20:00:00 <robcresswell> #startmeeting horizondrivers
20:00:01 <openstack> Meeting started Wed Jun 29 20:00:00 2016 UTC and is due to finish in 60 minutes.  The chair is robcresswell. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:02 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:00:05 <openstack> The meeting name has been set to 'horizondrivers'
20:00:46 <TravT> o/
20:00:51 <robcresswell> o/
20:00:58 <rdopiera> o/
20:01:00 <hurgleburgler> ʕ·ᴥ·ʔノ
20:01:01 <tsufiev> o/
20:01:29 <ediardo> o/
20:01:34 <robcresswell> Hey folks
20:01:50 <ediardo> Hello!
20:02:05 <robcresswell> Here's the agenda for today:
20:02:08 <robcresswell> https://wiki.openstack.org/wiki/Meetings/HorizonDrivers#Agenda_for_2016-06-29_2000_UTC
20:02:35 <robcresswell> I wanted to run through priorities and see where we're at
20:03:06 <jlopezgu> o/
20:03:16 <matt-borland> o/
20:03:49 <robcresswell> Richard isn't around, but I believe we're about ready to go on releasing xstatic. david-lyle also did some work on helping us not break everyone else when we do update.
20:04:12 <robcresswell> I'm still using https://etherpad.openstack.org/p/horizon-libs to track package updates
20:04:36 <robcresswell> And there are several in gerrit for testing: https://review.openstack.org/#/q/status:open+project:^openstack/xstatic.*
20:04:44 <lcastell> o/
20:04:57 <robcresswell> matt-borland: Where's the angular registry/images work at?
20:05:01 <david-lyle> o/
20:05:44 <matt-borland> registry is basically stable for what I think we need in Newton, 2 patches I know of at this point
20:06:05 <matt-borland> images work: I want to make an enablement patch similar to what Richard did for swift last time
20:06:23 <matt-borland> also, I'm adding in server-side faceted search...that patch is up for review
20:06:26 <tsufiev> enablement?
20:06:45 <matt-borland> em, a switch so people can enable it in place of the previous incarnation of the panle
20:06:48 <matt-borland> *panel
20:06:51 <tsufiev> oh
20:06:54 <tsufiev> got it
20:07:18 <david-lyle> by default?
20:07:24 <matt-borland> exactly
20:07:33 <matt-borland> so it's a simple switch
20:08:00 <matt-borland> I'd like people to be serious about reviewing the differences between the Angular and Django work in Images.
20:08:24 <matt-borland> The other status update is Instances...we have patches in GitHub that basically prepare that for switching to Angular as well.
20:08:43 <robcresswell> That'd be good to review switching at the midcycle, we'll see where we are though.
20:08:47 <TravT> even for a simple table based panel, the angular images panel is much nicer to use.
20:08:50 <matt-borland> exactly, it's a starting point. :)
20:09:07 <TravT> i think mid-cycle will be great to review.
20:09:26 <robcresswell> Cool, thanks matt-borland
20:10:28 <robcresswell> david-lyle: Anything need mentioning on the filtering improvements? I've been a bit slow on the uptake with those, but reviewed a couple today
20:11:10 <david-lyle> many patches in progress
20:11:24 <david-lyle> no real road blocks that I'm aware of
20:11:42 <david-lyle> lcastell, ediardo sound right?
20:11:50 <robcresswell> Go team intel
20:11:55 <ediardo> we've been working on several views
20:12:04 <jlopezgu> haahha
20:12:04 <ediardo> ~9 views out of ~26
20:12:15 <ediardo> I think we are making good progress there
20:12:19 <lcastell> yup, just struggling with swift ui filtering and magic search :P
20:12:53 <robcresswell> Really appreciate the work on that stuff. Long overdue, and very valuable.
20:13:04 <lcastell> need to define the way to go there. find some middle ground
20:13:04 <ediardo> I''ve been using the novaclient for a while, looks like we don't have many filtering choices in flavors, keypairs,hosts aggregates, etc
20:13:44 <david-lyle> I think those aren't as much of a pain point, so we should be ok
20:14:22 <robcresswell> Don't worry about things the API doesn't provide. If the filter isn't there, there's not much we can do.
20:14:40 <robcresswell> tsufiev: How we doing in integration tests and file upload?
20:15:19 <tsufiev> robcresswell, integration tests are stable for now, but we haven't time yet to investigate running them in parallel
20:15:58 <tsufiev> also, I managed to run them on new Marionette FF API, but Webdriver for Marionette is too immature yet
20:16:12 <tsufiev> so I think we'd better continue using FF46
20:16:35 <tsufiev> perhaps the bug in FF47 that caused Selenium to fail will be fixed in FF47.01
20:16:56 <tsufiev> I read some announcements, but have to try it myself once it gets to trusty repos
20:17:51 <clarkb> tsufiev: firefox ships precompiled binaries for all the platforms too, makes it really easy to swap in a specific version for testing. (Not sure if that helps but you shouldn' tneed to wait for ubuntu)
20:18:12 <tsufiev> file upload is good, I'm trying to lure some OpenStack Security folks into reviews so they would share their opinion on passing Keystone token into JS
20:18:42 <tsufiev> clarkb, AFAIK the problem with FF is that once, say FF47 is released, one cannot install FF46 in ubuntu
20:18:51 <tsufiev> it just disappears from the repo
20:18:57 <tsufiev> or perhaps I'm doing it wrong?
20:19:44 <clarkb> tsufiev: ubuntu does clean out old pacakges yes
20:19:55 <clarkb> just mentioning that you can go to firefox directly for early testing
20:20:11 <tsufiev> robcresswell, if I understand correctly, security concerns about CORS and Keystone tokens is a kind of blocker for these upload patches
20:20:39 <tsufiev> clarkb, thanks for the info :)
20:21:19 <robcresswell> tsufiev: Understood. Can you keep us updated when we get some security opinions?
20:21:40 <tsufiev> robcresswell, I can cc you in the private letter I sent to some security folks
20:22:01 <robcresswell> That'd be great, just trying to keep an eye on various efforts.
20:22:22 <robcresswell> hurgleburgler: How's things going on the theming side of things?
20:22:30 <tsufiev> oh, and I have some good news about 'quotas sanity' things in Horizon
20:22:31 <ediardo> :)
20:22:41 <tsufiev> sorry to interrupt, hurgleburgler
20:22:53 <robcresswell> tsufiev: One moment, I'll come back to ya :)
20:22:57 <tsufiev> k
20:23:01 <hurgleburgler> Going well
20:23:14 <hurgleburgler> Making progress, but no one really reviewing anymore
20:23:57 <tsufiev> seems all reviewers are tied implementing their own patches :/
20:25:38 <robcresswell> Actually, we should discuss that a little. We really need to encourage more review work, especially from non-cores too.
20:25:53 <robcresswell> Please don't think that just because you cannot +2, that your review is not meaningful.
20:26:15 <matt-borland> +1.9999
20:26:20 <ankur-gupta-f1> robcresswell: any recommendations on what to start reviewing? or just review all we can?
20:26:34 <robcresswell> It's very important for us to get continuous commentary and review work across the board, so please make sure to set time aside, or discuss with your managers to put time aside for review work too.
20:26:37 <matt-borland> I think prioritization is a problem robcresswell
20:26:55 <robcresswell> There's a few places to look
20:27:13 <robcresswell> https://launchpad.net/horizon/+milestone/newton-2
20:27:29 <robcresswell> Is a list of prioritised and targetted bugs, any In Progress have a patch up.
20:28:16 <robcresswell> You can also use the gerrit dash that david-lyle created which gives a nice overview of patches that have been ignored, un reviewed etc
20:28:53 <itxaka> uh, link?
20:28:55 <ediardo> where's that?
20:29:09 <ankur-gupta-f1> https://github.com/openstack/gerrit-dash-creator
20:29:26 <david-lyle> I have a better one not checked in will provide paste
20:30:13 <robcresswell> You can actually add those links in the Gerrit preferences too (which I only realised today) to have extra sections next to Changes/Draft/Starred at the top.
20:31:31 <robcresswell> The other item I've intermittently used is https://wiki.openstack.org/wiki/Horizon/WeeklyBugReport
20:31:44 <robcresswell> Though it tends to get time consuming to maintain on top of bug triage work :)
20:31:51 <robcresswell> So its fallen by the wayside.
20:32:31 <david-lyle> http://paste.openstack.org/show/524127/
20:32:32 <robcresswell> Happy to update it if that would help with review work
20:32:46 <robcresswell> thanks david-lyle
20:33:13 <lcastell> cool
20:33:24 <david-lyle> you can customize with the link ankur-gupta-f1 provided to target what's important to you
20:33:36 <ediardo> hey david-lyle, that looks like those urls hackers want dummy users to paste into the addres bar
20:33:39 <ediardo> jk
20:33:53 <david-lyle> I can already see the money rolling in
20:34:00 <ediardo> yeaah
20:34:02 <robcresswell> If someone logs into your fb next week, you know who it was...
20:34:12 <lcastell> =D
20:34:56 <itxaka> waaaa, that dashboard is pretty cool :o
20:35:22 <robcresswell> I had a couple of weeks holiday, but I'm back on actively reviewing and triaging, so the N-2/N-3 milestones should get fairly populated
20:35:50 <robcresswell> Help is welcome :D
20:35:58 <robcresswell> tsufiev: Quotas?
20:37:05 <tsufiev> so, the patch https://review.openstack.org/#/c/335524/ adds to novaclient an ability to return quota usages along with limits - similar to what we have already for cinder
20:37:41 <tsufiev> eric has already a patch to fetch cinder quota usages
20:38:22 <tsufiev> this is quite important because now to fetch quota usages horizon makes a bunch of calls to retrieve every possible list of nova/cinder/neutron entities
20:38:28 <tsufiev> which is quite time-consuming
20:38:55 <tsufiev> but of course there is still neutron, which doesn't have an API like this yet
20:39:20 <robcresswell> One step at a time :)
20:39:37 <tsufiev> that's unfortunate, because sequentially asking for networks, subnets and routers can be VEEERY SLOOOW
20:39:48 <robcresswell> haha
20:39:56 <robcresswell> Yeah, absolutely.
20:40:06 <robcresswell> Thanks for the update
20:40:13 <tsufiev> but the last issue can be worked around by parallelizing different calls to not send them sequentially
20:40:32 <robcresswell> In Python?
20:40:34 <tsufiev> that's what I've been doing the last week and gained 4x performance
20:40:35 <tsufiev> yes
20:40:45 <tsufiev> 20 seconds rendering time -> 5 seconds
20:40:58 <robcresswell> Eric might try and hug you
20:41:26 <tsufiev> I'd like to first propose nova and cinder sane usages, then parallelizing some stuff in case of neutron
20:41:27 <robcresswell> tsufiev: Is there a patch for this?
20:41:33 <robcresswell> Sure
20:41:56 <tsufiev> robcresswell, yes, but I want to rebase it onto usages unification work
20:42:03 <tsufiev> bp is still to be written
20:42:14 <tsufiev> I think that's all news about quotas
20:42:18 <robcresswell> Okay, so we're not looking at a Newton timescale for this then
20:42:30 <robcresswell> Thanks tsufiev
20:42:31 <tsufiev> yep, that would be too optimistic
20:42:44 <robcresswell> Any other updates? Otherwise I'd like to move onto docs
20:43:12 <robcresswell> #topic UI Guidelines
20:43:47 <robcresswell> So, we have someone who has thrown themselves under a bus/ kindly volunteered to help improve on http://docs.openstack.org/contributor-guide/ui-text-guidelines.html
20:43:59 <robcresswell> asettle: o/
20:44:24 <robcresswell> It'd be great to work on this content and improve it, but I've asked that Horizon is a bit more involved than it was in the original draft.
20:44:25 <hurgleburgler> thanks asettle
20:44:43 <ediardo> o/
20:44:53 <robcresswell> Looking for volunteers to help answer questions, offer reviews etc.
20:45:12 <hurgleburgler> Its good to point out that these are just guidelines … meant to help give some direction when desired, but not meant to be prescriptive or conforming
20:45:34 <ankur-gupta-f1> robcresswell: I can help in whatever capacity I can.
20:45:53 <ediardo> robcresswell: me too.
20:45:55 <ankur-gupta-f1> I have some free time
20:46:03 <robcresswell> Yeah, just a reference point. A lot of the content we have in Horizon isn't often built around anything other than "what was already done" which may not be the best, from a writing or design point of view.
20:46:19 <robcresswell> Awesome! Thanks
20:46:41 * robcresswell closes trap
20:46:54 <ediardo> it was a trap???????????????????????
20:47:10 <ediardo> https://www.youtube.com/watch?v=4F4qzPbcFiA
20:47:26 <hurgleburgler> I think the purpose to to try and determine whatever patterns currently exist, not necessary to create new ones?
20:47:59 <robcresswell> Well, both to an extent; if we're doing things 3 ways, lets pick one and document why
20:49:01 <robcresswell> Things like, how to phrase help text, what type of error to display, are all useful. For example, only recently a patch altered many of the alerts to be Info instead of Danger; but this wasn't really documented in anything other than a patch
20:49:38 <robcresswell> I think documenting that kind of data as a guideline has value, and saves load on developers answering questions for new folk or plugins, IMO.
20:50:08 <hurgleburgler> Right, but there is a push now to start documenting stuff like page layout and more UI type of elements
20:51:21 <robcresswell> Yeah, we can decide as we go if we agree or not
20:51:36 <robcresswell> The point is to avoid having UX make that decision for us, and make sure we're involved in the process.
20:51:54 <hurgleburgler> k
20:52:31 <robcresswell> Any other questions?
20:53:25 <robcresswell> Sweet, lets move on
20:53:41 <robcresswell> #topic 800UTC meeting
20:54:21 <robcresswell> So, the 1200UTC meeting was moved to 800UTC, which resulted in a 0% increase in the number of attendees
20:54:49 <robcresswell> Its generally only myself, Richard and tsufiev that are around at that time.
20:54:49 <itxaka> nice
20:54:52 <itxaka> :)
20:54:57 <rdopiera> at least it's not negative
20:55:26 <robcresswell> I'd like to consider dropping the early meeting slot until there is demand for it. I'm happy to be awake and run it (9am for me anyway)
20:55:47 <robcresswell> but at the moment it seems like a waste of time to demand horizon core attention for a meeting with so little value.
20:55:52 <robcresswell> Any thoughts/
20:55:54 <robcresswell> *?
20:56:00 <hurgleburgler> Does that means we'd just have a meeting every other week and toggle with Drivers?
20:56:17 <robcresswell> We'd just have one meeting each week at 2000UTC
20:56:19 <david-lyle> I think we could collapse the two back into one
20:56:26 <robcresswell> david-lyle: Agreed
20:56:38 <robcresswell> There isn't  as much blueprint review demand either.
20:56:52 <hurgleburgler> robcresswell: sounds ok to me … but I guess its really up to you, jones and tsufiev
20:57:20 <tsufiev> I don't mind to continue attending this meeting
20:57:34 <robcresswell> Yeah, naturally I'll put it on the mailer and discuss it more openly
20:57:59 <robcresswell> Just wanted to see if anyone had any input here
20:58:24 <tsufiev> but... it's funny that meeting that was moved after richard's mail is cancelled while he's on a vacation :)
20:59:00 <hurgleburgler> maybe we should bring it back up next then?
20:59:13 <TravT> i think ML is the right place for it
20:59:23 <robcresswell> As I said, I'll go on the ML, this isn't a formal cancellation of the meeting.
20:59:31 <tsufiev> k
20:59:31 <TravT> robcresswell bringing it up now and following it up in the ML is the right process
21:00:02 <robcresswell> Well, we're at time
21:00:09 <matt-borland> thanks!
21:00:10 <robcresswell> Thanks everyone
21:00:17 <robcresswell> #endmeeting