20:00:57 #startmeeting horizon 20:00:58 Meeting started Wed Nov 2 20:00:57 2016 UTC and is due to finish in 60 minutes. The chair is r1chardj0n3s. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:01:00 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:01:02 The meeting name has been set to 'horizon' 20:01:03 yeah bot must have disconnected 20:01:04 \o/ 20:01:06 woohoo 20:01:11 hello folks :-) 20:01:28 * ducttape_ better late than never 20:01:42 OK, got a few things to mention 20:01:52 #topic Selenium has been removed from the gate (was non-voting) 20:01:59 #link https://review.openstack.org/#/c/391342/ 20:02:11 These jobs were a burden on the gate systems and since they are broken and not voting, there is no reason for them to run. 20:02:29 We still have the non-voting check jobs to test out fixes. 20:02:39 sounds good! 20:02:51 +1 20:02:56 I think I saw tsufiev already has a patch up to test out Firefox version shenanigans 20:03:40 r1chardj0n3s, had a patch already for pinning marionette, and recently uploaded single browser fix 20:04:02 cool, will keep an eye on them then 20:04:03 now we've another occurrence of an old issue - mozilla ftp is not responding in time 20:04:13 oh :-/ 20:04:18 thus our pinned firefox is unable to download 20:04:36 will recheck it shortly after to see if it's a temporary issue 20:04:47 but it still is downloadable by hand 20:05:02 just like the phantomjs downloading, we need that thing to be cached or something. 20:05:13 wait, we are downloading and installing firefox for every test? 20:05:25 rdopiera, nope, just for the whole suite 20:05:34 so for every patch 20:05:36 doing it for every test would be too much even for us 20:05:42 yep, that's true 20:05:49 for every patch it is 20:06:24 ok, well, we will keep an eye on it, and how about we talk to #infra about possible caching options for the firefox download? 20:06:28 rdopiera, comparing to all the stuff happening during devstack deploy (for every patch) that's not a big dieal 20:06:42 r1chardj0n3s, it would be awesome if they agreed 20:06:52 if we could add it to the base image used for testing, that'd be nice 20:07:11 #action Richard to talk to #openstack-infra about browser install file caching 20:07:20 (I think #action is a thing ;-) 20:07:34 Does infra have its own requirements mirror, or just use pypi? 20:07:44 yup, it is 20:07:44 robcresswell: yes, we mirror pypi 20:07:48 r1chardj0n3s, well, may be we need a special builder macros, something like legacy-firefox 20:08:01 there is already 'firefox' builder macro 20:08:10 tsufiev: I'm sure we can sort out the detail with them :-) 20:08:20 cool. I've been mirroring things locally too because otherwise testing out multiple tox things takes forveerrrr 20:08:35 * robcresswell is off topic 20:08:56 robcresswell: oh, yes, definitely run your own local mirror, I shall provide clues for that later if you like :-) 20:08:59 ok, moving on 20:09:02 #topic Summit recap 20:09:03 * tsufiev thought for a moment about containers hotness 20:09:08 #link http://lists.openstack.org/pipermail/openstack-dev/2016-November/106520.html 20:09:14 Please have a read! 20:09:24 And discuss, if you find anything controversial :-) 20:09:27 r1chardj0n3s, that was a very nice recap :) 20:09:59 thanks tsufiev, I also hope it was accurate :-) 20:10:30 #topic New Keystone/Horizon meeting 20:10:39 The kickoff for these meetings will be 8th November 2000UTC in #openstack-meeting-cp 20:10:47 \o/ 20:10:50 every week? 20:10:52 ohai stevemar :-) 20:10:53 -cp? 20:11:01 cross-project? 20:11:02 ah, x-project 20:11:18 not openstack-meeting-copy 20:11:19 yeah, should have named it *-xp :) 20:11:20 :) 20:11:23 It's a temporary channel until we agree on a more permanent meeting time/place 20:11:44 :o need logs 20:11:52 yes, it's a logged channel 20:12:00 oh, neat 20:12:08 TIL 20:12:13 tsufiev: r1chardj0n3s we ended up with a list of things that need to be fixed in keystone/horizon and i imagine the meetings will go on until we've got patches up for all the broken things :) 20:12:25 yep, what stevemar said :-) 20:12:40 they aren't broken, they are opportunities 20:12:50 haha, I guess we're going to have these meeting for a long time ;) 20:12:51 ;) 20:12:53 newest d_o_a release has fixed token revocation btw. 20:12:58 I've not tagged new stable yet. 20:13:57 #topic Priority patches for review 20:14:28 Our priorities discussion at the summit ended up with a number of patches already in-flight that would be super awesome to land by Ocata-1 20:14:37 ... which is just under two weeks away :-) 20:14:47 https://review.openstack.org/#/c/389866/ is a simple one, if I can advocate for ediardo 20:14:49 wow that comes round fast 20:15:01 For your convenience, I have tagged them in gerrit for you: 20:15:03 #link https://review.openstack.org/#/q/starredby:r1chardj0n3s%20AND%20status:open 20:15:22 :) 20:15:30 * robcresswell feels the pressure 20:15:58 Guys, I'll need extra eyes on this patch that 'tries 20:16:18 to fix a few bugs in magicsearch 20:16:30 https://review.openstack.org/#/c/388200/ 20:16:58 Of course please do look at bug patches also, those ones I've starred are the features / ground work for the priorities we've tagged for Ocata goals 20:17:36 There is also ui-router, which I need to get moving through the xstatic packaging machinery and into requirements 20:17:38 #link https://blueprints.launchpad.net/horizon/+spec/ui-router 20:17:59 (and I also need to update some POC code, I think) 20:18:23 Made myself Approver so I can hassle you about its progress 20:18:30 thanks robcresswell :-) 20:18:35 would that help with extensibility issues ? 20:18:40 seems like maybe 20:18:59 If anyone has a patch that they feel is missing from the starred list please let me know. 20:19:23 ducttape_: ui-router is necessary for more complex angularjs panels with nested information in them 20:19:38 like Swift UI with container -> objects 20:19:46 sure, I just didn't know if it might also provide extension hooks to change certain behaviors 20:20:07 that's an interesting angle I hadn't considered 20:20:12 * ducttape_ needs to stay on topic, sorry 20:20:17 I have this one that's been around for a while https://review.openstack.org/327224 (server side filtering for swift) 20:20:33 ducttape_: please do mumble such things in the BP whiteboard :-) 20:21:13 lcastell: we will definitely be looking to land that in Ocata, though I have been talking to swift folks about getting more useful filtering than just prefix 20:21:33 yeah, right now is pretty lame :( 20:21:48 still useful though 20:21:49 r1chardj0n3s: I notice you said re panels again... is there any change to preference re people developing panels vs plugins? 20:22:16 betherly: we still prefer for non-"core" services to be plugins 20:22:18 you will get a mix of both betherly 20:22:41 we definitely like people to be working on both 20:22:53 ok cool. thanks! just wanted to check :) 20:23:06 if you are doing a plugin, chances are high it also contains a panel or 2 20:23:22 I smell sarcasm 20:23:39 none from me 20:23:48 :p 20:23:58 never 20:24:00 #topic Priority bugs 20:24:02 nor me 20:24:08 #link https://bugs.launchpad.net/horizon/+bug/1624743 20:24:08 Launchpad bug 1624743 in OpenStack Dashboard (Horizon) "Project image table: admin user sees images which are not shared with me" [High,In progress] - Assigned to Brad Pokorny (bpokorny) 20:24:25 Oh, we should remove bpokorny probably 20:24:38 (this is a short list, because I ran out of time ;-) 20:24:43 though it seems like he is lurking in the meeting if my autocomplete is right 20:25:01 yep, this one needs more eyes/brains on it 20:25:14 I think this is really a glance bug too 20:26:35 I haven't looked yet, but will do 20:26:51 I've also not had a chance to look into it in any detail 20:27:12 Also we should really fix the default LI source, its changed to volume snapshot or something for some reason 20:27:29 weird 20:28:39 #topic Open Discussion 20:28:51 robcresswell you wanted to have a local pypi mirror 20:29:01 https://review.openstack.org/#/c/389352/ << Patch for Horizon 10.0.1 20:29:02 I recommend http://doc.devpi.net/latest/quickstart-pypimirror.html 20:29:14 * ducttape_ if only anyone knew how to setup and run pypi 20:29:14 r1chardj0n3s: Excellent, thats what I've been using 20:29:20 hahaa 20:29:30 :p 20:29:36 that would be a handy person to have around, ducttape_ 20:29:37 there is also devpi 20:29:44 by the author of tox and pytest 20:29:49 rdopiera: that's what I linked ;-) 20:29:55 * robcresswell wonders if rdopiera can read 20:29:59 ouch 20:30:01 indeed 20:30:03 I'm blind 20:30:07 haha 20:30:13 forgot to tell you 20:30:17 :-) 20:30:35 I've got post conference illness so I'm just grumpy 20:30:40 :-( 20:30:49 shows the advantage of js over python -- all those indentations are difficult to touch-read 20:30:51 jamon-flu 20:31:08 I'm having a long weekend this coming weekend, so I'm away from tomorrow on 20:31:16 it's PAX AU weekend :-D 20:31:32 I think I will have a long weekend due to a sick-leave, if this continues 20:31:41 but it means I won't be around much after today until next week. 20:32:09 I forsee many reviews ahead 20:32:10 "today" being Thursday in my timezone 20:32:21 ah, the future. 20:32:34 good to know, let's prepare to merge all the stuff we don't want you to see 20:32:35 in the future, breakfast is needed 20:32:37 if you need anything done yesterday, lemme know 20:32:55 quotas? 20:32:58 heh 20:33:02 touché 20:33:18 incidentally, quotas planning has been postponed due to flu 20:33:19 I think I will hack up the tests for those things more significantly 20:33:20 btw 20:33:39 I mean that landing page pagination stuff 20:33:53 what's about throwing away all existing quota tests first, ducttape_? 20:33:55 On a serious note, we really do need eyes on those XStatic updates. This cycle is short already, and for stability sake we cant really merge important libs past b2 20:34:08 would be less work to do ;) 20:34:08 yep, what robcresswell said 20:34:19 yep, lets get those merged / in and find all the bugs 20:34:29 https://review.openstack.org/#/q/project:^stackforge/xstatic.*+status:open,n,z 20:34:37 ^-- useful link 20:34:47 tsufiev: I will likely change the tests so they look like what our future vision would be, not the old confusing past with quotas 20:34:54 rdopiera: stackforge? 20:35:02 actually, if throwing them away allowed us to rewrite tests using mock instead of mox, that would be really good 20:35:05 https://review.openstack.org/#/q/project:^openstack/xstatic.*+status:open,n,z is corrected link 20:35:10 ohai 2014 :p 20:35:16 argh 20:35:24 explains why it was always empty 20:35:27 :-D 20:35:35 haha 20:35:49 make typo. no reviews. job done. 20:35:52 * r1chardj0n3s notes those are already in the priorities starred list ;-) 20:36:11 yeah, I was just emphasising the point 20:36:19 emphasizing for those in the US 20:36:21 that's cool :-) 20:36:39 ok, any other business? 20:36:46 or can I go have breakfast? 20:36:50 most of them I've updated MAIN too, so we can steadily remove those hacks from our codebase 20:36:58 I'm all done, you may eat 20:37:09 \o/ robcresswell for more MAIN 20:37:24 thanks all 20:37:30 by the way, what is there to review in an xstatic 20:37:43 other than metadata and comparing the files with the upstream repo? 20:37:54 rdopiera: I tend to install it and make sure horizon doesn't fall about laughing, mostly 20:38:00 I'd think you want to download patch and deploy it rdopiera 20:38:03 and then double-check the metadata is sane 20:38:22 r1chardj0n3s: how do you check horizon? 20:38:23 reviewing line by line is going to be painful / not helpful I'd think 20:38:26 r1chardj0n3s: smoke tests? 20:38:40 grenade? 20:38:42 install it locally and run tests, then just try it 20:38:49 there's the JS unit tests for one 20:38:49 rdopiera, that what some used integration tests for 20:38:57 yep, what robcresswell said 20:39:09 and any big JS like bootstrap, jquery, angular, will likely console log as soon as its evaluated 20:39:30 ok, so check the logs 20:39:39 by the way, we still have that compatibility layer in there? 20:39:42 for jquery? 20:39:52 if tests pass and there are no obvious breakages, i.e. workflows are okay, nothing visually catastrophic, then check metadata and content, merge away 20:40:01 We do indeed 20:40:26 We could probably go straightto jquery 3 without much trouble since we basically only use it for selectors and simple manipulation... 20:40:47 but probably not in this release :) 20:41:22 well, depends on workload. Also I think jquery is still owned by someone else iirc 20:41:30 also depends on reviews :) 20:42:37 anyway we can discuss this another time 20:42:42 ok, I think we're done. thanks everyone for coming, and catch you in #openstack-horizon :-) 20:42:50 good night 20:42:50 \o thanks r1chardj0n3s 20:42:54 #endmeeting