> 
> Maybe that helps some better Perl coder to find out why the sequence
> gets twisted :)

Actually, the worst part is that the build sometimes simply blocks.

But I agree that it's very likely to be a load issue.
-- 
| Lucas Nussbaum
| lucas at lucas-nussbaum.net   http://www.lucas-nussbaum.net/ |
| jabber: lucas at nussbaum.fr             GPG: 1024D/023B3F4F |