16:00:40 #startmeeting Horizon 16:00:41 Meeting started Tue Sep 2 16:00:40 2014 UTC and is due to finish in 60 minutes. The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:42 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:45 The meeting name has been set to 'horizon' 16:00:48 Hello everyone 16:00:51 howdy 16:00:52 Hello! 16:00:54 hi 16:00:57 hello 16:00:59 hello 16:01:16 hello 16:01:18 hi 16:01:20 hi 16:01:22 hi 16:02:07 The tag for Juno-3 happens on Sept 4, this Thursday 16:02:09 hello 16:02:46 if things aren't approved and in the pipeline in the next <24 hours, they likely aren't making Juno 16:02:57 hi o/ 16:02:58 I do have a couple of FFEs in mind 16:03:16 and I've moved those BPs to high to indicate that 16:03:18 hi all 16:03:34 hi 16:03:54 hello all 16:04:02 #link https://launchpad.net/horizon/+milestone/juno-3 16:04:14 Blueprints: 16:04:14 18 Needs Code Review, 20 Implemented 16:04:25 that's +11 over a week ago 16:04:36 so that's great progress 16:04:48 I think 2-3 can easily land today 16:05:12 https://blueprints.launchpad.net/horizon/+spec/context-selection 16:05:27 https://blueprints.launchpad.net/horizon/+spec/horizontal-form 16:06:01 https://blueprints.launchpad.net/horizon/+spec/evacuate-host 16:06:20 all these have been reviewed several times and look ready 16:06:48 are there any others that people think are ready that just need the reviews 16:06:49 ? 16:07:06 mostly asking cores 16:07:25 as that will decide what makes it 16:08:35 https://review.openstack.org/#/c/94450/ 16:08:41 david-lyle: this is kinda new, but I think that we will need it anyways, and it's short: https://blueprints.launchpad.net/horizon/+spec/font-icons 16:08:56 This one has been sitting there for a while, would be nice if we could get some cores to take a look at it as well 16:09:00 (I know that it's not properly tagged) 16:09:37 rdopieralski: the review indicates a partial implementation 16:09:43 are we saying this is completed? 16:09:49 if so, I'll update 16:10:15 david-lyle: ive got code pending review for this bug now: https://bugs.launchpad.net/horizon/+bug/1322258 , not a bp, but should get attached to juno 16:10:17 I think further work can be in follow on bps 16:10:22 david-lyle: it introduces the icons and makes them possible to be used, but it doesn't replace any yet 16:10:37 david-lyle: it will be complete when we get rid of all old icons 16:10:39 jomara_: done 16:10:44 Hi David, can you pls approve this one https://review.openstack.org/#/c/83863/? review was done weeks ago and our neutronclient main patch was merged already. Thanks! 16:10:49 david-lyle: danke 16:10:59 here is the approved bp for it https://blueprints.launchpad.net/horizon/+spec/add-attributes-to-form-field-choices 16:11:30 mxu, as discussed in email, I cannot. I need python-neutronclient to be released first 16:11:35 jomara_: merge failed :( 16:12:09 once it is released, then I can look to approve it 16:12:30 I'm not approving potentially deadcode into Horizon 16:12:49 Hi David, python-neutronclient release is not alligned with 9/4 deadline 16:13:14 mxu, additionally an openstack/requirements patch needs to move the min version of neutronclient to the newly released version 16:13:21 rdopieralski: WHAT i rebased 10seconds before review 16:13:31 rdopieralski: (ill look at it now, thanks) 16:13:33 otherwise I have a potentially broken Horizon 16:14:01 rdopieralski: I don't see any patches to replace the icons 16:14:12 so that seems like work for Kilo 16:14:17 did I miss one? 16:14:17 david-lyle: yes 16:14:59 david-lyle: hmm, actually, if you don't merge any replacing of icons, we don't need the new icons to be available either, and the whole thing can go in in Kilo 16:15:14 david-lyle: ok, I retract my proposal :) 16:15:41 Hi David - https://blueprints.launchpad.net/horizon/+spec/cinder-qos-specs is ready to be reviewed 16:16:05 rdopieralski: I don't mind if the bp was just for getting font-icons into horizon and we have another bp to use them 16:16:23 splitting hairs, I'm fine with either 16:18:00 rhagarty: I'll try to get to that, there's a lot of good work there 16:18:06 david-lyle: ok, I will split it 16:18:19 rdopieralski: works for me, thanks 16:18:50 https://blueprints.launchpad.net/horizon/+spec/cinder-volume-transfer could also use some review love 16:19:37 https://blueprints.launchpad.net/horizon/+spec/reveal-icon-for-password-field have also been sitting there for a while, and could use some love 16:19:38 gary-smith_: acknowledged 16:20:02 so for the cinder folks, of volume transfers and QOS, which is the higher want? 16:20:13 QOS! 16:20:16 sophie's choice, eh? 16:20:35 trying to be realistic about the time left 16:20:58 probably QOS 16:21:31 tqtran: that is completed, the follow on patch should be treated as a bug 16:22:37 ok, so there's 4-5 BPs to review and see if we can get in the pipeline before tomorrow 16:22:50 I'll start bumping the rest to Kilp 16:22:54 *kilo 16:23:36 The agenda for today can be seen here: 16:23:44 #link https://wiki.openstack.org/wiki/Meetings/Horizon 16:23:59 #topic Making calls to APIs on every request (rdopieralski) 16:25:26 -Checking which panels to display 16:25:26 -Mocking it properly 16:25:56 rdopieralski: is this regarding the mailing list thread? 16:27:26 seems rdopieralski has gone missing 16:28:17 let's come back to it when he returns, because I'm not entirely sure the context 16:28:36 david-lyle: yes 16:28:37 sorry 16:28:50 so the problem that was described in the mailing list thread 16:28:53 just a dramatic pause 16:28:57 actually came back in a different patch 16:29:04 by amotoki, who couldn't come 16:29:32 the thing is, many panels need to check whether they should be displayed or not in an external API 16:29:49 that makes the API call be executed on every page rendering 16:30:02 apart from being inefficient, it's also very hard to mock 16:30:14 and it actually broke all tests in tuskar-ui 16:30:19 for instance 16:30:24 I'm wondering if we could cache the api call results by session id rather than request? 16:30:33 but that doesn't fix the testing problems 16:30:39 yes, that was our thinking with amotoki 16:30:53 and then mock that caching system in testcase's setup 16:31:46 the caching should probably be at the level of list of panels, since a lot of panels might need that 16:31:53 (and dashboards, of course) 16:32:42 still, the first request for a given user can potentially take a lot of time 16:32:59 rdopieralski: not sure I'm following that last part, rather than the API call you want to cache the accessible nav elements? 16:33:24 david-lyle: yes, basically I want to remember what menu to display 16:33:46 david-lyle: then in tests, mock that and make all of them always displayed 16:33:57 if I rescope my token, that info needs to be refreshed 16:34:08 cache invalidation 16:34:15 but that's an implementation detail 16:34:31 yeah, so maybe the cache keys are session_id and project_id 16:34:56 this sounds reasonable to me 16:35:03 I will write a blueprint 16:35:30 if we want it in Juno, it should be a High defect 16:35:46 I wonder if there is anything similar to that, that we could stumble upon in the near future, though 16:36:40 like the list of visible projects? 16:37:06 right now that list is only cached per request 16:37:56 Basiclaly anything in the "static" part of the horizon ui is a potential candidate 16:38:04 it would be a good candidate, but we can live without that until Kilo 16:38:28 rdopieralski: agreed, but there will be several invalidation triggers 16:38:35 that will need to be worked through 16:38:46 ok, so a bug for this now 16:38:55 and maybe a blueprint for a more general stuff in kilo 16:39:03 rdopieralski: sounds good 16:39:31 #topic Open Discussion 16:39:59 I would like to ask everyone to please also review the patches for all the xstatic-packages that we have 16:40:08 https://review.openstack.org/#/q/status:open+project:%255Estackforge/xstatic.*,n,z 16:40:25 right now those repos are behind what is actually on pypi 16:40:26 rdopieralski: thanks, overlooked those 16:40:55 (since at the time they were packaged, there were no stackforge repos yet) 16:41:14 there is some mess in there, but I hope we can keep it ordered from now on 16:42:29 I updated / added Horizon's JavaScript best practices. If I could get some eyes and comments on https://review.openstack.org/#/c/117595/ that would be appreciated. I also added the information to the wiki at https://wiki.openstack.org/wiki/Horizon/Javascript 16:43:09 asahlin: thanks for putting it together! 16:43:45 rdopieralski: to be clear, accepting the updates will make them the default version used in virtualenvs and devstack, however we still need to maintain backward compatibility in Horizon for the oldest version supported in openstack/requirements, this can get confusing quickly 16:44:13 rdopieralski: not so much directed at you as a reply 16:44:52 david-lyle: it's not exactly like that -- I didn't get all the automation to work in those repositories, so far I have been publishing them to pypi myself (but the openstackci user is set as owner) 16:45:01 thanks asahlin, I'll take a look soon 16:45:24 david-lyle: so they are not used until there is an actual (manual) release, and later I hope to make the releases automatic with tagging 16:45:42 rdopieralski: ok, that's good to know :) 16:45:55 but the same issue applies manual or automated 16:46:13 david-lyle: of course, for any non-backwards-compatible changes, we have to update global-requirements and requirements.txt normally 16:46:26 david-lyle: just as if those were external projects 16:46:34 rdopieralski: just trying to raise awareness of this fact 16:47:32 I hope to get some help setting all the automation in there properly in the coming weeks. 16:48:06 I'm sure once j-3 is out there will be more time for that 16:51:12 anything else? 16:52:32 Thanks everyone. Please get review the remaining list of prioritized bps. We've gotten a lot of great functionality into Horizon in Juno. 16:52:37 #endmeeting