[08:58:52] hi [09:05:09] New patchset: Hashar; "no submodule fetch for phpcs jobs" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/45557 [09:05:10] New patchset: Hashar; "git-zuul scm macro" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/45556 [09:08:16] New patchset: Hashar; "no submodule fetch for phpcs jobs" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/45557 [09:08:16] New patchset: Hashar; "git-zuul scm macro" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/45556 [09:08:55] New review: Hashar; "PS2: fix duplicate 'scm', add some new macros" [integration/jenkins-job-builder-config] (master); V: 0 C: 0; - https://gerrit.wikimedia.org/r/45556 [09:09:07] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/45556 [09:09:18] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/45557 [09:12:18] New review: Hashar; "fix up bug https://bugzilla.wikimedia.org/show_bug.cgi?id=44250" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/45557 [09:15:29] New review: Hashar; "That is more or less the idea :-]? I need to get Jenkins Job Builder to create the phpcs job, not su..." [integration/zuul-config] (master); V: 0 C: 0; - https://gerrit.wikimedia.org/r/45744 [14:01:04] hashar: sorry for the mess at https://gerrit.wikimedia.org/r/#/c/11137/ ^^' [14:14:58] Nemo_bis: I sorted it out :-] [14:15:21] Nemo_bis: you removed some global $wgFoo; while rebasing :-] Luckily that caused the tests to fail, so it has been a very easy fix [14:19:57] ^demon: Just a quick heads up ... is it ok to leave hooks-bugzilla for now until commit rights have been sorted out and start attacking your "gerrit LDAP group removal" problem? [14:20:46] <^demon> Yeah, I think BZ looks good for now. I started skimming your commits, but haven't reviewed them in full yet. [14:20:53] <^demon> I'll get to those this afternoon. [14:21:12] Great :-) [14:21:25] <^demon> I think Luca has submit permissions on hooks-its now, so hopefully the its base can go in today. [14:21:51] That's good to hear [14:22:06] I hope he accepts my patches without much problems [14:35:25] ^demon: hi Tom. Can you possibly abandon https://gerrit.wikimedia.org/r/#/c/30636/ ? That change is already in core :-] [14:35:58] <^demon> What commit? [14:36:04] no idea [14:36:13] I should git blame in the history [14:36:29] that changes added a div id=loginend in includes/templates/Userlogin.php [14:36:36] which is there nowadays [14:36:43] I guess munaf took another way [14:44:32] <^demon> Yeah, it would just be nice to reference the actual change before we abandon. [14:44:35] <^demon> :) [14:44:53] not willing to waste time on blaming the history :-] [14:50:17] Change abandoned: Hashar; "would restore / apply whenever labs/nagios-builder.git pass pep8 tests." [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39694 [15:53:56] New patchset: Hashar; "(bug 39950) deps for mw/ext/TranslationNotifications" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/46251 [15:57:29] New patchset: Hashar; "test pipeline for mw/ext/TranslationNotifications" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/46252 [15:57:55] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/46252 [16:04:16] Gullom, we're preparing a GeoData rollout blog post, currently drafted at https://www.mediawiki.org/wiki/Extension:GeoData/blogpost . To which date can we place it in the calendar? [16:04:24] I am out, be back later tonight [16:36:16] MaxSem: Try using guillom ;) [16:36:31] argh [16:36:34] heh, indeed :) [16:37:23] MaxSem: when would you like it to be published? [16:37:40] the strangest part of it is that I usually try to autocomplete ordinary words as I type - how the hell did I manage not to autocomplete there? [16:37:41] I'm actually editing the calendar right now, so I can add it [16:38:49] guillom, when there will be a slot for it - the software's already deployed and I hope we'll finish the post today [16:40:48] MaxSem: if you finish it today, I can review and publish it tomorrow or on Wednesday [16:40:58] guillom, awesome, thanks [16:41:06] I'll add it to the calendar for tomorrow and we can push it if it's not ready [16:41:09] np [16:41:38] MaxSem: FYI for future posts it's better to draft them on meta: https://meta.wikimedia.org/wiki/Wikimedia_Blog/Drafts (but it's not a big deal) [16:42:28] heh, I'm used to use mw.o as an operations base [17:06:18] Reedy: MaxSem: could I get you to review this? https://gerrit.wikimedia.org/r/#/c/44832/ [17:06:33] I'm not sure who to ping... I'll try Tim tonight [17:07:44] <^demon> I'm afraid of how it might behave on windows. [17:08:03] ^demon: it makes windows work [17:08:11] at least some windows [17:08:38] still working with a guy to track down his problem on windows (but he has it w/o the patch) [17:09:32] ^demon: but if there is something you'd like me to check I have access to a couple different windows systems that I could test on [17:09:43] hexmode, I wonder if for some of these calls a failure should result in an exception instead of relative path being returned (eg FileRepo) [17:09:46] and I only control one of them ;) [17:10:11] <^demon> Hmm, I see now. [17:10:27] MaxSem: good point... there was no error checking on the original code, though. [17:11:13] well, for installer it was ok - the wiki worked (just not maintenance scripts) [17:14:11] realpath('.') (from WebStart.php) is what gave the other guy problems. Still is, even with that fix. I'm not sure why, but, in any case, realpath('.') will cause problems if the parent dir can't be read. [17:16:28] <^demon> I think we should just add proper error checking on a case-by-case basis with realpath(). [17:16:43] <^demon> I agree--always returning a relative path isn't necessarily the best behavior. [17:17:22] k, I'll update the patch [17:17:23] <^demon> (and makes the error checking for callers kind of silly. $path == wfRealpath( $path )) [17:17:54] <^demon> If you wanted to say throw an exception instead of silently failing. [17:18:15] well, that check works only if we know for sure that the path is relative [17:18:33] <^demon> Indeed. [17:22:52] heh... just found a way to reproduce the realpath problem on a linux box... it took me some effort, though [17:23:07] who chmod's parent directories with 000? [17:41:23] <^demon> qchris: Initial commits of hooks-its and hooks-bugzilla have been merged. Yay, now we've got a nice common base to work from :) [17:41:35] Yay \o/ [17:42:08] So I'll go ahead and rebase the commits this evening [17:42:31] (just wrangling gerrit-dev.wmflabs setup) [17:42:33] <^demon> Did you and Luca come to an agreement on the "parent with only one child" issue? [17:42:43] More ore les [17:42:54] ^=More or less [17:43:07] For hooks-its he has plans to add further artifacts [17:43:26] So I guess it's best to leave it for now and bikeshed later if no furtehr artifacts pop up [17:43:35] For hooks-bugzilla he did not comment [17:43:36] <^demon> *nod* [17:43:55] So I guess it's fine to drop the additional artifact for hooks-bugzilla [17:44:14] <^demon> Maybe. There's something to be said for being consistent. [17:44:28] replication pulign does not have this additional layer [17:44:38] <^demon> This is true. [17:44:59] I mean I am fine either way, but why add unneded gruft? [17:45:10] <^demon> Yeah, that's fine. We can always undo that later if we change our mind. [17:45:18] <^demon> *minds [17:45:20] Ye.s [17:45:28] Yes. :-) [17:45:50] <^demon> Like, groups is not an issue. We're not going to use BZ groups for Gerrit. That's what ldap's for :) [17:46:09] :-) [17:46:41] But I guess we will have to discuss the generated comments of hooks-bugzilla [17:46:54] They are not in best shape [17:47:07] Sometimes way too long, or unstructured. [19:39:27] matthiasmullie: ping [19:39:40] preilly: pong [19:39:56] am in meeting, so there might be huge translatlantic lag ^^ [19:40:28] matthiasmullie: should this change-set https://gerrit.wikimedia.org/r/#/c/24264/ be abandoned? [19:41:00] matthiasmullie: that way you can just create a new one with DataModel removed that doesn't have 56 patch sets [19:43:13] AaronSchulz: what do you think of that suggestion? [19:43:42] ori-l: ping [19:43:49] probably makes sense since there will be no similarity [19:44:04] AaronSchulz: yeah that's what I was thinking [19:44:10] it can be done by just fudging the change-id before git-review [19:44:24] AaronSchulz: yeah, totally [19:44:27] or whatever is easiest [19:45:03] AaronSchulz: now we just need to wait for his reply post meeting e.g., account for transatlantic lag [19:45:08] there will be a lot of similarity; all "controller" code is still legit [19:45:34] but most of the history is no longer relevant, so killing the current commit does make sense [19:45:56] matthiasmullie: okay cool — do you mind doing that or should I? [19:46:05] i'll do it right away [19:46:20] matthiasmullie: Okay cool — thanks [19:46:41] dschoon: have you used Neo4j at all? [19:46:51] yeah, a little. [19:46:53] * preilly — http://www.neo4j.org/ [19:46:57] it was 2ish years ago [19:47:02] dschoon: what did you think of it? [19:47:29] i used it during a hackathon at addthis.com to build the graph of shared content and run pagerank [19:47:40] AaronSchulz: have you used Neo4j [19:47:49] good interfaces. at the time it wasn't distributed, so there were serious limits to what you could do [19:47:56] i think they've been working on sharding etc [19:47:57] dschoon: ah, I see [19:48:11] * AaronSchulz has only read about it [19:48:33] dschoon: the developers at Yerdle used it this weekend and it seems to fit the use case pretty well [19:48:35] the interface is indeed interesting, but I know little about performance characteristics [19:49:47] *nod* [19:50:07] binasher: re the centralnotice schema change -- peter and I have a deploy window from 1 to 3 today to push some things; and if that schema change is pushed we can also deploy some other stuff! (which would be cool...) Do you have any time in that window to push the change? [19:50:08] ultimately i ended up implementing the Gremlin interfaces on top of our datastore so i didn't have to import everything [19:50:12] * AaronSchulz reads http://jim.webber.name/2011/02/on-sharding-graph-databases/ [19:50:14] https://github.com/tinkerpop/gremlin/ [19:57:14] * preilly — wishes that he knew how to read... [19:59:19] dschoon: you've got to love things like, "—–oOOo-(_)-oOOo—–"…. [19:59:31] :) [19:59:44] it works quite well though [20:00:02] and it's nice to see a query api abstracted out from the storage details ^^ preilly [20:01:27] ori-l: have you read: ? [20:02:53] dschoon: https://github.com/nuvolabase/orientdb#scalability-the-database-is-the-bottleneck-of-most-of-applications [20:03:01] 125? lol [20:03:11] * AaronSchulz has some skepticism, heh [20:03:44] I'm sure one construct a case where that happens though [20:03:50] *one could [20:03:55] AaronSchulz: suffers from incredulity [20:07:12] bah [20:09:48] dschoon: are you using at all? [20:09:56] preilly: so anything using neo4j would need Cypher queries, hrm [20:10:12] ! [20:10:12] ok, you just typed an exclamation mark with no meaning in the channel, good job. If you want to see a list of all keys, check !botbrain [20:10:19] * AaronSchulz looks at the rest api [20:10:19] i have not seen that, preilly [20:10:25] (i suspect it is ... two days old.) [20:10:37] 11 days [20:10:37] he [20:10:38] h [20:22:23] <^demon> hashar: phpcs is eating up a lot of cpu on gallium, btw :\ [20:23:09] :( [20:23:11] ^demon: checking [20:23:34] maybe I should stop having it run after each merge hehe [20:26:20] cancelled them manually [20:28:22] New review: Hashar; "deployed earlier today" [integration/jenkins-job-builder-config] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/46251 [20:28:22] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/46251 [20:29:01] ^demon- Feel like reviewing !g 40998 quick? [20:29:37] <^demon> Oh yeah. I had cherry picked that and ran the tests. [20:29:47] <^demon> I didn't know if you wanted to submit it w/o the change to luasandbox. [20:31:09] It'll work without the change to luasandbox (it has to for luastandalone, after all). The change to luasandbox just does it faster. [20:32:16] And I've got a mw.uri module written, but it needs 40998 and three of the 4 patchsets left before ustring can be merged. [20:32:38] <^demon> Gotcha. [20:33:09] And we learned Gerrit really does not like a patchset that depends on multiple unmerged patchsets ;) [20:33:26] (unless it's a linear A depends on B depends on C) [20:35:43] <^demon> anomie: Bah. I'm getting test failures I didn't get before. [20:35:57] <^demon> http://p.defau.lt/?cK5xhnD_Q3APFT5feai8yA [20:36:04] <^demon> (#5 was self-inflicted and unrelated) [20:37:21] That's odd, that's what you'd get if you applied the unit tests but not the patch to mw.lua [20:37:41] <^demon> I'm running origin/master. [20:38:42] <^demon> Oh, whoops. [20:39:00] <^demon> User error. [20:47:33] New patchset: Hashar; "optimize phpcs by ignoring certain files" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/46291 [20:48:26] ^demon: so I made phpcs to ignore languages/messages/Messages files which are slow to parse [20:48:28] preilly: that document provides low-level details about the kafka wire protocol; it's only necessary for writing bindings for kafka in a new target language. If you use Java, you can simply make high-level calls to Kafka APIs without having to worry about the minutae of encoding data. [20:48:32] ^demon: and also only lint .php and .inc [20:48:40] hashar: siebrand will like you for that [20:48:53] I am pretty sure of it :-] [20:49:01] for when Jenkins decides to go lolno for no real reason :D [20:49:15] <^demon> :) [20:49:23] !google define lolno [20:49:23] https://www.google.com/search?q=define+lolno [20:49:33] New review: Hashar; "deployed" [integration/jenkins-job-builder-config] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/46291 [20:49:34] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/46291 [20:49:59] Reedy: ^demon whenever you see something awkward, please please fill a bug / mail me / ping me :-] [20:52:31] https://integration.mediawiki.org/ci/job/gerrit-plugin-hooks-bugzilla/2/ :-] [20:52:47] ^demon: sounds like you have started working on the bugzilla extension :-] [20:53:03] <^demon> Yep, qchris has done a lot of work on it. [20:53:15] <^demon> I've just been setting up the building stuff in jenkins and reviewing his code :) [20:53:23] oh qchris is back around [20:53:24] Hi [20:53:25] awesome [20:53:31] Salut hashar :-) [20:54:17] Servus Christian [20:54:22] <^demon> qchris: I just finished reviewing + merging everything up through https://gerrit-review.googlesource.com/#/c/41652/ [20:54:36] (my austrian is all about googling stuff) [20:55:09] ^demon: Great! (I'm still stuck at reproducing the LDAP problem :-( ) [20:55:33] <^demon> Don't beat yourself up over it. I've been pulling hair for the past week over it. [20:55:36] hashar: you're googling austrian stuff? :-)) [20:55:52] qchris: when I want to say "hello" to someone, yup :-] [20:56:30] ^demon: can I setup a second gerrit on the labs instance, so I can debug there, orr will this kill something if I screw up the instance? [20:57:02] hashar: I took you for an Austrian. Servus, grias di! [20:57:07] <^demon> You have to do some manual messing around with it. hashar knows a lot about setting up other gerrit installs on labs. [20:57:23] <^demon> s/a lot/has done it before outside of gerrit-dev/ [20:58:38] I'll just compile it in my users home directory and let it run on a different port, but i'll have to escalate my rights on the db, so I can test removing people from groups [21:00:01] ^demon: I think I updated some doc somewhere [21:00:11] my labs instance is missing the verified+2 though [21:00:28] qchris: https://wikitech.wikimedia.org/view/Gerrit that gives some hints [21:00:33] <^demon> Well for the things we're testing that doesn't matter as much. [21:00:37] <^demon> We mainly care about ldap. [21:00:53] there a few culprit though like having to download the gerrit war from somewhere [21:01:00] chad has some being build in Gerrit [21:01:00] * qchris takes a look [21:01:25] to create the mysql db make sure to use the https://wikitech.wikimedia.org/view/Gerrit#Database_install [21:01:31] that makes sure the DB uses utf-8 [21:01:34] the upstream doc does not use that [21:01:43] <^demon> https://integration.mediawiki.org/nightly/gerrit/wmf/?C=M;O=D - latest build is 45a3621e [21:02:07] and there is a puppet class [21:02:31] New patchset: Stefan.petrea; "Adding new pageview reports mobile (ready-to-merge)" [analytics/wikistats] (master) - https://gerrit.wikimedia.org/r/41979 [21:02:32] I guess I'll manage to get it built [21:02:54] But the database on gerrit-db.pmtpa.wmflabs is not used live anywhere. is it? [21:04:04] no idea [21:04:16] I also looked at http://ci.openstack.org/gerrit.html [21:04:43] So this means I better not touch the db then :-( [21:05:21] well gerrit takes care of creating a db for you [21:05:30] I guess you just need to install mysql-server [21:05:53] Yes, but with a clean install, I cannot reproduct the problems that we see in wmflabs [21:06:40] qchris: I have added you to the labs "integration" project. That would give you access to integration-jenkins2.pmtpa.wmflabs instance [21:06:52] Reedy: can you do database schema updates? [21:07:03] Thanks hashar [21:07:16] there is a repo cloable via ssh://qchris@integration-jenkins2.pmtpa.wmflabs:29418/mediawiki/core.git [21:07:28] I think the credentials are the same as in labs [21:07:37] you will have to log on http://integration.wmflabs.org/gerrit/ [21:07:42] and upload your ssh public key there [21:07:56] feel free to play with it [21:08:05] if you screw up something there, I will just recreate one [21:08:05] mediawiki/core ...? We're just debugging LDAP issues on gerrit, aren't we? [21:08:11] Eithor way thanks. I'll have alook [21:08:24] well that let you submit changes [21:08:25] That sounds great [21:08:27] if you need :-] [21:08:35] Ok wonderful thanks. [21:08:41] that is setup more or less like in production [21:08:49] gerrit is in /var/lib/gerrit2/ [21:08:57] it is home dir being /var/lib/gerrit2/review_site [21:08:57] <^demon> qchris: gerrit-dev is currently using data copied from production (I was testing upgrading from 2.4.2 -> master). [21:09:07] <^demon> So presumably, if we upgrade prod, we'll get hit with the same bug. [21:09:13] I can't remember the version IIRC a 2.4.2 with a custom patch from openstack [21:10:04] integration-jenkins2 is a Wikimedia Zuul on labs! (role::zuul::labs). [21:10:05] hmm [21:10:16] no Gerrit puppet class applied on it :/ [21:10:19] I must have set it up manually [21:10:44] np. Don't worry. I'll give it a try and if I get stuck I'll know whom to ask :-) [21:10:52] I think I gave you sudo right too [21:11:00] but please please don't screw up the other instances :-] [21:11:09] * qchris has sudo :-))) [21:11:12] <^demon> gerrit-dev is also running "jenkins-gerrit-core-64" [21:11:30] <^demon> Which relates to https://integration.mediawiki.org/ci/view/Java/job/gerrit-core/64/ (the last build I promoted) [21:11:38] I'll do my best to not set the place on fire. [21:13:25] in conf call [21:13:47] qchris: if the instance die, it is not a big deal [21:13:53] I can recreate it in a couple hour [21:13:55] s [21:14:04] Sounds wonderful. [21:14:13] (But hopefully we don't need it) [21:21:10] [21:21:27] again! [21:21:58] Heh [21:38:46] mwalker: I can, yeah [21:39:09] Reedy: it's cool -- I've got mr schultz doing it. but thanks :) [21:49:29] New patchset: Ottomata; "Now installing packet-loss as part of udplog .deb package" [analytics/udplog] (master) - https://gerrit.wikimedia.org/r/46303 [21:49:45] Change merged: Ottomata; [analytics/udplog] (master) - https://gerrit.wikimedia.org/r/46303 [21:52:00] ori-l: I was just curious if you read it [21:55:13] <^demon> qchris: Do you know when j2bugzilla 2.2 plans to make it out? [21:55:27] Thomas told me "today in the evening" [21:55:36] but it'll be 2.1.2 [21:55:40] i'll have to update the pom [21:55:43] <^demon> Ah, ok. [21:55:57] <^demon> I was just curious. I saw your changes worked fine with 2.2-SNAPSHOT. [21:56:19] I can send you the patches that I sent to Thomas, so you can see yourself (if you want) [22:16:18] boo [22:16:29] the sartoris repo has been created at the root of our git hierarchy :/ [22:17:02] <^demon> Yes, we know. [22:17:06] <^demon> It doesn't belong anywhere else. [22:17:22] <^demon> We bikeshedded over this at length. [22:17:23] wikimedia or operations/software maybe [22:17:26] hehe [22:17:30] <^demon> No, it's not going to be wmf-specific. [22:17:34] ahh [22:17:37] good to know [22:18:01] will create some jobs for it I guess [22:18:10] AaronSchulz: refreshLinks jobs are still piled up on enwiki, 3x where it was on friday. the graphite graphs look normal again though, is the dedupe issue fixed? [22:22:39] New patchset: Hashar; "pep8 job for sartoris" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/46443 [22:23:50] New review: Hashar; "deployed" [integration/jenkins-job-builder-config] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/46443 [22:23:51] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/46443 [22:32:26] New patchset: Hashar; "pyflakes builder macro and job template" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/46446 [22:32:26] New patchset: Hashar; "pyflakes job for sartoris" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/46447 [22:32:36] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/46446 [22:32:43] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/46447 [22:36:09] New patchset: Hashar; "triggers for sartoris jobs" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/46449 [22:37:00] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/46449 [22:53:22] New patchset: Hashar; "wrap pyflakes in tox" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/46453 [22:54:16] New review: Hashar; "there is no python-tox yet :(" [integration/jenkins-job-builder-config] (master); V: 0 C: -2; - https://gerrit.wikimedia.org/r/46453 [22:58:05] New patchset: Hashar; "pyflakes jobs do not vote by default" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/46455 [22:58:06] New patchset: Hashar; "sartoris-pep8 is now voting" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/46456 [22:59:06] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/46456 [22:59:06] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/46455 [23:01:38] preilly: hi :-] I have made Jenkins to run pep8 on sartoris repository [23:01:58] preilly: so non pep8 change will cause jenkins-bot to [23:02:36] vote verified -1 [23:34:09] hashar: thanks! [23:34:28] preilly: :-D [23:35:11] hashar: awesome. thanks [23:35:29] good to see us pep8'ing in a votiing way in at least one spot :) [23:37:45] ahh [23:37:47] i managed to make it work [23:37:49] sending patches [23:39:44] <^demon> Ryan_Lane: https://gerrit.wikimedia.org/r/#/c/45330/ fixes some pep8 violations :) [23:40:03] * Ryan_Lane twitches [23:40:20] <^demon> Review with Ignore Whitespace -> All. [23:40:24] <^demon> It's simple then. [23:41:15] how do I make that preference stick? [23:42:00] <^demon> Press Save? [23:42:08] I did [23:42:19] then I hit ] and it goes away [23:42:30] <^demon> Update and save? [23:42:33] <^demon> I dunno, wfm. [23:42:39] damn gerrit [23:43:33] grmblblb [23:43:40] sartoris spams #mediawiki :D [23:43:43] ooohhh [23:43:46] save is off the page [23:43:50] I was hitting update [23:43:51] <^demon> lol. [23:44:03] <^demon> you need a larger screen. [23:45:39] sartoris/sartoris.py:307: undefined name 'json' [23:45:41] bala