[03:43:23] <TimStarling>	 sometimes git gui blame is almost too clever
[03:43:40] <TimStarling>	 I asked it where a line in parserTest.inc came from, and it said Setup.php
[04:28:21] <TimStarling>	 which is true, a lot of this global setup code has been blindly copied from Setup.php
[04:28:28] <TimStarling>	 		if ( $wgStyleDirectory === false ) {
[04:28:28] <TimStarling>	 			$wgStyleDirectory = "$IP/skins";
[04:28:28] <TimStarling>	 		}
[13:17:50] <anomie>	 tgr|away: Reviewed 307893. Also, re your question from two days ago, ApiMain's "permissions" (mostly useless IMO), "data types", and "credits" aren't included in ApiSandbox.
[15:35:25] <paladox>	 hashar im wondering could you help me with https://phabricator.wikimedia.org/P3964 please?
[15:35:29] <paladox>	 im getting this error
[15:35:43] <paladox>	 [02-Sep-2016 16:17:46 Europe/London] PHP Notice:  Array to string conversion in /home/randomwi/public_html/en/includes/skins/Skin.php on line 872
[15:35:43] <paladox>	 [02-Sep-2016 16:17:46 Europe/London] PHP Warning:  strpos() expects parameter 1 to be string, array given in /home/randomwi/public_html/en/includes/OutputPage.php on line 3598
[15:36:08] <paladox>	 from patch https://gerrit.wikimedia.org/r/#/c/193434/
[15:37:36] <paladox>	 it isent from that patch
[15:37:43] <paladox>	 i did some minor changes i have uploaded yet
[15:37:49] <paladox>	 due to it showing errors in the log
[15:37:59] <paladox>	 which i did not notice before i uploaded it.
[15:41:52] <paladox>	 or legoktm ^^ please?
[15:58:51] <AaronSchulz>	 hashar: how hard would it get jenkins in mysql strict mode?
[16:01:10] <paladox>	 or ostriches do you know how i can fix that error ^^ please?
[16:01:15] <paladox>	 I want it so we can do
[16:01:48] <paladox>	 $wgLogo = [ 'png' => "path/to/png", 'svg' => "path/to/svg" ];
[16:02:22] <paladox>	 and also stay backword compat with wgLogo so we can still do $wgLogo = "path/to/logo";
[16:05:02] <ostriches>	 Why does it need fallback? Just do svg or png and be done with it?
[16:05:12] <ostriches>	 Transforming $wgLogo into an array seems overly complicated
[16:05:43] <paladox>	 ostriches because for internet explorer 8 users
[16:05:47] <paladox>	 who doint support svg
[16:05:53] <paladox>	 and old android os.
[16:05:56] <ostriches>	 Then you shouldn't use an svg logo :)
[16:06:02] <paladox>	 But you can
[16:06:04] <ostriches>	 A wiki should pick one or the other :)
[16:06:08] <paladox>	 since you can do an svg with png
[16:06:12] <paladox>	 like we support in less
[16:06:33] <paladox>	 .background-image-svg( @svg, @fallback ) {
[16:07:27] <ostriches>	 I honestly don't see the utility.
[16:07:47] <paladox>	 oh
[16:07:57] <paladox>	 ostriches what about this https://gerrit.wikimedia.org/r/#/c/193434/62/includes/resourceloader/ResourceLoaderSkinModule.php
[16:07:58] * ostriches removes self as reviewer
[16:39:17] <hasharAway>	 AaronSchulz: there is task about switching to mysql strict mode 
[16:39:24] <hasharAwayy>	 https://phabricator.wikimedia.org/T119371 
[16:39:32] <hasharAwayy>	 and parent has a bunch of sub tasks as well https://phabricator.wikimedia.org/T108255
[16:39:51] <hasharAwayy>	 i guess if we get the proper mysql/mariadb package  and one find out the proper setting to enable it, that is doable ;]
[16:43:40] <AaronSchulz>	 hasharAwayy: you can do it per-session via SET
[16:43:56] <AaronSchulz>	 we issue some stuff via SET already...I wonder how much would break :)
[17:04:41] * AaronSchulz makes a patch
[17:32:11] <ostriches>	 AaronSchulz: https://gerrit.wikimedia.org/r/#/c/308205/ and https://gerrit.wikimedia.org/r/#/c/308200/, latter is super trivial
[17:49:58] <ostriches>	 AaronSchulz: I have a feeling this is gonna break shit :p https://gerrit.wikimedia.org/r/#/c/308208/
[17:58:08] <AaronSchulz>	 ostriches: did you check LQT?
[17:58:16] <ostriches>	 Nerp
[17:58:35] <AaronSchulz>	 that broke legoktm's patch with one of the globals
[18:09:39] <legoktm>	 ostriches: let's not...I have a patchto re-do everything. Someone just needs to review SemanticForms and FlaggedRevs to make sure it won't break them
[18:10:06] <ostriches>	 Ugh, I'm neck deep in SemanticForms right now
[18:10:11] <ostriches>	 It's such a pile of crap
[18:11:39] <legoktm>	 https://phabricator.wikimedia.org/T144366#2597458
[18:13:23] <ostriches>	 Yeahhh
[18:14:12] <ostriches>	 legoktm: Relatedly, if we find a way to pass a title to transformMsg() in SF_AutoeditAPI.php, I can kill the optionality of $title in transformMsg()
[18:14:28] <ostriches>	 That's the *only* caller without a title passed
[18:15:50] <legoktm>	 I hear you can use global $wgTitle; ;-)
[18:16:01] <ostriches>	 Hehehee, make it SemanticForms' fault?
[18:16:03] <ostriches>	 I can do that ;-)
[18:16:06] <ostriches>	 https://gerrit.wikimedia.org/r/#/c/308212/ lol
[18:16:31] <legoktm>	 AaronSchulz was too fast.
[18:16:54] <legoktm>	 ostriches: the class extends ApiBase...would $this->getTitle() not work?
[18:17:10] <legoktm>	 oh nvm, that's Special:BadTitle blah blah
[18:17:20] <ostriches>	 It's only used for parsing context of a message.
[18:17:26] * ostriches shrugs
[18:17:39] <ostriches>	 eh fuck it, wgtitle it is
[18:21:19] <legoktm>	 still reading that code...
[18:21:20] <legoktm>	 > Uncomment once compatibility to pre1.21
[18:21:24] <ostriches>	 legoktm: https://gerrit.wikimedia.org/r/#/c/308213/
[18:21:28] <legoktm>	 how far back does it try to support?
[18:21:50] <ostriches>	 Whatever the answer, the real answer is "too far"
[18:22:04] <AaronSchulz>	 legoktm: wait what about flaggedrevs?
[18:22:41] <legoktm>	 AaronSchulz: it creates an EditPage object and I didn't have time that day to trace how it is used and whether it depended upon globals
[18:23:48] <AaronSchulz>	 I can put it on my todo list to look
[18:24:02] <AaronSchulz>	 I used to remember off the top of my head...now, not so much ;)
[18:24:43] <legoktm>	 thanks :)
[18:25:33] <ostriches>	 Yay I win! https://gerrit.wikimedia.org/r/#/c/308215/
[18:28:50] <ostriches>	 legoktm: It kinda sucks tbh. SemanticForms seems very widely used by people, but it's basically abandonware.
[18:28:58] <ostriches>	 How it works and doesn't explode is beyond me...
[18:32:18] <ostriches>	 Heh, adding the type hint made it go boom
[18:36:06] <legoktm>	 ostriches: the sad part is that its not really abandoned...its still getting new features and stuff...just not following modern mw development
[18:41:25] <hashar>	 AaronSchulz: did you get a patch to SET strict mode ? :D
[18:41:47] <AaronSchulz>	 mostly passing
[18:42:14] <hashar>	 neat
[19:04:02] * robla starts looking for where to merge this page: https://www.mediawiki.org/wiki/Parsoid/test_cases
[19:14:47] <legoktm>	 robla: https://www.mediawiki.org/wiki/Parser_tests ?
[19:15:33] <robla>	 cool, that looks like a good spot!  thanks!
[19:16:48] * robla had started yak shaving :-)
[19:17:57] <paladox>	 legoktm please could you merge https://gerrit.wikimedia.org/r/#/c/193434/ ?
[19:19:02] <legoktm>	 no, I'll post on the task in a bit
[19:19:09] <paladox>	 Ok
[19:19:41] <paladox>	 legoktm did you manage to work on it and get it to work? This from last night?
[21:16:34] <AaronSchulz>	 hashar: https://gerrit.wikimedia.org/r/#/c/308220/
[21:19:33] <hashar>	 oh
[21:27:45] <hashar>	 AaronSchulz: you renamed the variable :)  Pretty sure you can add a small assertion in   tests/phpunit/includes/page/WikiPageTest.php   testDoDeleteArticle()
[21:27:59] <hashar>	 then I am not going to obsess over it :]
[21:28:58] <AaronSchulz>	 well after my STRICT patch lands, that kind of thing will fail a bunch of article delete tests
[21:29:16] <AaronSchulz>	 all I had to do was fix some RC fields...I thought it would be harder
[21:29:37] <hashar>	 it is almost midnight so not going to push that / CR+2 
[21:29:58] <hashar>	 note the live hack on wmf.17 would conflict whenever your patch land
[21:30:11] <hashar>	 hey great!
[21:30:21] <AaronSchulz>	 well that RE's job ;)
[21:30:35] <AaronSchulz>	 I assume we don't keep them around too long anyway
[21:30:58] <hashar>	 yeah :]
[21:31:02] <AaronSchulz>	 hashar: I mentioned renaming the variable, you had the chance to make me do it!
[21:31:09] <AaronSchulz>	 *do it then
[21:31:11] <AaronSchulz>	 heh
[21:32:04] <hashar>	 for STRICT patch, I dont know :]
[21:32:18] <hashar>	 if it lands and is enabled by default that is going to cause major havoc on other extensions
[21:32:35] <hashar>	 would be nice to feature switch it so we can enable it on beta and then CI
[21:32:38] <AaronSchulz>	 not by default in core
[21:32:40] <AaronSchulz>	 just CI
[21:33:07] <AaronSchulz>	 though jynus wants to get there *eventually* maybe
[21:33:11] <hashar>	 yeah
[21:33:12] <hashar>	 definitely
[21:33:27] <hashar>	 might want to enable it on beta a bit before turning it on for CI
[21:33:48] <hashar>	 or the other way around I dont know.  I guess it is too late for me 
[21:33:54] <AaronSchulz>	 it already passes CI, so isn't it fine?
[21:34:06] <hashar>	 on mediawiki/core ?
[21:34:08] <AaronSchulz>	 I had to base it on top of my deletion fix plus an RC fix
[21:34:13] <AaronSchulz>	 to get it to pass
[21:34:17] <hashar>	 the trick is
[21:34:36] <hashar>	 once enabled , wide range of extensions will start to fail
[21:34:50] <hashar>	 cause we do not run all extensions tests when a mediawiki core patch is proposed
[21:35:11] <AaronSchulz>	 which -extensions do we run for the extensions tests run on core changes?
[21:35:15] <AaronSchulz>	 just wmf extensions?
[21:35:20] <hashar>	 not to say that test coverage is far from complete  . A recent example is a serie of patches that broke LQT    :(
[21:35:36] <hashar>	 that only got caught on prod (since nothing exercise LQT on beta
[21:35:44] <AaronSchulz>	 well the test well NOT get us ready for prod, we will do logging first
[21:35:47] <hashar>	 (CI only uses a subset of wmf extensions)
[21:36:00] <AaronSchulz>	 I'm thinking CI, then logging (after CI to not spam logs so hard), then eventually prod
[21:36:03] <AaronSchulz>	 but it's quite far out
[21:36:16] <AaronSchulz>	 someone has a logging patch for warnings from '15 or something
[21:36:42] <AaronSchulz>	 well beta logging in there somewhere too ;)
[21:36:45] <AaronSchulz>	 maybe before prod
[21:36:52] <AaronSchulz>	 might save some log spam
[21:37:39] <hashar>	 then
[21:37:59] <hashar>	 maybe the strict mode in test suite will be unnoticed
[21:38:07] <hashar>	 might want to announce it
[21:38:25] <hashar>	 then if we can get a patch to flip it easily, we can get it enabled on beta and catch a wide range of stuff
[21:39:12] <hashar>	 AaronSchulz: also reach out to DBAs :]
[21:39:15] <hashar>	 sure they will love that
[21:40:12] <hashar>	 I am crashing to bed.  Well done Aaron!
[21:41:27] <AaronSchulz>	 see you
[21:43:04] <ostriches>	 AaronSchulz: https://gerrit.wikimedia.org/r/#/c/308283/