20:01:26 #startmeeting Horizon 20:01:28 Meeting started Wed Dec 17 20:01:26 2014 UTC and is due to finish in 60 minutes. The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:01:29 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:01:31 The meeting name has been set to 'horizon' 20:01:43 hello 20:01:47 hi 20:01:47 o/ 20:01:52 hi 20:01:54 o/ 20:01:55 hola 20:02:00 o/ 20:02:04 *\o/* 20:02:09 Hi 20:02:10 sorry. g'day 20:02:42 Evening 20:02:49 Hello everyone 20:03:31 We are in the final hours of Kilo-1 20:03:44 Let's take a quick look 20:04:16 #link https://launchpad.net/horizon/+milestone/kilo-1 20:04:52 two items left after I bumped a couple to kilo-2 20:05:09 https://review.openstack.org/#/c/104063/ 20:05:26 has been around since Juno 20:05:33 but we didn't have time to get that in 20:05:38 the other 20:06:02 https://review.openstack.org/#/c/110888/ 20:06:37 I found languishing in the review queue when I was going to purge old patches, seemed useful and needed a little straightening 20:07:07 any way if the first made it, that would be nice, but if it doesn't start merging soon, I'll push to k-2 20:07:33 david-lyle: do you know if anything is wrong with zull. 20:07:35 zuul 20:07:52 we only have an hr or so here 20:08:04 ha 20:08:04 20:08:27 it doesn't seem to be picking up that patch for testing. 20:08:45 well the queues are quite deep 20:08:53 129 in check and 45 in gate 20:09:02 yay rush 20:09:19 so I'm going to just push the other two to k-2 and pass on the SHA on master for k-1 20:09:20 TravT, I have a cinderclient build...it finished...but doesn't seem to get out of the queue.. 20:09:47 tomorrow would be a great day to have the metadata admin functionality added to Horizon 20:10:01 that seems like the soonest 20:10:04 any way 20:10:36 https://launchpad.net/horizon/+milestone/kilo-2 is quite large 20:11:12 david-lyle: 20:11:15 like all milestones some of that will likely slip as well 20:11:16 now I see it. 20:11:17 TravT: 20:11:25 lol 20:11:26 :) 20:11:37 took forever for zuul browser page to load 20:11:50 anyway milestones, yay 20:12:25 the agenda for today can be found at #link https://wiki.openstack.org/wiki/Meetings/Horizon 20:12:43 #topic packaging Thunderdome 20:12:50 fight 20:13:25 uhm, rdopieralski is on pto for the rest of the year. 20:13:29 so I thought that we had pretty much reached a point where we could discuss a concrete plan; zigo is still a little nervous but I attribute that to unfamiliarity with bower 20:13:45 so, maybe we should defer this to next year? 20:13:50 I'm not likely to be able to do anything until next year anyway 20:14:19 so yeah, I guess so. Thunderdome postponed I suppose (the crowd *will be disappointed*) 20:14:25 booo 20:14:34 * david-lyle throws rotten fruit 20:14:46 and btw. there shouldn't be a reason to add another package manager.... 20:14:48 but I think there's solid support from everyone else? :) 20:15:08 I guess, we're still not convinced 20:15:27 r1chardj0n3s is there any pre-reading for the Tunderdome we can do? 20:15:29 I think we really need to schedule something with the appropriate parties outside of the Horizon meeting and hash it out 20:15:38 we need to get something resolved, because we need some new packages for angular development 20:15:41 david-lyle: that makes sense 20:15:43 mattfarina: just a small email thread 20:15:47 will there be a fight in said meeting? 20:15:56 david-lyle i'll finish it some day... maybe 20:16:21 A fight can only take place, when folks are physically at the same place 20:16:22 mattfarina: yeah, that thread that hit a measly 120 messages ;) 20:16:29 alright, let's table that until we have a quorum 20:16:33 deferred till Vancouver then? 20:16:35 yep 20:16:36 mattfarina: i read it. i recommend bringing something to gouge your eyes out halfway through. 20:16:51 Hmmm... my email filters are failing me. 20:16:59 mrunge: gods no, we should be able to figure it out without postponing *months* 20:17:14 TravT I didn't finish it because i got more than halfway though 20:17:36 r1chardj0n3s, I agree! 20:18:47 http://lists.openstack.org/pipermail/openstack-dev/2014-November/050093.html 20:18:55 in the meantime, if we can get xstatic for newer angular pieces through. we need them for launch instance. 20:19:02 may be the email that killed hope 20:19:16 #topic REST API: should we be rewriting property names? 20:19:33 david-lyle: Thankyou. Going to have to alter my mail filters -.- 20:19:35 vague and mysterious a good title 20:19:50 unfortunately, tqtran isn't here to defend the practise he proposed of rewriting the property names :/ 20:20:07 can you provide a point of reference? 20:20:15 so the current keystone REST API for the angular work rewrites "project" to be "project_id" for example 20:20:21 some of us have been happily sleeping 20:20:30 but that turns out to be fraught, very very fraught 20:20:36 seems like a bad idea to change the names 20:20:38 why would we do that 20:20:41 yeah 20:21:01 haha, I think that reaction is your answer to this topic... 20:21:09 right :) 20:21:18 that's like refactoring all the horizon code in a private branch to make it prettier 20:21:30 I'll PM tqtran directll to nut it out with him 20:21:32 then trying to merge regularly with master 20:21:51 r1chardj0n3s: how's this compare to the more passthrough approach I put into the glance rest api? 20:22:03 * TravT still needs to revise it a bit further. 20:22:08 yeah, let's avoid renames 20:22:14 +1 20:22:20 TravT: basically we end up with what you've written :) 20:22:49 ok. i only have to do a tiny bit of massage on it because of v1 glanceclient 20:23:08 if we move to v2 glanceclient, i can drop even more 20:23:52 a clear example of the rewriting is right at the bottom of https://review.openstack.org/#/c/139532/15/openstack_dashboard/api/rest/keystone.py,cm - project/project_id 20:24:42 anyway, gonna kill it off 20:25:22 I 'd be curious to hear tqtran's thoughts behind it, he was defending it? My immediate reaction would be to stop doing it, though. 20:25:48 it was his idea ;) 20:26:08 I think it mostly came out of the front-end work 20:26:24 and the confusion around things that are ids not being clearly labelled as such in the API 20:26:46 r1chardj0n3s: I think we should try to go for as much passthrough as possible. 20:26:54 are we expecting that we'd be passing objects? 20:26:54 TravT: agreed 20:27:07 what's the other choice than id 20:27:16 name, which is not unique 20:27:29 david-lyle: the ids inside objects are not labelled whatever_id 20:27:40 and those objects are passed around, yes 20:27:48 also, if he refactored the calls to the api into a service that is injected into the controller, then he could make the function params in the service say "project_id" 20:27:50 as parameters? 20:27:57 (*json* objects that is) 20:28:07 yes 20:28:44 ok, need to catch up non k-1 reviews 20:28:56 #topic REST API: Death by a million dependencies. 20:29:03 that one is mine 20:29:04 there's enough pain dealing it the is-it-a-project-or-tenant mess without also dealing with is-it-project-or-project_id 20:29:16 brb, sorry 20:29:32 on behalf of tqtran, r1chardj0n3s, other 20:29:38 re: https://review.openstack.org/#/c/136676/40 20:30:13 that is a decorator richard wrote that we're using in a few places. we're all really wishing we could get a base version landed 20:30:20 and then iterate on it 20:31:32 . 20:32:07 +1 20:32:10 I agree we need to start landing some the dependencies 20:33:17 did tihomir ever continue the conversation on the ML to reviews? 20:33:59 no. he hasn't added to the specific reviews, but some of the property re-writing / passthrough speaks to his points. 20:34:20 in any case, I'm not sure the base decorator is contentious. 20:34:58 sure 20:35:40 so reviews folks, please. says the most guilty 20:36:13 #Open Discussion 20:36:19 #topic Open Discussion 20:36:27 messy words 20:37:15 I've updated #link https://wiki.openstack.org/wiki/Horizon/Blueprint_Reviews with some more blueprints to review 20:37:47 If you have a blueprint you are championing and don't see it in a milestone or on that list, target it to a milestone, so that I see it 20:38:22 I can only sift through so many before my mind starts swimming 20:38:37 * david-lyle need to obsolete so many blueprints 20:38:49 Is this list only BP tageted for kilo-2? 20:39:10 https://blueprints.launchpad.net/horizon/+spec/volume-type-description is targeted for kilo-2, but not on your review list...the status you marked is review 20:39:30 r1chardj0n3s: I'll review the REST API patch tomorrow and ask colleagues to do the same - needs more attention. 20:39:41 robcresswell: ok thanks 20:40:05 Let me know when I can mention some UX stuff 20:40:11 go 20:40:26 Need feedback on the table pagination designs from Chris. http://invis.io/C71W9EAEX 20:41:00 If you haven't already, please reach-out to me if you want an account in Invision to review mocks 20:41:14 please note, you can turn on adding comments directly to the design 20:41:23 by toggling switch on bottom right 20:42:05 Just send an email to pkruithofr@gmail if you want an account 20:43:10 * TravT wonders how much spam Piet is about to get since this IRC meeting is logged. 20:43:14 david-lyle: could you please add https://blueprints.launchpad.net/horizon/+spec/volume-type-description to your review list? Thanks 20:43:23 If you have time, please complete the online card sort for end users or forward the link to someone else http://ows.io/os/0v46l867 20:43:31 gugl2: was just looking at it 20:43:43 david-lyle: k...thanks :) 20:43:52 and yes 20:43:54 We are running usability next week on the Launch Instance workflow 20:44:23 I think that's it from a UX perspective. Thanks to everyone for your feedback! 20:44:54 Thanks Piet 20:45:10 np 20:47:42 I will be offline the 19-27 Dec. I think most people will be for some portion of that or more. 20:47:54 additionally, no meeting the next two weeks 20:48:20 david-lyle: Can you also add https://blueprints.launchpad.net/horizon/+spec/admin-network-configuration to the review list? It is currently targeted for kilo-3, but would like some feedback if anyone feels this is the right direction / a good addition. 20:49:13 asahlin: done 20:49:23 thanks 20:51:25 anything else? or we can end early 20:51:36 good idea :D 20:52:27 Alright, I'll take end early. Have a great couple of weeks everyone. Thanks! 20:52:31 #endmeeting