12:02:17 <david-lyle> #startmeeting Horizon
12:02:18 <openstack> Meeting started Wed Jun 24 12:02:17 2015 UTC and is due to finish in 60 minutes.  The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:02:20 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
12:02:22 <openstack> The meeting name has been set to 'horizon'
12:02:33 <r1chardj0n3s> o/
12:02:36 <david-lyle> Hello, who's here
12:02:43 <tsufiev> hi
12:03:38 <kzaitsev_mb> o/
12:03:40 <neillc> o/
12:04:24 <kzaitsev_mb> not as crowded as alternate-time meeting it seems. )
12:04:27 <robcresswell> o/
12:04:42 <david-lyle> kzaitsev_mb: not usually
12:04:43 <r1chardj0n3s> all the sensible americans are asleep ;)
12:05:18 <r1chardj0n3s> all the non-sensible aussies are here; some came prepared with whiskey
12:05:19 * david-lyle feels there's an implication there
12:05:22 <neillc> Lol it's 10pm here and I'm sitting in a park because I forgot the meeting and went for a walk
12:05:31 <robcresswell> ha!
12:05:52 <neillc> Whiskey! That's what this walk needs!
12:05:56 <david-lyle> Ok bookkeeping stuff first
12:06:13 <david-lyle> Liberty-1 was tagged yesterday
12:06:34 <david-lyle> #link https://launchpad.net/horizon/liberty/liberty-1
12:06:42 <robcresswell> Nice
12:06:56 <r1chardj0n3s> ... which included a bunch of file renaming in horizon, mostly ;)
12:07:07 <david-lyle> 3 bps and 153 bug fixes
12:07:26 <r1chardj0n3s> huh, not bad
12:07:33 <david-lyle> less bps than usual
12:08:04 <david-lyle> which I blame on the endless quagmire of coding style changes
12:08:13 <r1chardj0n3s> yeah :/
12:08:15 <david-lyle> amongst other things
12:08:41 <david-lyle> There was a lot of progress on other items that just haven't landed yet too
12:09:20 <r1chardj0n3s> there's *ten* "JSCS Cleanup" patches in flight at the moment
12:09:37 <r1chardj0n3s> er, twelve
12:09:45 <r1chardj0n3s> I wish gerrit had a search function
12:09:56 <r1chardj0n3s> 13
12:10:02 <r1chardj0n3s> lol
12:10:06 <david-lyle> for the love of ...
12:10:06 <tsufiev> r1chardj0n3s, spawning right now?
12:10:08 <r1chardj0n3s> and that's not all of them
12:10:30 <r1chardj0n3s> though it's hard to guage the size of the problem because no-one doing the work is communicating about it grrrr
12:10:57 <robcresswell> This needs an etherpad or mailer thread
12:11:18 <david-lyle> I feel like cleanup is something you do after the fact, not to block all other things
12:11:24 <r1chardj0n3s> I think it's just cindy's big plan to be the #1 committer in Liberty ;)
12:11:41 <david-lyle> I think some people must get paid by the line
12:11:46 <robcresswell> ...or you pick a style before you write thousands of lines of code
12:11:50 <r1chardj0n3s> I think the JSCS stuff *should* be able to be done while we do other things, yes
12:12:03 <r1chardj0n3s> robcresswell: yeah, nice idea, but ...
12:12:05 <r1chardj0n3s> ;)
12:12:10 <david-lyle> there are downsides to having great contribution tracking
12:12:24 <robcresswell> Problem is that JSCS has been merged with "reorganise everything into components"
12:12:32 <r1chardj0n3s> for the record: I totally did not mean that thing I said about cindy
12:12:39 <robcresswell> JP style guide stuff, right? Seems to be a lot of crossover
12:12:45 <r1chardj0n3s> yeah
12:12:45 <robcresswell> haha
12:13:32 <r1chardj0n3s> I think a large part of the current efforts is acknowledgment that there's significant tech debt in the codebase, and trying to address some of that
12:14:19 <robcresswell> I'll try and get on top of reviewing those faster, so we can move on from just moving code around
12:14:52 <r1chardj0n3s> I've been focusing on sean's patch chain, but I will try to throw some +1s at the jscs cleanups as well
12:15:26 <david-lyle> so looking at the priority list for liberty
12:15:40 <david-lyle> #link https://etherpad.openstack.org/p/YVR-horizon-liberty-priorities
12:16:22 <david-lyle> most of the top priority items are in progress, not done
12:16:46 <david-lyle> translation appears to be close
12:16:54 <david-lyle> couple of patches to review
12:17:04 <neillc> Yep. Getting very close I think
12:17:22 <david-lyle> I'm still slugging it out with plugins
12:17:37 <david-lyle> testing config is my biggest holdup at this point
12:17:47 <robcresswell> I had a question actually, about that
12:17:56 <david-lyle> sure
12:18:07 <robcresswell> Do we want plugins to live under Horizon, like with Neutron?
12:18:22 <robcresswell> So Neutron actually covers all of the vendor plugins
12:18:51 <david-lyle> I think there is a difference
12:18:58 <robcresswell> and I think we cover some, but not all right now? Like manila and tuskar
12:19:15 <robcresswell> I was curious where <vendor>-ui should live.
12:19:27 <david-lyle> in Horizon we're supporting services that exist in the ecosystem already
12:20:05 <david-lyle> unlike neutron where the plugin is their only tie to the ecosystem
12:20:20 <david-lyle> I've left it up to the service teams
12:20:32 <david-lyle> they can own or Horizon can
12:20:51 <david-lyle> The issue being we need the reviewers for the plugin
12:21:03 <david-lyle> we can't handle all of that
12:21:09 <robcresswell> That stands for the component uis, but not for vendor extensions, I'd have thought.
12:21:17 <tsufiev> david-lyle, is it possible to grant people with per-directory +2 review rights?
12:21:24 <david-lyle> tsufiev: no
12:21:34 <david-lyle> gerrit is not that sophisticated
12:21:38 <tsufiev> :(
12:21:56 <david-lyle> So currently we have 3 arrangements
12:22:07 <david-lyle> or are working toward
12:22:16 <david-lyle> 1) plugin lives in contrib
12:22:42 <david-lyle> with that the service team members support and 1 horizon core could merge code
12:23:13 <david-lyle> 2) plugin is owned by horizon team but lives out of tree, like tuskar-ui and manila-ui
12:23:37 <david-lyle> 3) plugin is owned by the service team
12:23:52 <david-lyle> 2 and 3 aren't very different IMO
12:24:19 <david-lyle> and in both you can have a mixed core between service team and horizon
12:25:07 <david-lyle> the model we are shooting for is more like devstack, qa and docs than neutron
12:25:56 <david-lyle> where a vendor specific plugin to horizon lives is up for debate
12:26:28 <david-lyle> I don't think Horizon owns it though
12:27:21 <robcresswell> How does this tie into the governance/projects? Does a ui owned by a service team, count as its own project?
12:27:49 <robcresswell> Or does it have its own repo and cores, but still lives under horizon? or the original project, say trove-ui under trove
12:28:10 <robcresswell> Just trying to wrap my head around the organisation of it all
12:28:23 <sambetts> you can have completely independant projects under the stadium of another
12:28:29 <david-lyle> the gerrit ACLs can be set per repo
12:28:54 <david-lyle> so regardless of whether it lives in Horizon or service team the core is what you make of it
12:29:07 <david-lyle> it's ultimately up to who runs the show
12:29:26 <david-lyle> and we want to let the service teams do that where possible
12:29:27 <kzaitsev_mb> robcresswell: david-lyle: We in murano are comfortable with our current murano-dashboard plugin for horizon. Currently it lives under murano, has murano cores and plugs in by 1 file drop-in.
12:30:16 <david-lyle> we can't keep scaling to support all the things
12:30:26 <robcresswell> No, absolutely
12:30:51 <tsufiev> kzaitsev_mb, the only drawback of this approach is that you don't know in advance which horizon commit would break murano-dashboard
12:31:39 <david-lyle> we could possibly set up non-voting plugin test jobs as well
12:31:59 <kzaitsev_mb> tsufiev: that's true for dashboard, and heat and nova and oslo. and dozens of things, that's why we have gate jobs and stuff pointed at master horizon =)
12:31:59 <robcresswell> How its looking is that we'll end up with a Cisco extension in its own repo, with me as a core, but just wondering how to define that within openstack as a whole
12:32:23 <tsufiev> yet putting every possible test job against every plugin into horizon repo - I wander how long then will it took for the horizon commits to pass all checks?
12:32:23 <sambetts> the vendor repos in neutron are under neutron in the goverance big tent scheme even though they are managed completely indepentantly, just so that their relationship to neutron is documented
12:32:46 <david-lyle> tsufiev: good point
12:32:52 <robcresswell> sambetts: Right, which is what I'd like to do with our cisco dashboard stuff
12:33:11 <robcresswell> Just trying to understand if that link is acceptable, or they should be unrelated
12:33:21 <tsufiev> david-lyle, I saw a typical commit into Cinder repo, they said it takes ~6 hours for all vendor-specific jobs to pass
12:33:27 <david-lyle> I think it's acceptable
12:33:32 <david-lyle> robcresswell: ^^
12:34:05 <robcresswell> david-lyle: Ah, great. I'll link patches etc. when they're up
12:34:33 <robcresswell> Will make sure nothing merges without your approval.
12:34:51 <kzaitsev_mb> what about putting jobs, that can be launched by comment? like run plugin-foo? Although I'm not sure if anyone would ever run them in that case...
12:35:46 <tsufiev> kzaitsev_mb, or run nightly jobs - i.e. on a time periodic basis - inside every standalone plugin repo
12:36:04 <robcresswell> tsufiev: That doesn't really solve the issue though, just gives you a faster warning of breakages
12:36:12 <robcresswell> rather than prevent them
12:36:17 <tsufiev> robcresswell, agree
12:36:55 <david-lyle> let's solve that, but not right now
12:36:57 <david-lyle> :)
12:37:02 <kzaitsev_mb> =)
12:37:22 <david-lyle> moving on..
12:37:40 <david-lyle> Midcycle is July 21-23 in Fort Collins, CO, USA
12:37:51 <david-lyle> there was an email to the mailing list
12:38:07 <david-lyle> it would be nice to get a preliminary head count
12:38:21 <david-lyle> trying to figure out where would be best
12:38:30 <david-lyle> probably on the etherpad TravT set up
12:38:40 <robcresswell> Perhaps one of the existng etherpads?
12:38:42 <r1chardj0n3s> yep
12:38:44 * david-lyle now has to find that link
12:38:59 <robcresswell> I can't make it. No approval :(
12:39:08 <r1chardj0n3s> https://etherpad.openstack.org/p/horizon-liberty-midcycle
12:39:19 <r1chardj0n3s> robcresswell :(
12:39:25 <david-lyle> :(
12:40:08 <mrunge> yeah, same here.
12:40:19 <mrunge> 3000 $US for 3 days is a bit expensive
12:40:46 <r1chardj0n3s> etherpads aren't loading for me
12:40:54 <robcresswell> david-lyle: The indecisiveness on how to phrase that made me laugh. "Attendies" ... "Planned attendies" .... "Who's coming"
12:42:17 <david-lyle> I'm here to amuse
12:42:19 <robcresswell> It'd be great if someone could leave a hangout on air running, or similar
12:42:37 <david-lyle> We will try to do that
12:42:46 <david-lyle> shouldn't be a problem
12:43:34 <robcresswell> Cheers
12:44:01 <david-lyle> There was no official agenda items posted, so
12:44:09 <david-lyle> #topic Open Discussion
12:45:42 <r1chardj0n3s> there's an open question on neillc's translation thing
12:45:54 <r1chardj0n3s> so babel really wants its plugins to be pip installed
12:46:21 <r1chardj0n3s> but he's got a hack in there so we can include the angular plugin we're developing sit inside the horizon codebase
12:46:25 <r1chardj0n3s> while we develop it
12:46:33 <r1chardj0n3s> he's also pushed it to pypi
12:46:59 <r1chardj0n3s> but to keep development pace alive, it was felt better to also keep it in horizon while we polish it off
12:47:22 <r1chardj0n3s> once it's stable, push to pypi and reference it from requirements.txt instead of vendoring
12:47:24 <r1chardj0n3s> thoughts?
12:47:39 <neillc> Yep. My plan is to do just that
12:47:44 <david-lyle> that sounds reasonable
12:47:53 <r1chardj0n3s> neillc: sorry, figured your fingers were frozen ;)
12:48:13 <neillc> In particular not try to switch to oypi version until plurals working
12:48:13 <david-lyle> is the plugin horizon specific or angular/babel specific?
12:48:27 <r1chardj0n3s> angular-specific
12:48:33 <neillc> Fingers are a bit frozen!
12:48:38 <david-lyle> ok, yeah that makes sense
12:49:10 <tsufiev> david-lyle, so far as part of Merlin toolkit evaluation activity Dan has created the page https://wiki.openstack.org/wiki/Horizon/MerlinEval What's the next step here?
12:49:38 <neillc> I'm not sure if there will be too many other projects that will use it though
12:50:15 <david-lyle> neillc: should it need many updates?
12:50:34 <neillc> Not once plurals are working
12:50:36 <david-lyle> trying to decide if it should be a repo in openstack
12:51:02 <david-lyle> that can be updated after you come to your senses and find a better gig :)
12:51:25 <neillc> For use with babel integration with pip is important
12:51:41 <neillc> So pypi seems to make sense
12:52:29 <neillc> I seem to be doomed to work on i18n. Don't understand why :p
12:52:46 * doug-fish points and laughs
12:52:54 <david-lyle> neillc: no argument on pypi, I guess the question is does the source for that live in a personal github repo or openstack/neillsplugin that openstack can publish too also
12:52:54 <neillc> Pffff
12:52:56 <robcresswell> haha
12:53:25 <neillc> Ah OK. I don't mind either way.
12:53:50 <robcresswell> second one seems more sensible
12:54:28 <david-lyle> I think so, but start as your own, and we can source an openstack repo from it
12:54:39 <neillc> Cool
12:54:44 <robcresswell> Isn't that how the xstatic packages are supposed to be updated? Though I'm not sure how many actually are.
12:54:54 <david-lyle> yes
12:55:15 <david-lyle> turns out there was a missing step in xstatic repo creation by the authors
12:55:52 <david-lyle> which was including a tox.ini file which allows for the publish job to run
12:56:06 <david-lyle> that should be fixed as of yesterday
12:56:57 <david-lyle> tsufiev: I need to look into that more
12:57:12 <david-lyle> I don't have a clear next step at the moment
12:57:19 <robcresswell> Oh nice fix.
12:57:38 <tsufiev> david-lyle, okay, just pinging on that to make sure it wasn't forgot :)
12:58:02 <david-lyle> tsufiev: always good to prod with a stick occasionally
12:58:22 <robcresswell> Btw, several bugs on launchpad have been marked as fix commited by infra without bug links. Not sure whats going on.
12:58:50 <david-lyle> that was the L-1 tagging
12:59:01 <robcresswell> Ah, okay
12:59:29 <robcresswell> Should I just mark them L-2, or dig up the relevant bugs?
12:59:30 <david-lyle> wait committed or released?
12:59:39 <robcresswell> commited
12:59:50 <robcresswell> wait let me check
12:59:50 <david-lyle> do you have an example?
13:00:02 <robcresswell> https://bugs.launchpad.net/horizon/+bug/1463838
13:00:03 <openstack> Launchpad bug 1463838 in OpenStack Dashboard (Horizon) "[data processing] Stack trace for invalid cluster details page" [Medium,Fix committed] - Assigned to Tatiana Ovchinnikova (tmazur)
13:00:24 <robcresswell> Also here:
13:00:33 <robcresswell> #link https://bugs.launchpad.net/horizon/+bug/1466744
13:00:35 <openstack> Launchpad bug 1466744 in OpenStack Dashboard (Horizon) "Integration test test_image_register_unregister failing gate" [Undecided,Fix committed] - Assigned to Luigi Toscano (ltoscano)
13:01:18 <robcresswell> We can continue this in other channel anyway, hour is up
13:02:06 <david-lyle> sure, I think that may be the magic of automation
13:02:13 <david-lyle> thanks everyone
13:02:16 <david-lyle> #endmeeting