20:01:08 <r1chardj0n3s> #startmeeting horizon
20:01:08 <openstack> Meeting started Wed Feb  1 20:01:08 2017 UTC and is due to finish in 60 minutes.  The chair is r1chardj0n3s. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:01:09 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:01:12 <openstack> The meeting name has been set to 'horizon'
20:01:34 <robcresswell> o/
20:01:39 <r1chardj0n3s> hullo robcresswell :-)
20:01:41 <lcastell> o/
20:01:54 <rdopiera> o/
20:01:58 <robcresswell> if david-lyle is here then we have a full house
20:02:00 <david-lyle> o/
20:02:06 <robcresswell> speak of the devil...
20:02:13 <r1chardj0n3s> you summoned him!!
20:02:27 <robcresswell> :o
20:02:34 <ediardo> o/
20:02:39 <r1chardj0n3s> just a few things I need to mention
20:02:40 <r1chardj0n3s> #topic Ocata RC1 this week
20:02:40 <r1chardj0n3s> #link https://releases.openstack.org/ocata/schedule.html
20:02:56 <robcresswell> \o/
20:03:14 <asettle> Woah look here, we have a horizon meeting
20:03:18 * asettle waltzes in
20:03:21 <r1chardj0n3s> I'll be tagging RC1 tomorrow. That basically means a hard string freeze. No patches should be merged that change strings unless you've got an exemption.
20:03:24 <r1chardj0n3s> ohai alex
20:03:30 <asettle> r1chardj0n3s: WASSUP
20:03:50 <r1chardj0n3s> robcresswell: you've still got a patch in play under FFE - how's that looking?
20:03:55 <robcresswell> access & security reorg is done btw, pending reviews. There's probably bit I've missed/ done wrong
20:03:58 <robcresswell> https://review.openstack.org/#/q/status:open+topic:bp/reorganise-access-and-security
20:04:04 <robcresswell> thats the last 2 patches for it.
20:04:31 <r1chardj0n3s> right, I'll review those as my top priority today, hopefully nothing too outrageously bad in them :-)
20:04:41 <ediardo> I'll review 'em too
20:04:52 <robcresswell> I do have a tendency to write outrageously bad code
20:05:25 <r1chardj0n3s> yeah but as long as it's not *too* outrageously bad, you'll be ok this time :-D
20:05:28 <david-lyle> I'm looking at the floating IP one now
20:05:36 <robcresswell> floating_ips is annoyingly bloated because git won't acknowledge that the views.py is the same
20:05:56 <robcresswell> So its added a -100 +100 for a moved file -.-
20:06:01 <robcresswell> ah well.
20:06:18 <r1chardj0n3s> poor git
20:06:36 <robcresswell> Indeed
20:06:57 <r1chardj0n3s> we'll come back to those patches in this meeting if we've time, but for now I'll move on
20:07:01 <r1chardj0n3s> #topic Pike PTG planning
20:07:01 <r1chardj0n3s> #link https://etherpad.openstack.org/p/horizon-ptg-pike
20:07:08 <r1chardj0n3s> Please continue to add topics of discussion.
20:07:51 <r1chardj0n3s> Two things tho: 1) congrats robcresswell for Pike PTLship :-) and 2) I won't be attending the PTG.
20:08:10 <robcresswell> 1) :( 2) :'(
20:08:34 <r1chardj0n3s> Horizon is in good hands
20:08:57 <david-lyle> quitter
20:09:05 <david-lyle> ;)
20:09:07 <r1chardj0n3s> lol
20:09:12 <robcresswell> FYI I've added an Ocata-RC1 milestone and a ocata-backport-potential tag to LP
20:09:34 <robcresswell> Well, should add an rc2 one now I suppose
20:10:11 <r1chardj0n3s> thanks robcresswell
20:11:03 <r1chardj0n3s> looks like another name dropped off the PTG list also, gonna be an exclusive little club...
20:11:20 <ediardo> membership is quite expensive
20:11:37 <lcastell> for mortals
20:11:55 <lcastell> companies should have the money though hehe
20:12:01 <robcresswell> It'll be just be me and dave arguing
20:12:07 <robcresswell> So... business as usual :)
20:12:13 <r1chardj0n3s> anyhoo, that's the formal project-ey stuff I needed to bring up
20:12:34 <r1chardj0n3s> I've just been informed that it's Groundhog Day
20:12:38 <r1chardj0n3s> that may be important
20:12:54 <robcresswell> isnt that on the 2nd
20:12:54 <r1chardj0n3s> So, patches:
20:13:01 <robcresswell> Oh, timezones
20:13:29 <r1chardj0n3s> time travel is a thing on Groundhog Day, you know
20:13:45 <robcresswell> ofc
20:14:01 <r1chardj0n3s> apart from Rob's two patches, which should be our priority, if folks have a moment I'd appreciate some reviews on https://review.openstack.org/#/c/426984/ since it fixes a significant error in the profiler panel
20:14:09 <robcresswell> So I had an item on the agenda
20:14:15 <robcresswell> Though it seems to have been deleted
20:14:23 <r1chardj0n3s> I looked!
20:14:25 <r1chardj0n3s> oh!
20:14:31 <r1chardj0n3s> please, robcresswell, take the floor
20:14:35 <robcresswell> :D
20:14:41 <robcresswell> We need a new docs liaison
20:14:42 <r1chardj0n3s> #topic Rob's Deleted Agenda Item
20:15:06 <r1chardj0n3s> oh, *that's* why alex is here :-)
20:15:08 <robcresswell> It was me a while ago, but I've been pretty crap at keeping that up
20:15:26 <robcresswell> It would be helpful if someone could take that over and work with the docs team
20:15:27 <rdopiera> what those the docs liason do?
20:15:30 <r1chardj0n3s> what are the responsibilities of a docs liaison?
20:15:35 <rdopiera> what does*
20:15:39 <robcresswell> asettle ^^
20:15:52 <asettle> Yo
20:15:53 <asettle> So
20:16:12 <asettle> Basically you're an SME for horizon related bugs, questions, problems with the guides etc
20:16:21 <asettle> It's a relatively not-high demand job
20:16:30 <asettle> But we need someone there to access and be able to respond
20:16:51 <asettle> Previously Rob has done the role, and he has answered my PMs and pings regarding bug triaging etc
20:16:58 <asettle> Basically, you're an SME
20:17:19 <r1chardj0n3s> I tend to end up doing that thing because I'm in the rcbau IRC so I could continue to do that thing :-)
20:17:50 <asettle> In fairness r1chardj0n3s we don't need you in the rcbau wiki (that doesn't exist anymore?) we need you in openstack-doc
20:17:51 <r1chardj0n3s> so: sign me up, asettle!
20:18:01 <asettle> Thanks r1chardj0n3s :)
20:18:04 <robcresswell> sweet
20:18:06 <asettle> Will do!
20:18:06 <robcresswell> thankyou
20:18:18 <asettle> Cheers, okay, I bow out now. I'm halfway through eating my dinner and I'd like to continue ;)
20:18:27 <asettle> Peace out people :) thanks r1chardj0n3s (now, join openstack-doc and never leave ;) )
20:18:37 <r1chardj0n3s> o/ asettle
20:18:44 <robcresswell> \o/ cool thats my deleted agenda item resolved
20:18:48 <r1chardj0n3s> \o/
20:18:53 <r1chardj0n3s> #topic Patch Discussion
20:19:06 <r1chardj0n3s> so, let's look at robcresswell's two patches
20:19:25 <robcresswell> oh no, public shaming
20:19:35 <r1chardj0n3s> #link https://review.openstack.org/#/c/427882/ Move Security Groups into its own panel
20:19:40 <david-lyle> we really can't just do the git mv for views?
20:19:51 <r1chardj0n3s> #link https://review.openstack.org/#/c/425783/ Move Floating IPs from Access & Security to panel
20:19:54 <robcresswell> david-lyle: I did exactly that
20:20:01 <robcresswell> you're welcome to try again
20:20:02 <david-lyle> how many attempts?
20:20:05 <robcresswell> haha
20:20:11 <robcresswell> ...twice :(
20:20:22 <david-lyle> I hate to lose the history
20:20:27 <robcresswell> Thats why patch set 3 and 4 are almost identical
20:20:30 <robcresswell> yeah same
20:20:50 <robcresswell> Like I said, feel free to pull it down and try again. I also hate losing history
20:21:49 <david-lyle> sure get me to do your work for you ;)
20:22:25 <r1chardj0n3s> so robcresswell one of the things I did with the WIP patch to split up the tabbed Volumes panel was to leave the code in place and just move the tabs to be new panel.py (mostly)
20:22:34 <r1chardj0n3s> hardly any file movements
20:23:02 <r1chardj0n3s> https://review.openstack.org/#/c/427568/ for reference
20:23:25 <robcresswell> Doesn't that make the file structure a little confusing for new people?
20:23:53 <r1chardj0n3s> in the case of volumes it's all under a panel group still so ... a little? :-)
20:24:08 <robcresswell> I just did it tab by tab so that it was easy to review, I think its kept it all fairly consistent
20:24:12 <r1chardj0n3s> I did do the big moves but I found git had trouble tracking the files ;-)
20:24:25 <robcresswell> The only real changes are urls/file paths, and then the index view.
20:24:33 <robcresswell> everything else is just a rename.
20:25:28 <david-lyle> I'm trying the move, if it fails, your patch is fine
20:25:39 <robcresswell> anyway I'll get to any comments first thing in the morning
20:26:49 <robcresswell> thanks david-lyle
20:26:55 <r1chardj0n3s> I'll await the word from david-lyle
20:27:12 <ducttape_> that review worked well / was a nice change from my pov robcresswell
20:27:41 <david-lyle> yeah it works well
20:27:54 <r1chardj0n3s> brb
20:28:15 <robcresswell> \o/
20:30:25 <david-lyle> rdopiera: I looked into the domain scoping issue
20:30:26 <david-lyle> I see you did too
20:30:48 <rdopiera> david-lyle: I wrote two trivial patches, one of them is confirmed to at least fix the issue we ran into
20:31:01 <david-lyle> which was that?
20:31:05 <david-lyle> which patch?
20:31:10 <rdopiera> david-lyle: but not sure I got them all -- there are some uses of get_effective_domain that I'm not sure about
20:31:32 <rdopiera> https://review.openstack.org/#/c/427125/3
20:31:48 <rdopiera> this one fixes the bug with only one domain displaying on the domains list
20:32:56 <david-lyle> rdopiera: I'm wondering if it's more involved than that
20:33:15 <rdopiera> it probably is
20:33:20 <david-lyle> because if the domain_context is not set and you are domain admin, you would want to use the effective domain id
20:33:21 <rdopiera> I'm just sliding on the surface
20:33:34 <david-lyle> for the get
20:33:54 <rdopiera> but that would be the default domain, no>
20:33:58 <rdopiera> which is None anyways
20:34:06 <david-lyle> only if I'm scoped to the default
20:34:34 * david-lyle looks at the API
20:34:45 <rdopiera> ok, so get the context and if it's None, get the effective domain?
20:35:33 <david-lyle> rdopiera: I think so
20:36:17 <david-lyle> otherwise we are attempting to get the domain None
20:36:37 <david-lyle> which will likely not work very well, unless keystoneclient has some magic in it
20:36:53 <rdopiera> ok, I will update the patches to do that
20:37:06 <rdopiera> we probably want an utility function that encapsulates the logic
20:37:15 <rdopiera> any idea where to best put it?
20:38:07 <david-lyle> good question
20:38:41 <david-lyle> maybe dashboards/identity/utils.py ?
20:38:57 <rdopiera> ok, I will do that
20:39:03 <david-lyle> because there are cases where one of the panels may not be loaded
20:39:11 <rdopiera> I have one more thing, not related to the domains
20:39:31 <david-lyle> have at it
20:39:42 <rdopiera> I forgot to bump the novaclient in the requirements for the simple-tenant-usages pagination
20:40:01 <rdopiera> is it too late to still do it?
20:40:09 <david-lyle> in horizon?
20:40:16 <rdopiera> yes
20:40:32 <david-lyle> it's >= 6.0.0
20:40:42 <david-lyle> is that not high enough?
20:40:56 <david-lyle> excluding 7.0.0
20:41:07 <rdopiera> we need 7.1.0
20:41:41 <david-lyle> and otherwise we just don't use the pagination ?
20:41:54 <david-lyle> and use the support microversion?
20:42:21 <rdopiera> I'm afraid that if you have new nova and old novaclient, it will try to use the pagination and crash
20:42:48 <rdopiera> of course we never test such a combination
20:43:08 <r1chardj0n3s> upper-constraints.txt has python-novaclient===7.1.0
20:43:16 <david-lyle> hmm, you could try to get a g-r patch through, but I don't know the odds
20:43:48 <r1chardj0n3s> so if you always deploy honoring upper-constraints.txt (which you should do) then you're ok
20:45:17 <r1chardj0n3s> I think this meeting is done. Please get onto those reviews. Rob, I left you a present in https://review.openstack.org/#/c/427882 (you're welcome)
20:45:19 <david-lyle> ducttape_: do you honor upper-constraints when deploying?
20:45:50 * david-lyle wondering if u-c is just a dev tool in reality
20:46:07 <ducttape_> no
20:46:11 <r1chardj0n3s> yikes
20:46:17 <ducttape_> we do for some things, but not horizon
20:46:17 <david-lyle> it's just a dev tool
20:46:26 <r1chardj0n3s> it's not supposed to be just a dev tool
20:46:26 <ducttape_> I freeze all the reqs
20:46:39 <ducttape_> so a given deploy has exact versions
20:46:48 <r1chardj0n3s> but freezing reqs could also work, yeah
20:46:50 <david-lyle> r1chardj0n3s updating brings pain
20:46:51 <robcresswell> it probably won't affect ducttape_  as much because of edge Horizon though
20:46:57 <ducttape_> upper-reqs is generally useful though
20:47:08 <robcresswell> If you deploy Newton without u-c now you will break horribly
20:47:16 <r1chardj0n3s> yep
20:47:28 <ducttape_> yep, I know.  nova client is very exciting
20:47:39 <david-lyle> that seems counter productive
20:48:10 <ducttape_> it seems counter productive to do something like that to the client in the first place imo
20:48:35 <david-lyle> ducttape_: I meant that requirements ranges are mutually incompatible
20:48:44 <r1chardj0n3s> without upper-constraints limits, Horizon's JS dependencies break newton a whole lot
20:49:08 <david-lyle> but I will only glance into that rathole today, and walk away having spent too much time there before
20:49:09 <ducttape_> yep, the upper reqs are needed / useful
20:49:33 <david-lyle> because g-r is flawed
20:49:35 <ducttape_> I just don't use them, but I freeze everything and ship
20:49:36 <david-lyle> anyway
20:50:18 <r1chardj0n3s> thanks everyone for coming along to this meeting, catch you in #openstack-horizon :-)
20:50:22 <r1chardj0n3s> #endmeeting