meeting
LOGS
14:13:35 <mvollmer> #startmeeting meeting
14:13:35 <zodbot> Meeting started Mon Dec 21 14:13:35 2015 UTC.  The chair is mvollmer. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:13:35 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:13:36 <zodbot> The meeting name has been set to 'meeting'
14:13:40 <mvollmer> .hello mvo
14:13:41 <zodbot> mvollmer: mvo 'Marius Vollmer' <marius.vollmer@gmail.com>
14:13:55 <dperpeet> .hello dperpeet
14:13:56 <zodbot> dperpeet: dperpeet 'None' <dperpeet@redhat.com>
14:13:58 <mvollmer> #topic Agenda
14:14:10 <mvollmer> * Debian
14:14:50 <dperpeet> * avocado tests
14:16:26 <stefw> * splitting out test suites
14:17:21 <mvollmer> okay?
14:17:32 <mvollmer> #topic Debian
14:17:43 <mvollmer> Once again, I think I am done. :)
14:18:01 <mvollmer> Now we use virt-builder to upgrade to unstable
14:18:15 <mvollmer> and then boot with the new kernel for the setup script
14:18:28 <mvollmer> I think that's a good pattern for all our images
14:18:45 <mvollmer> virt-builder is quite powerful, it probably can do the whole setup script as well
14:18:59 <dperpeet> mvollmer, yes, it can do a lot
14:19:06 <dperpeet> vm-create is already much nicer
14:19:12 <mvollmer> but I am not sure if that would be a improvement
14:19:51 <mvollmer> anyway, I think it is ready to merge
14:19:54 <dperpeet> I don't think we should squeeze everything into virt-builder
14:20:09 <stefw> mvollmer, it's merged
14:20:11 <petervo> didn't it get merged?
14:20:22 <dperpeet> I was about to ask :)
14:20:25 <stefw> unlike some of my pull requests ...
14:21:09 <stefw> we need more test machines
14:21:16 <stefw> now that we have debian by default
14:21:18 <stefw> which is a good thing to have
14:21:26 <dperpeet> yeah, they're lagging behind
14:21:30 <mvollmer> oh!
14:21:48 <mvollmer> thanks for merging! :-)
14:22:13 <mvollmer> now we need to figure out the next steps
14:22:25 <mvollmer> let's talk to mbiebl etc
14:22:47 <mvollmer> we should work to reduce the number of skipped tests
14:23:17 <mvollmer> okay, next topic?
14:23:24 <dperpeet> great work, mvollmer!
14:23:28 <mvollmer> thanks
14:23:30 <mvollmer> took a while
14:24:20 <mvollmer> #topic avocado tests
14:24:42 <dperpeet> well, they finally got merged
14:24:48 <mvollmer> nice!
14:24:53 <dperpeet> next up is splitting them out into their own test suite
14:25:03 <mvollmer> (they are skipped on debian, does that make sense?)
14:25:09 <stefw> yup
14:25:16 <dperpeet> as part of our effort to split the behemoth our integration tests have become into bite-sized chunks
14:25:17 <stefw> they only need to run (by default) on one operating system
14:25:27 <mvollmer> okay
14:25:41 <dperpeet> especially if they increase in size/complexity
14:25:54 <dperpeet> which jscotka will probably work on
14:26:01 <stefw> right
14:26:04 <dperpeet> stefw has started the split in #3380
14:26:23 <stefw> which is blocking on #3335
14:26:44 <dperpeet> which is waiting for a test
14:26:51 <dperpeet> and currently in my hands to push soon
14:26:56 <dperpeet> end of topic :)
14:27:14 <mvollmer> #topic splitting out test suites
14:27:25 <stefw> This is related to #3380
14:27:35 <stefw> and what dperpeet just talked about
14:27:55 <stefw> these pull requests allow us to hav emore than one test suite scheduled on the verify machines
14:28:11 <stefw> and allow things like scheduling koji builds against github, etc.
14:28:43 <stefw> #3335 needs to be merged before we can even test #3380
14:29:05 <github> [cockpit] mvollmer opened pull request #3384: Handle unsupported domains (master...realms-handle-unsupported) http://git.io/vEq2c
14:29:49 <stefw> eot
14:30:34 <mvollmer> okay
14:30:43 <mvollmer> #topic any other business
14:31:29 <mvollmer> i guess: no weekly meeting until the next year, right?
14:31:38 <stefw> makes sense
14:31:44 <stefw> although i hope we can do one more release
14:32:08 <stefw> i was really hoping the fixes to make things easier for people hacking on cockpit could go in
14:32:13 <stefw> before everyone gets their time off
14:32:21 <stefw> because we may have people contributing during that time
14:32:51 <stefw> but obviously blocking on review etc.
14:32:56 <mvollmer> culd you mark them as priority?
14:33:00 <stefw> we may not end up having time, i guess
14:33:02 <mvollmer> *could
14:33:31 <mvollmer> I have some time still today and tomorrow, at least
14:33:50 <github> [cockpit] stefwalter closed pull request #3335: Split check-verify into github-task (master...split-verify) http://git.io/v0VzN
14:34:39 <stefw> marked them priority
14:34:45 <dperpeet> can you assign the pull request to yourself if you start reviewing?
14:34:51 <dperpeet> that way we don't overlap
14:34:55 <mvollmer> dperpeet, yes
14:35:00 <stefw> the problem
14:35:07 <stefw> is that someone goes away for days
14:35:10 <stefw> and then it's still assigned to them
14:35:14 <stefw> and everyone else backs away
14:35:20 <dperpeet> well, you can unassign after you've marked it
14:35:23 <dperpeet> needswork
14:35:29 <dperpeet> or merged it, then it doesn't matter
14:35:39 <dperpeet> just use it as a mutex
14:35:50 <dperpeet> spin lock :)
14:35:53 <mvollmer> stefw, good point, nevertheless
14:38:05 <mvollmer> one little thing
14:38:10 <mvollmer> I made a new label, "patch"
14:38:20 <stefw> i was wondering about that
14:38:26 <mvollmer> I put that on issues that I have made a PR for
14:38:39 <stefw> is this because we're so slow at reviewing and merging?
14:38:43 <mvollmer> so that I can quickly ignore them when checking what needs doing
14:39:04 <stefw> yeah, we used to have a patch label
14:39:09 <mvollmer> ahh
14:39:12 <stefw> if it works for you, sounds good
14:39:27 <mvollmer> yes, it's only for me
14:39:55 <stefw> but yes, it's nice to see some of the older bugs being solved
14:40:04 <stefw> mvollmer and petervo have been knocking away at them
14:40:09 <stefw> so whatever helps with tracking those lists is great, i think
14:40:21 <github> [cockpit] stefwalter closed pull request #3376: Update Vagrantfile to new COPR location (master...vagrantfile-copr) http://git.io/vEend
14:40:30 <mvollmer> i have been neglecting my backlog much too long
14:40:46 <mvollmer> i guess that's one downside of continous delivery
14:41:18 <mvollmer> one doesn't get into the "let's kill all our bugs before the release" mode
14:41:40 <stefw> yeah, true
14:42:42 <mvollmer> okay, done?
14:42:57 <stefw> ++
14:43:02 <mvollmer> #endmeeting