[01:20:34] <anomie>	 MaxSem: I only used the class_exists check in extensions that don't specify the 'compatibility policy' infobox parameter on their mediawiki.org extension page.
[18:42:20] <Reedy>	 MaxSem: https://github.com/ExoticComet/AutoWikiBrowser/commits/master
[18:42:20] <Reedy>	 lol
[18:43:05] <MaxSem>	 godspeed
[18:44:53] <MaxSem>	 I like how everything required to make it run on WikiHow was a readme tweak
[18:46:36] <Reedy>	 xD
[18:46:38] <subbu|lunch>	 Reedy, MaxSem what is involved to have AWB work with Linter output?
[18:46:45] <Reedy>	 What do you mean?
[18:46:50] <subbu|lunch>	 this came up when i was talking with folks at wikimania
[18:47:01] <subbu|lunch>	 as in getting its list of pages from a linter category
[18:47:12] <Reedy>	 Is it just a standard MW category?
[18:47:17] <subbu|lunch>	 no. it comes from an api.
[18:47:20] <MaxSem>	 the thing is you're probably looking at fixing templates mostly
[18:47:37] <subbu|lunch>	 not always. regular pages are involved too.
[18:47:44] <subbu|lunch>	 the larger context is this
[18:48:12] <subbu|lunch>	 project wikicheck folks has asked us to absorb those error categories into linter .. and i got affirmations from an enwiki editor at wikimania.
[18:48:16] <Reedy>	 Writing an AWB listmaker to pull from the linter api module should probably be trivial
[18:48:39] <subbu|lunch>	 so, if more stuff moves into linter, it seems that awb should be able to hook into linter output.
[18:48:41] <subbu|lunch>	 ok.
[18:49:04] <subbu|lunch>	 linter also provides page offsets to precisely identify what needs fixing. can awb be fixed to use those offsets for targeted editing?
[18:49:16] <Reedy>	 https://en.wikipedia.org/w/api.php?action=help&modules=query%2Blinterrors
[18:49:36] <Reedy>	 It could be. That'd be more work obviously
[18:49:49] <subbu|lunch>	 yup
[18:50:44] <subbu|lunch>	 https://en.wikipedia.org/w/api.php?action=query&list=linterrors&lntcategories=deletable-table-tag 
[18:53:34] <legoktm>	 anomie: after reviewing extension patches, I find the naming of CommentStore::insert() confusing, because when it's next to $dbw->insert() calls, I expect it to actually do insertion, but it only prepares it.
[18:54:05] <anomie>	 legoktm: It does insert the comment into the comment table, if necessary.
[18:55:02] <legoktm>	 but not in MIGRATION_OLD mode right?
[18:55:55] <anomie>	 True, but eventually everything will be migrated and MIGRATION_NEW will be the only code path kept.
[18:56:07] <legoktm>	 right
[19:03:31] <legoktm>	 anomie: how is https://gerrit.wikimedia.org/r/376076 ? 
[19:28:48] <no_justification>	 Bleh, why is Warning: parsing XML declaration: '?>' expected in Entity, line: 1 in /srv/mediawiki/php-1.30.0-wmf.16/includes/resourceloader/ResourceLoaderImage.php on line 302 still a thing?
[19:30:37] <MatmaRex>	 no_justification: presumably because i've never seen that before and presumably there is no traceback :)
[19:32:33] <no_justification>	 We saw it last week :)
[19:32:36] <no_justification>	 And fixed it
[19:32:44] <no_justification>	 Somebody must've uploaded another bogus SVG :)
[19:55:42] <no_justification>	 Super spammy: Warning: Using deprecated fallback handling for comment rev_comment [Called from CommentStore::getCommentInternal in /srv/mediawiki/php-1.30.0-wmf.17/includes/CommentStore.php at line 200] in /srv/mediawiki/php-1.30.0-wmf.17/includes/debug/MWDebug.php on line 309
[19:56:35] <MaxSem>	 anyone knows any reason why shouldn't I merge https://gerrit.wikimedia.org/r/#/c/349457/ ?
[20:01:05] <legoktm>	 MaxSem: does it work if the schema change hasn't been deployed yet?
[20:02:06] <MaxSem>	 legoktm, it's already deployed :P
[20:02:20] <legoktm>	 oh woot
[20:02:25] <legoktm>	 is the table populated too?
[20:02:37] <legoktm>	 oh, that script is in this change
[20:05:32] <legoktm>	 MaxSem: lgtm :)
[20:05:37] <legoktm>	 I'm excited to see that land
[20:07:18] <legoktm>	 anomie: would it be beneficial to run Scribunto tests on luasandbox patches?
[20:07:24] <legoktm>	 (as part of CI)
[20:07:36] <anomie>	 legoktm: It probably would be.
[20:07:55] <MaxSem>	 heh, I just broke gerrit with this one comment
[20:10:04] <anomie>	 no_justification: Sigh. No stack trace so I can look at what caller is hitting that?
[20:23:30] <Reedy>	 | CrossNamespaceLinks     | 20091016041335 |
[20:23:30] <Reedy>	 | Mostlinkedcategories    | 20091020025923 |
[20:23:53] <Reedy>	 | DisambiguationPages     | 20130716141819 |
[20:23:54] <Reedy>	 | Disambiguations         | 20130910070914 |
[20:24:20] <James_F>	 Reedy: Timestamps of those tables last getting touched?
[20:24:26] <Reedy>	 Yeah
[20:24:35] <Reedy>	 Were some of these special pages removed?
[20:24:42] <James_F>	 Plausibly.
[20:24:53] <Reedy>	 Special:MostLinkedCategories is obviously still there
[20:24:56] <James_F>	 Or disabled in miser mode?
[20:24:59] <MaxSem>	 miser moded?
[20:25:15] <Reedy>	 I can't find a source of Disambiguations or DisambiguationPages in core
[20:25:24] <James_F>	 Reedy: Ext:Disambiguation
[20:25:33] <Reedy>	 You mean Disambiguator ?
[20:25:39] <James_F>	 Sure?
[20:26:01] <Reedy>	 That has DisambiguationPages and DisambiguationPageLinks
[20:26:07] <Reedy>	 It doesn't have Disambiguations...
[20:26:38] <Reedy>	 https://github.com/wikimedia/mediawiki/blob/master/includes/specialpage/QueryPage.php#L66-L111
[20:26:40] <MatmaRex>	 it used to be in core
[20:26:49] <MatmaRex>	 but it was broken so we killed it
[20:26:51] <James_F>	 Yeah, and maybe it got renamed when it moved out?
[20:26:52] <MatmaRex>	 and wrote the extension
[20:26:56] <James_F>	 Right.
[20:27:00] <Reedy>	 So the shit is still in the table
[20:27:08] <Reedy>	 Soo...
[20:27:09] <James_F>	 Kill it with fire?
[20:27:15] <Reedy>	 That's vaguely what I'm thinking
[20:27:32] <Reedy>	 if I run QueryPage::getPages()
[20:27:39] <Reedy>	 Then get [0] from each array...
[20:27:49] <Reedy>	 we can delete where qc_type not in []
[20:27:50] <James_F>	 What about the miser mode ones? It's probably better to have no data than have eight-year-old data.
[20:27:58] <Reedy>	 Well, that's a different case :)
[20:28:05] <Reedy>	 But yes
[20:28:09] <Reedy>	 I think we should be clearing those up too
[20:28:16] <James_F>	 For DisambiguationPages just DROP TABLE DisambiguationPages;?
[20:28:40] <Reedy>	 there's no table :)
[20:29:34] <James_F>	 Oh, right, meh.
[20:30:08] <Reedy>	 it's crap in querycache, querycachetwo, querycache_info
[20:30:17] * James_F nods.
[20:30:28] <James_F>	 MediaWiki: Why do anything consistently?
[20:30:41] <Reedy>	 https://phabricator.wikimedia.org/P5961
[20:31:10] <Reedy>	 I am confused why DisambiguationPages hasn't been updated recently though
[20:31:10] <Reedy>	 https://github.com/wikimedia/mediawiki-extensions-Disambiguator/blob/master/specials/SpecialDisambiguationPages.php
[20:31:41] <Reedy>	 Or is it because it's a reused key?
[20:32:20] <Reedy>	 https://github.com/wikimedia/mediawiki-extensions-Disambiguator/blob/master/specials/SpecialDisambiguationPageLinks.php has code for actively pushing into querycache
[20:34:37] <Reedy>	 Yeah.. This seems like it's gonna be the case
[21:07:06] <Reedy>	 https://gerrit.wikimedia.org/r/#/c/376072/4
[21:17:15] <no_justification>	 Reedy: -1
[21:17:23] <Reedy>	 Yeah
[21:17:28] <Reedy>	 I think we've got some others we should move
[21:35:20] <no_justification>	 Indeed
[21:39:26] <Reedy>	 There's some reimplementations of logged maintenance or whatever it's called
[21:50:50] <no_justification>	 anomie: Nothing I see in logstash (sorry for the lag, missed the ping somehow)
[21:57:47] <no_justification>	 Reedy tbh we should adjust those scripts if possible to just detect their run conditions so they're no ops and dont depend on update log. 
[21:57:57] <Reedy>	 heh, that too
[22:05:07] <no_justification>	 anomie: rev_comment && rc_comment, seeing errors for both
[22:17:50] <Reedy>	 https://phabricator.wikimedia.org/T174513
[22:18:04] <Reedy>	 So... Unless I'm mistaken, any non expensive pages should be in that list
[22:20:14] <Reedy>	 That makes no sense
[22:20:32] <Reedy>	 the querycache table should only have rows for expensive query pages?
[22:21:06] <Reedy>	 no_justification: Anything exciting?
[22:21:35] <James_F>	 Isn't there a cron job for expensive query pages? Or… some of them?
[22:21:49] <Reedy>	 James_F: That's my point. Only those should be in the query cache
[22:21:57] <Reedy>	 the non expensive ones aren't cahced
[22:22:01] <Reedy>	 So they shouldn't be in it
[22:22:22] <Reedy>	 https://github.com/wikimedia/mediawiki/blob/master/includes/specialpage/QueryPage.php#L220-L228
[22:22:26] <Reedy>	 return $this->isExpensive() && $this->getConfig()->get( 'MiserMode' );
[22:22:29] <James_F>	 I guess. Maybe they used to be expensive?
[22:22:45] <James_F>	 (Or in some cases, used to exist.)
[22:23:02] <Reedy>	 Actually.. I guess I shouldn't be using isExpensive().. I should be using isCached
[22:23:07] <Reedy>	 As that's what it's explicitly there to tell us
[22:23:40] <James_F>	 When do the expensive ones ever get run?
[22:23:40] <no_justification>	 Reedy: Nope, nothing 'citing
[22:23:55] <Reedy>	 Aren't some never run?
[22:25:13] <Reedy>	 https://github.com/wikimedia/puppet/blob/production/modules/mediawiki/manifests/maintenance/updatequerypages/cronjob.pp
[22:26:07] <James_F>	 Should we re-assess which ones are never run?
[22:26:35] <Reedy>	 Probably
[22:26:38] <no_justification>	 We should re-assess which ones we even have ;-)
[22:26:42] <Reedy>	 And work out if we should be running
[22:26:42] <no_justification>	 E_TOOMANYFEATURES
[22:26:43] <James_F>	 That one too.
[22:28:59] <James_F>	 https://phabricator.wikimedia.org/T175088
[22:54:35] <Reedy>	 sigh
[22:54:43] <Reedy>	 oojs ui breaking core tests again
[23:20:46] <Reedy>	 DELETE FROM `querycache` WHERE qc_type NOT IN () LIMIT 1000
[23:20:58] <Reedy>	 Why is it only breaking on the qunit tests?
[23:23:16] <Reedy>	 well, that's fun. works in eval.php not in maintenance script
[23:29:55] <Reedy>	 Oh, fscking miser mode is off
[23:56:37] <Reedy>	 TimStarling: About? Any idea about the lua test failures on https://integration.wikimedia.org/ci/job/mw-testskin/1603/console ?
[23:59:30] <TimStarling>	 looking
[23:59:50] <Reedy>	 thanks