16:31:49 <djmitche> #startmeeting weekly
16:31:49 <bb-supy`> Meeting started Tue Dec  8 16:31:49 2015 UTC and is due to finish in 60 minutes.  The chair is djmitche. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:31:49 <bb-supy`> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:31:49 <bb-supy`> The meeting name has been set to 'weekly'
16:32:23 <djmitche> #topic Introductions
16:32:37 <djmitche> https://titanpad.com/buildbot-agenda
16:32:37 <infobob> http://paste.pound-python.org/show/uN2UUBiKoECrCvTLKPt7/ (repasted for djmitche)
16:32:46 <djmitche> Pierre's here, /me waves to bdbaddog
16:33:12 <djmitche> #topic MOSS award
16:33:47 <djmitche> the official announcement of the Mozilla Open Source Support program awards is tomorrow
16:34:05 <bdbaddog> Gotcha
16:34:06 <djmitche> bill, do you want to outline what our prpoosal is?
16:34:27 <djmitche> so yeah, we don't technically know whether we've been awarded, but we should probably talk about what to do if we are
16:35:34 <bdbaddog> so basically we've ask for funding of two bounties: $10k to remove slave from all code/docs.  We should probably talk a bit about how to retire "slave" terminology should we be awarded the $$'s. Likely we'd do it over a few releases.
16:35:54 <bdbaddog> $5k to work on stability/handling shutting down EC2 slaves when buildbot master crashes/exits/restarts
16:36:46 <bdbaddog> We'd need to either self publish/manage or decide on a platform for publishing/managing such. I think we mentioned a few options before (which of course (I think) take a cut)
16:36:59 <tardyp> can you restate on which was the criteria to select those 2 projects in particular?
16:37:26 <bdbaddog> those were the two mentioned as pain points by one of th developers inside mozilla.
16:38:02 <djmitche> the terminology bit is a good fit for Mozilla in a lot of ways
16:38:04 <bdbaddog> Also they aligned with issues which have had a reasonable amount of complaints from non-mozilla buildbot users
16:38:09 <djmitche> putting people first
16:38:26 <djmitche> yes - I've heard comments quietly from quite a few people now about the terminology
16:38:31 <bdbaddog> plus both are unlikely to get done (well) without some carrot.
16:38:34 <tardyp> looks good!
16:38:40 <djmitche> so I think getting it solved without requiring one of those people to put in the work is a good fit
16:38:43 <djmitche> yep
16:38:54 <djmitche> so bdbaddog what do you mean by spreading that over releases?
16:38:56 <tardyp> I think the two are very valuable
16:39:19 <djmitche> #info Buildbot's MOSS application: $10k to remove slave from all code/docs, $5k to work on stability/handling shutting down EC2 slaves when buildbot master crashes/exits/restarts
16:39:21 <bdbaddog> should we get awarded the moneys we'll have to write up a bit more detail on how such should be done.
16:39:58 <bdbaddog> Maybe put a page on trac for the bounties and a page for each item to fill in the details?
16:40:08 <djmitche> sounds good - I can write those up
16:40:26 <djmitche> #action djmitche to build bounties page
16:40:39 <djmitche> I think we should probably follow the GSoC model as much as possible, since we've walked that path before
16:40:58 <bdbaddog> also should we get the award, likley we'll deposit such funds with SFC to hold until payment.  Who is the best contact there? I've lost track.
16:41:37 <djmitche> buildbot@sfconservancy.org
16:41:46 <djmitche> it's a ML with everyone rleevant to the project on it
16:41:54 <bdbaddog> gotcha.
16:42:10 <djmitche> so we should probably have mentors -- I can mentor the terminology project
16:42:18 <djmitche> although I'd love a co-mentor ;)
16:42:27 <djmitche> and an application process, although not an application deadline
16:42:45 <djmitche> but some way for a potential bounty-claimer to write down some details about what they'll do, on what timescale
16:43:03 <djmitche> then we don't have to write out detailed criteria in advance -- we just agree to the proposal's criteria
16:43:15 <djmitche> make sense?
16:43:52 <bdbaddog> Yes. Likely remove the terms, provide migration path for code using old terms. details left to proposor
16:44:06 <tardyp> I can also candidate as a mentor
16:44:12 <tardyp> whichever project
16:44:24 <djmitche> #info djmitche offered to mentor terminology project
16:44:31 <djmitche> #info tardyp offered to mentor either project
16:44:50 <djmitche> bdbaddog: yep, and something similar for the other
16:45:48 <djmitche> we should probably also think ahead about any trolling we may receive
16:46:24 <djmitche> worth noting that we do have a code of conduct https://github.com/buildbot/botherders/blob/master/policies/conduct.md
16:47:14 <verm__> yay for a terminology change
16:47:22 <djmitche> I'll try to be on top of that, but I'm not always up-to-date, so please do seek me out if you see anything that needs to be addressed
16:47:36 <djmitche> glad to hear such support, thanks!
16:47:50 <djmitche> anything else to discuss wrt MOSS?
16:47:52 <verm__> i think it's a fantastic project for a bounty, as is the EC2
16:48:44 <djmitche> #topic Nine Update
16:48:45 <djmitche> http://trac.buildbot.net/query?status=accepted&status=assigned&status=new&status=reopened&milestone=0.9.0&group=status&order=priority
16:48:57 <djmitche> 31 open bugs for nine
16:49:14 <tardyp> not much progress on my side :( pretty busy those days..
16:49:22 <djmitche> I haven't heard much from andras lately -- is #3231 still in motion?
16:49:27 <djmitche> me too :(
16:50:01 <djmitche> I think we've already pared these bugs back to the bare minimum to ship 0.9.0
16:50:02 <verm__> is removing builders that are no longer used a ticket?
16:50:12 <tardyp> we still have big issues in the UI yes
16:50:13 <djmitche> so I'm not sure we have much choice but to wait
16:50:21 <tardyp> basically the builds are not updating live anymore
16:50:33 <djmitche> verm__: I don't think it's a 0.9.0 ticket, but yes I believe so
16:50:35 <verm__> other than the dangling builders the UI works great i use it in production
16:50:36 <tardyp> there is not much to do to fix that, but I've loss travk of it
16:50:42 <verm__> djmitche: ah, cool ok
16:51:33 <djmitche> verm__: although I'm not seeing it in a search
16:51:41 <djmitche> might be worht filing another - we can dupe
16:51:43 <djmitche> if it shows up
16:51:44 <verm__> hmm, ok
16:51:48 <myheadhurts> What is the difference between setting `max_builds` on a slave vs `maxCount` on SlaveLock?
16:52:04 <verm__> it's a real problem when you're first setting up lots of dangling builders..or if yo change a name
16:52:04 <myheadhurts> Are they two different ways of doing the same thing, or is the usecase different?
16:52:28 <djmitche> myheadhurts: the former refers to the # of builds on that slave, while the latter refers to # of bulids using that lock on the slave
16:52:47 <djmitche> verm__: yeah, definitely
16:53:03 <djmitche> ok, anything else to bring up?
16:53:03 <myheadhurts> djmitche: thanks, makes perfect sense!
16:53:14 <djmitche> hopefully your head hurts a little less? :)
16:53:28 <myheadhurts> Tends to get better after 5pm :)
16:53:33 <djmitche> :)
16:54:08 <tardyp> We have travis failure
16:54:13 <tardyp> on all the PRs
16:54:23 <djmitche> apropos of nothing, if anyone's in/around Orlando this week and wants to say hi, get in touch -- Mozilla's all here
16:54:36 <djmitche> #topic Consistent Travis failures
16:54:42 <tardyp> I am not sure why, but looks like there is a test in www that fails
16:54:43 <djmitche> tardyp: are they intermittent or constant?
16:54:55 <bb-trac_> [trac] #3386/undecided (v:master) created by verm (Removed builders show up in status displays) http://trac.buildbot.net/ticket/3386
16:55:01 <djmitche> thanks verm__
16:55:04 <verm__> sure
16:55:18 <tardyp> it looks like intermittent but 100% failure for all the PRs
16:55:25 <tardyp> mean that inside a build, some of the jobs reproduce about 50%
16:55:33 <djmitche> https://github.com/buildbot/buildbot/pull/1930#partial-pull-merging
16:55:48 <djmitche> so it's different tests different times?
16:56:02 <tardyp> no, looks the same
16:57:00 <djmitche> ok
16:57:06 <djmitche> looks like it's a string vs. int think?
16:57:09 <djmitche> *thing
16:57:17 <djmitche> twisted.trial.unittest.FailTest: '403' != 403
16:57:18 <djmitche> and similar
16:57:41 <djmitche> should I get a bug on file for this?
16:57:42 <tardyp> buildbot.test.unit.test_www_rest.V2RootResource_CORS.*
16:57:42 <tardyp> yes
16:57:48 <djmitche> ok
16:57:59 <djmitche> it looks not too difficult to fix -- maybe an issue with dependency versions
16:58:06 <tardyp> I just wanted to know if somebody started to look at it
16:58:26 <tardyp> if not we need to find time to try and reproduce
16:58:46 <bb-trac_> [trac] #3387/defect (v:master) created by dustin (www tests failing due to string vs int problems) http://trac.buildbot.net/ticket/3387
16:58:56 * djmitche hadn't
16:59:12 <djmitche> but I'm increasingly disengaged from the day-to-day (just turned off mail from users@)
16:59:44 <djmitche> I'll add that to the list for next week if nobody's taken it by then
17:00:09 <tardyp> ok
17:00:17 <tardyp> gtg
17:00:19 <djmitche> anything else before we close?
17:00:20 <djmitche> me too :)
17:00:28 <djmitche> #endmeeting