20:33:04 <prometheanfire> #startmeeting requirements
20:33:05 <openstack> Meeting started Wed Dec 19 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'
20:33:19 <prometheanfire> #topic rollcall
20:33:21 <prometheanfire> tonyb, prometheanfire, number80, dirk, coolsvap, toabctl, smcginnis, dhellmann ping
20:33:24 <prometheanfire> o/
20:33:27 <tonyb> \o
20:34:56 <ssbarnea> tonyb: the gate is blocked aparently due to a newer ansible-lint version. This review was fixing the linting issue (pre-commit involves pinning), but got stuck due to two issues: ansible being listed in requirements (unexpected and unrelated to that change) and 2nd missing pre-commit. Now both are fixed and that change will pass all jobs.
20:35:31 <tonyb> ssbarnea: Okay
20:35:41 <ssbarnea> tonyb: short story: "requirements-check" runs only when you touch requirements, guess what: the master was broken. D
20:36:02 <prometheanfire> #topic Any controversies in the Queue?
20:36:13 <prometheanfire> ssbarnea: :P
20:36:34 <tonyb> ssbarnea: Possibly someone would have to look at the git and gate histories to work how that happened
20:36:45 <prometheanfire> https://review.openstack.org/626383 seems ok at first glance, but I'd like stable team to look as well
20:36:49 * prometheanfire pokes tonyb :P
20:37:20 <smcginnis> o/
20:37:32 <ssbarnea> tonyb: this can happen when people over-optimize conditons when jobs are run. also adding a new job without running it is another source of such issues.
20:37:40 <ssbarnea> anyway sorted, and thanks for the help.
20:38:10 <prometheanfire> anyone have anything else?
20:38:32 <smcginnis> No
20:38:55 <prometheanfire> #topic open floor
20:38:58 <tonyb> Not from me
20:39:17 <prometheanfire> will leave open for a couple min then close
20:40:31 <tonyb> prometheanfire: Why's you abandon https://review.openstack.org/#/c/626096/ ?
20:48:54 <dhellmann> ssbarnea , tonyb : that happened because before the rocky cycle the requirements-check job only looked at the lines in the changed files, and did not validate the whole file on each change. we modified that last cycle.
20:49:57 <tonyb> dhellmann: but how'd we add the line to the file without runnign the job, and therefore not validate it?
20:50:07 <ssbarnea> dhellmann: here is one idea: why not making requirements-check a linter and including it in the linters jobs which always run?
20:50:11 <dhellmann> tonyb : maybe it was valid when it was added
20:50:13 <tonyb> dhellmann: not that it really matters now
20:50:31 <tonyb> ssbarnea: Nothing stopping you from doign that
20:51:03 <dhellmann> ssbarnea : it's configured now to check the whole file when it's touched now, but the tools to do that require access to the requirements repo so you can't just check it in place
20:51:14 <tonyb> ssbarnea: the current situation is a setup to cover what we consider the possible issues and reduce nodepoll useage
20:51:15 <ssbarnea> i had few similar problems in the past, all caused by skips of jobs on some conditions. running too many jobs can be resource intensive but grouping inexpensive ones inside linters could do the trick.
20:51:20 <dhellmann> I guess the point is that the way things are now this problem shouldn't come up again
20:54:59 <tonyb> prometheanfire: ping?
20:57:43 <prometheanfire> pong
20:57:46 <prometheanfire> #endmeeting