16:00:44 <djmitche> #startmeeting weekly
16:00:44 <bb-supy`> Meeting started Tue Aug 25 16:00:44 2015 UTC and is due to finish in 60 minutes.  The chair is djmitche. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:44 <bb-supy`> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:44 <bb-supy`> The meeting name has been set to 'weekly'
16:01:21 <djmitche> https://etherpad.mozilla.org/buildbot-agenda
16:01:40 * sa2ajj waves
16:02:00 * sa2ajj wonders if anybody else is online...
16:02:27 <djmitche> #topic action item followup
16:02:33 <djmitche> #info djmitche complete fail
16:02:54 <djmitche> almost complete - I think I added the version to trac
16:03:06 <bb-trac_> [trac] #3339/defect (v:master) created by sa2ajj (buildbot.test.integration.test_trigger.TriggeringMaster.test_trigger is flaky) http://trac.buildbot.net/ticket/3339
16:03:23 <bb-trac_> [trac] #3339/defect (new) updated by sa2ajj (empty comment) http://trac.buildbot.net/ticket/3339
16:03:38 <sa2ajj> djmitche: that's what vacations are for :)
16:03:45 <djmitche> yep
16:03:53 <djmitche> hah, indeed!
16:03:58 <djmitche> but I've been back for almost a week
16:04:18 <djmitche> #topic swag for contributors
16:04:25 <djmitche> nothing new from me
16:05:04 <djmitche> I'll figure something out this week
16:05:29 <djmitche> #topic mailing list migration
16:05:34 <djmitche> verm__: skelly - any update?
16:06:45 <bb-meta> build #225 of py27-sa086 is complete: Failure [4failed 2 tests]  Build details are at http://buildbot.buildbot.net/builders/py27-sa086/builds/225  blamelist: Mikhail Sobolev <mss@mawhrin.net>, Pierre Tardy <pierre.tardy@intel.com>
16:07:04 <bb-meta> build #219 of py26-tw1400 is complete: Failure [4failed 2 tests]  Build details are at http://buildbot.buildbot.net/builders/py26-tw1400/builds/219  blamelist: Mikhail Sobolev <mss@mawhrin.net>, Pierre Tardy <pierre.tardy@intel.com>
16:07:12 <djmitche> #topic Trac slowness
16:07:23 <djmitche> sa2ajj: have you noticed anything related?
16:07:30 <djmitche> I don't see updates on http://trac.buildbot.net/ticket/3328
16:07:38 <tardyp> I haven't played with trac that much since
16:07:41 <djmitche> or anyone :)
16:07:48 <sa2ajj> not really: it works pretty much the same way for me: not really snappy, but it's the same since i started to actively use it
16:07:52 <tardyp> but no problem so far
16:08:09 <djmitche> ok
16:08:36 <djmitche> #info no change - still usable
16:08:39 <tardyp> We need to do some serious bug scrubbing to really close
16:08:50 <djmitche> #topic nine update
16:08:56 <djmitche> regarding nine bugs specifically?
16:09:13 <tardyp> I dont remember anything significant
16:09:25 * sa2ajj nods
16:09:34 <sa2ajj> bugs scrubbing is usually helpful :)
16:09:39 <tardyp> we had a guy confused by the fact he had buildbot eight installed by distro
16:09:53 <tardyp> and pip install installed in /usr/local, and did not remove the other package
16:09:56 <djmitche> yeah, that /latest/ vs. /current/ thing is annoying
16:09:58 <djmitche> oh, ugh
16:10:11 <tardyp> no that another
16:10:26 <tardyp> he had 2 buildbot package installed in syspath
16:10:36 <tardyp> and the plugin system did not like it.
16:10:52 <tardyp> with an error that is not very intuitive
16:11:01 <tardyp> so we recommended to always use venv
16:11:02 <djmitche> that's tricky
16:11:07 <bb-trac_> [trac] #3340/defect (v:master) created by sa2ajj (buildbot.test.integration.test_customservices.CustomServiceMaster.test_customService is flaky) http://trac.buildbot.net/ticket/3340
16:11:17 <djmitche> did that end up in a bug? I remember seeing some of the irc conversation
16:11:37 <tardyp> not from me. I just followed that from my shitty connection
16:11:44 <sa2ajj> and not from me
16:11:55 <tardyp> I'll create
16:11:56 <djmitche> ok
16:12:05 <djmitche> #info no updates, lots of vacations
16:12:15 <djmitche> #topic GSoC updates
16:12:30 <djmitche> this is the last week before evals are due
16:12:48 <tardyp> from my side, both projects are well advanced
16:13:22 <djmitche> excellent
16:13:29 <tardyp> They will not finish before pencil up, strickly speaking. but we knew both projects were very difficult to finish
16:13:40 <tardyp> so this is a great success for me.
16:13:42 <djmitche> pencils-down was a week ago :)
16:13:43 * sa2ajj really likes the progress
16:13:43 <djmitche> yeah!
16:14:10 <djmitche> how's the Python-3 project wrapping up?
16:14:18 <djmitche> jaredgrubb: ? bobhood ?
16:14:29 <bobhood> We are pretty much stopped due to 'spread'.  Everything else is in place.
16:14:43 <sa2ajj> spread?
16:14:43 <djmitche> great
16:14:43 <bobhood> So, either twisted updates 'spread', or Buildbot migrates away from it.
16:15:01 <tardyp> bobhood: how much work to port spread?
16:15:03 <djmitche> is there an interim solution so that we can prevent backsliding in Buildbot?
16:15:08 <bobhood> We'd still like to have a regression lane for this code so it doesn't bit rot.
16:15:41 <bobhood> Jorn would have to speak to porting 'spread', but I doubt the twisted team is going to do it.
16:15:43 <djmitche> I think that needed more info on what to run, exactly
16:15:50 <bb-trac_> [trac] #3339/defect (new) updated by sa2ajj (empty comment) http://trac.buildbot.net/ticket/3339
16:16:02 <bobhood> Jorn has some regression tests that can be run by that lane.
16:16:06 <djmitche> http://trac.buildbot.net/ticket/3287#comment:6
16:16:20 <djmitche> great
16:16:31 <sa2ajj> oh, it's still not done.  sorry about that :(
16:16:42 <djmitche> heh, no worries
16:17:13 <djmitche> #info python-3 needs a metabuildbot builder added to keep improvements-to-date in place (#3287)
16:17:17 <jaredgrubb> (fwiw, i am here, but bob's said anything i would say)
16:17:21 <djmitche> :)
16:17:25 <djmitche> how about the stats project?
16:17:38 <sa2ajj> documentation is the current PR
16:17:42 <sa2ajj> mostly cosmetics
16:17:51 <sa2ajj> i reviewed it before the meeting
16:17:55 <djmitche> so pretty much wrapped up?
16:18:05 <sa2ajj> i can't say before i try it :)
16:18:11 <djmitche> #info stats has an outstanding PR for documentation
16:18:13 <djmitche> fair enough
16:18:25 * sa2ajj wanted to read the docs and then follow them to see what's missing
16:19:03 <djmitche> well, this is on track to be *by far* the best year of GSoC Buildbot has had
16:19:18 <sa2ajj> excellent :)
16:19:20 <djmitche> and that's squarely down to some great mentors and great students
16:19:43 <djmitche> so good job to everyone
16:19:52 <bobhood> Definitely some great students.  :)
16:19:59 <tardyp> agreed
16:20:02 * sa2ajj nods
16:20:16 * sa2ajj did not have a chance to show himself as a great mentor though
16:20:26 <djmitche> we can take some credit there too -- the selection process was important
16:20:33 <djmitche> next year :)
16:20:49 <djmitche> (and don't sell yourself short for ths year either!)
16:20:58 <djmitche> ok, last two items in AOB
16:20:59 <sa2ajj> re try client: we do not use it currently, but that's in plans
16:21:13 <djmitche> try client?
16:21:26 <sa2ajj> the topics of one gsoc projects
16:21:41 <sa2ajj> or try scheduler (these are really difficult to distinguish)
16:21:43 <djmitche> ah
16:21:53 <djmitche> two sides of a coin :)
16:22:01 <sa2ajj> :D
16:22:09 <djmitche> #topic maintainer availability
16:22:32 <djmitche> hopefully I can be around more and spend a bit more time hacking and reviewing and whatnot, now that I'm back from my holiday
16:22:46 <sa2ajj> i'm trying to look at open PRs in the morning (right before i leave home)
16:22:46 <djmitche> so if you need to fade out a little, sa2ajj, that should be fine
16:22:54 <sa2ajj> thanks!
16:23:14 <tardyp> I'm also back and fully refreshed
16:23:24 <djmitche> :thumbsup:
16:23:29 <sa2ajj> i hope i will not really "fade out", but 21 PRs is a bit too much...
16:23:33 <djmitche> yeah
16:23:53 <djmitche> #topic flaky tests
16:24:05 <sa2ajj> first time contributors would definitely appreciate quick feedback, that's my worry
16:24:14 <sa2ajj> re flaky tests: i'm about to submit a PR
16:24:20 <sa2ajj> i just filed two tickets
16:24:45 <djmitche> ok, great
16:24:50 <tardyp> I though I fixed it with the log semaphore
16:24:55 <tardyp> but it looks it did not
16:25:09 <bb-github> [13buildbot] 15sa2ajj opened pull request #1827: Disable flaky test cases (06master...06disable-flaky-test-cases) 02http://git.io/vs7vV
16:25:09 <djmitche> PR with a fix, or PR marking them flaky?
16:25:12 <sa2ajj> yay ^^
16:25:13 <djmitche> ah
16:25:20 <sa2ajj> PR making them flaky :( sorry
16:25:39 <djmitche> no, that's great
16:25:49 <sa2ajj> my intention is to move from Travis CI to meta BB for checking PRs
16:25:51 <djmitche> it means no failures distracting other contributors until we get it fixed
16:25:57 <djmitche> sweet
16:26:08 <sa2ajj> and flaky tests do not really help there
16:26:29 <djmitche> right
16:27:08 <djmitche> I don't think we're there yet, but we could add a "flaky" builder that runs flaky tests
16:27:23 <djmitche> at least it would be apparent quickly if only that builder failed that it didn't invalidate the PR
16:27:26 * sa2ajj likes the idea
16:27:37 <tardyp> a nine metabb?
16:28:00 <sa2ajj> and that as well, we need to make the metabbs comparable
16:28:33 <djmitche> so much to do
16:28:53 <djmitche> btw, some of the older flaky tests may no longer be flaky
16:29:07 <djmitche> I think a lot of the flaky was down to issues with the service hierarchy, which has changed quite a bit
16:29:10 <bb-trac_> [trac] #3341/undecided (v:0.8.12) created by tardyp (plugin system fails with two installs of buildbot) http://trac.buildbot.net/ticket/3341
16:29:28 * sa2ajj envisions nine is used for checking PRs
16:29:28 <djmitche> #action djmitche to re-run flaky tests a few hundred times to see if any are no longer flaky
16:29:37 <bb-trac_> [trac] #3341/defect (new) updated by tardyp (empty comment) http://trac.buildbot.net/ticket/3341
16:29:39 <djmitche> I like that idea
16:30:05 <sa2ajj> tardyp: it's _B_uildbot :P
16:30:08 <djmitche> #info want to start using the metabuildbot (running the nine beta) to test PRs
16:30:36 <djmitche> any other business to bring up?
16:30:43 <tardyp> sa2ajj: uh??
16:31:02 <sa2ajj> i believe Buildbot is written as Buildbot
16:31:12 <sa2ajj> while i might write it as 'buildbot' in IRC
16:31:27 <sa2ajj> i tend to put proper capitalisation for docs and tickets
16:31:38 <djmitche> picky picky (but correct) :)
16:31:52 <sa2ajj> sorry :/
16:31:54 <bb-trac_> [trac] #3341/defect (new) updated by tardyp (empty comment) http://trac.buildbot.net/ticket/3341
16:32:24 <djmitche> #endmeeting