[01:57:25] <legoktm>	 MaxSem: did you see my last comment on https://gerrit.wikimedia.org/r/#/c/376874/ ?
[02:02:17] <legoktm>	 maybe it's just me, but it feels like there's less release blockers this time around?
[02:10:43] <MaxSem>	 bleh, I wrote a long answer for you then decided that fuck it, I'll just leave MWS alone for now
[02:10:58] <MaxSem>	 thanks for poking :)
[02:13:04] <MaxSem>	 legoktm, that patch is still unfinished because it introduces slight changes to how memory options fall back on each other
[02:14:02] <legoktm>	 MaxSem: tgr said that those changes didn't really matter in practice?
[02:15:09] <MaxSem>	 mmm, if we only had a databse with settings of all wikis in existence... :)
[02:16:57] <legoktm>	 heh
[02:17:11] <legoktm>	 MaxSem: are you going to keep working on it, or should I pick it up? 
[02:17:33] <MaxSem>	 yep, I'm undoing some of my doing right now
[02:17:35] <legoktm>	 I have it marked as a 1.30 blocker because of the removal of Shell::command()
[02:18:49] <tgr>	 the difference is that in the old code if you set time but not walltime, it sets walltime as well
[02:19:11] <tgr>	 does not seem reasonable to me, but not hard to preserve
[02:19:23] <MaxSem>	 in practice, nobody overrides these per call anyway
[02:19:35] <MaxSem>	 yeah, I guess not a problem if we document it
[02:20:01] <tgr>	 I wouldn't mind dropping it as long as its explicit
[02:20:41] <tgr>	 and the stdout patch is done I think if you want to get that into 1.30 as well
[02:21:40] <tgr>	 Seb35's timeout patch is also done I think, the comments are just nitpicks
[02:22:35] <MaxSem>	 tgr & legoktm, cgroup() or cGroup()?
[02:23:47] <legoktm>	 cGroup() just looks...weird to me. I like cgroup()
[02:24:35] <tgr>	 +1 for cgroup
[02:26:04] <MaxSem>	 created the variable before the function, gave it a cooler name:P
[02:27:00] <Skizzerz>	 xXcgroupXx()
[02:27:16] <MaxSem>	 hehe
[02:28:14] <MaxSem>	 oops, also noticed an outdated comment
[02:31:18] <MaxSem>	 x2
[02:36:59] <MaxSem>	 https://gerrit.wikimedia.org/r/376874
[02:38:13] <legoktm>	 ah, you just got rid of the factory
[02:38:25] <MaxSem>	 yup:P
[02:38:36] <MaxSem>	 the factory is needed only for MWS
[02:39:12] <MaxSem>	 eh, still need relase notes
[02:50:41] <MaxSem>	 legoktm & tgr|away, doned
[02:52:25] <legoktm>	 MaxSem: I think you're missing injection of the logger?
[02:52:37] <MaxSem>	 duh
[02:55:05] <MaxSem>	 fixed
[03:00:19] <legoktm>	 you have a phpcs error, but other than that, lgtm
[16:36:09] <legoktm>	 anomie: on https://gerrit.wikimedia.org/r/#/c/378261/ you meant "without this patch" right?
[16:36:50] <anomie>	 legoktm: ... Yes.
[16:40:00] <legoktm>	 anomie: are there alternative ways to work around this that don't require bumping the minimum postgres version? like could we check for the enum already existing in MW before trying to use the patch file?
[16:41:49] <anomie>	 legoktm: Probably, using PG's system catalogs (the rough equivalent of MySQL's information_schema).
[16:44:08] <legoktm>	 are you up for implementing that? :) 
[16:45:10] <legoktm>	 https://phabricator.wikimedia.org/T164898 (which is archcom approved) only proposed bumping the minimum version to 9.2, not 9.3
[16:47:07] <anomie>	 I think the query would be something like `select 1 from pg_type join pg_enum on(enumtypid = pg_type.oid) where typname='media_type' and enumlabel='3D'`, if it returns a row then the patch should be skipped.
[18:17:21] <Krinkle>	 Reedy: Just to confirm - https://gerrit.wikimedia.org/r/#/c/355541/ was fixed, right?
[18:17:49] <Reedy>	 Krinkle: Yeah, Roan and I had a few followup patches, but it works now
[18:17:55] <Krinkle>	 k
[18:18:02] * Krinkle unstars
[18:18:06] <Reedy>	 https://github.com/wikimedia/mediawiki/commits/master/maintenance/populatePPSortKey.php
[18:18:11] <Krinkle>	 Wish we had the followup/fixme label 
[18:18:19] <Krinkle>	 that can be untagged
[18:18:49] <Reedy>	 long live Special:CodeReview