10:00:00 #startmeeting requirements 10:00:01 Meeting started Wed Dec 14 10:00:00 2016 UTC and is due to finish in 60 minutes. The chair is tonyb. Information about MeetBot at http://wiki.debian.org/MeetBot. 10:00:02 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 10:00:04 The meeting name has been set to 'requirements' 10:00:13 #topic rollcall 10:00:17 sigmavirus, prometheanfire, number80, dirk, coolsvap, toabctl 10:00:17 o/ 10:00:26 hey 10:00:29 o/ 10:00:53 o/ 10:00:57 woot! 10:01:11 Getting the band back together :) 10:01:24 #topic Any controversies in the Queue? 10:02:20 nothing major 10:02:23 gate's been a little wonky lately, bot don't think so 10:02:28 The kombu change is a little complex 10:02:29 https://review.openstack.org/410176 10:02:38 the only topic I have is whether or not we use upper-constraints for our own gating 10:02:51 Yeah thet gate has been bad lately 10:03:18 there was the docutils issue, then kombu + lots of other random failures 10:03:46 sdague was right abiyt landing stuff in openstack/requiremenst geting hard with the more testing we add 10:04:31 dirk: I'd like to here youre thoughts on that in open discussion 10:04:37 tonyb: wfm 10:04:56 #topic Requirements priorities for Ocata 10:05:03 #link https://etherpad.openstack.org/p/requirements-track-constraints-usage 10:05:28 dirk's been doign the good work of adding constraints support to projects that are missing it 10:05:47 we also need to work on the "incubated code" idea 10:05:53 and tools around that 10:06:45 dirk's also been working hard on geeting the constratints updates landed. 10:07:01 yeah, surprisingly easy 10:07:13 dirk: can you give us aun update on the state of docker-py and pillow 10:07:15 I was expecting more complaints, but the docutils breakage greatly helped 10:07:18 dirk: what can we do to help you 10:07:28 pillow is blocked on sphinxmark 10:07:41 dirk: okay I'll poke bmoss 10:07:41 https://github.com/kallimachos/sphinxmark/pull/21 10:07:52 dirk: did you PR pass unit tests? 10:07:55 if anyone knows the maintainer and can poke him, this PR needs to get in and then we need release 10:08:14 for docker-py we're still facing issues with conflicts, I think it was tripleo last time I looked 10:08:32 #action tonyb to talk to bmoss about sphinxmark 10:08:32 I can take a closer look today at that, maybe that is solved with the new tripleo-common release 10:09:05 #action dirk take a closer look at docker-py upper-constraints 10:09:11 dirk: that'd be awesom if you could look at that and drop an update into #openstack-requirements 10:09:28 the new Pillow release is imho just a minor fix 10:09:44 I don't think anyone is waiting for it so its annoying that the bot proposes it constantly but I don#t think there is any issue due to that 10:09:59 dirk: okay but it keeps coming up and causing you pain so we shoudl try to fix it 10:10:01 docker-py is more urgent as there are teams waiting for the uncapping 10:10:21 ok 10:10:22 tonyb: well, my secret weapin is that the bot will search for all open reviews with a magic topic and reverse-apply those patches 10:10:25 and then I don't care anymore :) 10:10:49 dirk: Huh really? 10:10:51 so far this secret weapon isn't implemented yet though, although I splitted the problematic components into separate reviews 10:11:05 #action tonyb go looking at the code for the constartints update bot 10:11:19 nice 10:11:49 tonyb: so you'll implement that? or you just educate yourself about the code ? :) 10:12:08 tonyb: the docker-py is isolated in https://review.openstack.org/#/c/406442 10:12:14 I just rechecked that now, if it passes, we're good 10:12:20 dirk: I mis understood. I thought you wre describing the state of the current code, not a new feature 10:12:33 dirk: Thanks 10:12:40 tonyb: so far its only an idea, 10:12:53 dirk: okay I see 10:13:59 Can we get some people to look at that link I posted above and take a few projects 10:14:18 We're kinda blocked on that for the rest of the "fun" work this cycle 10:14:30 ya :( 10:14:34 Perhaps I've over estimated what we can do in one cycle 10:14:41 not really 10:14:43 prometheanfire: Thanks. 10:14:49 I'll try, now that I have some time 10:15:01 prometheanfire: \o/ 10:15:04 tonyb: I think we can still do it. there is endless amount of time between christmas and new year, you know ;) 10:15:06 https://etherpad.openstack.org/p/requirements-track-constraints-usage right? 10:15:16 prometheanfire: ACK 10:16:03 dirk: Sure there is time but we're at milestone 2 and radically changing the infra of the whole openstack project after milestone3 is just a bad idea 10:16:28 dirk: so I tink we might just need to focus on adding the support and tools this cycle 10:16:34 then start again in pike 10:16:48 with slightly more lead time as the difficult design work is done 10:17:09 maybe I'm misinformed, but wasn't the goal for ocata to just establish upper constraints everywhere and increase test coverage so that we can do the 2nd steps then? 10:17:34 maybe thats why I'm relaxed because I still think we can totally reach *that* goal :) 10:17:41 we wanted it all 10:17:47 all the the todo :P 10:17:49 dirk: the goal was to hit divergent requirements in ocata 10:18:17 dirk: but I underestimated the pre-conditions of that work 10:18:48 I see. ok 10:19:19 don't get me wrong more projects using constraints is a good thing and a reasonable outcome for the cycle 10:19:49 We have until Jan 26th 2017 to hit that target 10:20:15 Anything else for goals/priorities? 10:21:10 #topic Tasks from Etherpad 10:21:18 #link https://etherpad.openstack.org/p/requirements-tasks 10:21:46 prometheanfire: I see you've been movign items from the tracking etherpad to open bugs/issues 10:21:52 very slowly 10:22:11 prometheanfire: slow forward progress is still forward progress 10:22:17 deps between 'bugs' is hard on etherpad 10:22:49 prometheanfire: Yeah I *think* that's one of the things in StoryBoard 10:23:09 s/etherpad/launchpad 10:23:17 prometheanfire: before you do too many more it might be worth testing Storyboard for our task tracking 10:23:39 One of the open issues is defining our electorate. 10:23:53 tonyb: can you add me to the launchpad group= 10:23:59 it seems I can not close bugs.. 10:24:04 Does anyone object to me adding all the PTLs of managed projects as extra-atcs? 10:24:14 dirk: I can try 10:24:23 dirk: i can do that for you 10:25:07 coolsvap: Thanks 10:25:11 coolsvap: thanks! 10:25:19 tonyb: sgtm 10:25:22 #action coolsvap to add dirk as a requirements-driver 10:25:28 I agree a bugtracker is much better than an etherpad :) 10:25:48 for the record I seem to have permission to do that thing I just can't find your LP account 10:25:50 tonyb: atm, I'll not be a core at least because of my lack of commits to this repo, at least I think 10:26:15 prometheanfire: s/core/atc/ ok 10:26:42 prometheanfire: we can verify and fix that 10:26:45 tonyb: i also could not find the correct acount for dirk 10:26:48 just mentioning 10:27:06 of course it's only an issue if we're havign an election ;P 10:27:44 lol 10:27:47 bdfl 10:28:11 coolsvap, dirk: dmllr and done 10:28:14 prometheanfire: :) 10:28:39 tonyb: i think we can keep the electorate to lower priority 10:28:39 moving on? 10:29:02 coolsvap: Sure but if we're going to do it we have a deadline to do it RSN 10:29:38 Jan 6th 10:29:50 #topic Next meeting? 10:29:59 that is likely right after next meeting 10:30:19 So with Christmas coming up I'm thinking that we shoudl skip meetings until Jan 4th 10:30:28 fine by me :D 10:30:35 that gives people a little more flexibility to take leave etc 10:31:36 i am fine either way (with or without meetings) :-) 10:31:55 wfm 10:32:40 okay I think we'll skip'em 10:32:57 #action tonyb to setup meeting skips ofr requirements meetings 10:33:05 #topic Open Discussion 10:33:32 dirk: Tell me about using constraints in our own tox targets 10:34:08 tonyb: so I was thinking that it would make sense to use our own proposed upper-constraints for testing 10:34:17 currently our own check jobs are not using upper-constraints 10:34:41 (so not use upper-constraints from master git like everywhere else) but the current git version in the working copy 10:34:55 dirk: Oh you'd like to add openstack/requirements to the gate-cross group alongside nova/glance etc? 10:35:19 https://review.openstack.org/#/c/406505/ 10:35:40 adding golang technical requirements (email) to the list of things to discuss/mention 10:35:41 tonyb: just use the same way like we use it in other projects, but against the local copy so that we're selfgating changes to those files 10:36:00 dirk: Ahh okay 10:36:22 so we can't do that because many of out tools rely on the fact they're unconstrained 10:36:47 dirk: if we only install the stuff in u-c how will the bot find new versions to suggest? 10:37:26 dirk: what we *can* do is assess which tox environment need to be unconstrained and constrain the rest 10:38:22 dirk: which might just mean adding "install_commnd = pip install -U {opts} {packages}" to the [testenv:update] tox environment 10:38:44 dirk: so from my POV as long as we do it carefully we shoudl totally do it 10:38:55 tonyb: yeah, good point, so the bot jobs need to be checked. 10:39:05 dirk: Yup 10:40:20 Actually s/update/generate/ and I think that's probably enough 10:40:40 dirk: Are you keen to look into this? 10:40:55 tonyb: sort of, yeah 10:40:59 lower prio 10:41:14 dirk: Or do you think you can convince number80 or tobctl to help ? 10:41:19 dirk: Sure. 10:41:57 dirk: if notthing else when we hit openstack/requirements in the constartints etherpad we can look at it then 10:42:41 dirk: You also suggested we add openstack-requirements to global-requirements ... is that still somethign you think we shoudl do? 10:43:38 sounds like a good idea to me (adding us to gr), wonder who our downstream would be for that 10:49:13 tonyb: yeah, I still thought its logical, but you -1#ed 10:49:16 I forgot the reason 10:49:27 prometheanfire: but g-r is about co-installability and no project shoudl depend on openstack-requirements that way 10:49:38 ya 10:50:01 like I said, wonder what our downstream would be for that (to require us being added to gr) 10:50:31 tonyb: for me the reason was to control what version of openstack-requirements ends up in uc, but you said we can't restrict that anyway due to bootstrap issues 10:50:41 tonyb: so for me it is also okay to just abandon that review 10:51:10 tonyb: I don't see a major point in adding it in g-r 10:52:38 dirk: Are we seeing it in u-c? 10:53:04 tonyb: no 10:53:14 dirk: phew! 10:53:15 tonyb: its not in there, and as you explained it shouldn't be in there 10:54:06 dirk: okay 10:54:49 Changing tac befoer we close. It seems that we need to alter our tox_install.sh in line with https://review.openstack.org/410018 10:55:25 I'll do that tomorrow and as long as noone minds I'll update all the open reviews adding tox_install.sh 10:55:42 and hope that we don't get more -1's 10:55:48 :) 10:56:20 so with that in mind please keep your open updates listed in the etherpad 10:57:31 Anything else? 10:58:17 Thanks everyone 10:58:21 #endmeeting