Description of problem: The tests run during a build of python3-3.3.0-0.1.b1.fc18 fail on power. https://ppc.koji.fedoraproject.org/koji/taskinfo?taskID=659743 build.log snippet: [167/369/2] test_iterlen test_invariant (test.test_iterlen.TestRepeat) ... ok test_no_len_for_infinite_repeat (test.test_iterlen.TestRepeat) ... ok test_invariant (test.test_iterlen.TestXrange) ... ok test_invariant (test.test_iterlen.TestXrangeCustomReversed) ... ok test_invariant (test.test_iterlen.TestTuple) ... ok test_immutable_during_iteration (test.test_iterlen.TestDeque) ... ok test_invariant (test.test_iterlen.TestDeque) ... ok test_immutable_during_iteration (test.test_iterlen.TestDequeReversed) ... ok test_invariant (test.test_iterlen.TestDequeReversed) ... ok test_immutable_during_iteration (test.test_iterlen.TestDictKeys) ... ok test_invariant (test.test_iterlen.TestDictKeys) ... ok test_immutable_during_iteration (test.test_iterlen.TestDictItems) ... ok test_invariant (test.test_iterlen.TestDictItems) ... ok test_immutable_during_iteration (test.test_iterlen.TestDictValues) ... ok test_invariant (test.test_iterlen.TestDictValues) ... ok test_immutable_during_iteration (test.test_iterlen.TestSet) ... ok test_invariant (test.test_iterlen.TestSet) ... ok test_invariant (test.test_iterlen.TestList) ... ok test_mutation (test.test_iterlen.TestList) ... ok test_invariant (test.test_iterlen.TestListReversed) ... ok test_mutation (test.test_iterlen.TestListReversed) ... ok test_invalid_hint (test.test_iterlen.TestLengthHintExceptions) ... ok test_issue1242657 (test.test_iterlen.TestLengthHintExceptions) ... /var/tmp/rpm-tmp.GRX5xh: line 32: 1093 Alarm clock WITHIN_PYTHON_RPM_BUILD= LD_LIBRARY_PATH=$ConfDir $ConfDir/python -m test.regrtest --verbose --findleaks
Created attachment 603117 [details] full build.log
As noted on IRC this happened only 13 minutes into a build, and we don't enable the timeout feature within regrtest. How reproducible is this failure?
It seems to fail in roughly, but not exactly, the same place: http://ppc.koji.fedoraproject.org/koji/taskinfo?taskID=660375
Hi Dave. Any chance you had to look into this? Thanks & regards, Phil
(summarizing IRC discussion from #fedora-ppc) The latest python3 for f18 and f19 built again: http://ppc.koji.fedoraproject.org/koji/buildinfo?buildID=112593 The f16 one failed though, FWIW: http://ppc.koji.fedoraproject.org/koji/buildinfo?buildID=97307 as the f16 one The last failed one was http://ppc.koji.fedoraproject.org/koji/buildinfo?buildID=104051 which was python3-3.3.0-0.1.b1.fc18 Not sure how to reproduce the failure; seems to be some kind of transient timeout. Closing for now as WORKSFORME.