{"thread":{"id":"46954","subject":"v2.15.0-rc1 test failure","startedAt":"2017-10-11T22:16:07Z","lastAt":"2017-10-12T09:06:56Z","messageCount":6,"participants":["Ramsay Jones","Adam Dinwoodie","Jonathan Nieder"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"330212","messageId":"8dd12465-0dc2-340e-8db0-e37719f34efc@ramsayjones.plus.com","threadId":"46954","inReplyTo":null,"subject":"v2.15.0-rc1 test failure","fromName":"Ramsay Jones","fromEmail":"ramsay@ramsayjones.plus.com","sentAt":"2017-10-11T22:15:57Z","receivedAt":"2017-10-11T22:16:07Z","isPatch":false,"sender":{"key":"ramsay@ramsayjones.plus.com","avatar":"https://avatars.githubusercontent.com/u/33702710?v=4"},"body":"Hi Adam,\n\nI had a test failure on the v2.15.0-rc1 build tonight.\nThe test in question being t0021-conversion.sh #15\n('required process filter should filter data'). I didn't\nhave any test failures on v2.15.0-rc0, and I don't see\nany change that would have affected this test.\n\nAlso, I ran this test by hand (well, in a shell loop) at\nleast 70 times tonight (after the test-suite run), without\nany failures, so ... (unfortunately, I don't have a trash\ndirectory to look at. :( )\n\nI have just kicked off another full test-suite run.\n\nJust a heads up! ;-)\n\nATB,\nRamsay Jones\n\n"},{"id":"330213","messageId":"edfe891b-1aa1-0b20-6af7-8b8bc47d69c9@ramsayjones.plus.com","threadId":"46954","inReplyTo":"8dd12465-0dc2-340e-8db0-e37719f34efc@ramsayjones.plus.com","subject":"Re: v2.15.0-rc1 test failure","fromName":"Ramsay Jones","fromEmail":"ramsay@ramsayjones.plus.com","sentAt":"2017-10-11T22:25:34Z","receivedAt":"2017-10-11T22:25:40Z","isPatch":false,"sender":{"key":"ramsay@ramsayjones.plus.com","avatar":"https://avatars.githubusercontent.com/u/33702710?v=4"},"body":"\n\nOn 11/10/17 23:15, Ramsay Jones wrote:\n> Hi Adam,\n> \n> I had a test failure on the v2.15.0-rc1 build tonight.\n> The test in question being t0021-conversion.sh #15\n> ('required process filter should filter data'). I didn't\n> have any test failures on v2.15.0-rc0, and I don't see\n> any change that would have affected this test.\n> \n> Also, I ran this test by hand (well, in a shell loop) at\n> least 70 times tonight (after the test-suite run), without\n> any failures, so ... (unfortunately, I don't have a trash\n> directory to look at. :( )\n> \n> I have just kicked off another full test-suite run.\n> \n> Just a heads up! ;-)\n\nOops, for mailing list reader, I should have made clear that\nthis failure is only on cygwin. :-D\n\nATB,\nRamsay Jones\n\n\n"},{"id":"330214","messageId":"20171011223433.GA6079@dinwoodie.org","threadId":"46954","inReplyTo":"8dd12465-0dc2-340e-8db0-e37719f34efc@ramsayjones.plus.com","subject":"Re: v2.15.0-rc1 test failure","fromName":"Adam Dinwoodie","fromEmail":"adam@dinwoodie.org","sentAt":"2017-10-11T22:34:33Z","receivedAt":"2017-10-11T22:34:49Z","isPatch":false,"sender":{"key":"adam@dinwoodie.org","avatar":"https://avatars.githubusercontent.com/u/1397507?v=4"},"body":"On Wed, Oct 11, 2017 at 11:15:57PM +0100, Ramsay Jones wrote:\n> Hi Adam,\n> \n> I had a test failure on the v2.15.0-rc1 build tonight.\n> The test in question being t0021-conversion.sh #15\n> ('required process filter should filter data'). I didn't\n> have any test failures on v2.15.0-rc0, and I don't see\n> any change that would have affected this test.\n> \n> Also, I ran this test by hand (well, in a shell loop) at\n> least 70 times tonight (after the test-suite run), without\n> any failures, so ... (unfortunately, I don't have a trash\n> directory to look at. :( )\n> \n> I have just kicked off another full test-suite run.\n> \n> Just a heads up! ;-)\n\nHi Ramsay,\n\nI assume, given you're emailing me, that this is a Cygwin failure?\n\nt0021.15 has PERL as a requirement, and I see semi-regular failures from\nGit tests that are Perl-based in one way or another (git-svn tests are\nthe most common problems).  I've not spotted t0021 failing in that way,\nbut it sounds like the same class of problem.\n\nI dig into these failures when I see them, mostly by running the script\na few hundred times until I get the failure again, and they've always\nbeen Perl itself segfaulting.  That points to the problem being in\nCygwin's Perl package rather than Git, and it's very unlikely to be\nanything that's got worse in v2.15.0.\n\nI've never managed to get further than pinning the blame on Perl,\nthough.  If you manage to reproduce the failure and it turns out to be\nsomething different, or you manage to dig into the failure in Perl\nitself, that'd be some very interesting news.\n\nAdam\n"},{"id":"330215","messageId":"20171011230830.GA47665@aiede.mtv.corp.google.com","threadId":"46954","inReplyTo":"20171011223433.GA6079@dinwoodie.org","subject":"Re: v2.15.0-rc1 test failure","fromName":"Jonathan Nieder","fromEmail":"jrnieder@gmail.com","sentAt":"2017-10-11T23:08:30Z","receivedAt":"2017-10-11T23:10:33Z","isPatch":false,"sender":{"key":"jrnieder@gmail.com","avatar":"https://avatars.githubusercontent.com/u/281595?v=4"},"body":"Hi,\n\nAdam Dinwoodie wrote:\n\n> t0021.15 has PERL as a requirement, and I see semi-regular failures from\n> Git tests that are Perl-based in one way or another (git-svn tests are\n> the most common problems).  I've not spotted t0021 failing in that way,\n> but it sounds like the same class of problem.\n>\n> I dig into these failures when I see them, mostly by running the script\n> a few hundred times until I get the failure again, and they've always\n> been Perl itself segfaulting.  That points to the problem being in\n> Cygwin's Perl package rather than Git, and it's very unlikely to be\n> anything that's got worse in v2.15.0.\n\nThat reminds me of https://bugs.debian.org/868738, which I tracked down\nto perl's \"die\" helper using errno to determine the exit status instead\nof deterministically using 128.  I wasn't able to track it down further\nthan that.\n\nt/t0021/rot13-filter.pl doesn't have any similar suspect constructs, but\nthought I should mention it anyway.\n\nThanks,\nJonathan\n"},{"id":"330217","messageId":"338656d4-182b-563f-9195-515d213d4b6d@ramsayjones.plus.com","threadId":"46954","inReplyTo":"20171011223433.GA6079@dinwoodie.org","subject":"Re: v2.15.0-rc1 test failure","fromName":"Ramsay Jones","fromEmail":"ramsay@ramsayjones.plus.com","sentAt":"2017-10-12T00:27:57Z","receivedAt":"2017-10-12T00:28:03Z","isPatch":false,"sender":{"key":"ramsay@ramsayjones.plus.com","avatar":"https://avatars.githubusercontent.com/u/33702710?v=4"},"body":"\n\nOn 11/10/17 23:34, Adam Dinwoodie wrote:\n[snip]\n> Hi Ramsay,\n> \n> I assume, given you're emailing me, that this is a Cygwin failure?\n\nYes, sorry, I should have made that clear.\n\n> t0021.15 has PERL as a requirement, and I see semi-regular failures from\n> Git tests that are Perl-based in one way or another (git-svn tests are\n> the most common problems).  I've not spotted t0021 failing in that way,\n> but it sounds like the same class of problem.\n\nYep, many moons ago, I used to run the svn tests (on Linux and cygwin)\nwhich would fail intermittently on cygwin. I didn't notice any problem\nwith perl though.\n\n> I dig into these failures when I see them, mostly by running the script\n> a few hundred times until I get the failure again, and they've always\n> been Perl itself segfaulting.  That points to the problem being in\n> Cygwin's Perl package rather than Git, and it's very unlikely to be\n> anything that's got worse in v2.15.0.\n\nSince I stopped running the svn tests, the number of intermittent test failures on cygwin have dropped significantly, but haven't gone away\ncompletely.\n\nI just finished the second test-suite run and, of course, t0021 ran\nwithout problem this time. Hmm, I don't think I have time to chase\nthis down at the moment. I will keep your 'perl hypothesis' in mind\nfor next time, however.\n\nThanks.\n\nATB,\nRamsay Jones\n\n\n"},{"id":"330253","messageId":"20171012090648.GB6079@dinwoodie.org","threadId":"46954","inReplyTo":"338656d4-182b-563f-9195-515d213d4b6d@ramsayjones.plus.com","subject":"Re: v2.15.0-rc1 test failure","fromName":"Adam Dinwoodie","fromEmail":"adam@dinwoodie.org","sentAt":"2017-10-12T09:06:48Z","receivedAt":"2017-10-12T09:06:56Z","isPatch":false,"sender":{"key":"adam@dinwoodie.org","avatar":"https://avatars.githubusercontent.com/u/1397507?v=4"},"body":"On Thu, Oct 12, 2017 at 01:27:57AM +0100, Ramsay Jones wrote:\n> On 11/10/17 23:34, Adam Dinwoodie wrote:\n> [snip]\n> > Hi Ramsay,\n> > \n> > I assume, given you're emailing me, that this is a Cygwin failure?\n> \n> Yes, sorry, I should have made that clear.\n> \n> > t0021.15 has PERL as a requirement, and I see semi-regular failures from\n> > Git tests that are Perl-based in one way or another (git-svn tests are\n> > the most common problems).  I've not spotted t0021 failing in that way,\n> > but it sounds like the same class of problem.\n> \n> Yep, many moons ago, I used to run the svn tests (on Linux and cygwin)\n> which would fail intermittently on cygwin. I didn't notice any problem\n> with perl though.\n> \n> > I dig into these failures when I see them, mostly by running the script\n> > a few hundred times until I get the failure again, and they've always\n> > been Perl itself segfaulting.  That points to the problem being in\n> > Cygwin's Perl package rather than Git, and it's very unlikely to be\n> > anything that's got worse in v2.15.0.\n> \n> Since I stopped running the svn tests, the number of intermittent test failures on cygwin have dropped significantly, but haven't gone away\n> completely.\n> \n> I just finished the second test-suite run and, of course, t0021 ran\n> without problem this time. Hmm, I don't think I have time to chase\n> this down at the moment. I will keep your 'perl hypothesis' in mind\n> for next time, however.\n\nEvidence for my Perl hypothesis, which I offer at least as much so other\npeople can check my logic as anything else:\n\nHere's a fairly typical set of verbose output from a test failing in\nthis way [0].  The critical bit is the line \"error: git-svn died of\nsignal 11\".  Since git-svn is a Perl script, and Perl is the sort of\ninterpreted language that would throw its own errors if it encountered a\nscript bug, the fact that it's hitting a segfault means there's a\nproblem of some ilk with the Perl interpreter itself.\n\n[0]: https://github.com/me-and/Cygwin-Git/issues/13#issuecomment-211372448\n\nAdam\n"}]}