[09:01:39] <tgr>	 what's the best practice for renaming special pages?
[09:01:59] <tgr>	 in https://gerrit.wikimedia.org/r/c/mediawiki/extensions/ElectronPdfService/+/527203 I recommended using an alias but I'm having second thoughts about it
[09:02:23] <tgr>	 there isn't any way to define an alias that would work in all languages, is there?
[09:03:45] <revi>	 Let's say there's abusefilter hit and a spam blacklist hit at the same time, which one gets priority?
[09:08:42] <legoktm>	 revi: spamblacklist should check before AF does
[09:08:52] <revi>	 meh thank
[09:08:54] <revi>	 thanks*
[09:09:14] <legoktm>	 not the answer you were hoping for?
[09:09:18] <revi>	 I was thinking of an abusefilter to tell people "use full url not shortener"
[09:09:19] <revi>	 yup
[09:11:08] <legoktm>	 SB is faster and cheaper than AF (especially when hitting the condition limit is an issue), so a few years ago billinghurst had requested that it go before AF
[09:11:53] <legoktm>	 I would like to see us have a better experience around url shorteners, like auto expanding them as part of PST or something
[09:12:27] <revi>	 Flood of youtu.be and naver.me is could've been avoided by abusefilter injecting "USE FULL URL BOI instead of "You've hit spam blacklist"."
[09:12:33] <revi>	 (speaking of kowiki)
[09:13:11] <legoktm>	 why is youtu.be a problem? don't all the urls go to youtube.com?
[09:13:28] <revi>	 Shortener gets blanket blacklist
[09:13:40] <legoktm>	 but it's not an arbitrary/open shortener?
[09:13:54] <revi>	 there's a discussion on SBL 
[09:14:01] <revi>	 [[m:Talk:Spam blacklist#Youtu.be]]
[09:14:02] <revi>	 @link
[09:14:02] <wm-bot>	 https://meta.wikimedia.org/wiki/Talk:Spam_blacklist#Youtu.be
[09:14:14] <revi>	 meh lowercase y
[09:14:21] <revi>	 @link [[m:Talk:Spam blacklist#youtu.be]]
[09:14:21] <wm-bot>	 https://meta.wikimedia.org/wiki/Talk:Spam_blacklist#youtu.be
[09:15:25] <legoktm>	 I see
[09:15:51] <legoktm>	 Fixing bots/software seems a lot easier than fixing humans IMO
[09:16:06] <legoktm>	 but I'll have to look into it a bit more
[09:16:49] <revi>	 true :P
[14:55:01] <snitch>	 [[Tech]]; 2600:1:C204:56ED:0:52:C2D0:8C01; [none]; https://meta.wikimedia.org/w/index.php?diff=19267680&oldid=19266732&rcid=13938770
[14:55:13] <snitch>	 [[Tech]]; Hasley; Revert to revision 19266732 dated 2019-08-06 07:55:37 by ArchiverBot using [[:en:Wikipedia:Tools/Navigation_popups|popups]]; https://meta.wikimedia.org/w/index.php?diff=19267681&oldid=19267680&rcid=13938774
[16:27:56] <ebernhardson>	 how to print all configuration for an extension? Prior to extension.json you could load the hhvm debug with `mwrepl testwiki` and then type `g wgCirrus` and it would print all configuration related to Cirrus for that wiki
[16:28:20] <ebernhardson>	 it seems we've changed things now (maybe because we switched cirrus to extension.json) and that no longer returns any configuration. Any idea how to get it?
[16:36:35] <Lucas_WMDE>	 that should still work, shouldn’t it? in the end the configuration still ends up in globals, at least currently
[16:36:46] <Lucas_WMDE>	 and when I try mwrepl testwiki, g wgCirrus on mwdebug1002 I get loads of output
[16:37:08] <Lucas_WMDE>	 (well, some output and then “do you really want 2424 more variables printed” because HHVM is friendly enough not to dump all at once ^^)
[16:37:42] <ebernhardson>	 Lucas_WMDE: it turns out, mwscript on deployment.eqiad.wmnet is broken, but it works on mwdebug1002
[16:37:47] <ebernhardson>	 s/mwscript/mwrepl/
[16:38:33] <Lucas_WMDE>	 huh
[16:39:33] <Lucas_WMDE>	 strange
[16:39:39] <Lucas_WMDE>	 (I usually use shell.php anyways)
[16:49:46] <ebernhardson>	 Lucas_WMDE: how do you set breakpoints with shell.php? Or maybe you have a different use case?
[16:50:04] <ebernhardson>	 my main use of mwrepl is to inspect config and set breakpoints in production
[16:50:42] <Lucas_WMDE>	 different use case, I’ve never used debugger functionality in production
[16:50:52] <Lucas_WMDE>	 just dumped different bits of config, mostly
[16:51:27] <ebernhardson>	 interesting.
[16:52:01] <ebernhardson>	 maybe i'm just lazy, it seems so much easier to set a breakpoint, feed in the api params of a call that failed, and immediately see what happened
[16:53:54] <Lucas_WMDE>	 maybe it is easier, once you know how to set breakpoints :D
[16:53:57] <Lucas_WMDE>	 I just never looked into it so far
[16:54:11] <ebernhardson>	 :)
[17:21:59] <_joe_>	 ebernhardson: mwrepl is going away :/
[17:22:23] <_joe_>	 php-fpm has no such functionality :/
[17:23:12] <ebernhardson>	 _joe_: i know :( i'll have to dig out ancient scripts i had pre-hhvm that spun up php-fpm+xdebug on a custom port
[17:23:29] <_joe_>	 ebernhardson: oh I was about to say
[17:24:00] <_joe_>	 if you have something, we could work on having an xdebug enabled pool 
[17:24:08] <_joe_>	 with 1 worker
[17:24:50] <ebernhardson>	 _joe_: hmm, that would actually make sense for mwdebug hosts
[17:25:02] <_joe_>	 and switch to it based on a header that we filter at the edge layer, or just for the mwdebugs yeah
[17:26:39] <_joe_>	 but if you just need it occasionally to spun up, we can just make a script
[17:26:58] <_joe_>	 you want to connect to the console and that's just going to be on localhost
[17:28:15] <ebernhardson>	 my use is pretty occasional, usually its some bug that wont reproduce locally. I probably need to play around with php 7, i haven't really done anything there since 5.2 any things are probably quite different.
[17:31:41] <_joe_>	 me neither
[18:19:57] <Platonides>	 I didn't do anything serious recently either