14:01:34 <tflink> #startmeeting fedora-qadevel
14:01:34 <zodbot> Meeting started Mon Jun 26 14:01:34 2017 UTC.  The chair is tflink. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:34 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:01:34 <zodbot> The meeting name has been set to 'fedora-qadevel'
14:01:34 <tflink> #topic Roll Call
14:01:44 * mkrizek is here
14:01:50 * lbrabec is here
14:01:51 <jskladan> yay
14:01:53 <tflink> #chair kparal mkrizek lbrabec jskladan
14:01:53 <zodbot> Current chairs: jskladan kparal lbrabec mkrizek tflink
14:02:04 * kparal is here
14:02:23 * roshi is here
14:02:39 <tflink> #chair roshi
14:02:39 <zodbot> Current chairs: jskladan kparal lbrabec mkrizek roshi tflink
14:03:26 <tflink> I have a sneaking suspicion that I forgot to do something for the meeting today but I suspect I'll find that out soon :)
14:03:38 <tflink> let's get-a-started!
14:03:48 <tflink> #topic Announcements and Information
14:04:26 <tflink> I see nothing on the wiki page, does anyone have something to add?
14:04:33 <tflink> taskotron-dev being re-deployed?
14:04:46 <kparal> #info the switch from depcheck to rpmdeplint has been announced. if no issues, I'll obsolete depcheck repo and projects in a week or so
14:05:45 <kparal> #info upgradepath no longer reports bodhi_update type results
14:05:55 <tflink> what type does it use?
14:06:01 <kparal> doh, that's actually not yet deployed to production, sorry
14:06:04 <kparal> #undo
14:06:04 <zodbot> Removing item from minutes: INFO by kparal at 14:05:45 : upgradepath no longer reports bodhi_update type results
14:06:04 * tflink forgot about or missed that ticket
14:06:18 <kparal> tflink: it reports results per koji build
14:06:26 <kparal> and bodhi collates that
14:06:55 <kparal> I need to test it in dev first, I got ahead of me
14:07:12 <tflink> ok
14:07:28 <tflink> anything else?
14:07:33 <kparal> #info task-modularity-testing-framework will get a new "modulelint" subtask this week
14:07:51 <kparal> that's all from me
14:07:52 <tflink> cool
14:08:17 <tflink> ok, moving on
14:08:24 <tflink> #topic Standard Interface Compatibility
14:08:40 <tflink> Where are we on this?
14:08:52 <tflink> IIRC, the review was merged into a new branch
14:08:56 <mkrizek> yep
14:08:58 <kparal> tflink: are you porting all the directives?
14:09:04 <tflink> all of them, no
14:09:22 <tflink> I have the ones ported which made sense to me to actually port, yes
14:09:30 <kparal> what about the rest?
14:09:44 <tflink> they don't make sense to me to port
14:09:53 <tflink> don't/didn't
14:10:00 <kparal> are all of them unused? or how are we going to use them?
14:10:18 * tflink attempts to clear the dust out of his brain
14:10:52 <tflink> stuff like the python or shell directive don't make sense with ansible
14:11:53 * tflink had forgotten about this task somehow
14:11:57 <kparal> some tasks will need to be re-adjusted for that
14:12:03 <kparal> just stating
14:12:06 <tflink> agreed
14:12:19 <kparal> tflink: does it make sense to merge your code to the ansiblize branch?
14:12:38 <kparal> so that we have a single place from which to test it all
14:12:40 <tflink> I just remember going through the directives and realizing that several of them weren't needed anymore
14:12:57 <tflink> kparal: depends on when we want to start splitting libtaskotron into 3
14:13:19 <kparal> it might be easier to bundle all together now and split later
14:13:31 <kparal> let's not do too many changes at once?
14:13:33 <tflink> fair enough
14:13:43 <roshi> splitting as we swap to ansible makes sense to me
14:13:57 <jskladan> +1
14:13:59 <roshi> I mean, we're already in there, might as well get it all done
14:14:04 <tflink> assuming that we're not making things more complicated by putting ansible modules in a place that's difficult to import
14:14:06 <roshi> instead of having to touch it twice
14:14:16 <tflink> those'll be some big patches
14:14:28 <jskladan> tflink: let's do those after we toss phab
14:14:31 <roshi> but I'm not married to the idea
14:14:35 <jskladan> will be more fun to review that way :)
14:14:56 <tflink> :-P
14:15:13 <kparal> jskladan: since you're so funny, you get the task of writing the export filters for phab tickets
14:15:23 <jskladan> kparal: like anybody cared...
14:15:44 <kparal> I don't want to lose them all
14:15:46 <tflink> #action jskladan to figure out way to archive and redirect phabricator tickets
14:15:52 <tflink> :-D
14:16:02 <jskladan> tflink: I have an idea - keep phab around
14:16:07 * jskladan badum tsss, problem solved
14:16:14 <tflink> that problem, yes
14:16:24 <jskladan> :)
14:16:29 <tflink> #undo
14:16:29 <zodbot> Removing item from minutes: ACTION by tflink at 14:15:46 : jskladan to figure out way to archive and redirect phabricator tickets
14:16:51 * tflink doesn't much care when we do it
14:17:12 <tflink> it seems silly to set up multiple new projects in phab just to depricate the whole system soon
14:17:33 <roshi> so we put th enew projects in pagure
14:18:04 <kparal> but we haven't decided the new project location yet! :)
14:18:14 <tflink> that would be the next topic
14:18:37 <jskladan> tflink: I agree, if we're about to kill phab, I'd not put much work into using it "right" just now
14:18:46 <roshi> whatever we decide, it makes sense to just put the new projects in the new place
14:18:48 <tflink> #info start splitting libtaskotron into 3 as we ansible-ize things
14:18:49 <roshi> imo, anyways
14:18:52 <tflink> er
14:18:54 <tflink> #undo
14:18:54 <zodbot> Removing item from minutes: INFO by tflink at 14:18:48 : start splitting libtaskotron into 3 as we ansible-ize things
14:19:16 <tflink> #info we've decided to start splitting libtaskotron into 3 as we shift over to the new standard interface
14:20:05 <tflink> #action tflink to push directives ported to ansible modules
14:21:01 <tflink> #undo
14:21:01 <zodbot> Removing item from minutes: ACTION by tflink at 14:20:05 : tflink to push directives ported to ansible modules
14:21:15 <tflink> #action tflink to finish porting directives to ansible modules
14:21:24 <tflink> current work: https://pagure.io/ansible-fedoraqa-modules
14:22:19 * roshi wishes pagure did namespaces
14:22:31 <tflink> roshi: ?
14:22:41 <roshi> pagure.io/fedora-qa/foo
14:22:46 * tflink chose to put it there, figuring it'd be renamed later
14:22:49 <tflink> it does that, no?
14:22:51 <roshi> and /fedora-qa is a landing page or something
14:22:59 <tflink> just not the second part
14:23:00 <roshi> the /fedora-qa has to be a repo
14:23:14 <roshi> not just a grouping of related things inside a namespace
14:23:14 <jskladan> tflink: well, it is kind of unusable...
14:23:23 <tflink> jskladan: what is?
14:23:23 <roshi> unless something changed and I missed it
14:23:28 <jskladan> the "namespaces"
14:23:42 <jskladan> (even the pagure's search bar fails if it encounters one)
14:23:43 * tflink suggests that we get back to the topic at hand
14:23:49 <jskladan> +1
14:23:57 * roshi apologizes
14:24:15 <tflink> any thoughts on how we want to go about doing the split?
14:24:15 <kparal> roshi: https://pagure.io/group/fedora-qa
14:24:49 <kparal> runner, library and ansible modules
14:25:12 <tflink> I was assuming that had been pretty much agreed upon
14:25:20 <kparal> so what is the question?
14:25:21 <tflink> I meant who's doing the work and when
14:25:27 <kparal> ah :)
14:25:41 * tflink wonders if it'd be wise to start with a single throw-away repo with subdirectories
14:25:54 <kparal> I guess I'm the packager nowadays
14:25:55 <jskladan> +1
14:26:08 <tflink> once we get that working well enough, start merging the code in to existing repos and create new ones
14:26:18 <kparal> sounds good
14:26:46 <jskladan> we will most probably lose some history, but it is deemed to happen anyway, with stuff like this, IMO
14:26:49 <tflink> that ansible module repo has a bundled copy of libtaskotron
14:27:01 <tflink> jskladan: how so?
14:27:10 <tflink> the libtaskotron repo isn't going anywhere ATM
14:27:19 <tflink> the others are being added
14:27:32 <tflink> or do you meant that searching history will be more complicated for the bits we split out?
14:27:40 <jskladan> say you want to port code from libtaskotron to
14:27:47 <jskladan> the ansible stuff
14:27:54 <jskladan> you most probably lose the context of
14:28:00 <jskladan> "this once was a directive in libtaskotron"
14:28:14 <jskladan> but I don't see it being a problem, was just saying
14:28:35 <tflink> yeah, some of that will be obscured at best
14:29:36 <tflink> unless someone else wants to do it, I can get the temporary repo started
14:30:00 <tflink> I'll probably delete the ansible-modules repo while I'm at it
14:30:03 <kparal> go for it
14:30:43 <tflink> where are we on the functionality?
14:31:03 <tflink> waiting for testing? waiting for ansible modules and I wasn't quite aware that it was blocked on me?
14:31:34 <tflink> mkrizek: IIRC, the ansible stuff is working but there are some hacky things which need to be worked out?
14:32:32 <mkrizek> my plan was to deploy it to dev for testing before doing any "polishing"
14:32:48 <tflink> ok
14:33:02 <tflink> that means that we are blocked on me
14:33:03 <kparal> maybe we can use stg for it?
14:33:09 <kparal> because it's not used atm
14:33:20 <tflink> or take the time to move dev to cloud
14:33:25 <tflink> and free up a virthost
14:33:32 <roshi> fedorainfracloud?
14:33:34 <kparal> or make stg listen to production messages. either way, let's have at least one machine for development
14:34:58 <tflink> roshi: yeah
14:35:24 <tflink> we can move stg back to listening to production
14:35:36 <tflink> IIRC, we changed that when the modularity stuff was all happening in stg, not prod
14:36:55 <kparal> right, now it's in prod
14:37:07 <kparal> we also experimented with distgit on stg
14:37:16 <kparal> but that seems stalled now
14:37:35 <tflink> I think the focus changed more than it stalled
14:37:53 <tflink> IIRC, it's due to land in prod before much longer
14:38:37 <tflink> either way, sounds like it might be wise to move stg back to listening on the prod bus
14:39:01 <roshi> +1
14:39:02 <tflink> which sounds like a good task for folks learning how stuff is set up
14:39:26 <tflink> *cough* jskladan, kparal, lbrabec *cough*
14:39:45 <jskladan> tflink: yeah, yeah
14:39:52 <kparal> a terrible cough you have
14:39:54 <jskladan> I'm sure I can't do _much_ damage :D
14:40:18 <tflink> kparal: the only cure is you doing more sysadmin work :-P
14:40:42 <tflink> who's taking the task?
14:41:15 * kparal will rather allow jskladan and lbrabec to volunteer
14:41:30 <tflink> and I'm sure they'd say pretty much the same thing
14:41:38 <jskladan> :D
14:41:52 <jskladan> I can break stuff! Pick me!
14:41:58 <kparal> jskladan++
14:41:59 <jskladan> (also, kparal should do it..)
14:42:09 <roshi> lol
14:42:19 <roshi> I can chip in as well
14:42:22 <tflink> #action jskladan and kparal to switch taskotron-stg back to listening to prod fedmsgs
14:42:41 <tflink> ok, anything else on this topic?
14:42:43 <jskladan> tflink: just to be sure - it means "change one string in one place", right?
14:42:55 <tflink> jskladan: IIRC, it's a few strings in a few places
14:42:59 <kparal> jskladan: mkrizek will know ;)
14:43:36 * tflink assumes not and moves on since we have 2 more topics and ~ 15 minutes
14:43:47 <tflink> #topic Project Storage
14:43:49 <jskladan> tflink: that's right, it is in all of the topics, is it not? (/me thought there would be a variable to just set all those up)
14:44:14 <tflink> jskladan: there should be but there's another change that says which fedmsg bus to listen to, IIRC
14:44:24 <jskladan> tflink: we'll see
14:44:24 <tflink> so there are configuration changes for trigger and fedmsg
14:44:27 <jskladan> let us move on
14:44:43 <tflink> everyone's favorite topic - project storage
14:44:48 <jskladan> yay!
14:45:27 <tflink> jskladan, kparal: I think you two are the most passionate about this, do you want to put together a plusses/minuses doc?
14:45:38 <kparal> pagure
14:45:42 <kparal> done
14:46:00 <jskladan> Let me just say - both github and pagure are "worse" solutions for me anyway. I'd go with the political decision, as we discussed on BJ
14:46:27 <jskladan> I don't really care which wrong solution is taken :)
14:46:28 <tflink> yeah, I'm not thrilled about spending time on this when we have enough other things to do, either
14:46:35 <kparal> I was thinking about this and since we intend to test pagure integration, it makes sense to be on pagure
14:46:40 <tflink> agreed
14:46:46 <kparal> we'll end up submitting PRs for pagure, I have no doubt
14:46:51 <tflink> yep
14:46:56 <kparal> but since I'm already a bodhi hacker, why not pagure hacker as well
14:47:03 <tflink> :)
14:47:11 <kparal> can't hurt to have few more projects on the plate
14:47:13 <jskladan> like, because we have not enough stuff to do, we could reimplement github, and gitlab, in pagure :)
14:47:28 <jskladan> seems reasonable :)
14:47:34 <kparal> exactly my thinking
14:47:36 <tflink> completely
14:47:48 <jskladan> at least I know who to ask for enhancements :)
14:47:52 <tflink> :)
14:48:03 <tflink> any more productive discussion on this, then?
14:48:10 * roshi has nothing
14:48:18 <jskladan> #info kparal is the dedicated pagure hacker, all RFE's are to be implemented by him in a timely manner
14:48:25 <tflink> +1
14:48:34 <kparal> also, I feel like my java skills are underutilized
14:48:46 <jskladan> kparal: that's why we also have Jenkins for you :)
14:48:49 <kparal> good, good!
14:48:59 <jskladan> but really - I don't really have strong feelings either way
14:49:07 <jskladan> my actual issue is with the code review process
14:49:19 <kparal> it will rely on pings...
14:49:41 <tflink> yeah, I'm not terribly thrilled about the move, either but unless we have better ideas ...
14:49:52 <tflink> we do only have 10 minutes left
14:49:56 <jskladan> well, if mailing-list reviews work for kernel...
14:50:19 * tflink will look into moving tickets from phab to pagure unless someone gets to it before he does
14:50:21 <kparal> and reporting a bug against kernel is such a wonderful experience...
14:50:58 <kparal> tflink: there's an existing github->pagure converter, so perhaps if there's phab->github converter already somewhere, we can make a roundtrip
14:51:15 <kparal> funny, right?
14:51:21 <tflink> I'll take a look but I don't think that's going to work for us going from 1 to many bug trackers
14:51:43 <tflink> which is the primary fscking reason we didn't use github or bitbucket in the first place
14:51:50 <tflink> but now I'm getting into the whining
14:52:03 <kparal> I'd assume the phab converter would only pick selected projects
14:52:13 <kparal> but I don't even know if something like that exists
14:52:18 <tflink> either way, worth looking into
14:52:27 <tflink> even if it's just for the api calls to phab
14:53:02 <tflink> kparal, jskladan: have you all had a chance to look through your RFE list for pagure to see if they're all still valid
14:53:03 <tflink> ?
14:53:20 <jskladan> tflink: nope
14:54:17 <tflink> we have 7 minutes left so unless we have anything else productive for this topic, I suggest we move on
14:54:29 <kparal> I haven't
14:54:44 <tflink> #topic Documentation/Examples
14:54:58 <tflink> in retrospect, this is kinda pointless until we have the splitting done
14:55:09 <kparal> and the ansible switch
14:55:10 <roshi> yeah
14:55:21 <tflink> I do think that we should move our docs to rtd, though
14:55:28 * tflink found some tricks to make that work
14:55:31 <kparal> it'd be nice to deploy at least the new libtaskotron docs. but that would require a new build
14:55:35 <jskladan> tflink: agreed, I don't think putting much work into examples/docs that become obsoleted next month is worth it
14:56:00 <tflink> I was more thinking of doing an example or two as we go
14:56:08 <tflink> but either way, I think we can hold off on this for a little bit
14:56:13 <tflink> which brings us on to
14:56:13 <roshi> for sure
14:56:15 <roshi> makes sense
14:56:16 <tflink> #topic Open Floor
14:56:23 <tflink> any other topics that folks want to bring up
14:56:24 <tflink> ?
14:57:13 <jskladan> not really. I still have that "how are we going to do code reviews" issue, but that's not to be solved in three minutes
14:57:22 * roshi heads out to get some more coffee
14:57:23 <tflink> anything cheery? unicorns or cute cats?
14:57:34 <jskladan> I have one!
14:57:42 <tflink> jskladan: yeah, that's a longer topic
14:57:52 <jskladan> https://www.facebook.com/9gag/videos/10155767949826840/?autoplay_reason=gatekeeper&video_container_type=0&video_creator_product_type=2&app_id=2392950137&live_video_guests=0
14:58:00 <jskladan> cuteness overload!
14:58:05 <tflink> because completely up-ending our workflow is always going to be productive in the short term
14:58:16 * tflink realizes that he's whining when he asked whining to stop
14:58:28 <tflink> gah! I'm part of the problem :-(
14:58:40 <kparal> try searching for "zombie unicorn". such pretty...
14:58:52 <jskladan> have a truckload of kittens, tflink! https://www.instagram.com/kote218/
14:58:54 <tflink> on that note, I'm setting the fuse for some amount of time
14:59:29 <tflink> ok, thanks for coming, everyone
14:59:34 * tflink will send out minutes shortly
14:59:41 <tflink> #endmeeting