20:33:24 <prometheanfire> #startmeeting requirements
20:33:25 <openstack> Meeting started Wed Dec 11 20:33:24 2019 UTC and is due to finish in 60 minutes.  The chair is prometheanfire. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:33:26 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:33:28 <openstack> The meeting name has been set to 'requirements'
20:33:29 <prometheanfire> #topic rollcall
20:33:31 <prometheanfire> tonyb, prometheanfire, dirk, smcginnis ping
20:33:33 <prometheanfire> o/
20:34:15 <dirk> o/
20:35:27 <prometheanfire> #topic Any controversies in the Queue?
20:35:38 <tonyb[m]> I'm kinda here
20:36:19 <prometheanfire> atm, I'd say we are moving along, sphinx has dropped py3 as well, so will need to uppdate gr for that (sphinxcontrib stuff)
20:37:31 <dirk> py2 you mean?
20:37:55 <prometheanfire> ya
20:39:57 <smcginnis> o/
20:40:19 <dirk> wfm
20:40:25 <smcginnis> Didn't we already handle that? Or did they change something else?
20:40:59 <prometheanfire> it's the contrib plugins this time
20:41:05 <prometheanfire> we have it handled for sphinx itself
20:41:21 <smcginnis> Ah, OK. The fun never stops.
20:41:35 <prometheanfire> yep, basically
20:41:42 <prometheanfire> did you want to mention the neutron thing?
20:42:21 <smcginnis> Oh right. So there are neutron plugins that use neutron itself as a library. Even though there is a neutron-lib, but I don't know enough to know why they don't have everything they need there.
20:42:30 <smcginnis> Likely going forward they should try to get it all in the lib.
20:42:38 <smcginnis> But there are some stable branches that are broken right now due to this.
20:43:07 <smcginnis> Basically, theirs no upper-constraint for neutron, so things like stable/stein plugins are blowing up when the latest release of neutron gets installed.
20:43:16 <tonyb[m]> We agreed they'd do that in Denver
20:43:25 <prometheanfire> yep, and then they didn't
20:43:53 <smcginnis> It would appear that plan was not followed through.
20:43:57 <dirk> so we just add it for stable/ ?
20:44:05 <dirk> I think its not fair to have stable broken over this
20:44:10 <smcginnis> I agree.
20:44:17 <dirk> master is a different story
20:44:20 <prometheanfire> ya, just to stable seems like the right move
20:44:25 <tonyb[m]> Then we said we'd treat them like any other library and add them to UC
20:44:37 <prometheanfire> so abandon https://review.opendev.org/697369 ?
20:44:44 <smcginnis> tonyb[m]: Add neutron? Or neutron-lib?
20:44:50 <prometheanfire> or the plugin
20:45:05 <tonyb[m]> Neutron
20:45:31 <smcginnis> So what https://review.opendev.org/#/c/697369/ is doing is actually the plan?
20:46:00 <smcginnis> I haven't had time to pay attention there. I see they all need updates to pick up the blacklist validation issue.
20:46:00 <prometheanfire> remove from blacklist, make sure it's in gr, add to uc
20:46:09 <tonyb[m]> And change the release type to cycle with intermediary
20:46:15 <prometheanfire> ya
20:46:53 <tonyb[m]> I'll try to do that today (and update smcginnis change)
20:47:04 <smcginnis> Thanks, that would be great.
20:47:08 <tonyb[m]> And do the appriate email etc
20:47:16 <prometheanfire> ack
20:47:50 <smcginnis> Slawek is the PTL now. Might be good to get his input on the plan.
20:48:07 <tonyb[m]> Yeah I'll cc him
20:48:08 <smcginnis> I would hope they could just put common things in neutron-lib, but don't know the details.
20:48:41 <tonyb[m]> I don't think he was in the room
20:49:04 <smcginnis> I see him in -dev.
20:49:30 <tonyb[m]> Akihiro agreed to that as the ideal but .... Time and person power is limited
20:49:43 <prometheanfire> as always :P
20:50:08 <smcginnis> Pinged slaweq in -dev.
20:51:13 <tonyb[m]> smcginnis: thanks
20:51:29 <smcginnis> We can probably continue on. Guessing this is out of his normal working hours.
20:51:39 <smcginnis> But maybe we can sync up out of meeting.
20:52:19 <prometheanfire> ya
20:52:23 <prometheanfire> #topic open floor
20:52:45 <smcginnis> There's the django issue.
20:53:14 <smcginnis> I have uncap patches out there, but then there's the larger issue that we do need it blacklisted or something so it doesn't get updated from the one blessed version that they need to move to.
20:53:22 <prometheanfire> I think the uncap was approved
20:53:37 <smcginnis> So I was wondering rather than uncapping, would it be better to recap it to the new version?
20:53:46 <prometheanfire> isn't that what constraints is for?
20:53:54 <smcginnis> Or do we have a way to prevent it from wanting to update since they apparently don't follow semver.
20:54:22 <smcginnis> Yeah. I think the concern is just that upper-constraints will automatically get proposed to a newer version and we will need to remember to not allow that.
20:54:24 <prometheanfire> constraints?
20:54:29 <prometheanfire> ah
20:55:00 <prometheanfire> is it just one or two versions that need masking? if not and there is a longer term new cap then a new cap seems ok
20:55:18 <prometheanfire> but if the goal is to not cap, I think uncapping and being carful is the way forward
20:56:44 <smcginnis> I believe longer term new cap.
20:57:14 <smcginnis> Even doing a minor version bump requires some rework on the consuming end since they make backwards incompatible changes all the time.
20:57:22 <smcginnis> At least from what I understand.
20:57:50 <prometheanfire> ya, in that case just changing the cap seems right
20:58:10 <smcginnis> There are some comments about it on https://review.opendev.org/#/c/697612/
20:58:54 <prometheanfire> yep
20:58:57 <prometheanfire> anything else?
20:59:04 <smcginnis> Nope
20:59:07 <tonyb[m]> We're supposed to recapp but also get horizon input
20:59:20 <prometheanfire> ya, for the value to cap at
20:59:51 <smcginnis> That started things with https://review.opendev.org/#/c/697431/3
21:03:07 <prometheanfire> anyone have anything else?
21:04:29 <smcginnis> Talking about django drove everyone away. ;)
21:11:25 <prometheanfire> #endmeeting