15:00:32 #startmeeting oslo 15:00:33 Meeting started Mon Apr 16 15:00:32 2018 UTC and is due to finish in 60 minutes. The chair is bnemec. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:34 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:36 The meeting name has been set to 'oslo' 15:00:47 o/ 15:00:55 courtesy ping for amotoki, amrith, ansmith, bnemec, dansmith, dhellmann, dims 15:00:55 courtesy ping for dougwig, e0ne, electrocucaracha, flaper87, garyk, gcb, haypo 15:00:55 courtesy ping for jd__, johnsom, jungleboyj, kgiusti, kragniz, lhx_, raildo 15:00:55 courtesy ping for redrobot, sileht, spamaps, sreshetnyak, stephenfin, stevemar, therve 15:00:55 courtesy ping for thinrichs, toabctl, zhiyan, zxy, zzzeek 15:01:01 o/ 15:01:02 o/ 15:01:05 o/ 15:01:13 o/ 15:01:32 Full house today. :-) 15:01:43 o/ 15:02:42 #topic Red flags for/from liaisons 15:03:10 We ran into a problem in oslo.versionedobjects last week. 15:03:16 #link https://review.openstack.org/#/c/560639/ 15:03:43 It was caught in the u-c requirements testing though so no major impacts. 15:03:55 But we do need to get that patch in to unblock ovo. 15:05:26 That's it from the oslo side. I don't see a lot of liaisons here so we'll move on. 15:05:34 #topic Releases for Rocky 15:05:43 We did the normal set of releases last week. 15:05:52 The most notable one was oslo.service with eventlet uncapped. 15:06:07 I'm not actually sure if that's in u-c yet. 15:07:03 1.31 is now, which I think is the uncapped one. 15:07:34 there are still caps in a bunch of the service projects, too 15:08:02 Yeah, I need to sit down and go through those reviews. 15:08:04 #link https://review.openstack.org/#/q/status:open++topic:uncap-eventlet 15:08:45 at least a couple of those are oslo libs (hacking, privsep, futurist, db) 15:08:50 #action finish merging eventlet uncapping patches 15:10:22 On a related note... 15:10:23 #topic Action items from last meeting 15:10:31 "Switch Oslo pep8 test jobs to python 3" 15:10:45 We actually took that one step further. 15:10:57 Doug has proposed patches to switch all the non-py27 jobs to python 3. 15:11:27 #link https://etherpad.openstack.org/p/converting-secondary-jobs-to-py3 15:12:01 I saw some of those started to merge, so I think we're making progress there. 15:12:01 the patches up against the oslo repos only affect the jobs that use tox. We will need another patch in project-config to convert the doc and release notes jobs. I will prepare that after we land the oslo patches. 15:12:43 Right, good point. 15:12:43 I encourage everyone to get into the habit of thinking "python 3 first" (run "tox -e py35" to run the unit tests, etc.) 15:13:00 We'll want to get those landed relatively close together so we don't accidentally break something in the meantime. 15:13:09 good point 15:13:11 seems like a trivial patch 15:13:21 so far they have been 15:13:29 can we just +2 if Zuul is happy? 15:13:38 I expected the pep8 changes to cause us to need to update some code, too, but none so far 15:13:48 after a quick eyeball of course :) 15:13:49 I'm good with a single +2 policy for these. 15:14:04 yeah, they're easy enough to revert if they do cause trouble 15:14:10 The only pep8 fix I ran into was in oslo.concurrency, and that was in code that was if'd for py2 only. 15:14:47 looks like taskflow has one 15:14:52 http://logs.openstack.org/41/561341/1/check/openstack-tox-pep8/285716a/job-output.txt.gz#_2018-04-13_20_24_45_586700 15:15:02 #link https://review.openstack.org/#/q/status:open+topic:pep8-python3 15:15:07 ^is the review queue for this. 15:16:02 Hmm, fun. That seems like a pretty legit complaint. 15:16:07 Will have to look at the taskflow code though. 15:17:05 Doug Hellmann proposed openstack/debtcollector master: set default python to python3 https://review.openstack.org/561310 15:17:05 Doug Hellmann proposed openstack/debtcollector master: remove obsolete tox environments https://review.openstack.org/561311 15:17:16 #action Oslo cores to single approve python 3 tox changes 15:17:40 if you have the time, I wouldn't mind help fixing the patches up as well 15:17:47 otherwise I'll work on them as I can this week 15:17:57 Sure 15:18:29 Seems like we've got a good start, just need to keep plugging away at this. 15:18:33 ++ 15:18:50 The joys of working on a project with 40-ish sub-projects. :-) 15:18:56 heh 15:19:32 #action fix up any python 3 tox changes that need additional work 15:19:34 I resisted the urge to make the tox files more consistent, but that may be another project for someone to take on 15:20:16 Yeah, that might be a good wishlist bug with a low-hanging-fruit tag. 15:20:27 good idea 15:20:32 Maybe we can sign somebody up at the onboarding session. 15:21:02 Although I'm not sure where to put the bug. Do we have an umbrella oslo project in lp? 15:21:12 I know we have a group, but I don't think you can open bugs against it. 15:22:22 Speaking of the onboarding session: 15:22:25 #topic Onboarding session moved from Thursday to Tuesday 15:22:40 This is mostly a heads up for anyone who was planning to attend. 15:22:55 Infra had a conflict with theirs and asked if we could switch. 15:23:10 Since ours wasn't near the project update anyway I said yes. 15:23:25 It's now Tuesday afternoon. 15:23:44 I didn't see any other sessions at that time that I thought would conflict for Oslo contributors, but let me know if it's a problem. 15:24:45 np for moi 15:25:29 sounds good 15:26:05 Hmm, I don't see the session on the schedule at all right now. 15:27:25 I'll have to follow up on that. 15:27:36 #action bnemec to check on missing onboarding session 15:27:57 #topic Split all optional oslo.messaging requirements except rabbit into extras 15:28:30 I skipped this action item because of the neat segue to the onboarding session. :-) 15:28:45 I see kgiusti just submitted a new patch set to this change. 15:28:53 #link https://review.openstack.org/525319 15:29:13 We decided to keep the zmq deps as well for now since we plan to remove it next cycle anyway. 15:29:30 That way we don't have any backward compatibility issues with the next release. 15:29:44 makes sense 15:29:46 yep 15:29:56 kgiusti: Anything else of note on this? I noticed it was failing ci last week. 15:30:20 as I mentioned to bnemec if we bumped tox to 2.4, we could eliminate the dups between setup.cfg and test-reqs.txt 15:30:47 no additional comments, but it did pass last run (I just rebased due to a merge conflict) 15:31:04 Cool. 15:31:11 tox >= 2.4 - only a handful of projects use it AFAIKT 15:31:26 I'm +1 on requiring tox 2.4. IIRC, that release is a year and a half old and there are other projects requiring more recent releases. 15:31:56 But I think I suggested making that a followup? 15:32:03 Friday was so long ago. :-) 15:32:15 agreed - a follow up 15:32:50 possibly unrelated: I am seeing an error running the pip-missing-req test locally: 15:33:11 from pip.commands.show import search_packages_info 15:33:11 ImportError: No module named commands.show 15:33:20 I don't think I've ever tried to run that test locally. 15:33:27 Oh, I bet that's pip 10. 15:33:40 pip 10 was just released and it blocks importing pip code and at the same time reorganizes some of the internals 15:33:49 Yeah, that^ 15:33:51 so if we have code that relies on pip internals we need to fix that 15:33:57 what does that job do? 15:34:38 https://git.openstack.org/cgit/openstack/oslo.messaging/tree/tox.ini#n116 15:34:41 I get the impression it ensures we don't import anything that's not in requirements. 15:35:40 bnemec: me also 15:36:30 git blame points to ... dims! 15:36:30 Sounds like that command will need some work to be compatible with pip 10. 15:36:38 Heh 15:36:42 haha 15:36:48 my it's an old one 2015 15:36:55 dims: no good deed man.... 15:37:09 Is that from an oslo project? 15:37:22 Oh, presumably pip_check_reqs, so no. 15:37:54 Whoa, pypi looks different today 15:38:20 this stuff is 2 years old https://github.com/r1chardj0n3s/pip-check-reqs/tree/master/pip_check_reqs 15:39:17 There's a bug open for this: 15:39:19 #link https://github.com/r1chardj0n3s/pip-check-reqs/issues/17 15:40:14 In fact it looks like there might be two issues open for this problem. 15:40:31 Anyway, if anyone has time to look into it that would be great. 15:41:02 bnemec: I'll pick at it a bit 15:41:16 kgiusti: Cool, thanks 15:41:38 do we run that in the gate? 15:41:38 #action kgiusti to look into pip-check-reqs with pip 10 15:44:08 I don't see a job for it. 15:44:20 ok, that's good 15:44:22 Unless it's part of the requirements-check job, but I don't find any references in that log either. 15:45:18 bnemec: I was just running it locally to be sure I didn't break anything since I was messing around with test-reqs, etc. 15:45:35 That project looks fairly dormant, so we'll see how it goes. 15:45:43 kgiusti: Yeah, seems like a good idea. 15:46:49 Okay, we already covered the other item on the agenda (eventlet uncapping), so 15:46:49 #topic Open discussion 15:50:12 I don't have anything else to bring up this week 15:50:38 Okay, we'll give everybody ten minutes back then. Thanks for joining! 15:50:41 #endmeeting