=== danilos_ is now known as danilos [12:08] bac frankban gmb call in 2 [12:08] * gmb fires up firefox [12:11] gary_poster: where we doing this? [12:11] nm [12:11] bac gmb frankban https://talkgadget.google.com/hangouts/_/edbf4dd708d7938ac8ac1f60fd51d01b84c936bc [12:37] * gmb -> Lunch [13:06] * gary_poster going to restart after new kernel [14:06] gary_poster: I was looking at bug 1003097. It seems that at the end increasing the query count by 1 is not a bad idea. The test fails because this query (http://pastebin.ubuntu.com/1003086/) is not cached. And I think that query is cached by Browser.open(TASK-URL). [14:06] <_mup_> Bug #1003097: lp.bugs.browser.tests.test_bugtask.TestBugTaskView.test_rendered_query_counts_constant_with_team_memberships fails when run by itself < https://launchpad.net/bugs/1003097 > [14:07] frankban, ok. I guess we could either increase by one, or make the test itself open that url and then reset the query count [14:07] either way, a comment will be important :-) [14:08] back in a few [14:25] hey gmb, I have a working parallel juju instance up so I'm going to try our benji's lpsetup branch. [14:25] if it works, I can let you know and you can do a MP and so on [14:25] is that ok? [14:25] gary_poster, That works for me. [14:25] cool thx [14:30] gmb, can you access that branch, like via http://bazaar.launchpad.net/~benji/lpbuildbot/bug-1003044 ? I can't [14:31] gmb, asking webops to make it public [14:31] gary_poster, Hmm. Me neither. Are lpbuildbot branches private by default (and if so, I guess they're private to canonical-sysadmins. [14:31] Ah right. [14:31] Yes, it'll need a rubber duckie. [14:31] :-) [14:31] gary_poster: do we want to try the triple jump from tasks/normal to qa/landing or is it too unfair? https://code.launchpad.net/~frankban/launchpad/bug-1003097-isolation-failure/+merge/107024 [14:33] frankban, :-) reflect reality. drag it to review and say that you went over the limit because you couldn't stand not fixing the bug :-P Then you can drag it to landing [14:34] approved frankban [14:34] thanks gary [14:34] thank you [15:20] gmb, I eventually got a patch for benji's branch, and applied that...but I was not up to date in my charm and so lpsetup's dependencies were wrong :-( [15:20] I'm redoing everything now [15:20] but gmb, if/when you do have a moment [15:21] if you could take a look at why tests fail in lp:~gary/charms/precise/buildbot-slave/add-max-builds-support that would be wonderful [15:21] if you don't get around to it that's fine [15:21] but it would be a great help [15:21] and I think you know the details there better than I [15:22] gary_poster, Actually, I can take a look in ten minutes or so. [15:22] So I will :) [15:22] awesome, thanks gmb [15:33] gary_poster, When you say "tests" do you mean the charm tests or do you mean LP tests when that charm's used as a slave? [15:57] gmb, sorry, I mean the charm tests [15:57] the LP tests are doing fine [15:57] gary_poster, Okay. Good - I didn't know if you mean that you'd also discovered that all the tests die horribly... [15:57] *meant [15:57] :-) no [15:58] Actually, that was my fault for not bootstrapping, but still. [16:00] Also, "ues" is not a valid response to a yes/no question. We live and learn. [16:28] heh [16:28] gmb, see my comment to https://bugs.launchpad.net/launchpad/+bug/1003206 , and shake your head and say "I knew it couldn't be that simple!" :-) [16:28] <_mup_> Bug #1003206: (parallel tests) worker output is missing from some tests < https://launchpad.net/bugs/1003206 > [16:32] gmb, and ow, we are doing something crazy with the subunit output now: various subunit bits are inserted within the last failure on the page of http://ec2-50-19-187-30.compute-1.amazonaws.com:8010/builders/lucid_lp/builds/0/steps/shell_9/logs/stdio , for instance. :-/ [16:33] holy smoke, the tests have gone haywire :-( [16:33] I'm hoing to go get some lunch :-/ [16:33] or going [16:34] you decide [16:47] Doubleyew tee eff. [16:52] Get someone else's branch landed means branch it and the ec2 landing it I suppose... could you confirm? [16:54] gary_poster, So, I think this is definitely a good time to revert to -p6 for a run. [16:54] frankban, Yes, that's right. You might be able to do `ec2 land $merge_proposal_url` though. [16:54] I can't quite remember. [16:54] gary_poster, Ah, crumbs, I think this is the problem: [16:56] When we come across an error, we pass it to the Subunit formatter's _subunit.addError() method. That passes it to _addOutcome(), which writes to _stream... which in this case means that it gets written to stdout. [16:56] Although the messiness makes no sense, since that's what _always_ happened. [16:57] Hmm. [16:57] gary_poster, I think we should do this: [16:57] revert to -p6 for a run, to confirm that nothing new is screwing with things [16:57] If that's confirmed, revert it in devel (and revert the -p7 change in zope.testing-3.9.4-fork) [16:57] ??? [16:57] Win. [16:58] * gmb -> grabbing some dinner but will BBL. [17:02] :-) [17:02] yeah will do after more lunch break [19:48] hey gary_poster, i haven't put it in canonicaladmin yet, but i plan to work on memorial day for a swap day tbd [20:05] bac, cool [20:06] gary_poster: i'm trying to run a ec2 buildbot test and am running into errors. you ever see permission denied errors re: /var/tmp/bazaar.launchpad.dev/mirrors [20:06] bac, yes. :-( today is not a good day for test runs. Maybe we should pair after I finish my tl call notes (I was chair) [20:07] oh, ok [20:07] so is it a busted branch for LP? [20:08] oh benji! :) [20:12] bac https://talkgadget.google.com/hangouts/_/860d9077559ceb3952f1dd8c072db17bf6d623a8 [20:22] ok [21:34] qbac you disappeared [21:34] bac [21:34] yesh [21:34] retrying [22:00] did the hail eat you bac?