[00:30] ugh [11:00] * gmb -> Lunch after a thoroughly frustrating morning. [12:10] bac benji frankban gmb call in 1 or 2 (asap essentially) [12:10] joining [12:11] Joining... [12:38] gary_poster: sorry about missing the last bit of the call; mechanics are such attention seekers [12:38] gary_poster, Oh, after all that: bug 924378 [12:38] <_mup_> Bug #924378: buildbot spurious failure UncleanReactorError for TestPullerMasterIntegration < https://launchpad.net/bugs/924378 > [12:38] benji, np :-) [12:39] gmb, oh! interesting, that's the exact same test? [12:39] Yes [12:39] By Deryck, funnily enough. [12:39] I'll add a comment. [12:40] * gmb forces a build [12:40] it must be a different sort of race condition [12:40] well, I guess what we had before was not a race condition [12:41] it was a repeatable problem that only came to light when tests were run out of order, or when you explicitly checked for a problem after each test [12:41] *this* may be more classically intermittent [12:41] * gary_poster stops muttering to himself [12:45] benji, will be ready for more @ 9 or so [12:46] gary_poster: sounds good [13:02] I'm filling my cup and I'll be ready. [13:29] benji, sorry. I'm in gokdenhorde [13:29] goldenhorde even [13:30] benji, also I've started an ec2, feeling optimistic :-) [13:38] gary_poster: my run is still up if you're interested in looking around: http://ec2-23-22-21-234.compute-1.amazonaws.com:8010/ [13:38] i'll tear it down soon [13:39] bac, can't get to it. If you put up the records, I'm ok with you tearing it down [13:39] ok [13:57] gary_poster: i pasted the wrong URL -- it should've been http://ec2-23-20-38-16.compute-1.amazonaws.com:8010/waterfall [13:57] what's interesting is it is reporting the last run had no errors [13:58] yay! [13:58] bac, it's a banner day! that's awesome! [13:58] have we seen a green run before? [13:58] I saw that just as you were saying it [13:58] no! [13:58] quick, screenshot time! [13:58] let's ship it! we're done! we're outta here! [13:59] some of those look like new failures to me bac, yeah? [13:59] lib/lp/code/vocabularies/tests/branch.txt [14:00] lp.services.config.tests.test_config.TestLaunchpadConfig.test_iter [14:00] lib/lp/bugs/tests/bugzilla-xmlrpc-transport.txt [14:01] bugzilla * 2 ... [14:01] lp.code.xmlrpc.tests.test_codehosting.CodehostingTest.test_recordSuccess(db) [14:02] those all look new to me bac. if you find them in LP please make sure we have cards for them and we have the "paralleltest" tag on the bug [14:03] gary_poster: ok [14:03] bac, I'm "happy" (i.e., I find it annoying but I think it is my responsibility :-) ) to file those [14:03] s/i.e./that is/ [14:03] gary_poster: i'll file them. [14:03] thank you bac [14:05] bac, we'll want subunit streams of those too, I'm afraid. I used ubuntu one just cause we have so much free space there [14:06] bac, I'm also happy to do half if that helps [14:18] gary_poster: how do you share with U1? i created the file and then selected 'publish' but i see no way to get the URL [14:19] bac, in web interface? [14:19] it is shown beneath the file when I do it [14:19] no, in desktop [14:19] * bac looks at web [14:19] how novel [14:29] gary_poster, So, random fun... if you rename that test (which is called test()) to test_anything_that_isnt_called_test(), the spurious zope:layer appears to vanish. [14:30] gmb, ugh [14:30] (Only just noticed this, since I was too busy trying to track down where the zope:layer actually came from and I got frustrated. [14:30] gary_poster, So, I can fix the bug, but not the root cause, right now. [14:30] right... [14:31] Which seems a bit like `bzr ci -m "Handwave"` [14:31] gmb, I don't like it because it leaves a landmine for the future. [14:31] gary_poster, I feel the same, if I'm honest. [14:32] gmb, +1 on dragging someone along to watch. :-) Hopefully that observation can give you an idea of what to do. Also... [14:32] So, I'd like to grab someone to pair with. My morning's been very interrupt-y, so I suspect that something obvious will leap out when two of us look together. [14:32] Ha. [14:32] do you have the yellow ppa installed? [14:32] Lemme check. [14:32] I'd suggest installing it and getting the newest versions of these things that we are using [14:32] Oh, ahh. [14:33] On the lucid container, no. [14:33] gary_poster, Okay, will do, then I'll start pinging people. [14:33] because it is sounding more and more like a potential problem in the bits and bobs that we have there [14:33] gmb, mm :-( some of that stuff is not built for lucid [14:33] I *think* because it didn't build [14:33] Oh. Bottom. [14:34] I mean, had errors [14:34] um [14:34] gmb, try installing the testing-cabal's ppa on lucid [14:34] they should be pretty close [14:34] gary_poster, Okay, I'll give it a shot. [14:34] cool [14:35] gmb, although, sorry... [14:35] does the zope.testing package have any subunit imports? [14:35] I think it does [14:35] if it does [14:35] Yes, I think so. [14:35] then that's what you need [14:35] if it doesn't [14:36] then that whole PPA question is probably irrelevant [14:36] (and it *may* be anyway now since you can dupe already) [14:36] Um. [14:36] sorry for just being confusing gmb [14:36] I'll check :) [14:36] gary_poster, That's okay; it's about par for my day today. [14:36] :-) [14:37] frankban, are you available anytime soon for pairing with gmb? [14:37] gary_poster: I am doing slack, so, sure! [14:37] frankban, Okay, cool. Let me see where I get to with this testing-cabal PPA and then I'll ping you when I'm ready. [14:38] May need to fill up on tea... [14:38] frankban, :-) it would be nice if you were able to get to a stopping place that didn't waste, but yeah, if that takes more than a few minutes you can just "pause" it [14:38] gmb: me too, no problem [15:04] gtsnkbsn, I'm trsfy. [15:04] Um. [15:04] frankban, I'm off by one. And ready. [15:04] I've shared a terminal per https://dev.launchpad.net/yellow/RemoteTerminalBroadcasting [15:04] (OTHER=graham) [15:05] Starting a hangout now. [15:07] frankban, I'm in https://plus.google.com/hangouts/_/extras/canonical.com/funwithsubunittags when you're ready [15:48] gary_poster: how do you want me to triage these new bug reports? high? [15:55] gary_poster: frankban and I are in a bit of a zope.testing/subunit maze. Do you have any knowledge of how tagging should work in zope.testing/subunit? If so, can you join our hangout? [16:12] gary_poster: https://dev.launchpad.net/ParallelTests/ResultsLog updated, bugs filed, cards created [16:42] gmb: crash? [16:48] bac, high, paralleltest [16:48] tag [16:48] sorry was lunching [16:48] gmb, yes I do. frankban would you like me to join still? [16:48] gary_poster, Hurrah, timing! [16:48] Yes, let's hang out... [16:49] cool [16:49] gary_poster: sure [16:49] gary_poster, frankban, goldenhorde [16:49] ? [16:49] gmb, oh ok was going to funwithsubunittags. Now going to goldenhorde [16:49] Ok. [17:50] To Whom It May Concern: [17:50] Fresh LP checkouts are borked [17:51] Because convoy deb packinging is borked [17:51] (Python bits are not installed) [17:51] and we need it for doing a make [17:51] so... [17:52] either we test a branch that has the change added to its buildout steps [17:52] or we wait for the packaging fix [18:03] exciting [18:04] I'm still trying to figure out how to properly format test failures, but that has devolved into figuring out why the testr tests don't pass using packages from our PPA [18:11] benji: are you ocr? [18:11] or was that yesterday? [18:12] bac: yesterday, but if you have an MP I won't mind looking at it [18:12] benji: cool, thx: https://code.launchpad.net/~bac/launchpad/bug-974617-2/+merge/103339 [18:15] benji, I see you there. :-) when will you be ready to reconvene? [18:15] gary_poster: after I'm done with this review, it'll just be a couple of minutes [18:18] bac: looks good, I did spot one docstring bug [18:19] benji: cool, thanks for the good eyes [18:19] gary_poster: the horde awaits [18:19] my pleasure [18:19] gary_poster: this appears to fix the /etc/mailname problem in the robert-approved way: http://pastebin.ubuntu.com/944504/ [18:20] make mp for it now [18:22] gary_poster: when you have a chance could you look at https://code.launchpad.net/~bac/launchpad/bug-981114-2/+merge/103342 [18:32] bac, approved with only mildly interesting and not strongly held comment [18:32] oops [18:39] gary_poster: hold on, firefox decided to crash [18:39] * gary_poster holds on to his treadmill desk [19:14] * bac afk. bbiab