14:03:49 <andreasn> #startmeeting Cockpit weekly IRC meeting 2016-01-25
14:03:49 <zodbot> Meeting started Mon Jan 25 14:03:49 2016 UTC.  The chair is andreasn. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:03:49 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:03:49 <zodbot> The meeting name has been set to 'cockpit_weekly_irc_meeting_2016-01-25'
14:03:55 <stefw> .hello stefw
14:03:55 <andreasn> .hello andreasn
14:03:56 <zodbot> stefw: stefw 'Stef Walter' <stefw@redhat.com>
14:03:59 <zodbot> andreasn: andreasn 'Andreas Nilsson' <anilsson@redhat.com>
14:04:22 <mvollmer> .hello mvo
14:04:23 <zodbot> mvollmer: mvo 'Marius Vollmer' <marius.vollmer@gmail.com>
14:05:33 <andreasn> #topic agenda
14:06:27 <andreasn> * tuned maybe
14:06:39 <mvollmer> * image refreshes
14:06:52 <stefw> Google Summer of Code
14:07:49 <andreasn> ok, lets go
14:07:53 <andreasn> #topic tuned
14:08:01 <andreasn> I saw this just got in
14:08:04 <andreasn> excellent!
14:08:12 <stefw> i'll be making a video of it
14:08:18 <mvollmer> okay!
14:08:24 <stefw> because i need to for the talks anyway
14:08:57 <mvollmer> the tuned people are working on improvements, right?
14:09:12 <mvollmer> localized descriptions for the profiles?
14:09:20 <mvollmer> better d-bus api?
14:09:28 <andreasn> yeah, apparently it's in fedora-testing https://github.com/cockpit-project/cockpit/pull/3279#issuecomment-169048948
14:09:29 <mvollmer> I am not really "in the loop".
14:09:32 <andreasn> or was 20 days ago
14:09:58 <stefw> one thing that's really missing is the polkit stuff
14:10:06 <stefw> it should be possible for anyone to query the profile in use, not just root
14:10:09 <andreasn> tuned-2.6.0 was the version
14:10:32 <mvollmer> alright
14:11:22 <mvollmer> auto activation and properties would be nice
14:11:57 <mvollmer> I can write a detailed proposal.
14:11:57 <andreasn> should we open bugs about that to keep track of it?
14:12:02 <andreasn> sure
14:12:14 <mvollmer> where should I send that text?
14:12:27 <mvollmer> I can figure it out
14:12:32 <stefw> bugzilla, i think
14:12:55 <mvollmer> #action mvollmer to write detailed d-bus enhancement proposal for tuned
14:13:44 <andreasn> sounds good
14:13:50 <andreasn> all, right, next topic?
14:14:27 <stefw> mvollmer, i think we should differentiate between nice to haves and musthaves
14:14:33 <stefw> for example non-root access, and activation
14:14:36 <stefw> vs. properties and API changes
14:15:09 <mvollmer> yes
14:15:19 <jskarvad> hi, tuned developer here ;)
14:15:24 <mvollmer> let's review what I'll write
14:15:31 <mvollmer> jskarvad, hi!
14:15:45 <jskarvad> I did some API extension to 2.6.0
14:15:51 <jskarvad> could extend it more
14:15:54 <stefw> polkit support by any chance?
14:16:12 <stefw> initial goal: non-root user should be able to retrieve the current profile
14:16:26 <jskarvad> regarding non root acess (polkit) I will address it in 2.6.1, hopefully this week (sorry I was busy with other tasks)
14:16:33 <stefw> nice, that's great news
14:16:52 <stefw> i would like us to be able to eventually display the main serverc summary screen without escalating privileges
14:16:59 <stefw> i think there is another blocker, but i need to do research on that
14:17:13 <stefw> so having tuned readable without root gets us in that direction
14:17:14 <jskarvad> ok, np, feel free to contact me directly
14:17:16 <andreasn> jskarvad: do you need karma for the update in updates-testing in F23?
14:17:31 <andreasn> so it's not stuck in updates-testing or so
14:17:44 <jskarvad> andreasn: that would be great
14:17:56 <andreasn> do you have a url handy?
14:17:59 <jskarvad> andreasn: I probably missed "updates in testing....." mail
14:18:02 <jskarvad> sec
14:18:14 <jskarvad> or I can push it to stable
14:18:22 <jskarvad> checking
14:18:50 <github> [cockpit] stefwalter closed pull request #3556: test: Clean up test image dependencies (master...avocado_deps) https://git.io/vz6oy
14:18:51 <jskarvad> yeh, bodhi fault, it didn't send me the notification mail
14:19:00 <jskarvad> sorry about that, I am maintaining too much packages :)
14:19:09 <jskarvad> 2.6.0 pushed to stable now
14:19:14 <aweiteka> stefw, FYI, for the registry interface I have an AD server integrated and cloud storage backend working.
14:19:37 <andreasn> jskarvad: nice!
14:19:38 <aweiteka> it appears to "just work"
14:19:49 <stefw> andreasn, should we move on to the next meeting topic?
14:19:56 <andreasn> yup
14:20:05 <andreasn> #topic image refreshes
14:20:06 <mvollmer> okay, so I'll update tuned to use the new API if available
14:20:16 <mvollmer> I mean "update cockpit tuned support"
14:20:16 <jskarvad> mvollmer: thanks
14:20:45 <mvollmer> jskarvad, thanks for implementing this!
14:20:56 <jskarvad> mvollmer: no problem
14:21:22 <mvollmer> #action mvollmer to update tuned support to use name/description API if available
14:23:14 <mvollmer> done with the topic, right?
14:23:27 <andreasn> yeah, now we're on image refreshes
14:23:41 <mvollmer> upps, missed that, sorry.
14:23:43 <mvollmer> okay
14:23:51 <mvollmer> so I have resumed working on that
14:24:04 <mvollmer> with a new plan, full of good ideas by stef
14:24:27 <mvollmer> the best thing would be to improve the sink as well, and why not
14:24:53 <mvollmer> so I am moving code around a bit
14:25:13 <mvollmer> instead of getting the feature working with the least amoount of changes
14:25:38 <mvollmer> there is a new label, "bot".
14:25:45 <mvollmer> you can ignore all issues with that label
14:26:13 <mvollmer> github-task will create them when working on new images
14:27:16 <petervo> how is it deciding when to create new images?
14:27:25 <mvollmer> once the image is done, the issue will turn into a pull request
14:27:37 <mvollmer> it scans all issues with the "bot" label
14:27:57 <mvollmer> and includes a task for a new image when the issues are too old
14:28:39 <mvollmer> nothing more fancy than that
14:29:07 <mvollmer> it wont retry automatically if a image couldn't be created, or its verification failed.
14:29:24 <mvollmer> initially, that is :)
14:29:44 <mvollmer> the sink will create the issues, actually, and add a comment to them with a link to the log.
14:30:28 <andreasn> next up?
14:31:00 <mvollmer> yep
14:31:12 <andreasn> #topic Google Summer of Code
14:31:50 <stefw> i was asked whether we're going a summer of code thing this year
14:31:53 <stefw> by a prospective student
14:32:02 <stefw> i think we did one in 2015 in the Fedora project
14:32:06 <stefw> is that something we're interested in doing again?
14:32:19 <stefw> do we have any requirements on the tasks or applicants?
14:33:26 <andreasn> the task should be solvable without fixing up 3-4 different subsystems
14:33:32 <andreasn> ideally
14:33:35 <petervo> i think we should participate, esp if someone is already interested
14:34:03 <andreasn> so maybe an interface to something that has dbus support already
14:34:10 <stefw> true, or maybe docker work
14:34:20 <stefw> rolekit could be one obviously
14:35:10 <andreasn> firewall?
14:35:28 <mvollmer> maybe the student has an idea/interest?
14:36:42 <andreasn> is there a potential student already? or am I mixing it up with something else in my head?
14:36:51 <andreasn> I recall seeing something
14:36:52 <stefw> we have a potential student
14:37:04 <stefw> i suggested to him that he start to contribute small patches to cockpit
14:37:11 <andreasn> yeah, that's good
14:37:14 <stefw> maybe 5 to 10 before we sign up for a GSoC
14:38:13 <andreasn> right
14:38:35 <stefw> does anyone have any thoughts on how we can do better than last year?
14:38:42 <stefw> it seemed that the students both failed, right?
14:39:03 <andreasn> yep
14:39:34 <andreasn> making sure the student is active and motivated is a good start
14:40:02 <mvollmer> why did they fail?  did they just disappear?
14:40:17 <andreasn> failed on delivery in time I think
14:40:20 <petervo> making sure the expectations are clear up front so they don't try to do too much at once
14:41:05 <stefw> failed to show up as well
14:41:22 <andreasn> so breaking it up in clear subtasks and have clear time estimates for all those subtasks?
14:41:35 <andreasn> "by week 3 this part needs to be done"
14:41:43 <mvollmer> well
14:41:52 <mvollmer> is that how we work? :-)
14:42:36 <andreasn> it's more sprinty, not sure if it would produce a better result or not
14:43:17 <mvollmer> right, if it's a sprint plan, like "next goal is to do ..."
14:43:28 <mvollmer> not a deadline, "meet this or else"
14:43:51 <mvollmer> I'd say let's just try again
14:43:56 <petervo> the main thing is the see regular progress from them
14:44:09 <stefw> right
14:44:22 <petervo> to*
14:44:39 <stefw> i would say communication trumps genius
14:44:41 <mvollmer> i was a mentor once, many years ago, and my conclusion was "not again". :-/
14:44:58 <mvollmer> have to search my memories to remember why
14:45:55 <mvollmer> or I can just be one again. :)
14:46:31 <mvollmer> in any case it's cool that it's still happening.  i think it was 2005 for me
14:47:11 <andreasn> it can be challenging with some students, but also really good when it just works
14:47:38 <andreasn> I'm helping mentoring some gnome interns right now, and both of them are really good
14:49:02 <andreasn> so I'm up for helping with a gsoc project I think
14:49:06 <mvollmer> it wasn't about the student for me, that was great, just the process around it.  it felt weird.
14:50:02 <sgallagh> Sorry, I meant to chime in here and my system crashed
14:50:19 <sgallagh> So, Fedora has tightened its GSoC policy this year, because we had so many bad apples last year.
14:50:31 <mvollmer> anyway, don't listen to me here, of course
14:50:40 <sgallagh> (People who tried to basically collect the cash without delivering work)
14:50:42 <mvollmer> I am happy to mentor.
14:51:02 <cockpitbot> Release complete: 0.93 - http://fedorapeople.org/groups/cockpit/logs/release-0.93/
14:51:05 <sgallagh> So this year, we will only accept students who have a strong pre-existing relationship with the project they are applying for.
14:51:25 <sgallagh> It's up to the project to decide what that means, but a good rule of thumb is that they have to have contributed something non-trivial before
14:51:26 <mvollmer> i don't feel like administering google's money
14:51:32 <stefw> sgallagh, ++
14:51:32 <mvollmer> i guess that's the part.
14:52:06 <andreasn> sgallagh: so more than just a spelling fix, but more like something where they had to put a bit more effort into it?
14:52:34 <sgallagh> andreasn: Yeah, something on the order of a new feature or significant bugfix.
14:52:54 <andreasn> I see
14:53:38 <sgallagh> andreasn: Unfortunately, GSoC has grown a culture of people who sign up, do a half-assed job for the first half, collect half of the
14:53:44 <sgallagh> money and then disappear completely.
14:54:20 <sgallagh> So to prevent that, we want to make sure that anyone who gets accepted already has a proven history of doing actual work
14:54:38 <sgallagh> Otherwise we're wasting Google's money and the mentor's time
14:54:41 <andreasn> yeah
14:56:06 <sgallagh> EOF
14:56:48 <andreasn> anything else on that subject?
14:57:05 <stefw> i just wanted to get the discussion started, i don't think we have to take any action yet
14:57:18 <andreasn> I guess we can gather up some project ideas on an etherpad or in the wiki
14:57:20 <sgallagh> stefw: Well, probably a good idea to reply to that student.
14:57:23 <stefw> except for highly recommend that would be students become contributors
14:57:29 <stefw> sgallagh, i already did, and CC'd you
14:57:32 <sgallagh> See if they're willing to pick up a Trello ticket and run through it
14:58:37 <sgallagh> OK, I think I remember seeing that.
14:58:41 <sgallagh> /me drowns in email
15:00:15 <andreasn> ok, open floor next, since we're up at the hour
15:00:19 <andreasn> #topic open floor
15:01:21 <stefw> we only have 9 pull requests open
15:01:22 <stefw> yay :)
15:01:45 <andreasn> nice!
15:02:22 <stefw> do we need to announce the hackfest at devconf, or our other talks?
15:02:30 <stefw> i think we did a bit, but may be worth doing more?
15:03:00 <andreasn> I posted about your fosdem talk last week (I think it was last week)
15:03:24 <stefw> i tweeted
15:03:41 <stefw> but i think we should probably all do a bit of promotion for the hackfest s ince it's so early
15:03:47 <stefw> the devconf guys gave us a 9:00 AM slot after a party
15:04:08 <mvollmer> i can close two more PRs...
15:04:36 <github> [cockpit] mvollmer closed pull request #3470: test: Support for image creation tasks (master...test-image-refresh) https://git.io/vzim1
15:04:40 <andreasn> what do we do with this one? https://github.com/cockpit-project/cockpit/pull/2062
15:04:49 <cockpitbot> 1 tests failed - http://fedorapeople.org/groups/cockpit/logs/master-ac7869c2-verify-debian-unstable/log.html
15:05:01 <github> [cockpit] mvollmer closed pull request #3474: test: Support for creating image refreshment pull requests (master...test-github-image-refresh) https://git.io/vzimS
15:05:19 <andreasn> I'll post on twitter and G+ about the hackfest at devconf
15:07:33 <cockpitbot> 1 tests failed - http://fedorapeople.org/groups/cockpit/logs/master-ac7869c2-verify-fedora-atomic/log.html
15:10:47 <andreasn> ok, I guess that's it
15:10:49 <andreasn> #endmeeting