16:02:07 #startmeeting Horizon 16:02:08 Meeting started Tue Jun 3 16:02:07 2014 UTC and is due to finish in 60 minutes. The chair is mrunge. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:02:10 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:02:12 The meeting name has been set to 'horizon' 16:02:21 good afternoon 16:02:24 hello 16:02:25 hiya! 16:02:26 hello everyone o/ 16:02:27 o/ 16:02:27 hey 16:02:30 hi 16:02:32 hi all 16:02:32 hi!! 16:02:33 everyone :) 16:02:34 hello all 16:02:35 hi 16:02:39 hello 16:02:46 hi 16:02:53 o/ 16:02:56 david lyle is currently on pto, so I'll chair the meeting 16:02:59 hi 16:03:10 https://wiki.openstack.org/wiki/Meetings/Horizon 16:03:39 #topic project news 16:03:47 any news to share? 16:03:50 o/ 16:04:27 the pyscss patch for global-requirements has been merged :) 16:04:40 The nova guys have been asking a lot about the default quota stuff they took out in icehouse. 16:04:53 hooray rdopiera 16:05:01 They must be getting a lot of grief about it, and it seems that its going back in 16:05:14 doug-fish: I've seen a patch that is supposed to restore some views, is that it? 16:05:14 good news as well 16:05:19 rdopiera: nice 16:05:57 of course this has led to questions like "Why don't you have better integration tests, so we can know this earlier?" 16:06:33 (I think this is where jpich is supposed to post the link to the horizon integration test wiki) 16:06:41 doug-fish, I assume you answered: known issue; we're working on this ;-) 16:06:49 I'm still working on the sahara merge. Next-up is adding some tests for each of the panels. 16:06:49 yep 16:07:08 doug-fish: :) 16:07:39 crobertsrh, yes, you have a lot of patches up for review, progress with reviews is still a bit slow 16:08:06 Yes, reviews have been a bit slow to trickle in, but those that we have got have been helpful. 16:09:00 let's move to the next topic 16:09:01 another kind of news about stable update. 16:09:08 stable icehouse update is now soft freeze. if any important backport, please send a mail to stable-maint. 16:09:22 oh yes 16:09:56 let's move on. 16:10:07 #topic Horizon split 16:10:28 rdopiera, I assume, you brought this up? 16:10:32 yes 16:10:46 https://etherpad.openstack.org/p/horizon-split-plan is the linked document 16:10:53 I just wanted to mention again that the plan is on the etherpad, at https://etherpad.openstack.org/p/horizon-split-plan 16:11:01 and that anybody is free to add to it 16:11:10 or raise concerns, etc. 16:11:25 another thing that we have to do before the split, is decide on the name 16:11:40 it has been raised on the mailing list that django_horizon is not the best choice 16:11:53 IMHO we should reach out to the mailing list for proposals 16:11:59 I wonder what the most effective method of picking the name would be 16:12:06 I agree here 16:12:32 so, an etherpad with the names, and people putting their names next to the ones they like? 16:12:32 call for proposals? make a list of 3-5 names? 16:13:02 we could have some an online poll after we get some name proposals? 16:13:02 and a week from now we count them, and if there are ties, we just pick randomly? 16:13:24 i highly doubt there will be a tie 16:13:38 I assume, we could make a poll? does anyone know, how other projects decided about a name? 16:13:39 tqtran: just in case, to make it quicker :) 16:13:44 neutron folks here? 16:13:47 if there is, we can have david flip a coin 16:13:56 tqtran: exactly what I meant 16:14:08 mrunge: ? 16:14:26 amotoki, how was the name change to neutron decided? 16:14:29 then we will potentially need a n-sided coin :) 16:14:34 I would hate to waste a week for gathering the proposals, and another for voting 16:14:51 quantum - (something I forgot) -> neutron 16:14:52 we voted a couple of candidates. 16:15:00 2 days for proposals, and until the next meeting for voting, maybe? 16:15:23 only core member voted at that time. 16:15:34 rdopiera, could you send a mail to the list about name proposals? 16:16:01 I suggest "mirokolitsa". No one uses this name before, that's win-win! ^^ 16:16:09 mrunge: should we collect them on an etherpad or wiki? because tracking them on e-mails would be error prone 16:16:20 #todo rdopiera to send a mail to the mailing list, asking for naming proposals 16:16:22 do we want other openstack projects to vote on the name too? 16:16:22 tmazur: lol what does it mean? 16:16:47 rdopiera, etherpad or wiki: works both for me 16:16:57 sorry for the noob question, where to subscribe to the mailing list? I looked around and couldn't find it 16:17:05 That's "horizon" in fake proto slavic :) 16:17:18 AFAIK, in neutron cases, we have candidate names without copyright problem. 16:17:21 tmazur: i like! haha 16:17:44 what i am not sure is do we need to check copyright issue before name vote. 16:17:46 gary-smith, http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev 16:17:48 gary-smith: You can find more information over there -> https://wiki.openstack.org/wiki/MailingLists#Future_Development 16:17:56 amotoki: you mean trademark 16:17:59 gary-smith, http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev 16:18:05 ah.... right. 16:18:06 thanks, got it 16:18:35 amotoki: do you know how one would go about it? 16:18:47 amotoki: does openstack have any lawyers who would help? 16:18:59 the foundation does 16:19:10 rdopiera: i am not sure. markmcclain, former neutron ptl, is the right contact. 16:19:29 If you need it, I can send a mail to ask. 16:19:33 otoh, if we use something like "horizon-lib", we can be pretty cofident it's safe 16:19:53 amotoki: that would be awesome 16:20:13 rdopiera: sure. 16:20:14 on the other side, it doesn't need to have horizon in the name 16:20:36 mrunge: right 16:20:40 as it should be more generic and consumable from the outside, totally unrelated to openstack 16:20:41 I just like boring names 16:20:57 something like django-generic (or so) ;-) 16:21:06 "skyline" is taken already, unfortunately :( 16:21:16 "firmament" 16:21:23 what's beneath a horizon? an ocean or seabed 16:21:27 nadir 16:21:49 proposals to the wiki or etherpad 16:21:53 yup 16:22:03 I will send the e-mail shortly 16:22:11 I think we are done with that 16:22:11 ok, then we could start a poll next week? 16:22:12 "sunrise" or something along the line of horizon? 16:22:29 mrunge: I wuld want to start it before the weekend 16:22:39 mrunge: and have the name decided on the next meetings 16:22:55 rdopiera, start the poll before next weekend? 16:22:56 but maybe that's unrealistic 16:23:00 might be a bit tight 16:23:04 yeah 16:23:05 ok 16:23:10 so poll next week 16:23:15 yes 16:23:18 which brings me to the next topic 16:23:22 dates for the split 16:23:31 yes please 16:23:33 skywave? 16:23:34 :) 16:23:48 j-1 will be cut on june 12th 16:24:05 I'm still waiting for some patches to merge, and we will have to depend on some more patches for the requirements 16:24:06 I think we should do the split shortly after 16:24:50 mrunge: I'm not sure we will be ready 16:25:09 mrunge: also, how long in advance do we need to set the date for the freeze? 16:25:22 and what is the latest date when it still makes sense 16:25:50 by June 12 we still won't have the name, for example 16:25:52 rdopiera, not putting any pressure, but j-3 would be too late (imho) 16:26:01 mrunge: when is that? 16:26:12 rdopiera, good question 16:26:23 anyone? 16:27:31 https://wiki.openstack.org/wiki/Juno_Release_Schedule 16:27:47 around Sept 4th 16:28:12 thanks 16:28:20 ok, so we have to do it before that 16:29:34 can someone explain the process of how this split works. For example, once this split happens how do we coordinate the changes that are still in review that depend on the old code structure? 16:29:36 yes 16:30:11 johnma: the patch queue stays with the current openstack_dashboard 16:30:29 johnma: so any changes that touch the current horizon part will become invalid 16:30:41 johnma: and will have to be broken up and re-submitted 16:30:44 in fact, most patches in the past have been for openstack-dashboard 16:31:29 the /horizon part is mostly a library 16:31:40 so the changes there are more general and "global" 16:31:51 rdopiera, do we have an overview on what needs to get merged before? 16:31:55 it makes sense to make them separately 16:31:56 aah ok. Thanks for clarifying that. 16:32:21 mrunge: I will add links to patches to the plan, good idea 16:32:33 great, thank you 16:32:40 perhaps we need to clarfiy the detail process on etherpad and review it in the meeting. 16:33:02 amotoki, good idea 16:33:16 anything else to add? 16:33:32 does it make sense to have a separate launchpad project too? 16:33:52 hmm, I don't think so 16:34:10 thoughts? 16:34:33 mrunge: if it will be used by other projects, they will want to report bugs and stuff 16:34:38 a person from the outside can't know, which component is failing/causing issues 16:34:42 framework vs content related? Right now, we do have a separate project for openstack_auth. 16:35:08 I think it may make sense to have separate launchpad 16:35:25 for the library/framework 16:35:43 would a tag work as well? 16:35:56 a tag #framework or #ui ? 16:36:14 or name-of-the-project-that-we-will-come-up-with 16:37:05 I mean, that would avoid copying bugs from there to there 16:37:48 for launchpad we can migrate to a separate launchpad project gradually. At first most reports should be from openstack-dashboard. 16:37:49 mrunge got a point... From user perspective, it does make it confusing figuring out which project to log the bug or request. 16:37:53 I second the idea of using a tag with the new name as the tag name 16:38:32 its easier that way to track bugs and bp 16:38:34 i also think it might be hard if a bug spans both projects 16:38:47 then where do you report it? two different reports? thats confusing.... 16:38:55 we don't really get bugs logged in openstack_auth right now, mostly bugs comes in horizon project and moved to openstack_auth after triage. 16:39:04 that's the reason, why we're moving away from launchpad 16:39:06 tqtran: Launchpad lets you open tasks so you can say a bug affects 2 projects within the same report 16:39:11 tqtran: we can add multiple projects for one bug. 16:39:36 oh right.... 16:39:40 it is a common way to track a bug across multiple projects in openstack. 16:40:43 can we move to the next topic? 16:40:57 #topic django-angular.js 16:41:24 who brought this up? 16:41:37 i think eric did, but hes not here.... 16:41:46 i'll speak on his behalf =) 16:42:21 thanks 16:42:26 so the current plan is to provide an alternative way for rendering, mainly client-side using angular 16:42:38 it will hook nicely into the current django implementation 16:42:57 and overtime, we can have people jump in to retrofit the django version wth the angular version 16:43:26 eventually, we end up with everything rendering on client-side, then we can go back and clean out the django rendering logic 16:43:52 that's about: why we're investigating angular, right? 16:44:00 Link to the current discussion on list as well, started by Veronica Musso: http://lists.openstack.org/pipermail/openstack-dev/2014-June/036488.html 16:44:03 how comes django-angular into the game? 16:44:22 the question is, does this make sense? or are people oppose to removing the django implementation over time? or do we want to keep 2 rendering path for new developers? 16:44:57 tqtran: I brought it up, because I've seen the discussion on the ML 16:45:11 but I don't really have anything to say about it 16:45:13 sorry 16:45:20 ah...=) 16:45:37 I thought it might just be good to compare notes 16:45:43 and agree where we stand 16:46:13 so, this topic is about django-angular (the python package), right? 16:46:13 * jrist is sitting 16:46:30 so tqtran any thoughts about that? 16:46:32 sorry, late comming 16:48:06 i would have to read more about it 16:48:31 I am not sure django-angular is critical, but would likely be a good first thing to use. just having a json centric web layer would be helpful.... no matter if it's angular or something else 16:48:48 yes, that is my impression as well 16:49:26 and angular seems to have a lot of support among current devs too, which never hurts 16:49:38 sorry guys, I need to run out right now! 16:49:46 thanks mrunge! 16:50:43 anything else I could help on the subject? (apologies for the late arrival) ? 16:50:45 Looks like the conclusion on django-angular is "more reading required"? Was there any other items on the agenda? 16:51:23 ericpeterson: Did you see the discussion on list? Your input would likely be appreciated there as well 16:51:38 http://lists.openstack.org/pipermail/openstack-dev/2014-June/036488.html 16:51:57 #topic Open discussion 16:52:00 I see "Use mock?" in the agenda. Anything to discuss? 16:52:11 Nope, not working. I think only the chair will be able to end the meeting too 16:52:40 amotoki: yes, that's me too 16:52:42 amotoki: I think it was a question about the current status which I believe is "feel free to use mock in future tests, no need to rewrite all the current ones, mox3 should work with python 3 in the future"...? 16:52:59 the thing is, mock is not added to requirements.txt in Horizon 16:53:03 so can we add it? 16:53:18 jpich: it is like that in Nova 16:53:22 rdopiera: I guess so? I don't know where the current OpenStack direction is documented 16:53:23 jpich: but not sure about Horizon 16:53:24 IMO it is better to avoid two mocking methods in a single project. 16:53:55 the thing is, mox is not really suitable for unit tests 16:54:00 amotoki: mox is confusing enough on its own alright 16:54:05 more for acceptance testing and bigger stuff 16:54:23 it forces you to check everything 16:54:34 while with mock, you can just check the things you care about 16:54:40 and not lock the implementation 16:54:45 I assume mock is something we'd add to test-requirements.txt right? and not necessarily require it for production deploys 16:54:54 doug-fish: yes 16:54:59 super 16:55:03 jpich: rdopiera: totally agree with you. I have experince both. 16:55:06 it's also in the stdlib in recent pythons 16:55:17 Does someone fancy starting the discussion the list again, asking for the current status for mocking frameworks in the wider project and where we are documenting this? 16:55:25 *where it is documented 16:55:35 jpich: ok, I will do that tomorrow 16:55:40 good idea. 16:55:44 rdopiera: Thank you 16:56:26 Anything else? 16:56:26 is there anything we should talk about? 16:57:06 Just a minor issue - I think we had discussed using a "docimpact" tag at some point in the past 16:57:10 but I don't think we have been 16:57:19 and I can't recall exactly how we were supposed to use it 16:57:34 anyone of a link to a wiki describing what we intend? 16:57:49 Right... We should summarise what was discussed in a wiki page 16:58:07 https://wiki.openstack.org/wiki/Documentation/DocImpact 16:58:10 There is the generic page here: https://wiki.openstack.org/wiki/Documentation/DocImpact 16:58:13 Thanks amotoki 16:58:31 and we need more info about how we are using it in Horizon? 16:58:39 commits which affect documentation team should have DocImpact flag. 16:59:12 IMO commits which only affects developer docs don't need DocImpact flag. 16:59:18 amotoki: Because nearly every commit affects the UI in some way it's unclear what requires a DocImpact tag and what doesn't 16:59:35 We document settings in the dev docs that really should probably belong to the wider project documentation, it is useful to deployers 17:00:13 it is different from what other projects do. 17:00:23 our time is over... 17:00:24 Ok, tbc on that one, I'm not 100% clear on what is appropriate to flag or not either, I would need to reread the logs - I think annegentle had popped up to clarify a few things during that meeting 17:00:27 Right 17:00:28 #endmeeting 17:00:37 bye 17:00:41 thanks all 17:00:42 bye all! 17:00:43 fungi: We lost the chair, any chance someone in infra could help us force-end the meeting? 17:00:49 bye everyone 17:00:54 bye all! 17:00:54 I am afraid that only chair can send *endmeeting*..... 17:01:07 o/ 17:01:21 amotoki: try it :) 17:01:24 bye bye 17:01:36 bye all! 17:01:44 bye 17:01:52 Anyone here for Congress? 17:01:56 hi this is kudva 17:02:00 Hi thinrichs 17:02:02 i here :) 17:02:06 kudva, skn, banix: hi! 17:02:22 we are still in Horizon meeting topic 17:02:24 where is pballand 17:02:29 hi 17:02:29 thinrichs: Error: Can't start another meeting, one is in progress. Use #endmeeting first. 17:02:33 jpich: it will be endable after an hour 17:02:50 which is right about now 17:02:55 #endmeeting