20:33:04 <prometheanfire> #startmeeting requirements 20:33:05 <openstack> Meeting started Wed Jul 11 20:33:04 2018 UTC and is due to finish in 60 minutes. The chair is prometheanfire. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:33:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:33:08 <openstack> The meeting name has been set to 'requirements' 20:33:10 <prometheanfire> #topic rollcall 20:33:13 <prometheanfire> o/ 20:33:17 <prometheanfire> tonyb, prometheanfire, number80, dirk, coolsvap, toabctl, smcginnis, dhellmann 20:33:26 <tonyb> o/ 20:34:35 <openstackgerrit> Michal Arbet proposed openstack/requirements master: Update constraints for networkx to new release 2.1 https://review.openstack.org/578094 20:35:10 <tonyb> looks like just us two ;p 20:35:12 <prometheanfire> ya 20:35:34 <prometheanfire> tonyb: I'm trying to get this networkx thing finished, just need a release of taskflow and dib to get their shit together about it 20:35:43 <prometheanfire> http://logs.openstack.org/94/578094/5/check/requirements-tox-py27-check-uc/1f1eeb6/job-output.txt.gz#_2018-07-11_19_26_50_344225 20:36:04 <tonyb> prometheanfire: I'll take care of DIB 20:36:04 <prometheanfire> #topic Any controversies in the Queue? 20:36:17 <prometheanfire> tonyb: thanks, the taskflow patch may help 20:36:30 <tonyb> not controversies, but things I'd liekto highligh 20:36:53 <tonyb> https://review.openstack.org/581160 oslo.messaging 8.0.0 20:37:08 <prometheanfire> yes? 20:37:27 <tonyb> we want to make sure the API break isn't going to cause pain or at least call it out on the list with a reasonable set of instructions as to how to fix it 20:38:04 <tonyb> I wanted to write a script to find likley breakers and run a noop chnage against it and I may still do that 20:38:12 <prometheanfire> shouldn't oslo (and us) mail the list trying to get people to test/use lower-constraints to guard against it? 20:39:03 <tonyb> well we need to gett it into upper-consstraints first, and if the API break doesn't affect $project then they don't need to bump l-c either 20:39:22 <tonyb> but yes if we do find a project that breaks we do need to call that out in l-c 20:39:33 <prometheanfire> ok 20:39:47 <prometheanfire> so are you worried that the cross tests are not good enough? 20:39:57 <tonyb> prometheanfire: Yeah, basically 20:40:21 <tonyb> prometheanfire: I don't *know* but I don't think it covers all the oslo.messaging (direct) users 20:41:17 * tonyb runs his scripts to check. 20:41:27 <tonyb> wheil I do ... 20:41:55 <tonyb> we need to decide about our support for py34 on master 20:42:08 <tonyb> (backtrack)Package : oslo-messaging [oslo-messaging] (used by 51 projects) 20:42:33 <tonyb> so yeah I doubt our cross-gate will catch the breaks if they happen. 20:42:41 <prometheanfire> tonyb: ya, sounds good 20:42:49 <tonyb> anyway, the python3.4 stuff 20:42:55 <prometheanfire> sure 20:43:07 <tonyb> we have it in our constraints but techically we don't support master on py3.4 20:43:12 <prometheanfire> oslo-messaging is used by deps of deps 20:43:23 <prometheanfire> I suspect we cover at least the majority of that 51 20:43:31 <tonyb> but I feel like perhaps suse could be doing that thing 20:43:36 <prometheanfire> ya, now that we have py36 we should drop py34 20:44:07 <prometheanfire> I thought gentoo dropped py34, but we still have it, think we switched default to 36 20:44:26 <prometheanfire> suse could do what thing? 20:45:00 <tonyb> prometheanfire: have py34 in the OS and run rocky/master on that OS 20:45:26 <tonyb> ooo actually I know people other than dirk in suse I'll ping them 20:45:51 <prometheanfire> tonyb: lol 20:46:03 <prometheanfire> I thought rocky was targeting 35 20:46:21 <dirk> tonyb: could be doing what? 20:46:51 <tonyb> dirk: have py34 in the OS and run rocky/master on that OS 20:47:08 <dirk> yep, that could happen 20:47:21 <tonyb> dirk: We're trying to work out if having the py3.4 constrarints in u-c is helpful or can be removed 20:47:55 <prometheanfire> ok, in that case perhaps we should schedule the drop for stein? 20:47:56 <dirk> so opensuse-leap-423, which is supportd until end of 01/2019, contains python 3.4 20:48:07 <dirk> so it is potentially used somewhere 20:48:26 <dirk> (its available as a nodepool image) 20:48:26 <tonyb> dirk: Okay. 20:48:35 <dirk> we could add it to the testing if thats your concern :) 20:48:37 <prometheanfire> technically you could do it in gentoo too, but I've been switching things over to 35, might make 36 for rocky 20:48:48 <dirk> adding another tests job against a leap image shouldn't be difficult 20:48:52 <tonyb> dirk: Great, so wheil it's a nodepool image we shoudl keep the constraints 20:49:10 <tonyb> prometheanfire: Yup I think we shoudl keep them and schedule the removal fro m3 in stien 20:49:22 <dirk> tonyb: ever tried running against it? 20:49:23 <tonyb> dirk: I think you started doign that at one point 20:49:50 <tonyb> dirk: personally no, but are there *any* jobs that do? 20:50:12 <prometheanfire> tonyb: are we the ones that make that decision (remove support for a python from openstack)? 20:50:37 <tonyb> prometheanfire: Yes and no 20:51:10 <tonyb> There is a community statement, we're just extending it a little 20:51:32 <prometheanfire> ok 20:51:46 <dirk> tonyb: I tried py36 20:52:01 <prometheanfire> honestly I'd prefer removing 34 from stein right after we branch 20:52:16 <openstackgerrit> Dirk Mueller proposed openstack/requirements master: DNM: py34 constraints testing https://review.openstack.org/581882 20:52:26 <dirk> prometheanfire: works for me 20:52:26 <tonyb> prometheanfire: but that'll break opensuse-leap-423 nodepool images 20:52:34 <tonyb> dirk: nice one! 20:52:37 <dirk> leap-423 is going to be removed jan 2019 20:52:58 <dirk> ah, so proposal is to remove earlier 20:53:01 <dirk> sorry, was confused about tiemline 20:53:10 <dirk> well, lets see how far it works.. :) 20:53:41 <prometheanfire> dirk: ok, as long as it doesn't hurt any project (and it shouldn't) 20:53:59 <prometheanfire> honestly we'll probably only remove py34 once upstream stops supporting it 20:54:08 <prometheanfire> meaning I could add a py34 gentoo job til that happens 20:54:16 <tonyb> Ok so I think we need to do 2 things 1) ask the list about when we remove py34 constraints and suggest stien-m3 for ^^ reasons ; and 2) merge https://review.openstack.org/580241 + one more (unwritten) chnage so that we don't keep losing py36 stuff 20:54:19 <prometheanfire> at least for the simple checks 20:54:53 <prometheanfire> tonyb: sgtm 20:55:21 <tonyb> prometheanfire: cool. 20:55:25 <prometheanfire> my attempt at on freeze to others didn't go well :P 20:56:04 <tonyb> dirk: how doe your chnage run on leap-423? 20:56:12 <dirk> tonyb: it doesn't yet 20:56:31 <tonyb> dirk: Ahh okay glad I didn't just miss it 20:56:32 <dirk> tonyb: I was trying to figure out how to change the image whithout copying all of the job parent 20:56:54 <dirk> I was hoping the default image already contains python 3.4 20:57:04 <tonyb> dirk: Ahh okay, as long as they aren't 'final' you shoudl be able to do that 20:57:22 <tonyb> dirk: I don't think it does but we'll see 20:57:34 <prometheanfire> ok, other than those tasks, move on? 20:57:44 <tonyb> prometheanfire: +1 21:01:15 <prometheanfire> #topic open floor 21:01:31 <openstackgerrit> Dirk Mueller proposed openstack/requirements master: DNM: py34 constraints testing https://review.openstack.org/581882 21:01:55 <tonyb> chnages from the constraints bot ... is it time to add known bad exclusions to g-r? 21:02:16 <tonyb> do we know why the 'bad' things are bad? 21:02:27 <openstackgerrit> Dirk Mueller proposed openstack/requirements master: DNM: py34 constraints testing https://review.openstack.org/581882 21:03:36 <prometheanfire> tonyb: yes, generally, the google stuff is waiting on cinder 21:03:47 <prometheanfire> the pyldap stuff should be fixed in code to not be proposed 21:03:54 <prometheanfire> networkx is known about 21:04:08 <prometheanfire> eventlet is the only one off the top of my head that has a case to be capped 21:04:16 <tonyb> prometheanfire: are the cinder team workign on fixing the google stuff? 21:07:08 <prometheanfire> yes 21:07:12 <dirk> is this mental to do list noted down somewhere? 21:07:15 <prometheanfire> I talked with them yesterday iirc 21:07:20 <prometheanfire> dirk: nope 21:08:31 <tonyb> prometheanfire: Okay 21:09:24 <prometheanfire> anything else? 21:10:15 <tonyb> not from me 21:12:08 <prometheanfire> #endmeeting