16:29:59 <djmitche> #startmeeting weekly
16:29:59 <bb-supy`> Meeting started Tue Nov 15 16:29:59 2016 UTC and is due to finish in 60 minutes.  The chair is djmitche. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:29:59 <bb-supy`> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:29:59 <bb-supy`> The meeting name has been set to 'weekly'
16:30:11 <djmitche> #topic introduction
16:30:18 <djmitche> https://titanpad.com/buildbot-agenda
16:30:19 <tardyp> hi
16:30:24 <bdbaddog> Hi!
16:30:24 <djmitche> hi!
16:30:34 <djmitche> who else? rutsky? sa2ajj?
16:32:15 * sa2ajj waves
16:33:28 <tardyp> lets go?
16:33:54 * sa2ajj checks the agenda (let's go!)
16:34:19 * sa2ajj nods
16:34:41 <tardyp> djmitche: ?
16:35:03 <tardyp> I dont have the rights to start the topic.
16:35:14 <tardyp> #topic week in review
16:35:32 <sa2ajj> djmitche: you can make tardyp a chair...
16:36:01 <tardyp> This weeks was as usual, lots of activity in irc from new users
16:36:25 <tardyp> half a dozen PRs with small fixes and improvements.
16:36:33 <sa2ajj> which is very good, i must say
16:36:38 <tardyp> A noticable update of the tutorial tour
16:36:41 <tardyp> indeed.
16:37:06 <tardyp> This is very warming to see all those contributions from new users or returning upgrading users.
16:37:20 <sa2ajj> docs is hard and even small bits are apreciated very much
16:37:31 <tardyp> yes
16:37:58 <tardyp> So this week I have looked at gerrit 2.13, and the lost of new plugins available.
16:38:04 <djmitche> oh sorry
16:38:07 <djmitche> #chair tardyp
16:38:07 <bb-supy`> Current chairs: djmitche tardyp
16:38:17 <djmitche> this is the time of year I'm in this meeting during another
16:38:18 <sa2ajj> lost?
16:38:29 <djmitche> *lots
16:38:47 <tardyp> I just sent a PR with the support for the new gerrit 2.13 REST api CI status
16:39:03 <tardyp> which is cool as this is very similar to the other http reporter plugins
16:39:09 <sa2ajj> re the last pr for the verify-status: looks interesting, but it seems it might not be suitable for our cause :/
16:39:28 <tardyp> what do you mean?
16:39:33 <sa2ajj> sorry, not PR itself, the plugin :)
16:39:51 <tardyp> what is our cause?
16:40:00 <sa2ajj> many different labels i really need to have
16:40:11 <tardyp> ah you mean your own buildbot setup at work.
16:40:17 <sa2ajj> yup
16:40:31 <sa2ajj> the pr i welcome and i'd like to see it in
16:40:55 <tardyp> There will be another PR for a new GerritChangeSource that uses the Event-log plugin
16:41:02 <sa2ajj> if it matched our workflow, i'd be happy to contribute.
16:41:08 <tardyp> which allows to get the events from http instead of ssh
16:41:10 <sa2ajj> oh! that sounds interesting
16:41:40 <sa2ajj> [side note] we seem to experience "lost events", but it's not bb specific.
16:41:46 <sa2ajj> (so let's move on :))
16:42:05 <tardyp> event-log will help a lot with lost events, which are inevitable with ssh api
16:42:24 <tardyp> #topic week in review
16:42:37 <tardyp> #info  This weeks was as usual, lots of activity in irc from new users
16:42:42 <tardyp> #info This is very warming to see all those contributions from new users or returning upgrading users.
16:42:46 <sa2ajj> if you don't mind, i'll chat to you about these in parallel since this *is* a big issue for us (lost events)
16:42:55 <tardyp> ok
16:43:46 <sa2ajj> (and *after* this meeting is the good time)
16:43:52 <tardyp> #info support comming for new gerrit plugins helping to build a modern CI on top of gerrit
16:44:07 <tardyp> #topic botbotme
16:44:21 <tardyp> as rutsky is not there I  think we can move on
16:44:23 <sa2ajj> re modern ci: any links/refs to what enables it?
16:44:24 <tardyp> #info nothing new
16:45:22 <tardyp> well I was thinking about the CI statuses that all new git servers have
16:45:50 <bb-trac> [trac] #2853/task (assigned) updated by dustin (No response yet on https://github.com/BotBotMe/botbot-bot/issues/44) http://trac.buildbot.net/ticket/2853
16:45:54 <tardyp> which is common concept, while gerrit had this weird label concept, which is specific to gerrit, and not very scalable (imho
16:46:11 <tardyp> #topic trac2github
16:46:17 <tardyp> sa2ajj: want to update us?
16:46:23 <sa2ajj> sure
16:46:43 <sa2ajj> i'm trying to spend at least several hours a week (sorry) on finalising it
16:47:15 <sa2ajj> the latest code is available in a branch (one sec)
16:47:38 <sa2ajj> https://github.com/sa2ajj/trac2gh/tree/implement-it
16:47:55 <sa2ajj> during the testing i hit the rate limit for the github api
16:48:03 <sa2ajj> i need to accomodate it
16:48:07 <djmitche> oops :)
16:48:09 <tardyp> ah thats annoying
16:48:16 <djmitche> https://github.com/buildbot/sandbox/issues/4 was an example of a migrated issue -- looks good
16:48:34 <sa2ajj> it is indeed: i'll see *which* part of the migration hits it
16:48:39 <tardyp> maybe you could have the sandbox in another github account
16:48:45 <djmitche> sa2ajj: you might also ask github for a temporary rate-limit increase during the import?
16:48:48 <tardyp> so that we are not spammed when you do your tests
16:48:52 <bdbaddog> Can you batch api calls? usually that's how such limits are avoided..
16:48:53 <sa2ajj> the mitigations: mapping of *known* users/e-mail addresses to github nicks
16:49:04 <sa2ajj> caching the information that was obtained from github
16:49:19 <tardyp> this will also help speed..
16:49:25 <sa2ajj> bdbaddog: i need to look into it
16:49:49 <sa2ajj> so those mappings and the cache will help to get the things mapped properly
16:50:08 <sa2ajj> i *think* that version needs to be migrated as well
16:50:41 <sa2ajj> and the main "quirk" is that gh md treats new lines differently than just md :)
16:50:55 <bdbaddog> sa2ajj: looks like you can poll the current rate limit and use that to pace api calls.
16:50:58 <tardyp> #info good progress on trac2github. Some issues with gh API rate limitting. Some caching needs to be implemented
16:51:06 <sa2ajj> i'd really like to avoid unnecessary nls
16:51:17 <tardyp> agree
16:51:25 <tardyp> its already hard enough
16:51:32 <sa2ajj> bdbaddog: yes, i can, it's just makes the things a bit too complicated
16:51:51 * sa2ajj is inclined to just add "sleep something" now and then...
16:52:33 <bdbaddog> https://developer.github.com/v3/rate_limit/
16:52:36 <tardyp> ok shall we move forward?
16:52:44 * sa2ajj has nothing to add
16:52:58 <tardyp> #topic Pycon HKI
16:53:05 <tardyp> was it last week end?
16:53:11 <djmitche> two weeks ago I think
16:53:17 <djmitche> how was it?
16:53:18 <sa2ajj> it was two weeks ago
16:53:52 <sa2ajj> some presentations were good: people did not talk about Python, but what cool things they do/did w/ Python
16:54:03 <sa2ajj> one sec
16:54:28 <sa2ajj> http://fi.pycon.org/2016/#talks-data-scraping was very practical and educating
16:54:59 <tardyp> any change to talk about buildbot with attendees?
16:55:02 <tardyp> chance
16:55:09 <sa2ajj> http://fi.pycon.org/2016/#talks-twitter-discussions showed how people's interactions can reveal what is going on :)
16:55:34 <sa2ajj> (i did not submit any talks, i might next time: since i want)
16:55:50 <tardyp> #info some talks were interesting (see logs for urls)
16:56:01 <sa2ajj> http://fi.pycon.org/2016/#talks-zodb -- how you can use zodb for time travel :)
16:56:31 <sa2ajj> http://fi.pycon.org/2016/#talks-unit-testing-the-scientific-python-stack -- floating point, what problems you might encounter and how to handle them in testing
16:56:53 <sa2ajj> + some intersting things about various Pythong testing frameworks
16:57:06 <sa2ajj> less impressive:
16:57:29 <sa2ajj> http://fi.pycon.org/2016/#talks-fp-in-python -- "i want to do FP in Python like in Haskell"
16:58:09 <sa2ajj> http://fi.pycon.org/2016/#talks-attack-of-the-pythons -- "i never used PyLint or any similar tools"
16:58:14 <sa2ajj> the rest i did not attend
16:58:30 <djmitche> there's always a few duds
16:58:32 <sa2ajj> the videos should be available rsn (tm)
16:58:55 <sa2ajj> the organisation was *almost* ok :)
16:58:58 <djmitche> hah
16:59:08 <djmitche> hopefully next year I can make it to that or something similar
16:59:50 <sa2ajj> at the beginning each of the speakers presented a "pitch speach" for their talks, but the organisers did not manage to put them in *one* slideset, so it could be shown continously..
17:00:41 <sa2ajj> i skipped the social event the day before (i wanted a clear head for the talks) and the dinner on the day (i was already a bit exhausted as talks required some attention)
17:00:42 <djmitche> that's no good
17:01:01 <djmitche> those can be the best part :(
17:01:08 <djmitche> if you avoid the head-clouding stuff
17:01:13 <sa2ajj> :)
17:01:15 <djmitche> anyway, we are over on time
17:01:17 <djmitche> #topic     py3: StringIO vs BytesIO vs ??
17:01:19 <sa2ajj> i do, i do :D
17:01:45 <djmitche> I don't know the context here, but I feel like the case should usually be clear between BytesIO and StringIO?
17:01:46 <tardyp> So there are a bunch of py3 that are blocked by this
17:02:18 <tardyp> its not that simple as our current use of stringio module is not very clear
17:02:27 <sa2ajj> these need to be treaded carefully
17:02:33 <tardyp> so craig has some issues to find how to treat those
17:02:41 <sa2ajj> and these *are* important
17:02:43 <tardyp> and I am unsure how to guide him
17:02:48 <tardyp> yes
17:03:01 <tardyp> I have not much experience on py3 porting
17:03:17 <sa2ajj> the current code base, i must say, is quite py2ish
17:03:28 <tardyp> the PR is here: https://github.com/buildbot/buildbot/pull/2429
17:03:43 * sa2ajj checks it out
17:04:33 <sa2ajj> i _think_ it should be split into two: StringIO.StringIO (and companions) and *then* `str` and `bytes`
17:05:02 <sa2ajj> i'm *certain* though we need this kind of change to be integrated
17:05:08 <tardyp> I am not sure we want to decide now, but I would like that we can move this thread forward, and maybe decide next week.
17:05:57 <djmitche> ok
17:06:03 <djmitche> let's wrap up then
17:06:07 * djmitche will look at the PR too
17:06:42 <tardyp> #info if you have some experience on py3 transition, please have a look at the discussion in https://github.com/buildbot/buildbot/pull/2429 and help us to find out what is the best options to write portable stringio code.
17:07:01 * sa2ajj nods
17:07:12 <tardyp> #endmeeting