16:00:58 <djmitche> #startmeeting weekly
16:00:58 <bb-supy> Meeting started Tue Aug  4 16:00:58 2015 UTC and is due to finish in 60 minutes.  The chair is djmitche. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:58 <bb-supy> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:58 <bb-supy> The meeting name has been set to 'weekly'
16:01:16 <sa2ajj> what's the agenda's url btw? :)
16:01:21 <djmitche> https://etherpad.mozilla.org/buildbot-agenda
16:01:32 <tardyp> hi
16:01:41 <bb-meta> build #188 of py27-tw1220 is complete: Failure [4failed virtualenv setup]  Build details are at http://buildbot.buildbot.net/builders/py27-tw1220/builds/188  blamelist: Mikhail Sobolev <mss@mawhrin.net>, Prasoon Shukla <prasoon92.iitr@gmail.com>
16:01:44 <Lurimax> Hi all
16:01:52 * djmitche waves to everyone
16:02:00 <djmitche> #topic meeting coverage
16:02:11 <sa2ajj> what this is about?
16:02:14 <djmitche> First up, I'll be in .ie for the next two meetings, so I need a volunteer to run them during that time
16:02:21 <djmitche> on vacation, no laptop, etc.
16:02:36 <djmitche> probably a great deal of guinnes involved
16:02:46 * sa2ajj prefers murphy's
16:02:47 <bb-meta> build #193 of py27-sam072 is complete: Success [3build successful]  Build details are at http://buildbot.buildbot.net/builders/py27-sam072/builds/193
16:02:56 <djmitche> yeah, we'll see what the pubs have
16:03:02 <Lurimax> I can do it, if I too can have a Guiness?
16:03:16 <sa2ajj> pubs have both, iirc
16:03:17 <djmitche> if you're in ireland :)
16:03:24 <tardyp> I'm In vacation too... 3g roaming with terrible bw..
16:03:39 <djmitche> thanks for attending  then!
16:03:49 <djmitche> Lurimax: OK, I'll count you in and find some way to send you a guinness
16:03:51 <sa2ajj> so it's gonna be gsoc only?
16:04:07 <Lurimax> And any updates I'm guessing
16:04:19 <djmitche> yep
16:04:26 <djmitche> cool!
16:04:27 <sa2ajj> maybe, we'll just ask to send the gsoc updates to the ml?
16:04:35 <djmitche> that works too
16:04:44 <sa2ajj> (for these meetings)
16:04:52 <djmitche> sounds like a plan
16:04:55 <tardyp> agree
16:05:18 <djmitche> sa2ajj: can you take care of gathering and sending?
16:05:24 <sa2ajj> sure
16:05:45 <djmitche> great!
16:05:52 <djmitche> so I wrote up a new stalled PR policy
16:06:00 <djmitche> http://trac.buildbot.net/wiki/StalePullRequestPolicy
16:06:05 <djmitche> #topic stalled PR policy
16:06:12 <djmitche> stale/stalled meh
16:06:24 <djmitche> basically 7 days to stalled, then 30 days to closed
16:06:28 <djmitche> with discretion
16:06:31 <sa2ajj> sounds good
16:06:35 <Lurimax> looks good to me
16:06:36 <djmitche> k
16:06:40 <tardyp> +1
16:06:43 <sa2ajj> btw, i closed my stalled pr :)
16:06:50 <djmitche> I looked at cafepress
16:06:53 <djmitche> and in fact made a "store"
16:06:54 <tardyp> great initiative
16:07:08 <djmitche> http://www.cafepress.com/dd/100371708
16:07:14 <djmitche> but the products don't look too good
16:07:29 <Lurimax> hmmmmm
16:07:36 <djmitche> #topic making more swag
16:07:54 <djmitche> (apologies to those reading the minutes for the late topic changes!)
16:07:56 <Lurimax> could to with some falvour text
16:08:04 <Lurimax> the logo doesn't say much
16:08:07 <djmitche> yeah
16:08:09 <sa2ajj> re swag policy: i was thinking about 1 big feature/3-5 bugfixes -> a swag of choice :)
16:08:28 <djmitche> I like that
16:08:35 <bb-meta> build #185 of db-mysql is complete: Success [3build successful]  Build details are at http://buildbot.buildbot.net/builders/db-mysql/builds/185
16:08:44 <djmitche> we could limit it to things <= $20 or something
16:08:48 <sa2ajj> s/bugfixes/changes/
16:08:56 <sa2ajj> sounds good
16:08:57 <djmitche> yeah
16:09:03 <djmitche> but let's find some swag we like first
16:09:10 <djmitche> Lurimax: do you want to give that a try?
16:09:37 <sa2ajj> one bit: maybe first item would always be a t-shirt, and then of choice
16:09:38 * tardyp needs some context on this topic
16:09:48 <djmitche> #info suggestion of giving a swag item of choice for anyone contributing 1 big feature or 3-5 bugfixes
16:09:59 <Lurimax> not sure i have the time for it :/
16:10:01 <djmitche> sa2ajj: makes sense
16:10:02 <djmitche> ok
16:10:14 <Lurimax> and i suck at design
16:10:15 <djmitche> tardyp: the idea is to reward short-term contributors
16:10:37 <djmitche> #action djmitche to try to come up with better-looking cafepress swag, perhaps with some text (over the next few weeks)
16:10:38 <sa2ajj> tardyp: swag is about promoting buildbot + encouraging contributors
16:11:02 <tardyp> ok. I missed this from last meeting probably
16:11:08 <djmitche> yes, sorry
16:11:11 <sa2ajj> it was during the meeting before
16:11:12 <Lurimax> maybe a few pint will help with that action ;)
16:11:13 <tardyp> cafepress site is still loading on my shity connection
16:11:21 <djmitche> :(
16:11:34 <djmitche> I'll see if there are other vendors that will do a "site" like this with multiple swag items
16:11:47 <djmitche> I'll also check with the SFC about "selling" swag (can we count the income as a donation?)
16:12:01 <djmitche> #action djmitche to check with SFC about cafepress "sales"
16:12:18 <djmitche> I'd like to give people the opportunity to buy swag -- I'm less concerned about the income
16:12:25 <sa2ajj> they actually might have experience in the area and suggest some vendors...
16:12:30 <djmitche> true
16:12:48 <djmitche> #topic mailing list migration
16:13:03 <djmitche> I think skelly's not around.. verm__?
16:13:16 <djmitche> when last we left the topic, we were waiting for a tarball from SF
16:13:52 <sa2ajj> right
16:14:05 <djmitche> I guess no updates :(
16:14:06 <sa2ajj> i remember verm__ mentioned (around a year ago) that sf was not very fast
16:14:12 <djmitche> right
16:14:17 <djmitche> they're probably worse now :(
16:14:26 <sa2ajj> or maybe just different :)
16:14:41 <djmitche> we should probably consider just starting in on the new list at some point
16:14:44 <djmitche> without re-subscribing
16:14:48 <djmitche> if this turns out to be difficult
16:14:58 <sa2ajj> i think this is what skelly proposed
16:15:07 <djmitche> perhaps
16:15:13 <tardyp> we could setup a forward
16:15:18 <sa2ajj> not on sf
16:15:21 <djmitche> the downside is losing a lot of subscribers who don't heed the request to subscribe
16:15:24 <djmitche> right
16:15:26 <tardyp> e.g register the sf ml address to the bb ml address
16:15:56 <sa2ajj> we can also do it other way around:
16:16:22 <djmitche> #info if SF can't produce the tarball in a timely fashion, we may need to consider alternatives
16:16:29 <sa2ajj> create the new ml, announce it, and subscribe b-d@sf to this new list
16:16:48 <sa2ajj> those people who will subscribe to the new one most likely will unsubscribe from the old one
16:16:59 <djmitche> we'd have to open b-d to posts from anyone in that case -> more spam
16:17:13 <djmitche> maybe we could subscribe b-d just to digests?
16:17:17 <tardyp> sa2ajj: that was my proposal :)
16:17:22 <djmitche> that'd be a good reminder that replies on b-d are not invited
16:17:34 <sa2ajj> sounds good
16:17:38 <djmitche> #info maybe just exhort people to subscribe to teh new lists
16:17:40 <sa2ajj> tardyp: great minds think alike :)
16:18:01 <djmitche> #info maybe subscribe buildbot-devel to users@buildbot.net, maybe in digest mode, as a reminder
16:18:14 <djmitche> #topic 0.9.0b2!!
16:18:18 <sa2ajj> djmitche: we could add moderation and i volunteer to moderate the ml
16:18:48 <sa2ajj> i usually check things in the morning and that's not a big effort (i did this before for a number of lists)
16:18:51 <djmitche> #action djmitche to summarize mailing list options to the list itself
16:18:55 <tardyp> sa2ajj: I dont think we have that much spam issue (or gmail is filtering it for me)
16:19:15 <sa2ajj> we do not know how much spam sf filters (if any)
16:19:18 <djmitche> limiting the list to postings from subscribers keeps the spam down
16:19:45 <sa2ajj> true, however making everyone to subscribe might not be good to keep users :)
16:19:45 <tardyp> I dont think we need moderation. This just rise the latency
16:19:52 <Lurimax> I have to go now. Good luck with the rest of the meeting
16:19:58 <sa2ajj> Lurimax: thank you
16:19:59 <djmitche> thanks :D
16:20:08 <bobhood> Thanks for coming, pal. ;)
16:20:24 <sa2ajj> moderation of mails from non-subscribers
16:20:43 <sa2ajj> mailman, iirc, has an option to "approve" the non-subscriber for future postings
16:20:56 <djmitche> sa2ajj: what I'd like to avoid is folks living on forever on buildbot-devel
16:21:12 <tardyp> ok in that case
16:21:13 <djmitche> much less posting to buildbot-devel
16:21:14 <sa2ajj> that will be taken care of once we get the stuff from sf
16:21:33 <djmitche> oh, I see, so this is just temporary, then we automatically subscribe everyone
16:21:44 <sa2ajj> yep
16:21:57 <tardyp> +1
16:22:04 <djmitche> ok
16:22:19 <djmitche> now, on to 0.9.0b2 :)
16:22:28 <djmitche> I haven't seen any feedback yet, but it's only been out for two days
16:22:28 <sa2ajj> it was released :)
16:22:30 <djmitche> yay!
16:22:38 <tardyp> I got one via private mail
16:22:47 <djmitche> oh, good
16:22:49 <tardyp> some confusion about the use of --pre to install ui plugins
16:23:16 <tardyp> everything good apart from that
16:23:25 <sa2ajj> that's _really_ good
16:23:34 <tardyp> we are in summer period, so I dont expect tha tmuch feedback
16:23:36 <djmitche> indeed, because that goes away once we make a full release
16:23:38 <djmitche> true
16:23:44 * sa2ajj nods
16:23:48 <djmitche> we probably won't get much until mid-september
16:24:06 <djmitche> which works out well with vacation timing :)
16:24:12 <tardyp> that's a target for beta 3!
16:24:19 <sa2ajj> excellent
16:24:21 <djmitche> sounds good
16:24:32 <djmitche> #info beta 3 targetted for mid-september, after more feedback comes in
16:24:34 <tardyp> will probably mostly have gsoc stuff anyway
16:24:35 <sa2ajj> so when people are more active, we'll have even better beta
16:24:51 <djmitche> good points
16:24:56 <djmitche> and speaking of gsoc...
16:25:07 <djmitche> #topic GSoC updates
16:25:19 <djmitche> I saw some comments just now on prasoon's PR -- how's stats going?
16:25:39 <djmitche> merged, in fact
16:26:16 <sa2ajj> yes, i did it just before the meeting
16:26:29 <sa2ajj> he lists two todo items more in the last pr
16:26:49 <djmitche> but good to see code flowing in
16:26:56 <sa2ajj> i do not know what else is there
16:27:13 <sa2ajj> re code: i was a bit insisting on abc testing... but it looks pylint takes care of that
16:27:29 <djmitche> oh, wow, that's cool!
16:27:30 <sa2ajj> which is good for bb code
16:27:44 <djmitche> indeed, esp with more use of ABC throughout
16:28:16 <sa2ajj> i tried to convert the master to it, however there were some issues related to old style vs new style classes...
16:28:26 <djmitche> #info stats - prasoon's PR was just merged; a few more TODO items
16:28:36 <djmitche> yeah, maybe Lurimax's work will fis that
16:28:38 <djmitche> *fix
16:28:51 * sa2ajj really hopes it is so :)
16:28:54 <bb-trac> [trac] #3327/defect (v:master) created by vlovich (buildbot 0.9b1 tries to build deleted branch?) http://trac.buildbot.net/ticket/3327
16:28:54 <djmitche> next: material design
16:29:02 <djmitche> chasezhang: you around? what's new?
16:29:17 <djmitche> ^^ more 0.9.0b1 feedback :)
16:29:35 <sa2ajj> we need one for 0.9.0b2 :)
16:29:58 <tardyp> chasezhang: I merged its last PR
16:30:07 <tardyp> with build support
16:30:14 <tardyp> he is working on step + log support
16:30:23 <tardyp> and then will be working on forcesched forms
16:30:49 <djmitche> wow
16:30:53 <sa2ajj> that's great
16:30:58 <tardyp> andras has sent a working code for buildbot-data looks very promising
16:31:10 <djmitche> #info chase is working on porting steps and logs to material design, with force scheduler forms up next
16:31:24 <djmitche> #info andras has a buildbot-data client-side service up and running, looking promising
16:31:31 <tardyp> he said he will be working with chase to test it more heavily under real load
16:31:44 <djmitche> #action dustin to add a 0.9.0b2 version to trac, update default version to 0.8.12
16:31:49 <djmitche> awesome
16:32:03 <djmitche> it's great that they can work together like that
16:32:04 <sa2ajj> default version update will require configuration file update
16:32:12 <tardyp> I think he will be able to move soon the main UI
16:32:20 <tardyp> probably for beta3
16:32:35 <djmitche> to the buildbot-data service, you mean/
16:32:44 * sa2ajj needs to look at console view for the nine
16:32:58 <djmitche> #info 0.9.0beta3 may include the new buildbot-data service
16:33:01 <tardyp> move the main UI to the new data service yes
16:33:09 <sa2ajj> if it looks reasonable, i'll switch our installation to the latest beta
16:33:20 <djmitche> sweet!
16:33:26 <djmitche> last but not least, Lurimax's project
16:33:49 <djmitche> #info Twisted trial is now ported to python-3; spread remains un-ported
16:34:10 <djmitche> #info Lurimax is looking at porting spread and at some last bits of bytestring/unicode issues in porting Buildbot (aside from spread)
16:34:21 <djmitche> bobhood: anything to add? this is just summarized from our convo before the hour
16:34:33 <bobhood> I've had some problems meeting with Jørn lately (my fault, not his).  His latest update is "I'm still working on twisted.spread. Still hoping to have it done by Friday "
16:34:47 <djmitche> great :)
16:35:06 <djmitche> ok, the last bit then is trac slowness
16:35:10 <sa2ajj> wait
16:35:13 * djmitche waits
16:35:21 <sa2ajj> regarding PB, what options do we have?
16:35:25 <sa2ajj> buildbot-data?
16:35:36 <djmitche> the two are unrelated
16:35:44 <sa2ajj> are they really?
16:35:46 <djmitche> PB is master/slave, buildbot-data is between tabs in a browser
16:35:57 <sa2ajj> right...
16:36:02 <tardyp> We have my wamp poc
16:36:10 <tardyp> but it really needs a lot of works
16:36:20 <sa2ajj> do you think you could make a todo list of sort?
16:36:24 <tardyp> actually for me any change of slave/master proto require lot of work
16:36:25 <djmitche> ah, I see, yes, crossbar.io could definitely be used for master/slave communication
16:36:34 <tardyp> because it is not very well tested
16:36:41 <djmitche> yes, for most users!
16:36:48 <djmitche> to answer the first question
16:36:53 <tardyp> there are a lot of connectivity cornercase to deal with
16:37:06 <djmitche> we have a partly-completed factorization of master/slave communication into an abstraction
16:37:08 <tardyp> that was dealt with the several years of buildbot
16:37:24 <sa2ajj> i mean if pb stands in the way, we might want to invest a bit of time to look at alternatives
16:37:27 <djmitche> yeah -- handling reconnects, TCP keepalives, etc.
16:37:46 <tardyp> my poc worked for the normal case, but I wasn't able to make it realliable at all
16:37:46 <sa2ajj> djmitche: is it that branch?
16:37:59 <sa2ajj> slave-proto
16:38:01 <djmitche> it was tomprince's work, yeah
16:38:13 <djmitche> he was cleaning up after the GSoC project
16:38:19 <djmitche> then disapepared :(
16:38:25 <tardyp> I have a wampslave branch I think. Not enough bw to open github :-/
16:38:42 <djmitche> I remember that too
16:38:44 <sa2ajj> right. and that branch @ b/b is the last state, is it not?
16:38:54 <djmitche> the last state I've seen, yes
16:39:02 <tardyp> Yeah I cleaned up that part before the wamp poc
16:39:04 <djmitche> I think he said he had it taken apart in his working copy but that was 18mo ago at least
16:39:05 <sa2ajj> ok, i'll have a look then
16:39:08 <tardyp> and it is merged
16:39:25 <tardyp> so the abstraction is there.
16:39:31 <sa2ajj> aha
16:39:44 <djmitche> At best I think that might give us an excuse to skip spread-related tests on python3
16:39:49 <tardyp> that does not mean creating a new proto is easy
16:40:04 <djmitche> I don't know if this is a good time to start changing the protocols -- lots of other things in motion right now
16:40:06 <tardyp> certainly not in the area of possible for an end of gsoc
16:40:39 <tardyp> The thing is we will not claim python3 in september
16:40:45 * sa2ajj said he'd have a look :)
16:40:49 <tardyp> so we can work on other things
16:40:58 <tardyp> I already merged a null proto though
16:41:01 <djmitche> nothing wrong with looking
16:41:08 * sa2ajj looks around...
16:41:14 <djmitche> ok!
16:41:14 <tardyp> wich works when master and slave are in the same process
16:41:33 <djmitche> hio was asking about that yesterday
16:41:37 <tardyp> maybe that could be an option to move forward
16:42:59 <sa2ajj> anyway, back to trac :)
16:43:36 <tardyp> I was trying to move some bugs, for the beta release
16:43:46 <tardyp> the slowness just stopped my diligence..
16:43:52 <djmitche> #topic Trac Slowness
16:43:56 <tardyp> Its basically not usable for me
16:44:04 <djmitche> slow load time, in particular?
16:44:24 <tardyp> After I update one bug, then the UI is very slow for about a minute
16:45:17 <bb-trac> [trac] #3327/defect (new) updated by dustin (empty comment) http://trac.buildbot.net/ticket/3327
16:45:40 <djmitche> as in, when you try to make a subsequent change it takes a long time to load after clicking "submit changes"?
16:46:00 <tardyp> that my experience yes
16:46:09 <sa2ajj> it's not that bad for me, though trac is definitely not snappy
16:46:15 <bb-trac> [trac] #3327/defect (new) updated by dustin (empty comment) http://trac.buildbot.net/ticket/3327
16:46:16 <tardyp> maybe not in the first update
16:46:33 <tardyp> but after a few changes, then it starts to be slow, then I give up
16:46:34 <djmitche> same as sa2ajj for me
16:46:43 <djmitche> ok
16:46:57 <djmitche> I'll open a bug on that, but I think we'll need more data
16:46:59 <tardyp> there are a lots of bugs to triage, so it make it complicated
16:47:12 <djmitche> slow load times, slow JS, failing AJAX, ..
16:47:15 <djmitche> yeah, that sucks
16:48:12 <sa2ajj> static served over wsgi
16:48:20 <sa2ajj> sqlite db is too big
16:48:27 <tardyp> nobody but me is experiencing the slowness?
16:48:30 <bb-trac> [trac] #3328/defect (v:) created by dustin (Trac gets slow after modifying bugs) http://trac.buildbot.net/ticket/3328
16:48:45 <tardyp> its sqlite?!
16:48:47 <djmitche> well, not the two of us anyway
16:49:01 <sa2ajj> that kind of slowness -- no, however, as i said, a bigger tracs than that were snappier
16:49:13 <djmitche> no, mysql
16:49:54 <sa2ajj> is it really?!
16:50:07 <djmitche> according to trac.ini :)
16:50:24 <tardyp> it is basically a mono user server, so I don't know why this could be so slow at times
16:50:35 <djmitche> database = mysql://trac:***@mysql.int.buildbot.net/trac
16:50:43 <djmitche> yeah, I dont' know either
16:50:47 <sa2ajj> yes, it is
16:50:47 <tardyp> I'll update the trac with more data when I feel the slowness again
16:50:55 <djmitche> thanks!
16:51:07 <djmitche> some chrome diagnostics might help
16:51:13 <djmitche> and we can correlate timestamps with logs
16:51:14 <tardyp> indeed
16:51:18 <djmitche> I know there are some crontasks that consult git
16:51:38 <sa2ajj> i do not think that'd affect performace that badly :/
16:51:46 <djmitche> not if it's working correctly
16:51:51 <tardyp> I dont think that anybody is actually using the git integration
16:51:52 <djmitche> we had some trouble with it before, but that's been fixed
16:52:16 <djmitche> in fact it's not working
16:52:19 <djmitche> i don't see any commits in the timeline
16:52:42 <skelly> that part hasn't been working for a while
16:52:46 <tardyp> I've seen some when doing research. Was just noise for me as I was looking for bugs :)
16:52:59 <djmitche> http://trac.buildbot.net/changeset/7a8e8c2f13be6cac84481fe6d13d95de9e559ecc/ is the last commit
16:53:13 <sa2ajj> djmitche: skelly and i mentioned that already :)
16:53:24 <sa2ajj> hook does not work
16:53:29 <djmitche> yeah, just confirming :)
16:54:19 <djmitche> http://trac.buildbot.net/ticket/3243
16:55:00 <sa2ajj> anyway, we have a chance to finish this meeting before the allocated hour is over :D
16:55:03 <bb-trac> [trac] #3243/support-request (new) updated by dustin (This is preventing Git updates on Trac, and has been since January 2015.    We ...) http://trac.buildbot.net/ticket/3243
16:55:15 <djmitche> yeah, sorry, I stopped rushing things :)
16:55:21 <sa2ajj> :)
16:55:22 <djmitche> #info bugs filed for trac, will investigate
16:55:25 <djmitche> #endmeeting