[02:50:08] <legoktm>	 Krenair, Krinkle: how many copies of HotCat that don't pull from Commons do you think there are? https://phabricator.wikimedia.org/T67297#3641651
[05:33:13] <legoktm>	 MaxSem: maybe it's some stupid browser plugin?
[05:33:28] <legoktm>	 MaxSem: also your test edit was on en.wp which doesn't have the new branch yet I think?
[05:33:31] <MaxSem>	 on every browser in existence?
[05:33:44] <MaxSem>	 eh, riight:P
[05:34:50] <legoktm>	 I remember a while back Skype used to install a browser plugin that would mangle anything that looked even close to a phone number
[05:37:21] <legoktm>	 https://gerrit.wikimedia.org/r/#/c/163961/2/modules/ve-mw/init/ve.init.mw.Target.js ew
[05:37:28] <MaxSem>	 yaaay: https://test.wikipedia.org/w/index.php?title=Wikipedia:Sandbox&diff=326124&oldid=326123
[05:38:47] <legoktm>	 what setting did you change? just the encoding?
[05:39:47] <MaxSem>	 yup
[05:40:11] <MaxSem>	 interestingly, I couldn't repro the complete unicode mangling I did on enwiki
[05:41:17] <legoktm>	 so we need a better feature test :(
[05:41:47] <legoktm>	 lesson: trying to maintain a list of broken browsers is futile!
[05:47:06] <MatmaRex>	 legoktm: MaxSem: <input type=hidden> and <textarea>/<input type=text> might be handled differently by whatever browsers mangle this. try an invisible <textarea> perhaps?
[05:47:51] <MaxSem>	 at least in my testing in links it mangled it
[05:48:23] <MaxSem>	 and actually, the style of that mangling resembles what we're seeing now
[13:14:36] <anomie>	 Krinkle: thanks
[16:37:33] <Krinkle>	 legoktm: Presumably any outdated of HotCat still existing now are broken anyways.
[16:38:00] <Krinkle>	 Everytime we (volunteers) did a fix up, we tend to look for copies based on user reports and as part of fixing, replace with a call to Commons. 
[16:38:12] <Krinkle>	 and upstream any local translations if they are new/more complete. 
[20:23:18] <no_justification>	 Did anyone ever manage to dig into the RecentChanges/ORES/Closure issue?
[21:59:03] <legoktm>	 Yes I filed a ticket for it
[21:59:30] <legoktm>	 https://phabricator.wikimedia.org/T176236
[21:59:58] <legoktm>	 no_justification: ^
[22:24:49] <MaxSem>	 legoktm, Krinkle - any thoughts about our little Unicode issue?
[22:25:49] <legoktm>	 Did we actually regress?
[22:26:22] <MaxSem>	 that's what I'm trying to determine :P
[22:26:24] <legoktm>	 Based on your description I thought we're just catching people with weird settings that mangle Unicode?
[22:26:58] <MaxSem>	 Chrome doesn't even allow you to change page encoding anymore
[22:27:11] <legoktm>	 And if there's a way to bypass the check in some browsers that mangle unicode its not great but not a regression either
[22:28:21] <MaxSem>	 I'd still like a theory what's going on before we declare victory
[22:29:25] <legoktm>	 I think we need a user to complain to us so we can debug with them
[22:29:55] <legoktm>	 Because we're trying to find literally an edge case
[22:30:32] <MaxSem>	 cat AdHocDebug.log | wc -l
[22:30:32] <MaxSem>	 2243
[22:30:35] <MaxSem>	 eek
[22:35:46] <legoktm>	 All different users?
[22:37:49] <MaxSem>	 one bot running https://github.com/eldur/jwbf
[22:38:16] <MaxSem>	 I suppose it uses an old version cuz the current one appears to use the API
[22:38:36] <legoktm>	 Sigh
[22:38:39] <MaxSem>	 (holy crap, "proper" Java code is a pain to read)
[22:39:24] <MaxSem>	 another bot that just uses a CURL UA
[22:40:48] <MaxSem>	 704 attempts without hte check field total
[22:41:22] <MaxSem>	 one UA literally "Opera"
[22:45:14] <MaxSem>	 definitely some user scripts are broken
[22:46:57] <legoktm>	 I'm okay reverting it in the WMF branch and letting it go out next week with warnings sent out about the new parameter
[22:47:06] <legoktm>	 But I have little sympathy for screenscrapers
[22:48:53] <Krinkle>	 JobQueue problems solved. https://tideways.io/profiler/blog/using-php-fpm-as-a-simple-built-in-async-queue
[22:49:02] <Krinkle>	 Can we migrate to this? :D
[22:49:56] <MaxSem>	 uhh https://bs.wikibooks.org/wiki/MediaWiki:Monobook.js
[22:54:52] <legoktm>	 Poor MonoBook users
[22:58:09] <RoanKattouw>	 A while ago some people were talking about the idea of writing an authentication service (i.e. breaking the auth parts of MW out into a separate service to make it more difficult to compromise). Did that ever go anywhere?
[23:02:44] <legoktm>	 I think it was planned out but never implemented
[23:03:57] <legoktm>	 https://phabricator.wikimedia.org/T140813
[23:07:07] <MaxSem>	 we still got AuthManager though