16:31:50 <djmitche> #startmeeting weekly
16:31:50 <bb-supy> Meeting started Tue Apr 26 16:31:50 2016 UTC and is due to finish in 60 minutes.  The chair is djmitche. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:31:50 <bb-supy> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:31:50 <bb-supy> The meeting name has been set to 'weekly'
16:32:00 <djmitche> #topic Introduction
16:32:36 <djmitche> https://titanpad.com/buildbot-agenda
16:33:02 <djmitche> heh
16:33:08 <djmitche> #topic Development Week in Review
16:33:45 <djmitche> #info about 30 pull requests this week -- wow!
16:34:15 <djmitche> tomprince: I see you were involved on some of those -- care to summarize?
16:35:02 <tomprince> The ones I was involved with were mostly testing improvements, so that integration tests can be written without needing to spin the reactor.
16:35:10 <djmitche> cool
16:35:17 <tomprince> Or dump config files to disk just to read them back in.
16:35:44 <djmitche> #info testing improvements to avoid unnecessarily spinning the reactor (as begun last week) or writing files to disk
16:35:53 <djmitche> thanks
16:36:13 <djmitche> looks like some changes around the `buildbot` cli
16:36:58 <djmitche> http://trac.buildbot.net/query?status=accepted&status=assigned&status=new&status=reopened&milestone=0.9.0&group=status&order=priority
16:37:02 <djmitche> #info 17 remaining 0.9.0 bugs
16:37:36 <bb-trac> [trac] #3538/defect (new) updated by dustin (empty comment) http://trac.buildbot.net/ticket/3538
16:37:50 <djmitche> #topic Bug 2340 update - worker renaming
16:38:14 <djmitche> #info rutsky can't be here for the meeting today, but "next week is mostly holidays in Russia and I hope to complete bug 2340 during it"
16:39:02 <bb-trac> [trac] #3537/enhancement (new) updated by dustin (I think this is a good idea, and important, but let's not block 0.9.0 for it.) http://trac.buildbot.net/ticket/3537
16:39:34 <djmitche> any other topics?
16:39:35 <djmitche> frustrations?
16:39:43 <djmitche> anything I can unblock?
16:40:36 <gracinet> hi djmitche, I think I'd need advice / coordination for the bugs uncovered with postgresql
16:40:41 <bb-trac> [trac] #3534/defect (new) updated by dustin (#2027 has a patch that fixed this before.  Basically, this exception should be ...) http://trac.buildbot.net/ticket/3534
16:40:55 <gracinet> espectially http://trac.buildbot.net/ticket/3532
16:41:37 <bb-trac> [trac] #3533/defect (new) updated by dustin (Setting to 0.9.0 since there's a patch pending) http://trac.buildbot.net/ticket/3533
16:41:46 * djmitche looks
16:42:25 <gracinet> the question (probably not for the meeting) would be how deeply to fix this
16:43:02 <djmitche> gracinet: yeah, that seems like an incomplete change from using schedulerids to objectids
16:43:03 <djmitche> or the reverse
16:43:06 <djmitche> I don't remember which way we went
16:43:26 <gracinet> I think the objectid is the oldest of them
16:44:04 <djmitche> yeah
16:44:05 <djmitche> scheduler state should be in the scheduler_changes table - https://github.com/buildbot/buildbot/blob/master/master/buildbot/db/model.py#L484
16:44:20 <djmitche> ah, I see
16:45:21 <tomprince> gracinet:
16:45:34 <tomprince> I would assume there isn't a deep reason for things being they way they are.
16:45:45 <gracinet> nor would I tomprince :-)
16:45:46 <djmitche> that's usually wise
16:45:54 <djmitche> yeah, I don't ithink it should be passing objectid every time
16:45:58 <djmitche> I suspect that oculd just be fixed
16:46:02 <djmitche> in the 0.9.0 release
16:46:47 <gracinet> actually there's a findSchedulerId method (taking name) on the schedulers db API
16:47:15 <gracinet> but using it makes thing a bit more complicated (and multiplies nb requests by 2), I'd prefer to have the schedulerid as a direct attribute
16:47:41 <djmitche> I think that caches its result
16:47:49 <gracinet> or to issue directly UPDATE statements based on names
16:47:53 <gracinet> ah djmitche, maybe
16:49:20 <gracinet> I did a first experiment, just enough to see I'll have to do some test refactoring
16:49:30 <djmitche> anywya, using findSchedulerId is the right approach
16:49:42 <djmitche> if it isn't caching, then it should
16:49:49 * gracinet checks
16:52:44 <gracinet> djmitche: I don't see any obvious caching in it, nor in findSomethingId (that it calls)
16:52:47 <djmitche> #info discussed bug 3532, regarding schedulers using objectid where they should use schedulerid
16:53:43 <djmitche> oh, it should probably be here - https://github.com/buildbot/buildbot/blob/master/master/buildbot/schedulers/base.py#L102
16:53:56 <djmitche> #endmeeting