16:30:18 <djmitche> #startmeeting weekly
16:30:18 <bb-supy> Meeting started Tue Jun  7 16:30:18 2016 UTC and is due to finish in 60 minutes.  The chair is djmitche. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:30:18 <bb-supy> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:30:18 <bb-supy> The meeting name has been set to 'weekly'
16:30:38 <djmitche> #topic Introduction
16:30:46 <djmitche> https://titanpad.com/buildbot-agenda
16:31:12 <djmitche> #topic Week in Review
16:31:39 <tardyp> rutsky:  is our patch machine gun of the week
16:31:44 <djmitche> Quiet as far as trac goes this week
16:31:58 <djmitche> haha
16:32:01 <rutsky> :)
16:32:17 <rutsky> I really want 0.9.0 release :)
16:32:20 <djmitche> anything specific to highlight?
16:32:21 <gracinet> more items in in pad this week
16:32:31 <djmitche> #info Vladimir is the "patch machine gun of the week" --tardyp
16:32:57 <djmitche> I'll look at the merges after the meeting
16:33:02 <djmitche> there are several mergeable PRs
16:33:08 <rutsky> @seankelly submitted few useful fixes
16:33:29 <skelly> ty
16:33:44 <rutsky> mine changes mostly "slave"->"worker" renamings, and fix of simple 0.9.0 issues
16:33:47 <tardyp> There a few I plus 1, and as non trivial let another maintainer merge
16:33:57 <tardyp> I'll merge if no objection by eow
16:34:12 <djmitche> #info Sean has several PRs for OpenStack and git shallow clones
16:34:30 <djmitche> tardyp: sure, hopefully I get to them first :)
16:34:31 <tardyp> gracinet: thanks for your contribution to the agenda
16:34:44 <djmitche> #topic Bug 2340 update
16:34:53 <gracinet> tardyp: I contributed nothing to the agenda
16:34:53 <rutsky> I viewed at EC2 and OpenStack PRs but don't have expertise to properly review them
16:35:04 <djmitche> gracinet: aw, take the credit :)
16:35:09 <tardyp> gracinet: I can see buildbot <3
16:35:34 <rutsky> For 2340 few PRs awaiting review and merge
16:35:39 <djmitche> So it looks like pull 2234 needs some attention first
16:35:50 <rutsky> I really want this one to be merged: https://github.com/buildbot/buildbot/pull/2234
16:35:51 <djmitche> #action djmitche to review PR 2234 and other renaming-related PRs
16:35:58 <tardyp> rutsky: as you are starting to fix other bug I expect 2340 is done?
16:36:02 <gracinet> I've had some experience / issues with shallow clones in the past, I'll take a look at Sean's PR
16:36:15 <skelly> it's pretty simple
16:36:19 <djmitche> #info Starting the bounty contract process - in touch with Tony at SFC for details
16:36:21 <rutsky> djmitche marked it as :+1:, btw
16:36:32 <gracinet> skelly: don't doubt it ;-)
16:36:40 <skelly> conveniently python's True maps to 1 so it was easy to extend
16:36:42 <rutsky> tardyp: 2340 is almost complete (in part of BB codebase)
16:37:01 <tardyp> it has been almost complete for weeks ;)
16:37:09 <tardyp> how much almost?
16:37:23 <rutsky> contrib scripts left mostly untouched for 2340
16:37:31 <rutsky> master contrib scripts
16:37:38 <tardyp> (buildbot nine is almost complete as well :) )
16:37:46 <rutsky> I think I fixed all other issues with 2340
16:37:59 <bb-github> [13buildbot] 15djmitche commented on issue #2255: Yay, I've had a vendetta against apidocs since I started on Buildbot :) 02https://git.io/voLvs
16:38:04 <djmitche> tardyp: the two aren't blocking one another at this point, right?
16:38:29 <tardyp> I'm afraid yes
16:38:39 <rutsky> if pending 2340 PRs will be merged, 2340 will not block 0.9.0 release any more
16:38:44 <djmitche> ok, great
16:39:10 <djmitche> #info once the pending 2340-related PR's are merged, they will not block 0.9.0 anymore (remainder is in non-shipped code)
16:39:13 <djmitche> cool
16:39:30 <djmitche> #topic Switch to trac-github to authenticate users
16:39:40 <djmitche> I didn't see a ML thread about this -- did one get started?
16:39:40 <rutsky> I started to look at other 0.9.0-related bugs --- lets discuss them in "Bug triage for 0.9.0 release" topic
16:40:03 <djmitche> ok
16:40:17 <gracinet> out of curiosity, why aren't github issues used ?
16:40:19 <djmitche> tardyp: ^^ should we defer this until next week, or drop it?
16:40:33 <djmitche> gracinet: it's difficult to switch at this point since we have lots of poitners to trac bug #'s
16:41:05 <tardyp> and github bugs do not have all the features that we use from trac
16:41:05 <gracinet> djmitche: I see
16:41:12 <rutsky> tardyp: +1
16:41:22 <tomprince> tardyp: What do we use from trac that github doesn't provide?
16:41:25 <djmitche> the proposal here is just to use github to auth to trac, as an option
16:41:36 <tardyp> I raised the same point last year, and now that I have more experience with trac. I find it indeed useful
16:41:52 <tardyp> but I heard that github's bug is improving so..
16:42:01 <tardyp> we can revised that
16:42:05 <djmitche> yeah
16:42:08 <djmitche> maybe after 0.9.0 :)
16:42:27 <djmitche> so what should we do with the trac-github idea?
16:43:03 <gracinet> it looks good to me, but can't this also wait after 0.9.0 ?
16:43:07 * tomprince doesn't care but thinks it is a good idea.
16:43:17 <tardyp> Well I didn't change my mind about it. I didn't have the time to launch the poll as I said
16:44:13 <djmitche> ok
16:44:25 <djmitche> do you want to bring it up on the ML in a few weeks, then?
16:44:28 <djmitche> I don't think it's urgent
16:44:29 <rutsky> verm__ are you around?
16:44:58 <tardyp> yes that's fine.
16:45:11 <djmitche> awesome :)
16:45:13 <djmitche> ok
16:45:24 <djmitche> #topic Remove apidocs
16:45:30 <djmitche> I just merged this, so not much to talk about, other than
16:45:33 <djmitche> YAY! FINALLY!
16:45:36 <djmitche> :)
16:45:40 <rutsky> great :)
16:45:50 <djmitche> #topic v0.9.0b9 git tag missing
16:45:54 <rutsky> btw, I managed to build them yesterday... it was painful :)
16:45:58 <djmitche> yeah
16:46:12 <djmitche> my issue was they tended to expose things we didn't want to be part of the API
16:46:16 <tardyp> I'm sorry, I might have forgot to push them
16:46:27 <djmitche> ok, as long as we know what revision to tag
16:46:35 <rutsky> so action for tardyp?
16:46:36 <djmitche> #action tardyp to push v0.9.0b9 tag :)
16:46:40 <rutsky> thanks)
16:46:46 <djmitche> #topic rename GithubStatusPush to GitHubStatusPush
16:46:58 <tardyp> Its on my home machine, but I have the taxi in 10min, I don't think I will be able to push it before eow
16:47:00 <djmitche> I'm not too keen on this -- personally I can never remember which way the service capitalizes itself
16:47:05 <tardyp> please remind me on saturday
16:47:07 <djmitche> so I think changing it will just cause more confusion than it sovles
16:47:14 <djmitche> ok
16:47:31 <rutsky> previously it was named with proper capitalization
16:47:55 <djmitche> oh, it got changed?
16:48:02 <rutsky> it may confuse users of old GitHubStatus status reporter
16:48:04 * tardyp is the culprit
16:48:33 <djmitche> OK
16:48:34 <tardyp> go for the fix
16:48:39 <djmitche> I imagine supporting both is easy
16:48:46 <tardyp> no
16:49:03 <rutsky> I think they are not interchengable
16:49:06 <tardyp> let's just make one with the proper official case
16:49:46 <rutsky> do we want backward compatibility with previous 0.9 betas?
16:49:52 <tardyp> no
16:50:05 <rutsky> then it's easy
16:50:13 <tardyp> that's the point of the beta, we can break stuf until official release
16:50:19 <djmitche> #info the new GithubStatusPush is spelled differently than the old, but also incompatible
16:50:20 <rutsky> (but having old name with warning is easy to, btw)
16:50:20 <tardyp> but then, we have to keep it
16:50:30 <tardyp> so its the last chance to fix it. lets take it
16:50:34 <djmitche> #info the plan is to fix the spelling, without concern for compatibility within the betas
16:50:37 <rutsky> agree
16:50:42 <djmitche> cool
16:50:58 <djmitche> #topic Mention-Bot is missing in action
16:51:06 <gracinet> as long as it's in release notes (don't use it myself)
16:51:42 <tomprince> https://github.com/buildbot/buildbot/settings/hooks doesn't have mention bot.
16:52:09 <tomprince> I can add it back in.
16:52:15 <djmitche> OK, thanks
16:52:22 <djmitche> that's odd
16:52:40 <tardyp> +1
16:52:46 <djmitche> #info mention-bot was removed from the repository web-hooks; tomprince will add it back in
16:52:54 <rutsky> actually I saw yesterday that mention-bot was in list of participants of one of the PRs
16:53:06 <rutsky> but that PR haven't any comment from him
16:53:13 <djmitche> #topic "Completely remove `public_html` or resurrect its functionality?"
16:53:43 <rutsky> ticket has some pros/cons arguments: http://trac.buildbot.net/ticket/3547
16:53:43 <djmitche> I wonder if it's being quashed as spam
16:53:45 <djmitche> let's see what happens
16:54:19 <tardyp> I got to run..
16:54:29 <djmitche> tardyp: thanks!
16:54:40 <rutsky> tardyp: thanks, have a nice flight!
16:54:42 <djmitche> rutsky: what do you think about serving public_html as a plugin?
16:55:13 <tomprince> Builtin plugin, I assume?
16:55:26 <rutsky> djmitche: I tend to think so
16:55:42 <rutsky> public_html is already not working and if we want we need to write code for it
16:55:56 <djmitche> I was thinking as a different pypi package
16:55:59 <rutsky> and I think BB plugin is a good way to do so
16:56:04 <djmitche> but still in the same git repo like our other plugins
16:56:20 <djmitche> Let's not block 0.9.0 on that, though
16:56:21 <gracinet> sounds good indeed
16:56:31 <djmitche> in principle, it's something an interested user could implement
16:56:38 <rutsky> ok, so we don't want this for 0.9.0?
16:56:40 <rutsky> djmitche: yes
16:56:47 <djmitche> and if it's a separate component, someone running 0.9.0 could still add the plugin even if it's released later
16:57:06 <skelly> I think we would want it for 0.9.0, if someone were to do it by then but it shouldn't block
16:57:20 <rutsky> so lets remove notion of public_html and maybe reimplement it later
16:57:43 <tomprince> That seem reasonable.
16:57:43 <rutsky> skelly: I think those users can wait for 0.9.1 or implement this pluging themself
16:58:15 <djmitche> skelly: yeah, a plugin gets us a little wiggle room there :)
16:58:22 <djmitche> rutsky: can you summarize that in the ticket?
16:58:36 <rutsky> djmitche: ok
16:59:01 <djmitche> #topic web UI doesn't work behind a proxy
16:59:03 <djmitche> https://lists.buildbot.net/pipermail/users/2016-June/000519.html
16:59:49 <rutsky> not sure is this resolved, but looks like this is proxy misconfiguration (proxy doesn't work with websockets)
17:00:07 <djmitche> ok
17:00:15 <djmitche> let's skip and keep handling in the ML then
17:00:16 <rutsky> so static pages can be seend, but no updates or data which is gatheres through WS
17:00:22 <rutsky> yes
17:00:31 <djmitche> ok!
17:00:40 <djmitche> #info will continue to handle this on the mailing list
17:00:50 <djmitche> #topic Bug Triage for 0.9.0 release
17:01:08 <gracinet> web ui works fine for me with Apache + websockets module
17:01:17 <djmitche> http://trac.buildbot.net/query?status=accepted&status=assigned&status=new&status=reopened&milestone=0.9.0&group=status&order=priority
17:01:41 <rutsky> gracinet: but have you tried to enable auth in Apache?
17:01:54 <rutsky> I'm not sure how well proxy with auth plays with WS
17:02:01 <gracinet> basic auth, yeah that's what I'm doing
17:02:35 <rutsky> gracinet: can you reply on ML with info about your config?
17:02:36 <gracinet> (for the time being, plan to switch to some Oauth2/OIDC)
17:02:52 <gracinet> last mail I see, he says he missed the part about websockets
17:02:59 <rutsky> tardyp: Daniel Collin might be interested in your experience
17:03:19 <gracinet> Ill read the full thread first :-)
17:03:36 * djmitche got ahead of things, sorry
17:05:10 <rutsky> regarding 0.9.0 release --- we need to review bug-reports for 0.9.0 milestone
17:05:16 <djmitche> so I see 14 of them
17:05:20 <gracinet> rutsky: after reading the full thread, it seems tardyp answered all the questions
17:05:25 <djmitche> one will be done soon (2340)
17:05:34 <djmitche> and one we just moved out (public_html)
17:05:40 <rutsky> for assigned on me I submitted PRs for all od them
17:06:08 <djmitche> so responsible! very awesome!
17:06:25 <djmitche> so I haven't looked at this list for a while
17:06:32 <djmitche> I can commit to looking through it again and trying to move some more things out
17:06:39 <djmitche> unless you want to try to do it now?
17:06:40 <rutsky> djmitche: nice lifehack --- I first submit PR, and them assing ticket on myself :)
17:07:02 <djmitche> yeah, clever ;)
17:07:15 <rutsky> I raised this triage topic becaouse I don't understand some of very old tickets :)
17:07:25 <gracinet> I'll try and do #3504 tomorrow
17:07:26 <djmitche> ok
17:07:30 <djmitche> cool!
17:07:45 <bb-trac> [trac] #3559/defect (assigned) updated by rutsky (empty comment) http://trac.buildbot.net/ticket/3559
17:08:25 <gracinet> my lifehack, report, PR and then *someone else* assigns me to it
17:08:26 <djmitche> rutsky: want to go one-by-one?
17:08:42 <rutsky> djmitche: yes please
17:08:58 <rutsky> http://trac.buildbot.net/ticket/2995 - Get rid of all the broken / unwired elements from the web UI.
17:09:18 <rutsky> last time djmitche asked tardyp about it: is this a blocker for 0.9.0?  :)
17:09:28 <djmitche> heh
17:09:37 <rutsky> I think it's better to wait for tardyp for this one
17:09:43 <tomprince> I think that should be a blocker.
17:10:07 <djmitche> yes
17:10:12 <djmitche> post another query on it?
17:10:14 <tomprince> I don't think having non-working UI elements is a good first impression.
17:10:20 <djmitche> agree
17:10:22 <rutsky> agree
17:10:48 <djmitche> next?
17:11:08 <bb-trac> [trac] #2995/defect (new) updated by rutsky (@tardyp, ping) http://trac.buildbot.net/ticket/2995
17:11:22 <rutsky> http://trac.buildbot.net/ticket/3395 - document what is working with multi master in 0.9
17:11:33 <rutsky> I never tried BB multimaster at all
17:11:42 <tomprince> Has anybody used that?
17:12:34 <rutsky> I think we need tardyp comments here too...
17:12:37 <djmitche> there are a lot of 0.8.x users with multimaster
17:12:47 <djmitche> this is more of a "these 5 things are broken" documentation
17:12:51 <djmitche> so should be fairly straightforward
17:13:05 <djmitche> it might be a good start to just put that text in the docs and make a PR
17:13:09 <djmitche> tardyp can edit (or not)
17:14:56 <rutsky> current multimaster docs: http://docs.buildbot.net/latest/manual/cfg-global.html?highlight=multimaster#multi-master-mode
17:16:22 <rutsky> 0.8 multimaster docs: http://docs.buildbot.net/0.8.12/manual/cfg-global.html?highlight=multimaster#multi-master-mode
17:16:35 <rutsky> looks like they are identical...
17:16:52 <rutsky> I don't understand from the ticket what is working and what is not...
17:17:22 <djmitche> yeah
17:17:30 <djmitche> assign that one to me, actually -- I'll try to write somethign up for review
17:17:36 <rutsky> ok, thanks
17:17:43 <djmitche> sometimes I think I'm the only one who cares about multimaster
17:17:44 <bb-trac> [trac] #3395/enhancement (assigned) updated by rutsky (empty comment) http://trac.buildbot.net/ticket/3395
17:17:46 <djmitche> (and I don't, so..)
17:18:10 <rutsky> I never heard of multimaster BB usage before :)
17:18:25 <rutsky> http://trac.buildbot.net/ticket/3483 - rewrite usage of virtualenv in firstrun docs
17:18:36 <cmouse> rutsky: it's when CI outage cannot b ehad
17:18:37 <cmouse> be had
17:19:33 <rutsky> cmouse: yes, HA, and stuff like that, but I haven't seen much of this in my practice
17:19:49 <rutsky> 3483 is actually not blocking 0.9.0
17:20:07 <cmouse> rutsky: i was trying to be sarcastic
17:20:13 <cmouse> such hard thing to over irc
17:20:19 <rutsky> ah :)
17:20:29 <bb-trac> [trac] #3483/undecided (new) updated by rutsky (empty comment) http://trac.buildbot.net/ticket/3483
17:20:40 <rutsky> I moved 3483 to 0.9.+
17:20:51 <tomprince> cmouse rutsky: CI outage is quite expensive, in terms of developer time, though.
17:21:02 <rutsky> http://trac.buildbot.net/ticket/3484 - update screenshots in firstrun tutorial
17:21:04 <cmouse> tomprince: possibly.
17:21:27 <rutsky> this is not really 0.9.0 stopper, but it's nice to fix this :)
17:21:30 <cmouse> tomprince: although can't readily imagine what kind of outage that would be.
17:21:36 <cmouse> tomprince: since the master is rather thin component after all
17:21:54 <rutsky> I'll try to revisit firstrun tutorial and update it
17:22:01 <bb-trac> [trac] #3484/undecided (assigned) updated by rutsky (empty comment) http://trac.buildbot.net/ticket/3484
17:22:07 <bb-trac> [trac] #3483/undecided (assigned) updated by rutsky (empty comment) http://trac.buildbot.net/ticket/3483
17:22:07 <cmouse> tomprince: at least what kind of outage would only affect master
17:22:21 <rutsky> so 3483 and 3484 are mine now :)
17:22:49 <tomprince> Yeah ... it isn't clear that multi-master is a useful way to avoid CI outages.
17:22:53 <rutsky> http://trac.buildbot.net/ticket/3504 - PosgreSQL foreign key error upon reconfigure
17:23:14 <rutsky> gracinet agreed to try to fix it tomorrow, thanks!
17:23:33 <bb-trac> [trac] #3504/undecided (assigned) updated by rutsky (empty comment) http://trac.buildbot.net/ticket/3504
17:23:57 <rutsky> http://trac.buildbot.net/ticket/3525 - Fix customisations.rst
17:24:05 <rutsky> it's about outdated docs
17:24:20 <rutsky> actually, BB has *a lot* of outdated docs
17:25:11 <rutsky> this is not good from one side, but we can think about this in a way that: BB actually has docs, and BB has a lot of docs (thats why it's hard to update them) :)
17:25:28 * djmitche distracted by other meeting - ping me if you need me
17:25:30 <bb-trac> [trac] #3504/undecided (assigned) updated by gracinet (I'll add the cascading deletes on such many2many relationships  (heard of something ...) http://trac.buildbot.net/ticket/3504
17:26:14 <rutsky> anyone want to update http://docs.buildbot.net/latest/manual/customization.html ?
17:26:47 <rutsky> I got to go in few minutes
17:27:04 <rutsky> few more tickets left:
17:27:09 <rutsky> http://trac.buildbot.net/ticket/3529 - Add warning in trigger step if collapseBuild=True for the target builder
17:27:23 <rutsky> not sure is this blocking 0.9.0
17:27:30 <rutsky> need to read IRC chat there...
17:27:33 <tardyp> djmitche: you say a lot of 0.8 users are multimaster
17:27:46 <tardyp> do you have any example?
17:28:10 <tardyp> I never had anybody asking question about multimaster actually
17:28:10 <djmitche> none come to mind -- it's been a while
17:28:34 <tardyp> basically the story of multimaster nine is depending on the story of crossbar
17:28:59 <tardyp> I did make it work once for my dev, but then I never actually run it in prod
17:29:34 <gracinet> wouldn't it be wise then not to make too much promises and say that multimaster is beta within 0.9.0 ?
17:29:35 <tardyp> So its considered experimental
17:29:55 <tardyp> yes
17:30:07 <djmitche> that's a good way to put it
17:30:08 <gracinet> of course if nobody reports about it, we're back to current situation of 0.9.1
17:30:09 <tomprince> rutsky: re: #3529 I don't think it should block 0.9.0, the behavior isn't new in 0.9 and is just a documentation issue.
17:30:32 <tardyp> the whole multimaster doc is http://docs.buildbot.net/latest/manual/cfg-global.html#wamp
17:30:34 <gracinet> s/of/for
17:31:11 <tardyp> there is not so much information missing here
17:33:16 <gracinet> I have to run now
17:33:27 <rutsky> tomprince: agree
17:33:29 <gracinet> looks like 0.9.0 is really around the corner, that's great
17:33:36 <bb-trac> [trac] #3529/enhancement (new) updated by rutsky (empty comment) http://trac.buildbot.net/ticket/3529
17:33:49 <rutsky> I moved 3529 to 0.9.+
17:34:01 <rutsky> sorry, I got to run
17:34:21 <rutsky> there are few unassigned tickets for 0.9.0
17:34:24 <tardyp> gracinet: if you have the time to contribute your apache config in the doc. Would be a nice addition to http://docs.buildbot.net/latest/manual/cfg-www.html#reverse-proxy-configuration
17:34:26 <rutsky> and few questions for tardyp
17:34:35 <tardyp> aside from the nginx I contributed
17:34:53 <gracinet> tardyp: oh I didn't remember it was there
17:35:00 <gracinet> wasn't there
17:36:17 <tardyp> #2995 I think we have "almost"tm all the unwired element removed
17:36:52 <bb-trac> [trac] #3560/undecided (v:master) created by rutsky (implement plugin for serving static files (equivalent of 0.8 `public_html`)) http://trac.buildbot.net/ticket/3560
17:36:57 <tardyp> I keep it open, because we need to make a manual turn around of all the pages, just before the release, and see if everything is fine
17:37:13 <bb-trac> [trac] #3561/task (v:0.8.12) created by gracinet (provide sample Apache conf (websockets)) http://trac.buildbot.net/ticket/3561
17:38:11 <tardyp> btw djmitche I *do* care about multimaster, and use it in prod (eight)
17:38:26 <djmitche> haha, now I know one!
17:39:10 <gracinet> bye
17:40:14 <djmitche> thanks gracinet
17:49:21 <djmitche> #endmeeting