debhelper-9.20140228-1.fc22 failed to build for me because of tests: # Failed test 'DEB_BUILD_OPTIONS=parallel=5 with --parallel' # at t/buildsystems/buildsystem_tests line 561. # Structures begin differing at: # $got->[0] = 'make[1]: *** [FIRST] Error 10' # $expected->[0] = Does not exist # Failed test '(exit status=0) DEB_BUILD_OPTIONS=parallel=5 with --parallel' # at t/buildsystems/buildsystem_tests line 562. # got: '2' # expected: '0' # Looks like you failed 2 tests of 297. t/buildsystems/buildsystem_tests .. Dubious, test returned 2 (wstat 512, 0x200) Failed Anothewr scratch build in koji <http://koji.fedoraproject.org/koji/taskinfo?taskID=7171555> failed with: # Failed test 'autoconf/stamp_configure contains extra args' # at t/buildsystems/buildsystem_tests line 472. # Structures begin differing at: # $got->[0] = '--host=arm-linux-gnueabi' # $expected->[0] = Does not exist # Failed test 'autoconf/stamp_test exists' # at t/buildsystems/buildsystem_tests line 477. # Failed test 'bld/dir/stamp_configure contains extra args' # at t/buildsystems/buildsystem_tests line 472. # Structures begin differing at: # $got->[0] = '--host=arm-linux-gnueabi' # $expected->[0] = '--extra-autoconf-configure-arg' # Failed test 'bld/dir/stamp_test exists' # at t/buildsystems/buildsystem_tests line 477. # Looks like you failed 4 tests of 297. t/buildsystems/buildsystem_tests .. Dubious, test returned 4 (wstat 1024, 0x400) Failed 4/297 subtests Test Summary Report ------------------- t/buildsystems/buildsystem_tests (Wstat: 1024 Tests: 297 Failed: 4) Failed tests: 248, 252, 262, 266 Non-zero exit status: 4 Files=8, Tests=588, 24 wallclock secs ( 0.39 usr 0.04 sys + 15.35 cusr 3.15 csys = 18.93 CPU) Result: FAIL Similar issue has already been reported as bug #1106129. This looks like a race in the tests.
I report this to devel ML https://lists.fedoraproject.org/pipermail/devel/2014-February/195730.html but was not conclusive , maybe upstream (debhelper) the problem ?
Just for your information, the first failure is from local mock on x86_64. Try ask upstream how to debug it.
(In reply to Petr Pisar from comment #2) > Just for your information, the first failure is from local mock on x86_64. > Try ask upstream how to debug it. sorry I'm in holidays, the bug you mention, the problem was when builder is one arm , not happened when build was not a arm . So if you got tests fails on builder that is not an arm we got another bug . on my local mock I can't reproduce the problem because my laptop is an x86_64. your first fail test is not about --host=arm-linux-gnueabi , I will checkout when I'm back
(In reply to Petr Pisar from comment #2) > Just for your information, the first failure is from local mock on x86_64. > Try ask upstream how to debug it. Hi, today local mock on x86_64 builds well for rawhide, although I had a fail some days ago , but not now so I assume this problem is solved. Can you check it and close the bug if it is fixed ? Thanks,
The only change in code is: diff -Naur debhelper-9.20140228/Debian/Debhelper/Buildsystem/autoconf.pm debhelper/Debian/Debhelper/Buildsystem/autoconf.pm --- debhelper-9.20140228/Debian/Debhelper/Buildsystem/autoconf.pm 2013-09-03 20:44:37.000000000 +0200 +++ debhelper/Debian/Debhelper/Buildsystem/autoconf.pm 2014-06-13 17:49:04.000000000 +0200 @@ -37,6 +37,9 @@ push @opts, "--infodir=\${prefix}/share/info"; push @opts, "--sysconfdir=/etc"; push @opts, "--localstatedir=/var"; + if (defined $ENV{DH_VERBOSE} && $ENV{DH_VERBOSE} ne "") { + push @opts, "--disable-silent-rules"; + } my $multiarch=dpkg_architecture_value("DEB_HOST_MULTIARCH"); if (! compat(8)) { if (defined $multiarch) { diff -Naur debhelper-9.20140228/dh_installinit debhelper/dh_installinit --- debhelper-9.20140228/dh_installinit 2013-11-10 20:23:13.000000000 +0100 +++ debhelper/dh_installinit 2014-05-24 19:33:24.000000000 +0200 @@ -165,7 +165,7 @@ my $scriptsrc; my $jobfile=$package; if (defined $dh{NAME}) { - $jobfile=$script=$dh{NAME}; + $jobfile=$script=$scriptsrc=$dh{NAME}; } elsif ($dh{D_FLAG}) { # -d on the command line sets D_FLAG. We will I don't believe this could fix the race. It's not possible to prove the bug is fixed just by passing build. You can only prove by this method that the bug is still there.
(In reply to Petr Pisar from comment #5) > It's not possible to prove the bug is fixed just by passing build. You can > only prove by this method that the bug is still there. I said the bug was fixed, before update debhelper, the fix was not the debhelper update , but rawhide updates ... have you ever got this bug in F21 branch ?
(In reply to Sergio Monteiro Basto from comment #6) > (In reply to Petr Pisar from comment #5) > > It's not possible to prove the bug is fixed just by passing build. You can > > only prove by this method that the bug is still there. > > > I said the bug was fixed, before update debhelper, the fix was not the > debhelper update , but rawhide updates ... > That could be possible. > have you ever got this bug in F21 branch ? No.
OK , I'm closing this bug, as not a bug in debhelper , if problem persist please reopen it . Best regards