20:00:44 #startmeeting Horizon 20:00:45 Meeting started Wed Jun 17 20:00:44 2015 UTC and is due to finish in 60 minutes. The chair is david-lyle. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:46 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:00:49 The meeting name has been set to 'horizon' 20:00:58 o/ 20:01:00 0/ 20:01:01 hi o/ 20:01:01 Let's play Horizon 20:01:05 here's the players 20:01:05 good day all 20:01:08 o/ 20:01:12 o/ 20:01:16 o/ 20:01:17 o/ 20:01:22 O/ 20:01:25 o/ 20:01:29 o/ 20:01:42 o/ 20:01:59 next time, we should count instead of o/'ing 20:02:08 13 20:02:14 [_]/ 20:02:25 ducttape_, :D 20:02:35 (╯°□°)╯︵ ┻━┻ 20:02:48 krotscheck: you're gonna run out of tables 20:02:49 * ducttape_ sees that we are not linear thinkers 20:03:00 r1chardj0n3s: I have a woodshop. I'll just make more. 20:03:17 alright, first the general stuff 20:03:30 L1 is Tuesday 20:03:47 LOL-1 you mean 20:03:55 I believe we've implemented 2 blueprints 20:04:10 so not the fastest jump out of the gate 20:04:44 . 20:04:55 so I just want to do a quick status check on the top priorities 20:05:20 because the milestones really don't mean much, but 2 is much less than our past average of 10ish 20:05:30 #link https://etherpad.openstack.org/p/YVR-horizon-liberty-priorities 20:05:42 The reOrg is still WIP 20:05:52 a patch or two has landed 20:05:58 still a ways to go 20:06:08 sound right r1chardj0n3s? 20:06:12 fixing style kinda got inserted as a top top top priority 20:06:23 I have no idea why 20:06:23 david-lyle: yep, it's 95% done 20:06:36 Probably my fault. 20:06:42 * krotscheck accepts all blame 20:06:49 let's blame krotscheck 20:06:54 krotscheck: no it went in the weeds before you 20:07:05 I'd vote for plugin rework to be high priority 20:07:10 you just helped us steer further off course 20:07:15 david-lyle: That doesn't mean you can't blame me :) 20:07:23 true 20:07:37 ducttape_: ++ 20:07:38 I think that the enforcement of JP style just ended up being way, way more work than expected 20:08:20 r1chardj0n3s: and in the end it's just code shuffling 20:08:22 oh well 20:08:32 on the reorg note, I felt compelled to file another bug introduced by the API patch landing (my bad, I didn't notice in the reviews :/ ) https://bugs.launchpad.net/horizon/+bug/1465885 20:08:33 Launchpad bug 1465885 in OpenStack Dashboard (Horizon) "API reorg retained some framework/dashboard crossover" [Undecided,New] 20:08:41 yeah, it is 20:09:25 if we can get the plugin stuff more flushed out, then new features can be added / fixed without all this refactor stuff impeding other work 20:09:36 for the translation fix, I saw django-babel made global requirements 20:09:40 or babel-django 20:09:52 are we happy with that path as a group? 20:10:04 ducttape_: I'm working on the plugin piece 20:10:12 yeah, i proposed the requirement 20:10:23 so i think we are in agreement 20:10:24 I have most done, but testing is proving a little difficult to bootstrap 20:10:32 tqtran_: ok, good 20:10:34 I recall agreeing with r1chardj0n3s on that i18n thing 20:10:55 patch for babel-django is getting close I think 20:11:40 (in horizon that is) 20:12:00 neillc: thanks neillc is it ready to review? or close to ready for review? 20:12:15 neillc: in case I forgot to type that 20:12:18 It's ready for review 20:12:29 excellent 20:12:51 There is also a follow on patch for angular gettext 20:13:00 sessions storage, lhcheng? 20:13:02 https://review.openstack.org/#/c/188665/ 20:13:21 https://review.openstack.org/#/c/192035/ (angular follow-on) 20:13:32 neillc: that looks less painful than I anticipated :) 20:13:34 david-lyle: no progress, thought folks want to put everything in session 20:13:34 (hm, I had hoped a bot would do something with those links :/ 20:13:53 Yep :) 20:13:57 r1chardj0n3s: You have to prefix them with #link 20:13:58 r1chardj0n3s: I think for launchpady stuff 20:14:27 lhcheng - so that means we should abuse the session until it falls over ? 20:14:34 lhcheng: we need memcache either way 20:14:42 which is to say, the same stuff we have been doing ? 20:14:45 so put it in the cache 20:14:50 same difference 20:15:17 in the end, it's a memcache backed store 20:15:23 do what makes sense to you 20:15:39 noone reviews d-o-a code anyway :P 20:15:52 ducttape_: I think session data should be small, and static data should be in the cache 20:16:03 thats not exactly ideal for deployment, its a bit more work to use memcached 20:16:33 ducttape_: your cache could be db backed as well 20:16:36 lhcheng: are you talking about static data that would be cross session? 20:16:40 ducttape_: from the last summit, it seems the preference was just to use session db backend? and not to have extra setup 20:17:13 ducttape_: is it because you want the session expiry part? 20:17:25 yeah, thats all fine. I'm good with it, but like the install / setup docs will need to cover how to ensure this is setup correctly (thinking like HA specifically) 20:17:34 sure 20:17:47 and david-lyle I think will update devstack to use db/memcahce session backend 20:18:06 I've heard that for some time :p 20:18:06 let's come up with a solution after the meeting 20:18:08 doug-fish: static data per user 20:18:19 yeah, yeah 20:18:23 :D 20:19:02 ok and the style enforcement is well down the road 20:19:31 is that hurgleburgler's stuff/ 20:19:32 ? 20:19:42 no jscs enforcements 20:19:44 seems like we're generally making progress, we just didn't choose much low hanging fruit to start with 20:19:58 tqtran_: ? 20:20:10 sorry, was answering ducttape_ 20:20:14 ok, I see 20:20:32 no, but hurgleburgler has some patches to review I know regarding theming 20:20:39 that would be very nice to land 20:20:58 yes, she had a video explaining it too 20:20:59 I'll add that to the priority list 20:21:01 they seemed pretty straight-forward 20:21:08 came about after the summit 20:21:09 I do, any eyes on them would be great 20:21:16 hence the omission 20:21:18 hurgleburgler: linky? 20:21:38 https://www.youtube.com/watch?v=KMY-4pg8-o0 20:21:50 #link https://www.youtube.com/watch?v=KMY-4pg8-o0 20:21:52 :-D 20:21:55 I think we hijacked tqtran's question ;) 20:22:06 no, he was clarifying 20:22:11 see how well it worked 20:22:13 :D 20:22:16 lol 20:22:27 apparently my clarification needs further clarifying 20:22:43 I said style, but should have said coding style 20:22:47 style guide = no jscs enforcements 20:23:01 no, jscs enforcements 20:23:06 lol 20:23:08 omg 20:23:10 anyway 20:23:20 see, commas matter above ;) 20:23:32 next item, autodiscovery 20:23:38 ah yes 20:23:46 seems reasonable enough 20:23:50 https://review.openstack.org/#/c/191592/ 20:24:00 the first util patch already landed 20:24:11 my only concern was essentially embedding metadata in the filename 20:24:22 not very flexible 20:24:36 django does it, and we havent had problems with it 20:24:46 we've been developing patterns for filenames anyway 20:25:02 its just file structuring, plus its something that is in the JP style guide and generally agreed on in the angular community 20:25:31 sure and the 1500 exceptions show we're pretty close on matching the style :P 20:25:37 it's fine 20:25:39 we have taken that standard and basically developed a tool to discovery it 20:25:46 wow Mr Optimist here 20:25:49 :) 20:26:11 krotscheck: maybe something that can go into oslo at some point? 20:26:22 if other JS projects are interested in reusing it for their own 20:26:31 tqtran_: You mean autodiscovery? 20:26:35 yeah 20:26:53 tqtran_: Maybe. Let me look at the patch. 20:27:03 so we're doing things quite differently to more conventional modern Javascript-heavy web apps 20:27:13 I'm not sure our solution makes sense outside of Horizon 20:27:13 tqtran_: THough if it's python based I'm going to guess no. 20:27:22 Something like a jasmine plugin however would definitely be useful. 20:27:23 what krotscheck said 20:27:24 https://review.openstack.org/#/c/190028/ its python based 20:27:38 ok 20:27:56 anyway, its almost done, so progress is good 20:28:14 ok 20:29:16 I think that covers the top priorities, once we have a couple of those, the user's table can land 20:29:28 speaking of which 20:29:39 https://review.openstack.org/#/c/190852/ 20:29:52 is the starting point, i have broken up the patch into smaller consumable pieces 20:30:06 https://review.openstack.org/#/c/190865/ 20:30:23 still waiting on translation to go through, but its making really good progress 20:30:32 excellent 20:30:57 and we have a good standard to base the work off of, so this will fit nicely into how we can direct other projects that want to horizon compatible 20:31:24 like the new magnum and cinder ui brad and zhenguo are leading 20:32:04 magnum and ironic, but I've been trying to link up krotscheck and zhenguo on the latter 20:32:15 ops ironic lol 20:32:21 zhenguo: won't be on due to TZ 20:32:22 not sure why i read cinder 20:32:28 WHere's he based? 20:32:33 China 20:32:50 Ok, so likely sometime this evening. 20:33:06 not sure if Shanghai or Beijing 20:33:19 yes 20:33:32 If all else fails, we can talk on the list 20:33:53 sure 20:34:01 seems like all of china uses one time zone 20:34:07 tqtran_: It does. 20:34:14 sensible 20:34:14 wow 20:34:30 UTC +8 20:34:39 big country with one time zone, they should do it in the US, will save some headaches 20:34:44 western china sunrise at 2am? 20:34:50 hahaha 20:34:57 where are we again? 20:35:12 last general item midcycle 20:35:19 #topic Midcycle 20:35:45 so we are strongly leaning toward July 21-23 in Fort Collins 20:35:58 ugh 20:36:04 we won't be co-locating with Glance and Searchlight in that caes 20:36:07 case 20:36:12 :( ok 20:36:12 lolz 20:36:35 * krotscheck will be in Napa ;) 20:37:11 * mrunge can't attend then :( 20:37:31 the other choice is july 28-30 in Boston or Blacksburg, VA and that could be co-located 20:37:52 even worse 20:37:58 the third option is don't do a midcycle and save your travel karma for Tokyo 20:38:04 Boston is closer to EU :) 20:38:15 yes! 20:38:23 and it's mid holiday season 20:38:36 I'm open to suggestions 20:38:56 Maybe do a virtual sprint instead? 20:38:57 fort collins sounds nice too though 20:38:59 That's what infra's doing 20:39:05 Yup...my Birthday should be a holiday 20:39:06 The thing with a midcycle is, it won't work for everyone 20:39:14 never will 20:39:26 It will never work for everyone 20:39:28 I'll help everyone stay hydrated in FC 20:39:32 NOte that they're also recommending that individual subproject teams arrange for their own colocation, if they deem that's necessary for productivity 20:39:42 * krotscheck shuts up about infra. 20:39:43 and ANY place will be nicer than Tucson in july 20:40:00 we can do a virtual, but with the three corners of the TZ triangle covered finding common time is hard 20:40:01 hurgleburgler, don't be sure about that 20:40:19 any place on the list* 20:40:31 krotscheck: we did some virtual last cycle, TZs are hard 20:40:32 :P 20:40:32 nice correction there :P 20:40:40 TZs are hard :/ 20:40:48 except in China 20:41:03 haah 20:41:04 I welcome our imminent single-TZ Chinese overlords 20:41:11 david-lyle: Well, the alternative is meet 1/2 way. For infra, that meant Bora Bora (because half of us are Down Under) 20:41:17 yeah, we should strictly stick to UTC, nothing else. 20:41:21 (Australia has a dozen timezones, some of them on the half hour) 20:41:52 krotscheck: the Hawaii mid-cycle is *never* going to happen :) 20:41:58 Too be fair that's three TZ and one state that refused dst :) 20:42:01 joking aside: would a mid-cycle in September an option, too? 20:42:02 ooo hawaii.... 20:42:13 if we are going to do a midcycle we need to set the date and let people book travel 20:42:27 I think Sept is too close to Oct 27 20:42:28 September is very late in the cycle 20:42:38 yes, I agree 20:42:53 should we vote? 20:43:01 this cycle is a bit shorter than the other 20:44:15 lets vote 20:44:15 the spreadsheet is practically our votes 20:44:15 esp. since last summit was later as usual 20:44:16 but yeah, let's do this thing 20:44:16 hi cindy! 20:44:16 hi cindy! 20:44:16 :D putting the spot light on her 20:44:16 cindy arrived to rock the vote 20:44:16 :) 20:44:16 o/ 20:44:40 david-lyle: i'd say if the spreadsheet shows that there is enough interest, then lets do it 20:44:54 you decide :) 20:45:00 #startvote Midcyle? July 21, July 28, no 20:45:01 Begin voting on: Midcyle? Valid vote options are July, 21, July, 28, no. 20:45:02 Vote using '#vote OPTION'. Only your last vote counts. 20:45:03 I'd help out with Foco, but am not on the list fwiw 20:45:15 #abstain 20:45:22 * krotscheck thinks that's how it works. 20:45:24 #vote July, 21 20:45:25 r1chardj0n3s: July, 21 is not a valid option. Valid options are July, 21, July, 28, no. 20:45:31 I can help organize if in Fort Collins 20:45:32 lol 20:45:34 #vote no 20:45:37 oops 20:45:38 #vote July, 21 20:45:38 neillc: July, 21 is not a valid option. Valid options are July, 21, July, 28, no. 20:45:39 #vote July, 21 20:45:39 hurgleburgler: July, 21 is not a valid option. Valid options are July, 21, July, 28, no. 20:45:40 #vote 21 20:45:41 #vote 21 20:45:43 lol 20:45:50 #vote 28 20:45:54 so many commas 20:45:55 #vote 21 20:46:04 bonus commas! 20:46:20 #vote 21 20:46:22 #vote 21 20:46:30 #vote 21 20:46:53 #endvote 20:46:54 Voted on "Midcyle?" Results are 20:46:56 28 (1): lapalm 20:46:57 21 (6): neillc, ducttape_, r1chardj0n3s, tqtran_, doug-fish, hurgleburgler 20:46:58 no (1): mrunge 20:47:09 I forgot to vote (facepalm) 20:47:13 lol 20:47:16 #vote 21 20:47:22 I'd be 21 as well 20:47:27 david-lyle: .. :D 20:47:33 One man, one vote? David is the man... 20:48:08 2000 dollar US just for a flight... 20:48:24 but that's like 6 EUD, right? 20:48:37 ok that's 8 for July 21 and 1 for July 28, and 1 no 20:48:39 And just think of the # of carbon offsets you have to buy! 20:49:26 let's finalize July 21-23 in Fort Collins, CO. HP will host. 20:49:37 \o/ we made it! 20:49:41 DEN is the nearest airport 20:49:49 ok 20:49:57 Grats :) 20:50:00 there are a couple of hotels near the Fort Collins site 20:50:02 (the meeting seem to have come to a sudden halt..) 20:50:06 and more in town 20:50:20 ok will send email to the list 20:50:23 oh no. seems like it was my client =/ my bad 20:50:41 mrunge: if you are working at home we can try to sync you in 20:51:01 hangout or vidyo or bluejeans if we're feeling crazy 20:51:17 if on holiday, that won't work as well 20:51:25 oh, apparently the 10 person limit for vidyo has been removed 20:51:34 there were some agenda to cover, let's jump to that 20:51:55 #topic integration tests on the gate - enable Sahara? (svasheka, mpavlase) it will presumably make integration job green. 20:52:15 I voted +1 on the patch, don't have handy 20:52:23 makes sense while in tree 20:52:31 I'm still working on moving to contrib 20:52:32 thanks for highlight the topic 20:52:44 but that will still be in tree 20:53:15 basically, sahara is not enabled by default in devstack 20:53:24 this patch enables it for horizon test runs 20:53:45 current integration tests are failing on 4 tests, two of them are caused by disabled sahara in gate job... https://review.openstack.org/#/c/192645/ 20:53:47 so going back to what was formerly expected behavior 20:53:58 makes sense to have it! 20:54:06 this will help fix integration madness 20:54:15 thanks mpavlase 20:54:32 #topic Linter switch discussion, jscs/eslint 20:54:37 hi hi 20:54:49 So, turns out eslint is actually on the YVR Roadmap? https://etherpad.openstack.org/p/YVR-horizon-liberty-priorities 20:54:58 the rest are there because something is broken on horizon, let me find the patch... 20:55:02 And, well, linting is now running in the gate. 20:55:08 * david-lyle wonders why krotscheck likes lint so much 20:55:09 Sorry, in the check queue. It's not voting yet. 20:55:10 :) 20:55:14 this one: https://review.openstack.org/#/c/189792/ 20:55:22 david-lyle: It's my bellybutton. Don't judge. 20:55:49 aha.. next topic is already here.. 20:55:56 I've also put together this patch, which switches to ESLint, turns off a bunch of rules (flags the ones that were turned off), but activates the john papa style rules. 20:56:02 https://review.openstack.org/#/c/192327/ 20:56:33 most importantly, does the error count go down? no point switching tools if the error count doesn't go down :) 20:56:52 Note that there's some overlap there (which is why the # of failures is 1100 rather than 1500 with jscs). Notably, line length. 20:57:10 krotscheck: With the swtich to ESLint, does JSHint go away as well? 20:57:17 asahlin: It should. 20:57:18 why is line length turned off? 20:57:26 r1chardj0n3s: I can turn it back on. 20:57:26 we want line length 100, yes? 20:57:33 r1chardj0n3s: It's just not part of the papa rules. 20:57:40 sure, but it's part of our rules 20:57:45 r1chardj0n3s: we could do that subsequently 20:57:52 with the gettext/etc. caveats we built into jscs 20:57:57 fix the other 1100 say 20:58:03 and then tackle 400 more 20:58:07 and enable the check 20:58:15 so my problem is that it's hard to know what to work on in that 1100 20:58:16 r1chardj0n3s: The point is that we can reactivate all the rules we want (and argue about which rules we want) once the papa guides are in place. 20:58:26 that's what we've done with the pep8 changes in the past 20:58:27 krotscheck: my point is we *already* argued about that 20:58:47 don't make us argue it again, please 20:58:48 r1chardj0n3s: I really don't care which rules are activated or deactivated. 20:58:59 sure, then just duplicate the jscs set please 20:59:22 +1 ^^ 20:59:35 we won't have you hiding our disgrace 20:59:40 :) 20:59:49 this tool change should have zero functionlity change (like the translation tool change) 21:00:01 r1chardj0n3s:+1 21:00:05 I don't have a strong preference as to which linter 21:00:08 the jscs rules were papa plus our caveats 21:00:08 r1chardj0n3s: I don't think the zero functionality change expectation is reasonable 21:00:11 But we're out of time. 21:00:24 if we make the switch, i prefer to not see another 20+ patches style fixes 21:00:29 if eslint works more broadly and works for us, ok 21:01:29 I don't see any disagreements with switching to eslint. Do I have that right? 21:01:36 It's just the rules themselves that are under discussion. 21:01:49 In that case, can I ask for help trying to map the .jscs exceptions to eslint? 21:01:50 we pushed up to the end today, if we want to discuss more let's move to #horizon 21:01:58 kk 21:02:00 #endmeeting