17:04:28 <djmitche> #startmeeting weekly
17:04:28 <bb-supy> Meeting started Tue Jan 29 17:04:28 2019 UTC and is due to finish in 60 minutes.  The chair is djmitche. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:04:28 <bb-supy> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:04:28 <bb-supy> The meeting name has been set to 'weekly'
17:04:32 <djmitche> #topic Introduction
17:04:41 <djmitche> this usually starts 4 mintues earlier :)
17:04:44 <djmitche> 30 minute time limit
17:04:56 <djmitche> http://bit.ly/2rup31x
17:05:02 <rjarry> hi djmitche
17:05:07 <djmitche> hi!
17:05:12 <djmitche> anyone else around?
17:05:13 <p12tic> hi o/
17:05:17 <djmitche> hey!
17:06:11 <djmitche> #topic Week in Review
17:06:16 <djmitche> tardyp: ^^ around?
17:06:43 <tardyp> Hi
17:06:49 * djmitche waves
17:06:57 <djmitche> what's new this week?
17:07:03 <rjarry> looks like more kube related fixes
17:07:47 <tardyp> right, with some regressions :(
17:07:56 <tardyp> this morning I had to hotfix them..
17:08:08 <djmitche> ugh
17:08:37 <tardyp> apart from that lots of PR with py3 cleanups
17:10:10 <djmitche> #info fixes related to running on k8s
17:10:21 <tardyp> and then there is the GSoC activities we need to talk about
17:10:22 <djmitche> #info lots of Py3 cleanups
17:10:26 <djmitche> yep, let's
17:10:31 <djmitche> #topic GSoC 2019
17:10:48 <djmitche> sounds like we've been accepted by Python as a sub-org?
17:11:01 <tardyp> not sure yet
17:11:03 <rjarry> I didn't see any reply from them
17:11:19 <djmitche> #info still in progress applying as Python sub-org
17:11:23 <djmitche> #info looking for mentors
17:11:23 <tardyp> they have an acceptance process. So we need to make sure we have a strong ideas page
17:11:47 <p12tic> i think we need one more idea
17:11:52 <djmitche> where did that ideas page get made?
17:12:05 <p12tic> what about ensuring that multimaster actually works? integration tests, etc?
17:12:07 <rjarry> p12tic: that means 4 ideas == 8 mentors
17:12:24 <rjarry> djmitche: https://github.com/buildbot/buildbot/wiki/Buildbot-GSoC-Projects-2019
17:12:25 <p12tic> does it mean that we need to accept all 4 ideas?
17:12:29 <djmitche> not necessarily - we can accept any number of participants regardless of number of ideas
17:12:41 <p12tic> yes, so for acceptance i guess more ideas is better
17:12:47 <p12tic> and then we can actually see what we can do
17:12:52 <rjarry> no, but the GSoC guys said that we needed 2 mentors per project
17:12:56 <djmitche> #info ideas page: https://github.com/buildbot/buildbot/wiki/Buildbot-GSoC-Projects-2019
17:13:06 <tardyp> we can share mentors I think
17:13:13 <djmitche> yeah
17:13:21 <tardyp> we need a main mentor for each project and then a comentor shared
17:13:22 <p12tic> iirc their main worry was that 1 mentor is single point of failure
17:13:31 <djmitche> 2 mentors per is a good idea anyway -- they can sort of split the responsibilities (one does a lot of reviwe, one handles communication, for example)
17:13:44 <rjarry> understandable
17:13:53 <rjarry> but since we only are 4
17:13:54 <rjarry> :)
17:14:13 <djmitche> yeah
17:14:29 <djmitche> anything else about GSoC?
17:15:25 <tardyp> dont hesitate to enhance the ideas page
17:15:26 <djmitche> ok!
17:15:33 <tardyp> and to add more projects if you have some
17:15:34 <djmitche> #topic Python2.7 deprecation
17:15:36 <djmitche> yeah
17:15:43 <p12tic> ^^ what about ensuring that multimaster actually works? integration tests, etc?
17:15:47 <tardyp> we better have lots of projects even if we can't mentor all
17:15:54 <p12tic> idea for idea:)
17:15:56 <djmitche> p12tic: that seems like a good project :)
17:16:23 <p12tic> ok, i'll write an entry
17:17:20 <rjarry> djmitche: python2.7 and 3.4 deprecation :)
17:17:29 <djmitche> right!
17:17:57 <rjarry> https://github.com/buildbot/buildbot/milestone/7
17:18:03 <rjarry> 2 bugs left
17:18:09 <djmitche> awesome
17:18:11 <djmitche> and that will be 2.0?
17:18:49 <rjarry> I'm not sure, tardyp do you want to wait for https://github.com/buildbot/buildbot/pull/4554 to tag 2.0.0 ?
17:19:02 <rjarry> I am still working on it
17:20:09 <djmitche> #info ..and Python-3.5 deprecation
17:20:16 <djmitche> #info results of this will be in 2.0.0
17:20:19 <rjarry> not 3.5 !!!!
17:20:33 <tardyp> not really waiting for $4554
17:20:40 <tardyp> just waiting for february
17:20:45 <rjarry> ok
17:20:47 <tardyp> I think we are good for 2.0.0 now
17:21:02 <djmitche> #info .. I mean Python-3.4!!
17:21:26 <rjarry> after 2.0.0 we go async/await ?
17:21:36 <rjarry> or we "await" a bit
17:21:40 <rjarry> (sorry)
17:21:41 <djmitche> haha
17:22:05 <p12tic> async/await is an API break when used on any API
17:22:35 <rjarry> p12tic: not if you wrap all async def functions with a proper decorator
17:22:53 <p12tic> yes, but then performance is even slower than just @inlineCallbacks
17:23:01 <djmitche> yikes
17:23:15 <rjarry> how does that even work with twisted without these decorators ?
17:23:18 <djmitche> #info can consider async/await
17:23:51 <p12tic> for perf numbers you can see this PR https://github.com/twisted/twisted/pull/1090
17:24:01 <p12tic> also https://github.com/p12tic/gist-twisted-perf-tests
17:24:12 <p12tic> Deferred is awaitable, so you can await stuff returned by twisted
17:24:19 <p12tic> the other way round is not true
17:24:26 <p12tic> either the async function needs to be decorated
17:24:42 <p12tic> or you need to call defer.ensureDeferred whenever you want a Deferred out of a async function
17:25:08 <rjarry> that's what I had understood yes
17:25:24 <rjarry> so for all our internal code we can skip the decorators
17:25:29 <p12tic> yes
17:25:33 <djmitche> do we need to talk about k8s workers too?
17:25:57 <rjarry> but do we need to keep them for all our "public" API ?
17:26:02 <p12tic> yes
17:26:07 <rjarry> sad
17:26:16 <djmitche> #topic Kubernetes Workers
17:27:37 <rjarry> I have to go
17:28:43 <djmitche> haha, ok
17:28:46 <djmitche> soeems like we're done
17:28:47 <djmitche> *seems
17:29:01 <djmitche> tardyp: anything else?
17:29:42 <tardyp> nothing else. working fine now..
17:33:28 <djmitche> #endmeeting