14:02:20 <mvollmer> #startmeeting meeting
14:02:20 <zodbot> Meeting started Mon Jan  4 14:02:20 2016 UTC.  The chair is mvollmer. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:02:20 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:02:20 <zodbot> The meeting name has been set to 'meeting'
14:02:25 <mvollmer> .hello mvo
14:02:26 <zodbot> mvollmer: mvo 'Marius Vollmer' <marius.vollmer@gmail.com>
14:02:29 <stefw> .hello stefw
14:02:33 <zodbot> stefw: stefw 'Stef Walter' <stefw@redhat.com>
14:02:34 <dperpeet> .hello dperpeet
14:02:36 <zodbot> dperpeet: dperpeet 'None' <dperpeet@redhat.com>
14:03:07 <mvollmer> do we have andreas?
14:03:23 <dperpeet> doesn't look like it
14:04:06 <mvollmer> #topic Agenda
14:04:09 <dperpeet> * debian packages
14:04:15 <stefw> * CSP fixes
14:04:32 <stefw> * Test failures
14:06:15 <mvollmer> okay
14:06:22 <mvollmer> #topic Debian packages
14:06:36 <mvollmer> I asked our Debian contacts for help
14:07:02 <stefw> Great
14:07:13 <stefw> So the goal for us upstream is continuous delivery of each weekly release
14:07:17 <cockpitbot> 2 tests failed - http://fedorapeople.org/groups/cockpit/logs/master-0dfcac2d-fedora-22/log.html
14:07:22 <stefw> into a debian repo where people can try it out
14:07:27 <stefw> that *may* be debian experimental
14:07:43 <dperpeet> but any publicly accessible repo should do
14:07:48 <stefw> but if debian experimental cannot be automated, then it's likely a different package repo where we would push to
14:08:06 <mvollmer> we can either push sources or binaries
14:08:12 <stefw> likely both?
14:08:20 <stefw> on fedora we push both, no?
14:08:21 <dperpeet> do we package the source for debian?
14:08:28 <dperpeet> or just tarball the tree
14:08:31 <mvollmer> yeah, either sources only, or sources+binaries
14:08:42 <stefw> what's the advantage to sources only?
14:08:46 <mvollmer> sources only is easier for the release container, I think.
14:08:51 <stefw> true
14:08:54 <stefw> we could start with that
14:09:16 <mvollmer> but then we need a builder at the other end
14:09:28 <mvollmer> which could be our verify container, slightly modified
14:09:35 <stefw> openSUSE build service?
14:09:39 <stefw> Ubuntu PPA
14:09:49 <stefw> we could use any of those builders if we publish source packages from our Continuous delivery
14:09:51 <mvollmer> yep, or something that exists already
14:10:07 <stefw> but using our verify container isn't bad either
14:10:17 <stefw> is there a single command to build a debian package, given the sources?
14:10:34 <stefw> sorta like the rpmbuild rebuild stuff?
14:10:54 <mvollmer> ther is dpkg-buildpackage
14:11:05 <mvollmer> but that doesn't set up the build environment
14:11:22 <dperpeet> debuild
14:11:35 <dperpeet> apt-get has build-dep
14:11:49 <stefw> i guess my question is can we expect end users to try out a source package
14:11:53 <mvollmer> the mock equivalent would be pbuilder
14:12:00 <stefw> early adopters, etc
14:12:13 <mvollmer> I guess we could, actually.
14:12:20 <stefw> or put another way:
14:12:28 <stefw> what should our continuous delivery deliver for debian
14:12:31 <dperpeet> well, jpsutton specifically asked for the source packages today
14:12:41 <dperpeet> to build for ubuntu
14:12:49 <stefw> cool
14:12:54 <stefw> so i agree with mvollmer
14:12:59 <stefw> what if source packages in a repo is our next step?
14:13:10 <dperpeet> the goal should be both
14:13:13 <dperpeet> but we can do source first
14:13:19 <stefw> yup
14:13:29 <mvollmer> good idea, didn't think of that at all
14:13:42 <dperpeet> jpsutton can use what we have for debian and make it work on ubuntu
14:13:45 <stefw> we can then say we need a maintainer to get the source packages built into debian experimental
14:14:02 <stefw> and if that's not automatable, then it'll just be a routine job
14:14:08 <stefw> for that maintainer
14:14:09 <dperpeet> everyone I've spoken to agrees that having it work on ubuntu is a good step towards getting into debian
14:14:26 <mvollmer> should I work on the source repo?
14:14:30 <stefw> i think so
14:14:36 <mvollmer> okay
14:14:38 <dperpeet> mvollmer, maybe we should talk to jpsutton
14:14:45 <dperpeet> to see what exactly he needs
14:14:52 <dperpeet> and what makes sense to release
14:15:06 <mvollmer> sure
14:15:19 <dperpeet> do you want to reply to his e-mail on the list?
14:16:04 <dperpeet> otherwise I can do that and send a link to what we've done for debian
14:16:05 <mvollmer> hmm, did I miss the email?
14:16:10 <mvollmer> let me check
14:17:11 <mvollmer> okay, got it
14:17:14 <mvollmer> I'll reply
14:17:29 <dperpeet> thanks!
14:18:11 <github> [cockpit] stefwalter opened pull request #3410: Fix default test operating system "fedora-23" (master...fix-23-default) http://git.io/vuZ2i
14:18:39 <mvollmer> #action mvo talk to jpsutton about a 'official' Ubuntu PPA for Cockpit
14:20:08 <stefw> and we agreed to do the source repo right?
14:20:25 <mvollmer> yes
14:20:54 <stefw> #action mvo will add logic to the continous release tooling to create a source repo
14:21:02 <stefw> likely on files.cockpit-project.org ... i would imagine
14:21:05 <mvollmer> #action mvo let the release scripts maintain a repo with Debian sources
14:21:11 <stefw> or fedorapeople.org
14:21:12 <stefw> ha ha
14:21:27 <mvollmer> oh, now I have to do it twice
14:21:36 <stefw> the second time should be easier
14:21:41 <mvollmer> :)
14:21:50 <dperpeet> but the bar will be higher regarding code design
14:22:14 <mvollmer> maybe I do them the other way around, hmm.
14:22:24 <mvollmer> okay, next?
14:23:15 <mvollmer> #topic CSP fixes
14:23:33 <stefw> The CSP stuff included in 0.89 was broken in 3 different ways
14:23:37 <stefw> i'm surprised cockpit works at all
14:23:40 <stefw> especially on chrome
14:23:50 <stefw> the CSP syntax was an old syntax
14:24:04 <stefw> in addition, the websocket gets blocked from being created
14:24:15 <stefw> but apparently inspite of all that, it still "works" on Firefox
14:24:19 <stefw> and many versions of Chrome
14:24:29 <stefw> so much for security
14:24:30 <mvollmer> yeah, works fine here...
14:24:37 <stefw> anyway, just a heads up
14:24:43 <stefw> there's a pull request
14:24:48 <dperpeet> I couldn't reproduce the failure
14:24:52 <petervo> i'll review
14:24:54 <stefw> since it doesn't seem *that* broken ... i won't treat 0.89 as broken
14:24:56 <dperpeet> with my versions of firefox and chrome
14:25:14 <stefw> but if anyone runs into these issues
14:25:17 <stefw> that's probably the cause
14:25:22 <mvollmer> okay
14:25:30 <stefw> #info https://github.com/cockpit-project/cockpit/pull/3408
14:25:31 <dperpeet> petervo, I only added needswork because of the typos - I didn't properly review
14:25:47 <petervo> ok
14:26:42 <stefw> that's it on that topic
14:27:02 <mvollmer> #topic Test failures
14:27:13 <stefw> Everything seems to be red on pull requests
14:27:21 <stefw> Can we split up the work of getting it green again?
14:27:24 <stefw> i'll work on the IPA problems
14:27:39 * mvollmer hadn't really noticed yet.
14:29:32 <stefw> petervo, the rpm-ostree crash seems to be happening regularly
14:29:47 <petervo> did the new images get merged?
14:29:58 <stefw> no idea
14:30:06 <stefw> mvollmer, blank pages on every second run of debian
14:30:06 <stefw> https://fedorapeople.org/groups/cockpit/logs/pull-3408-37c385fe-debian-unstable/TestPages-testBasic-FAIL.png
14:30:08 <petervo> https://github.com/cockpit-project/cockpit/pull/3382
14:30:35 <stefw> so i guess that's waiting on mvollmer
14:31:02 <mvollmer> okay
14:31:07 <stefw> or who created those images?
14:31:16 <stefw> unfortunately one gets to debug the issues when one creates images
14:31:26 <dperpeet> peter's images got merged I think
14:31:27 * stefw is worrying about creating a new ipa image
14:31:31 <petervo> i created a new image
14:31:40 <petervo> but it's included in that PR
14:31:42 <dperpeet> ah ok
14:32:07 <petervo> i guess i can create another one
14:32:47 <petervo> without the network changes, if that PR is blocked
14:32:58 <stefw> i think the PR is just blocking on image related test failures
14:34:06 <dperpeet> mvollmer, will you create a new image for this?
14:34:23 <mvollmer> hmm, now I can't follow
14:34:36 <petervo> i can create another one, but i'm not sure that docker problem will be fixed by a new image
14:34:48 <dperpeet> I have no idea how old the image in #3382 is
14:34:49 <mvollmer> i agree
14:35:02 <dperpeet> it might "overwrite" a newer one
14:35:23 <mvollmer> that would be a merge conflict, no?
14:35:37 <dperpeet> ok, I checked
14:35:41 <petervo> dperpeet, it won't
14:35:43 <dperpeet> it wouldn't overwrite a newer one
14:36:15 <dperpeet> mvollmer, something with the images can quickly disappear during a rebase
14:36:30 <dperpeet> but no matter, this is a straightforward update
14:37:06 <mvollmer> in any case, I'll try to unblock that PR, of course
14:39:09 <mvollmer> so, action point for me is to check the debian failures, right?
14:39:19 <stefw> if you could, that would be the missing part
14:39:26 <mvollmer> okay
14:39:27 <stefw> everything else seems to be tracked down
14:39:32 <stefw> and just waiting on fixes
14:39:42 * stefw has tracked down the IPA problems to an expired password
14:39:44 <stefw> in the IPA image
14:39:46 * stefw regenerates
14:39:56 <mvollmer> ahh
14:40:06 <mvollmer> is ipa now based on f23?
14:40:17 * stefw ain't got time for that
14:40:33 <stefw> we're not testing ipa
14:40:40 <mvollmer> no
14:40:42 <stefw> so i figure leaving it on fedora-22 which has been working, is ok
14:40:50 <mvollmer> sure, just curious
14:40:51 <dperpeet> stefw, you can fix the image and upload that without creating an entirely new one
14:41:17 <stefw> dperpeet, i'll do that if it doesn't work to create a new one
14:41:22 <dperpeet> ok
14:41:36 <mvollmer> stefw, maybe we can stop the passwrod from expiring
14:41:44 <cockpitbot> 5 tests failed - http://fedorapeople.org/groups/cockpit/logs/master-0dfcac2d-fedora-testing/log.html
14:41:46 <mvollmer> or reset it from the tests
14:42:24 <stefw> maybe
14:43:10 <stefw> will see if i have time
14:43:11 <petervo> mvollmer, i think the blank pages might be my bug
14:43:30 <petervo> i don't think it's debian specific
14:43:41 <mvollmer> okay, tell us more. :)
14:43:57 <petervo> since i messed with index.js
14:44:16 <petervo> i can look into it
14:44:23 <mvollmer> okay, thanks!
14:45:22 <mvollmer> #topic other business
14:45:49 <stefw> Talk accepted for FOSDEM main room
14:46:00 <stefw> which is pretty exciting and  scary
14:46:07 <stefw> i have my work cut out for me
14:46:14 <mvollmer> wow!
14:46:23 <github> [cockpit] stefwalter closed pull request #3410: Fix default test operating system "fedora-23" (master...fix-23-default) http://git.io/vuZ2i
14:46:27 <stefw> https://fosdem.org/2016/schedule/track/enterprise/
14:46:45 <dperpeet> awesome!
14:46:56 <stefw> And the day before, Friday, i'll be talking at the CentOS Dojo at IBM in Brussels
14:47:00 <stefw> a little more low key
14:47:10 <stefw> but hopefully lots of sysadmins and nice folk
14:47:55 <mvollmer> good goo
14:47:57 <mvollmer> d
14:48:01 <dperpeet> there's a duplicate sentence on the details page: https://fosdem.org/2016/schedule/event/cockpit/
14:48:20 <stefw> strange
14:48:29 * stefw looks at my submission
14:48:51 <stefw> Ah, because it's in the abstract and in the description
14:48:56 <stefw> i guess i'll remove it from one of them
14:49:34 <stefw> k, i changed it
14:49:41 <stefw> probably will take a while to update
14:50:08 <dperpeet> do you need help in preparing?
14:50:20 <dperpeet> no, would you like help in preparing? =)
14:50:30 <stefw> yes
14:50:42 <stefw> i need to spend time to come up with a plan
14:50:59 <stefw> and so for starters, i probably need to focus less on keeping things like the test system running
14:51:13 <stefw> but will know more, if there's stuff we can work on together for the talks
14:51:18 <stefw> will know more soon
14:54:14 <mvollmer> okay, are we done?
14:55:01 <dperpeet> I believe so
14:55:21 <stefw> yup
14:55:28 <mvollmer> #endmeeting