[00:00:02] ebernhardson: explained here: http://www.gossamer-threads.com/lists/wiki/wikitech/391001#391001 [00:00:14] ori-l: did you mean to pass console pack around on top as well? It seems to be absent on top, only at the bottom: https://en.wikipedia.org/wiki/User:Ori.livneh/common.js [00:00:40] Krinkle: eep, yes [00:00:56] ori-l: you'll probably want the other one to be guarded to avoid a reference error exceptoin [00:01:05] e.g. typeof or window.console [00:01:11] (03CR) 10EBernhardson (WMF): "I started working on porting over to less, but didn't get very far. For reference there is good info in the wikitech thread here: http://" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/86856 (owner: 10Matthias Mullie) [00:02:20] legoktm: yay, go us [00:02:42] ori-l: Very nice, bookmarked also. On Commons I see HotCat is 75 KB [00:02:54] legoktm: (that flyout bug is filed already, btw) [00:03:04] which flyout bug? [00:03:13] oh [00:03:16] brackets showing? [00:03:21] yeah [00:03:33] !bug :echo section mention [00:03:34] https://bugzilla.wikimedia.org/buglist.cgi?quicksearch=%3aecho+section+mention [00:04:02] You marked it as easy and didn't fix it? :< [00:04:30] legoktm: i only have twenty-four hours a day. [00:04:44] Wow, at least I go to sleep. [00:04:44] Krinkle: my only reservation about it is that it's dangerous to have a one-dimensional notion of performance, i.e., one that starts and stops with byte size. it could lead to a witch hunt of larger modules when they might not be the worst offenders) [00:05:52] ori-l: Witch hunts I think might be more appropriate are of the size of things that have a 5-minute cache timeout (as opposed to 30 days) and the execution time of things in the top queue [00:07:22] (03PS2) 10saper: Generate PDFs if $wgScriptPath is empty [extensions/Collection] - 10https://gerrit.wikimedia.org/r/87293 (owner: 10Bpetty) [00:07:35] RoanKattouw: https://en.wikipedia.org/w/index.php?title=User%3AOri.livneh%2Fcommon.js&diff=575657908&oldid=575657279 [00:08:31] Thanks :) [00:08:34] (03PS1) 10MaxSem: Extract variant-related code into a separate function [core] - 10https://gerrit.wikimedia.org/r/87486 [00:08:41] ori-l: Those are just my pet peeves :) [00:08:45] (03PS1) 10MarkTraceur: Add file-link "parser" to mw.Title from commons [core] - 10https://gerrit.wikimedia.org/r/87487 [00:08:53] In general, I'm very happy that someone's working on RL performance analysis, thank you very much for that [00:09:05] ori-l: indeed, I'm more worried about execution performance after the code arrives. [00:09:28] a silly dom action can slow you down more than a few KB of download size up front [00:09:31] it's fun to work on [00:09:44] maybe less on mobile where download is more important, too, but even then there is very good caching. [00:09:59] The localStorage thing seems promising [00:10:09] Someone mentioned it around the office just the other day, I think it might have been Trevor [00:10:16] execution is more O(N), it happens over and over and contiously as you scroll and interact. [00:10:24] That too [00:10:32] But I'm more worried about init time [00:10:33] Krinkle: i'm hoping that http://ganglia.wikimedia.org/latest/graph.php?r=week&z=xlarge&hreg[]=client-side&mreg[]=%5Ebrowser.pageSpeed.%28desktop%7Cmobile%29_median%24>ype=line&title=pageSpeed%3A+domInteractive+to+loadEventEnd&aggregate=1 points out regressions [00:10:34] RoanKattouw: might be, but I know we talked about it in DC and in Israel, I even implemented some of it [00:10:49] that's why the size table Ori shows reminded me of it, (similar styling in chrome) [00:10:54] right [00:12:22] Krinkle: I make no claim to originality, I think the technique was popularized by Steve Souders in this 2011 blog post: http://www.stevesouders.com/blog/2011/03/28/storager-case-study-bing-google/ [00:12:46] but then there was that other blog post about the atrocious performance characteristics of localStorage and nobody wanted to touch that API for the next year [00:12:58] (03CR) 10Krinkle: [C: 04-1] "quick notes:" [core] - 10https://gerrit.wikimedia.org/r/87487 (owner: 10MarkTraceur) [00:13:05] then there was a follow-up that pointed out the awful performance manifests itself only in very specific circumstances [00:13:47] oh, wrong blog post, ermm [00:13:51] ori-l, FYI I put @import guidance in https://www.mediawiki.org/wiki/Manual:Coding_conventions/CSS#Import [00:13:58] here we go http://www.stevesouders.com/blog/2011/09/26/app-cache-localstorage-survey/ [00:14:03] ori-l: yeah, I think I might have been inspired by a souder's post (not sure which). I used to follow that blog a lot, not so much anymore though. proven him wrong a few times, and didn't have time to keep up with it either. looking forward to get back to it though. [00:14:23] Krinkle: I'm able to get a title object from a src on my local wiki, so it's probably OK [00:14:32] spagewmf: <3! appreciate it. i was just feeling guilty over not having done that yet. [00:15:35] Will add tests though [00:15:49] np. I didn't grok C.Scott's reply about media-query restrictions and the (reference) option, so left them out ;-) [00:16:01] though when he was wrong, it's not so much wrong as it was a bit too extreme. But I kind of relate to that (people wont do what you tell them anyway, and shouldn't) - it's just that some people take the extreme and do it extreme without knowing why, and then stuff breaks and dude X posts it on stackoverflow wondering what his script tag is doing inside a style tag in between and [00:16:02] :P [00:16:12] (03PS1) 10EBernhardson (WMF): Convert from SASS to LESS [extensions/Flow] - 10https://gerrit.wikimedia.org/r/87488 [00:17:20] yep [00:17:56] (03PS2) 10Physikerwelt: Try to introduce some error handling [extensions/Math] - 10https://gerrit.wikimedia.org/r/87429 [00:18:14] Krinkle: btw, https://gerrit.wikimedia.org/r/#/c/86616/ should be noncontroversial, I think [00:18:26] (03CR) 10EBernhardson (WMF): "The patch containing my work to convert from SASS -> LESS is at I5c4528b58477af09885f50ecbe8a8ecf8c2d430d. If you want to work off it fe" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/86856 (owner: 10Matthias Mullie) [00:22:08] (03PS4) 10Krinkle: Update lessphp to 261f1bd28 [core] - 10https://gerrit.wikimedia.org/r/86616 (owner: 10Ori.livneh) [00:22:13] (03CR) 10Krinkle: [C: 032] Update lessphp to 261f1bd28 [core] - 10https://gerrit.wikimedia.org/r/86616 (owner: 10Ori.livneh) [00:22:31] url makes it easier to verify without a local checkout [00:22:51] makes sense, thanks [00:22:55] ori-l: you're trending on lessphp pulse [00:23:15] https://github.com/leafo/lessphp/pulse/monthly [00:23:44] heh, i think he was pretty excited about mediawiki integration [00:24:03] which might be the reason he merged my patches rather than, say, their amazing quality [00:24:28] see also: https://github.com/leafo/lessphp/pull/480#issuecomment-25652509 [00:24:35] > I'm gong to release a new numbered version soon, are there any other changes you wanted to make first? [00:24:46] yeah, saw that [00:25:01] we're like, yeah, sure, but we're already using it in production without a release [00:25:06] xD [00:25:21] release number [00:25:23] heh, I just meant that if there are other things we wanted to do it might be a good time [00:25:52] i also have this idea of convincing him to create a lessphp organization and to deputize some folk to help with triaging issues and merging pull requests [00:26:50] (03Merged) 10jenkins-bot: Update lessphp to 261f1bd28 [core] - 10https://gerrit.wikimedia.org/r/86616 (owner: 10Ori.livneh) [00:26:56] was going to suggest you if i did, is that something you'd be interested in? [00:28:18] create a lessphp organization meaning create a github organization, so it's github.com/lessphp/lessphp and more than one person can merge stuff [00:28:37] i am envisioning a LESSPHP HQ skyscraper in manhattan :P [00:28:41] s/am/am not! [00:32:03] (03PS1) 10Siebrand: Support GENDER for "contribsub2" [core] - 10https://gerrit.wikimedia.org/r/87490 [00:33:06] greg-g: Argh, is it OK to push another change? Trivial file rename in VE so that the icon is actually picked up. :-( [00:33:13] greg-g: Cherry pick to wmf20 only. [00:33:23] (03PS2) 10Siebrand: Support GENDER for "contribsub2" [core] - 10https://gerrit.wikimedia.org/r/87490 [00:33:38] greg-g: ( https://gerrit.wikimedia.org/r/87491 ) [00:54:38] (03PS1) 10Ori.livneh: LESS embeddable(): Use lessc::toBool [core] - 10https://gerrit.wikimedia.org/r/87493 [01:02:06] (03PS1) 10Catrope: Update VisualEditor to wmf20 to fix syntaxhighlight.svg file name [core] (wmf/1.22wmf20) - 10https://gerrit.wikimedia.org/r/87498 [01:11:39] (03CR) 10Catrope: [C: 032 V: 032] Update VisualEditor to wmf20 to fix syntaxhighlight.svg file name [core] (wmf/1.22wmf20) - 10https://gerrit.wikimedia.org/r/87498 (owner: 10Catrope) [01:19:19] (03CR) 10Helder.wiki: [C: 031] Support non-English namespace names in DiscussionParser::getUserFromLine [extensions/Echo] - 10https://gerrit.wikimedia.org/r/87477 (owner: 10Legoktm) [01:24:15] (03PS2) 10MarkTraceur: Add file-link "parser" to mw.Title from commons [core] - 10https://gerrit.wikimedia.org/r/87487 [01:25:09] (03CR) 10MarkTraceur: "Tests done, will rebase when mw.Title refactor is done." [core] - 10https://gerrit.wikimedia.org/r/87487 (owner: 10MarkTraceur) [01:44:24] (03CR) 10Spage: [C: 04-1] "(3 comments)" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/87488 (owner: 10EBernhardson (WMF)) [01:45:57] (03PS2) 10Jackmcbarn: Allow appending text to redirects from moved pages [core] - 10https://gerrit.wikimedia.org/r/87345 [01:47:00] (03CR) 10Jackmcbarn: "I think it increases future flexibility to pass text rather than a message to makeRedirectContent. I've added inContentLanguage(), and I'v" [core] - 10https://gerrit.wikimedia.org/r/87345 (owner: 10Jackmcbarn) [01:56:37] (03CR) 10Parent5446: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/86883 (owner: 10BryanDavis) [02:06:05] (03CR) 10Parent5446: "Are these functions still used? Because if not we should also throw some wfDeprecated() in their rather than just relying on docs." [core] - 10https://gerrit.wikimedia.org/r/87479 (owner: 10Physikerwelt) [02:09:18] (03CR) 10Parent5446: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/87302 (owner: 10PleaseStand) [02:10:42] (03CR) 10PiRSquared17: "Thank you for working on this." [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/78946 (owner: 10Legoktm) [02:13:46] (03CR) 10Physikerwelt: "After this" [core] - 10https://gerrit.wikimedia.org/r/87479 (owner: 10Physikerwelt) [02:14:34] (03CR) 10Parent5446: "One day we'll actually ditch this function and merge the Uri class." [core] - 10https://gerrit.wikimedia.org/r/87482 (owner: 10saper) [03:20:42] (03CR) 10Catrope: [C: 032] mw.Title: Rewrite from scratch (porting logic from Title.php) [core] - 10https://gerrit.wikimedia.org/r/83047 (owner: 10Krinkle) [03:24:00] (03Merged) 10jenkins-bot: mw.Title: Rewrite from scratch (porting logic from Title.php) [core] - 10https://gerrit.wikimedia.org/r/83047 (owner: 10Krinkle) [03:34:35] Heh... guess it's a good thing that the .src of an doesn't take into account redirects. [04:09:36] (03PS1) 10Tim Starling: Wikipedia archive import script [extensions/WikimediaMaintenance] - 10https://gerrit.wikimedia.org/r/87503 [04:14:26] ori-l: do you have some time to talk about MassMessage? [04:16:15] i was going to do what i usually do, which is take credit it it works well but rely on reedy to debug it if there are issues [04:16:36] anyways, i have few minutes; whats up? [04:16:46] heh. [04:16:48] * ori-l actively attacks Elsie. [04:17:33] well right now the main functionality we're missing that EdwardsBot had was finding out when your job is done [04:17:51] I thought you'd decided to send an Echo notification about that? [04:18:08] https://gerrit.wikimedia.org/r/#/c/87001/ <-- my diea [04:18:10] idea* [04:18:28] Elsie: yes, but a) not all wikis have echo b) it relies on memcache, which ideally should not be depended upon [04:18:37] Both fair points. [04:19:09] I would rather not send a notification at all if we cannot rely on the notification getting to most users [04:19:18] because if memcache loses state, the notification will never be sent [04:20:11] I'm looking at your patch [04:20:29] oh and the other thing with using Echo, is that it doesn't support cross-wiki notifications, so if your last job finishes on xxwiki, and was started on metawiki, you need to go back into the job queue to end up on meta to send the echo notification [04:22:37] Which bug is this? [04:23:20] https://bugzilla.wikimedia.org/show_bug.cgi?id=54861 [04:23:53] How can I have 956 unread e-mails... [04:24:07] Bugzilla e-mails, I mean. [04:24:16] I have many thousand more unread e-mails generally. [04:25:38] Echo has an API [04:25:53] so why can't you just send the notification cross-wiki by using the API? [04:26:06] there are abstractions for making API calls from within MediaWiki code [04:26:38] do you mean like a /w/api.php endpoint? those are read only, not for sending notifications [04:27:05] are you certain? I could have sworn.. [04:27:36] there is action=echomarkread which marks your notifications as read, and meta=notifications which shows you the notifications [04:27:48] only extensions can create Echo notifications [04:27:59] or the core echo code like reverts, mentions, etc [04:31:45] the issue with using the internal PHP API designed for extensions, is that the user should receive the notification that they send the message from [04:32:07] however due to cross-wiki messages, the last message might end up being on some random wiki instead of metawiki [04:32:19] and theres no way to create an echo notification on xxwiki and have it delivered on metawiki [04:32:59] look at includes/DbEchoBackend.php in Echo [04:33:02] ori-l: I think you'll find there are a lot of reasons currently that Echo and MM are incompatible. [04:33:47] you see the line: $dbw = MWEchoDbFactory::getDB( DB_MASTER ); [04:33:51] Echo has no cross-wiki support, as I understand it. [04:35:24] ori-l: right, but ideally we wouldn't be dependent upon the fact that the WMF uses the db backend [04:35:27] you can replace that with $dbw = wfGetLB( 'enwiki' )->getConnection( DB_MASTER, array(), 'enwiki' ); [04:35:31] even though no other backend exists… [04:36:40] you could also patch Echo to add a sub-class [04:36:58] ForeignDbEchoBackend.php, say [04:37:12] that takes a wiki DB name as a parameter to the constructor [04:37:21] and uses that to acquire a database connection [04:37:38] normally people would be very reluctant to review that code because it's easy to screw up [04:38:13] My recommendation would be to go back a few steps. [04:38:18] The goal is notifying a user of a completed job. [04:38:27] Ideally that would not involve hacking another extension. [04:38:34] but if you were careful to reproduce (or invoke) DbEchoBackend and only modified the way a database handle is acquired then you'd be OK [04:38:41] Elsie: why? it sounds like a fun project [04:38:48] and it'd be useful [04:39:03] It's focus drift. :-) [04:39:24] maybe a little [04:39:30] I'm all for improving Echo, but I see that as mostly unrelated to getting MM deployed. [04:39:41] Especially if Echo is in such an undeveloped state. [04:39:54] Does the public Web API really not support sending notifications? [04:40:03] without Echo, there's no easy way to notify a user without a talk page notification or maybe a log entry [04:40:16] Elsie: Echo isn't designed for users messaging other users [04:40:16] Or an e-mail. [04:40:32] Or... a Growl alert! [04:40:39] IRC ping. [04:40:42] Tweet! [04:40:57] hehe [04:41:08] I've been working on an IRC bot that pings you when you get a new notification [04:41:25] https://gerrit.wikimedia.org/r/#/c/80710/ came out of it [04:41:41] as for being memcached-specific, just look at http://www.mediawiki.org/wiki/Manual:$wgObjectCaches [04:42:11] i have to run [04:42:25] technically its not memcache specific, its whatever $wfMemc is [04:42:47] legoktm: Can you file a bug for Erik's comment and cross-reference? [04:42:47] just that as I understand it, $wgMemc should not be depended upon as a persistent cache [04:43:03] job queue questions can be deferred to the job queue expert, Elsie [04:43:10] :P [04:43:15] legoktm: The bug would be to re-add restrictions for cross-wiki notifications. [04:43:17] back later [04:43:18] thanks for the help ori-l [04:43:24] (03PS1) 10Tim Starling: Fix revision table cleanup on delete [core] - 10https://gerrit.wikimedia.org/r/87505 [04:43:25] Elsie: sure [04:44:12] Related question: If you have a spamlist with cross-wiki targets, and try to send a message to it without the -global userright, what happens? [04:44:21] Does it warn you and then silently remove the ones you can't send to? [04:44:25] legoktm: Thx. Sorry about the back and forth. I don't really agree, but Erik is being super-reasonable and nice, so... [04:44:34] Does it throw an error and refuse to send until you remove those entries? [04:44:57] Do we need a separate user right? [04:45:18] You could also do it as a config switch [04:45:20] I don't think a -global user right is needed. [04:45:31] Yeah. [04:45:46] Let's create a Meta-Wiki variable or something. [04:45:49] We'll special-case Meta-Wiki. [04:45:52] So that it can post in the main namespace. [04:46:01] And it can post to wikis outside of meta.wikimedia.org. [04:46:06] How's that sound? [04:46:19] $wgPrimaryHostWiki = 'metawiki'; or something. [04:46:23] er, posting in the main namespace isn't handled in the extension [04:46:37] Hm? [04:46:37] thats just a simple thing in the wmf config [04:47:05] if dbname == metawiki { allow posting in NS:0 } [04:47:07] No? [04:47:13] 'wgNamespacesToPostIn' => array( '+metawiki' => array( NS_MAIN) ), [04:47:23] something like that. [04:47:36] What's the default of that variable? [04:47:49] NS_USER_TALK, NS_PROJECT [04:48:12] if i remember $wgConf correctly, the +metawiki will merge the arrays [04:48:21] Okay. [04:48:24] I'm pretty burned out. [04:48:29] You'll file that bug? [04:48:34] We can work out the other issues later. [05:42:33] (03PS1) 10Legoktm: Use PROTO_CANONICAL correctly. [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87509 [05:42:39] Elsie: ^ [05:54:43] (03PS2) 10Netbrain: Added "search and replace within selection" functionality [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/87322 [06:00:22] (03PS3) 10Siebrand: Support GENDER for "contribsub2" [core] - 10https://gerrit.wikimedia.org/r/87490 [06:00:23] (03PS1) 10Siebrand: Declare visibility on class properties [core] - 10https://gerrit.wikimedia.org/r/87510 [06:23:03] (03PS1) 10Legoktm: Link to specific revision of spamlist that was used [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87512 [06:25:03] (03CR) 10Ori.livneh: "Aside: toBool should really be static.. I should submit another patch upstream." [core] - 10https://gerrit.wikimedia.org/r/87493 (owner: 10Ori.livneh) [06:25:13] (03CR) 10Legoktm: "Is it safe to assume that a valid page title won't have any bad characters for an XSS vulnerability? Or should it be escaped somehow?" [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87512 (owner: 10Legoktm) [06:27:05] (03PS1) 10Yuvipanda: Rename UploadWizardCampaign class to Campaign [extensions/UploadWizard] - 10https://gerrit.wikimedia.org/r/87513 [06:49:26] (03CR) 10Eranroz: "[Though it is not directly related to this specific patch:]" [core] - 10https://gerrit.wikimedia.org/r/80492 (owner: 10Amire80) [07:07:15] (03CR) 10Siebrand: [C: 031] "L10n/i18n reviewed." [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/87322 (owner: 10Netbrain) [07:36:12] (03PS1) 10TTO: Add namespace names for Malayalam (ml) [extensions/Scribunto] - 10https://gerrit.wikimedia.org/r/87515 [08:03:20] (03PS8) 10Dan-nl: addressing brian wolff’s comments [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/83657 [08:06:38] (03CR) 10Siebrand: [C: 032] Add namespace names for Malayalam (ml) [extensions/Scribunto] - 10https://gerrit.wikimedia.org/r/87515 (owner: 10TTO) [08:07:59] (03Merged) 10jenkins-bot: Add namespace names for Malayalam (ml) [extensions/Scribunto] - 10https://gerrit.wikimedia.org/r/87515 (owner: 10TTO) [08:09:37] (03PS14) 10Dan-nl: addressing mark traceur’s comments [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/82619 [08:09:49] (03CR) 10Nikerabbit: "Doesn't MessagesEn.php need changing?" [core] - 10https://gerrit.wikimedia.org/r/87490 (owner: 10Siebrand) [08:11:45] (03CR) 10Liangent: [C: 04-1] "$wgContLang->armourMath( ) is still called in Extension:Math." [core] - 10https://gerrit.wikimedia.org/r/87479 (owner: 10Physikerwelt) [08:13:07] hmm anyone around here who's a hadoop lover, is it reasonable to assume that someone would want more than 10 thousand input files at once to it? (example: someone processing en wp xml dumps) [08:13:34] (03PS4) 10Siebrand: Support GENDER for "contribsub2" [core] - 10https://gerrit.wikimedia.org/r/87490 [08:14:08] (03CR) 10Siebrand: "MessagesEn.php updated in PS4." [core] - 10https://gerrit.wikimedia.org/r/87490 (owner: 10Siebrand) [08:16:05] (03PS9) 10Dan-nl: addressing brian wolff’s comments [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/83657 [08:17:47] yuvi [08:17:49] (03CR) 10Siebrand: "These patch sets are huge, and they're hard to review. I prefer smaller patch sets addressing fewer things." [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/82619 (owner: 10Dan-nl) [08:17:49] woops [08:18:43] apergos: what do you mean, at once? do what at once? [08:19:32] YuviPanda: [08:19:43] (03PS1) 10Liangent: Remove one $wgContLang usage. [extensions/Math] - 10https://gerrit.wikimedia.org/r/87519 [08:20:06] ori-l: nm, I can ask the person who wanted the xml dumps split into pieces and find out directly [08:20:13] if they are not afk... [08:20:27] (03CR) 10Liangent: "Hm Math got rewritten." [core] - 10https://gerrit.wikimedia.org/r/87479 (owner: 10Physikerwelt) [08:24:06] (03CR) 10Liangent: [C: 04-1] "Yeah armourMath is still needed." [core] - 10https://gerrit.wikimedia.org/r/87479 (owner: 10Physikerwelt) [08:25:30] (03CR) 10Liangent: "Ah those relevant lines got moved to the extension." [core] - 10https://gerrit.wikimedia.org/r/87479 (owner: 10Physikerwelt) [08:27:06] (03CR) 10Liangent: "Will be no longer needed once 85801 is merged." [extensions/Math] - 10https://gerrit.wikimedia.org/r/87519 (owner: 10Liangent) [08:28:10] (03PS10) 10Yuvipanda: Add caching & cache invalidation for Campaign pages [extensions/UploadWizard] - 10https://gerrit.wikimedia.org/r/86960 [08:28:11] (03PS2) 10Yuvipanda: Rename UploadWizardCampaign class to Campaign [extensions/UploadWizard] - 10https://gerrit.wikimedia.org/r/87513 [08:30:32] (03CR) 10Liangent: [C: 04-1] "(1 comment)" [extensions/Math] - 10https://gerrit.wikimedia.org/r/85801 (owner: 10Physikerwelt) [08:38:35] (03CR) 10Nikerabbit: "(4 comments)" [core] - 10https://gerrit.wikimedia.org/r/87479 (owner: 10Physikerwelt) [08:40:29] (03CR) 10Nikerabbit: [C: 04-1] "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/87482 (owner: 10saper) [08:41:35] (03CR) 10Nikerabbit: [C: 032] Support GENDER for "contribsub2" [core] - 10https://gerrit.wikimedia.org/r/87490 (owner: 10Siebrand) [08:44:37] (03Merged) 10jenkins-bot: Support GENDER for "contribsub2" [core] - 10https://gerrit.wikimedia.org/r/87490 (owner: 10Siebrand) [08:58:12] (03PS1) 10Raimond Spekking: Consistency tweaks in preparation for adding extension to translatewiki.net [extensions/MultimediaViewer] - 10https://gerrit.wikimedia.org/r/87522 [09:02:44] (03CR) 10Hashar: "Sorry, I should have tested it, and probably need an integration test to verify the Zuul configuration." [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/87475 (owner: 10MarkTraceur) [09:06:40] (03CR) 10TMg: [C: 04-1] "Why does this add an additional checkbox? Why not make the existing button aware of the selection? Work on the selection if something is s" [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/87322 (owner: 10Netbrain) [09:08:19] (03PS3) 10Bpetty: Generate PDFs if $wgScriptPath is empty [extensions/Collection] - 10https://gerrit.wikimedia.org/r/87293 [09:08:49] (03CR) 10Bpetty: [C: 04-1] "You forgot to declare the $wgServer global." [extensions/Collection] - 10https://gerrit.wikimedia.org/r/87293 (owner: 10Bpetty) [09:12:52] (03CR) 10Siebrand: [C: 031] "L10n/i18n reviewed." [extensions/CentralAuth] - 10https://gerrit.wikimedia.org/r/86874 (owner: 10Anomie) [09:15:41] (03CR) 10Siebrand: "Added a few more designers for feedback. Netbrain, can you please provide a screenshot of the feature somewhere? (Possibly in a bugzilla f" [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/87322 (owner: 10Netbrain) [09:30:10] (03PS3) 10Tpt: Implement ContentHandler facility for page pages [extensions/ProofreadPage] (pagePagesRefactoring) - 10https://gerrit.wikimedia.org/r/85418 [09:30:57] (03CR) 10Tpt: "Fix a bug with proofreadpage_qualityX_message" [extensions/ProofreadPage] (pagePagesRefactoring) - 10https://gerrit.wikimedia.org/r/85418 (owner: 10Tpt) [09:37:50] (03CR) 10Netbrain: "https://bugzilla.wikimedia.org/show_bug.cgi?id=54961" [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/87322 (owner: 10Netbrain) [09:40:22] (03CR) 10Netbrain: "I think it is better that the functionality is visible, instead of just assuming that it exists. Better usability, and the extra checkbox " [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/87322 (owner: 10Netbrain) [09:45:03] (03PS1) 10Addshore: Use covers tags in tests [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87525 [09:47:15] (03PS1) 10Addshore: Remove pointless override teardown method [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87526 [09:50:52] (03PS1) 10Tpt: Major refactoring of the JS part of the page edition system [extensions/ProofreadPage] (pagePagesRefactoring) - 10https://gerrit.wikimedia.org/r/87527 [09:53:04] (03Abandoned) 10Tpt: Major refactoring of the JS part of the page edition system [extensions/ProofreadPage] - 10https://gerrit.wikimedia.org/r/87044 (owner: 10Tpt) [09:56:46] (03PS1) 10Addshore: Add some limited scope to the special page [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87528 [10:16:52] (03CR) 10Siebrand: [C: 031] "i18n/L10n reviewed." [extensions/ProofreadPage] (pagePagesRefactoring) - 10https://gerrit.wikimedia.org/r/87527 (owner: 10Tpt) [10:35:40] (03PS5) 10Zfilipin: Moving code used in multiple repositores to a Ruby gem [qa/browsertests] - 10https://gerrit.wikimedia.org/r/87125 [10:39:20] (03Abandoned) 10Zfilipin: @language tag is now only used in ULS repository [qa/browsertests] - 10https://gerrit.wikimedia.org/r/86837 (owner: 10Zfilipin) [10:40:36] (03CR) 10Zfilipin: "Luis, can you please take a quick look and let us know if this can be merged into master?" [qa/browsertests] - 10https://gerrit.wikimedia.org/r/86355 (owner: 10Yuvipanda) [10:56:18] (03CR) 10Zfilipin: [C: 031] Adding new CREDITS and LICENSE files, and license headers to all files, per bug 49313. [qa/browsertests] - 10https://gerrit.wikimedia.org/r/86355 (owner: 10Yuvipanda) [11:00:16] (03CR) 10Siebrand: [C: 032] Consistency tweaks in preparation for adding extension to translatewiki.net [extensions/MultimediaViewer] - 10https://gerrit.wikimedia.org/r/87522 (owner: 10Raimond Spekking) [11:00:24] (03Merged) 10jenkins-bot: Consistency tweaks in preparation for adding extension to translatewiki.net [extensions/MultimediaViewer] - 10https://gerrit.wikimedia.org/r/87522 (owner: 10Raimond Spekking) [11:06:29] (03PS2) 10Zfilipin: Adding new CREDITS and LICENSE files, and license headers to all files. [qa/browsertests] - 10https://gerrit.wikimedia.org/r/86355 (owner: 10Yuvipanda) [11:06:31] (03CR) 10jenkins-bot: [V: 04-1] Adding new CREDITS and LICENSE files, and license headers to all files. [qa/browsertests] - 10https://gerrit.wikimedia.org/r/86355 (owner: 10Yuvipanda) [11:31:56] (03PS1) 10Ori.livneh: Use '//'-style comments in LESS doc blocks [core] - 10https://gerrit.wikimedia.org/r/87536 [11:37:54] (03PS3) 10Zfilipin: Adding new CREDITS and LICENSE files, and license headers to all files. [qa/browsertests] - 10https://gerrit.wikimedia.org/r/86355 (owner: 10Yuvipanda) [11:40:55] (03CR) 10Zfilipin: [C: 031] "I have:" [qa/browsertests] - 10https://gerrit.wikimedia.org/r/86355 (owner: 10Yuvipanda) [11:51:31] (03CR) 10Bartosz Dziewoński: "Won't the comments be stripped by CSSMin anyway? Or do we not run it?" [core] - 10https://gerrit.wikimedia.org/r/87536 (owner: 10Ori.livneh) [11:53:45] (03CR) 10Bartosz Dziewoński: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/85161 (owner: 10Krinkle) [11:53:49] (03CR) 10Bartosz Dziewoński: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/87165 (owner: 10Jdlrobson) [12:04:04] (03PS1) 10Zfilipin: Moving shared code to mediawiki-selenium Ruby gem [extensions/TwnMainPage] - 10https://gerrit.wikimedia.org/r/87542 [12:42:23] (03PS1) 10Stwalkerster: Add an API module for querying the AntiSpoof results without having to attempt to create an account. [extensions/AntiSpoof] - 10https://gerrit.wikimedia.org/r/87546 [12:46:29] (03CR) 10Bartosz Dziewoński: [C: 04-1] "(3 comments)" [core] - 10https://gerrit.wikimedia.org/r/87345 (owner: 10Jackmcbarn) [12:51:56] (03PS2) 10Bartosz Dziewoński: mw.config: Deprecate wgMonthNames and wgMonthNamesShort [core] - 10https://gerrit.wikimedia.org/r/76497 [12:52:02] (03PS3) 10Bartosz Dziewoński: mw.config: Deprecate wgMonthNames and wgMonthNamesShort [core] - 10https://gerrit.wikimedia.org/r/76497 [12:53:09] (03PS4) 10Bartosz Dziewoński: mw.config: Deprecate wgMonthNames and wgMonthNamesShort [core] - 10https://gerrit.wikimedia.org/r/76497 [12:55:18] (03Abandoned) 10Bartosz Dziewoński: mw.loader: Remove remnants of synchronous loading mode [core] - 10https://gerrit.wikimedia.org/r/73178 (owner: 10Bartosz Dziewoński) [12:55:40] (03Abandoned) 10Bartosz Dziewoński: mw.loader: Rip out broken code path from addScript() [core] - 10https://gerrit.wikimedia.org/r/73177 (owner: 10Bartosz Dziewoński) [12:57:26] (03CR) 10Bartosz Dziewoński: "Ping?" [core] - 10https://gerrit.wikimedia.org/r/82828 (owner: 10Bartosz Dziewoński) [12:57:43] (03CR) 10Bartosz Dziewoński: "Ping?" [core] - 10https://gerrit.wikimedia.org/r/80769 (owner: 10Bartosz Dziewoński) [12:58:07] (03CR) 10Bartosz Dziewoński: "Ping?" [core] - 10https://gerrit.wikimedia.org/r/82045 (owner: 10Bartosz Dziewoński) [13:02:58] (03CR) 10Nikerabbit: [C: 04-1] "Needs more explanation." [core] - 10https://gerrit.wikimedia.org/r/84138 (owner: 10Umherirrender) [13:05:29] (03CR) 10Wizardist: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/35871 (owner: 10Wizardist) [13:21:40] (03PS1) 10Zfilipin: Moving shared code to mediawiki-selenium Ruby gem [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/87552 [13:24:22] (03CR) 10Nikerabbit: [C: 04-1] "(3 comments)" [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/78946 (owner: 10Legoktm) [13:33:42] * YuviPanda pokes MaxSem with https://gerrit.wikimedia.org/r/#/c/86960/ [13:44:04] * MatmaRex pokes YuviPanda with https://gerrit.wikimedia.org/r/#/c/86382/ [13:44:28] (03PS1) 10Siebrand: Update docs and declare visibility on class props [core] - 10https://gerrit.wikimedia.org/r/87556 [13:46:51] (03CR) 10Krinkle: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/85161 (owner: 10Krinkle) [13:46:55] (03Abandoned) 10Krinkle: (DRAFT) mediawiki.action.view.postEdit: Convert to use LESS [core] - 10https://gerrit.wikimedia.org/r/85161 (owner: 10Krinkle) [13:47:33] (03CR) 10MaxSem: [C: 031] Add caching & cache invalidation for Campaign pages [extensions/UploadWizard] - 10https://gerrit.wikimedia.org/r/86960 (owner: 10Yuvipanda) [13:51:42] (03PS2) 10Siebrand: Declare visibility on class properties [core] - 10https://gerrit.wikimedia.org/r/87510 [13:52:20] (03PS2) 10Yuvipanda: Restrict formatting to where a title attribute is present [core] - 10https://gerrit.wikimedia.org/r/86382 (owner: 10Bartosz Dziewoński) [13:52:25] (03CR) 10Yuvipanda: [C: 032] Restrict formatting to where a title attribute is present [core] - 10https://gerrit.wikimedia.org/r/86382 (owner: 10Bartosz Dziewoński) [13:52:28] MatmaRex: ^ [13:52:53] thanks YuviPanda [13:53:20] yw [13:53:40] although merging CSS patches makes me feel a little itchy, no matter how trivial [13:53:44] MaxSem: thanks! [13:53:52] MaxSem: I'll get brion to test it and then merge, hopefully [13:54:14] * YuviPanda pokes brion [13:54:27] going to be slow. hand pain a bit too much [13:57:28] (03CR) 10Rillke: "(2 comments)" [core] - 10https://gerrit.wikimedia.org/r/87487 (owner: 10MarkTraceur) [13:58:10] (03Merged) 10jenkins-bot: Restrict formatting to where a title attribute is present [core] - 10https://gerrit.wikimedia.org/r/86382 (owner: 10Bartosz Dziewoński) [14:11:45] (03PS1) 10Nemo bis: Quote the alleged "edit" message directly in livetranslate-dictionary-* [extensions/LiveTranslate] - 10https://gerrit.wikimedia.org/r/87562 [14:11:59] (03CR) 10Umherirrender: "When the static function ContentHandler::getLocalizedName is called, no language is specified and the Message class is using $wgLang which" [core] - 10https://gerrit.wikimedia.org/r/84138 (owner: 10Umherirrender) [14:14:09] YuviPanda: so the hadoop xml thing... tell me more (slowly or with lots of breaks) [14:15:04] because eg there is https://github.com/whym/wikihadoop (for history files) [14:19:50] (03CR) 10Rillke: "(2 comments)" [core] - 10https://gerrit.wikimedia.org/r/87487 (owner: 10MarkTraceur) [14:20:00] apergos: taking enforced 5m breaks every 30m, sorry. [14:20:08] please do [14:20:15] take them. please don't be sorry. [14:20:25] apergos: the end goal was for me to play with hadoop and learn. [14:20:44] apergos: so I wanted to split the logging table for enwiki [14:20:56] and run it through some simple queries [14:21:19] so is this thing above not helpful for you then? I have no idea if it is too complex, doesn't handle logging files etc [14:21:37] apergos: that link is meant for one specific thing, and it is useful for that, yeah [14:21:53] ok [14:21:55] apergos: which is the full dump with full history. [14:21:59] oh. [14:22:22] ok well I have new crap c code that will split logging or page file into pieces with x items in each [14:22:34] need to tet gzip bit though [14:22:37] *test [14:22:46] apergos: i checked out ops/dumps y'day. will take a look at it [14:22:59] also trying to limit computer time until i can get my hand fixed [14:23:10] ok [14:23:32] it's been physically and mentally rather stressful, since 90% of my interactions seem to involve a keyboard of some sort :) [14:23:35] i'll try to get latet version checked in today [14:23:39] uh i can only imagine [14:24:02] (03PS3) 10Jackmcbarn: Allow appending text to redirects from moved pages [core] - 10https://gerrit.wikimedia.org/r/87345 [14:24:24] (03CR) 10Jackmcbarn: "Good question about the space; upon thinking about it, a newline does make more sense." [core] - 10https://gerrit.wikimedia.org/r/87345 (owner: 10Jackmcbarn) [14:33:10] (03CR) 10Anomie: "(2 comments)" [core] - 10https://gerrit.wikimedia.org/r/87345 (owner: 10Jackmcbarn) [14:36:40] (03PS4) 10Jackmcbarn: Allow appending text to redirects from moved pages [core] - 10https://gerrit.wikimedia.org/r/87345 [14:39:16] pokes siebrand [14:39:25] yes? [14:40:02] siebrand: saw you last comment on the patch set … do you have time to discuss the workflow? [14:40:18] siebrand: in a hangout? [14:40:30] dan-nl: Oh, sure. [14:40:56] ping me at smazeland@wikimedia.org [14:41:55] (hangout-wise, that is :) [14:44:47] (03PS11) 10Wizardist: Mouseover explanations for interlanguage links in native language [core] - 10https://gerrit.wikimedia.org/r/35871 [14:54:01] (03CR) 10Parent5446: [C: 031] Allow appending text to redirects from moved pages [core] - 10https://gerrit.wikimedia.org/r/87345 (owner: 10Jackmcbarn) [14:54:54] (03CR) 10Anomie: [C: 032] "Code works, looks fine." [core] - 10https://gerrit.wikimedia.org/r/87345 (owner: 10Jackmcbarn) [14:57:58] (03Merged) 10jenkins-bot: Allow appending text to redirects from moved pages [core] - 10https://gerrit.wikimedia.org/r/87345 (owner: 10Jackmcbarn) [15:08:45] (03PS2) 10Siebrand: Update docs and declare visibility on class props [core] - 10https://gerrit.wikimedia.org/r/87556 [15:12:29] (03CR) 10Siebrand: "PS2 aims to address whatever phpcs blabbered about." [core] - 10https://gerrit.wikimedia.org/r/87556 (owner: 10Siebrand) [15:14:21] (03CR) 10Anomie: [C: 04-1] "(3 comments)" [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87311 (owner: 10Legoktm) [15:18:38] (03PS1) 10Chad: Move maintenance scripts out of top level [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/87578 [15:19:02] (03PS3) 10Siebrand: Update docs and declare visibility on class props [core] - 10https://gerrit.wikimedia.org/r/87556 [15:29:52] (03CR) 10Siebrand: [C: 04-1] "(1 comment)" [extensions/LiveTranslate] - 10https://gerrit.wikimedia.org/r/87562 (owner: 10Nemo bis) [15:31:57] (03PS7) 10Matthias Mullie: [WIP] New design [extensions/Flow] - 10https://gerrit.wikimedia.org/r/86856 [15:33:05] brion: around? [15:39:52] (03CR) 10Nemo bis: "(1 comment)" [extensions/LiveTranslate] - 10https://gerrit.wikimedia.org/r/87562 (owner: 10Nemo bis) [15:40:02] YuviPanda: yo [15:40:33] brion: https://gerrit.wikimedia.org/r/#/c/86960/ :) [15:40:54] ok lemme look over it [15:41:52] (03PS2) 10Ori.livneh: Have EL depend on Geshi, related minor refactoring [vagrant] - 10https://gerrit.wikimedia.org/r/84488 (owner: 10Mattflaschen) [15:43:42] (03CR) 10Ori.livneh: [C: 04-1] "(2 comments)" [vagrant] - 10https://gerrit.wikimedia.org/r/84488 (owner: 10Mattflaschen) [15:44:31] looks good, lemme do a quick smoke check [15:45:02] oh btw -- final version of OS X 10.9 is available to registered mac devs [15:45:06] should release soonish i imagine [15:45:22] i'll update my main machine this weekend :D [15:47:01] brion: gives me anywhere from 4x - 2x perf improvements [15:48:08] (03CR) 10Brion VIBBER: [C: 032] "(1 comment)" [extensions/UploadWizard] - 10https://gerrit.wikimedia.org/r/86960 (owner: 10Yuvipanda) [15:48:15] (03Merged) 10jenkins-bot: Add caching & cache invalidation for Campaign pages [extensions/UploadWizard] - 10https://gerrit.wikimedia.org/r/86960 (owner: 10Yuvipanda) [15:48:23] looks good [15:48:25] siebrand: do you have any idea how one could pass username for GENDER to a message fed into Echo via hook like Thanks does? :[ https://git.wikimedia.org/blob/mediawiki%2Fextensions%2FThanks.git/cd51c766859c9413a7c719a26ffe8cec28447491/Thanks.hooks.php#L130 [15:48:52] brion: so, it *is* generalized. sortof [15:49:05] brion: you can return a ParserOutput object, and then the templatelinks stuff will be setup for you [15:49:12] siebrand: i18n issue involving two different extensions... 3 possible components for a bug and near 0 changes for someone to pick it up [15:49:13] brion: except that it only accepts you returning *one* ParserOutput object [15:49:18] brion: while we have like 4 or 5 [15:49:22] *chances [15:51:06] heh [15:51:18] YuviPanda: i think you can combine objects in some way….. but yeah that might not be pretty :D [15:51:29] brion: sure, you can construct those out [15:51:37] brion: but I think that'll blow up big time elsewhere [15:51:46] and i'll have to make sure that all the invariants and shit are maintained [15:51:59] heh [15:53:52] brion: did you make any updates to the license code? [15:54:11] brion: I'll be slow for a bit, since my hands have gone batshit insane. [15:54:32] (03PD6) 10Ori.livneh: Added a message about forced password resets [extensions/WikimediaMessages] - 10https://gerrit.wikimedia.org/r/87297 (owner: 10Aaron Schulz) [15:54:47] YuviPanda: i'm partway through, still have to tidy it up [15:54:51] got distracted with code review :D [15:54:54] brion: heh, okay [15:54:58] brion: i'll not review it right now then [15:55:06] i did update the patch set if you want to peek it over though [15:55:07] (03CR) 10Jeroen De Dauw: "(1 comment)" [extensions/LiveTranslate] - 10https://gerrit.wikimedia.org/r/87562 (owner: 10Nemo bis) [15:55:10] brion: think I can convince someone to get that patch to testwiki? [15:55:15] brion: nah, i'll wait till you're done [15:55:17] (03CR) 10Ori.livneh: [C: 032 V: 032] Added a message about forced password resets [extensions/WikimediaMessages] - 10https://gerrit.wikimedia.org/r/87297 (owner: 10Aaron Schulz) [15:55:23] thanks AaronSchulz [15:55:33] i try to avoid deployment questions, they get complicated ;) [15:55:39] or i just ask reedy and it becomes his problem ;) [15:57:30] ok i'm going to wander into the office today, so gotta get ready. see ya in a bit [15:59:23] greg-g: ping [16:00:01] Reedy? [16:00:11] i should stop pretending to be away, shouldn't i [16:00:57] ori-l: you're never away, are you? :) [16:01:18] i am [16:01:22] (03CR) 10Matthias Mullie: "(3 comments)" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/87488 (owner: 10EBernhardson (WMF)) [16:01:29] ori-l: hi there [16:01:52] morning greg-g [16:01:59] what's the match to testwiki? [16:02:21] https://dpaste.de/vA6R/raw/ [16:02:27] oh, that's not why i was pinging [16:02:32] ^^ that's why i was pinging [16:02:41] (that's from #wikimedia-tech earlier) [16:02:57] aaron published the draft and i merged it, but it still requires a sync of 1.22wmf20 and a l10nupdate [16:03:02] or a scap [16:03:12] hrm, actually [16:03:24] ahhh [16:03:28] yeah, fix the message, worth it now [16:04:52] (03PS1) 10Ori.livneh: Update WikimediaMessages to master [core] (wmf/1.22wmf20) - 10https://gerrit.wikimedia.org/r/87584 [16:05:26] i'm pretty groggy [16:05:36] Reedy isn't around to do the actual scap by any chance is he? [16:06:13] how dare he sleep [16:06:17] (03PS1) 10Aaron Schulz: Fixed silly unlock bug in LocalFile [core] - 10https://gerrit.wikimedia.org/r/87585 [16:06:20] scandal! [16:06:36] (03PS2) 10Aaron Schulz: Fixed silly unlock bug in LocalFile [core] - 10https://gerrit.wikimedia.org/r/87585 [16:07:11] * YuviPanda gives ori-l a pitchfork and Reedy's address [16:07:16] ori-l, if you don't feel well I can scap for you [16:07:29] MaxSem: yes, that would be very kind of you [16:07:41] (03CR) 10Ori.livneh: [C: 032] Update WikimediaMessages to master [core] (wmf/1.22wmf20) - 10https://gerrit.wikimedia.org/r/87584 (owner: 10Ori.livneh) [16:07:59] ^ have to wait for that to merge [16:08:25] it would take an ISP shutdown [16:08:40] ahem I'm late [16:09:06] Nemo_bis: we'll dock it from your paycheck [16:09:55] okay, i'm actually going to crash for a half hour or something [16:10:05] MaxSem: thanks again [16:10:14] sleep well ori-l [16:12:19] (03Merged) 10jenkins-bot: Update WikimediaMessages to master [core] (wmf/1.22wmf20) - 10https://gerrit.wikimedia.org/r/87584 (owner: 10Ori.livneh) [16:13:05] scapping... [16:13:28] bd808: https://gerrit.wikimedia.org/r/#/c/87585/2 silly fix [16:14:11] man, insomnia sucks [16:15:52] (03CR) 10BryanDavis: [C: 032] Fixed silly unlock bug in LocalFile [core] - 10https://gerrit.wikimedia.org/r/87585 (owner: 10Aaron Schulz) [16:16:48] (03PS1) 10Aaron Schulz: Fixed silly unlock bug in LocalFile [core] (wmf/1.22wmf20) - 10https://gerrit.wikimedia.org/r/87587 [16:17:10] (03CR) 10Aaron Schulz: [C: 032] Fixed silly unlock bug in LocalFile [core] (wmf/1.22wmf20) - 10https://gerrit.wikimedia.org/r/87587 (owner: 10Aaron Schulz) [16:19:34] (03Merged) 10jenkins-bot: Fixed silly unlock bug in LocalFile [core] - 10https://gerrit.wikimedia.org/r/87585 (owner: 10Aaron Schulz) [16:23:13] (03Merged) 10jenkins-bot: Fixed silly unlock bug in LocalFile [core] (wmf/1.22wmf20) - 10https://gerrit.wikimedia.org/r/87587 (owner: 10Aaron Schulz) [16:27:56] (03CR) 10Anomie: "(3 comments)" [core] - 10https://gerrit.wikimedia.org/r/87288 (owner: 10Liangent) [16:40:20] How do I get started with contributing to mediawiki ? [16:40:46] pallavi: you just did! next steps: https://www.mediawiki.org/wiki/How_to_contribute [16:40:58] and don't hesitate to ask if you have questions [16:41:23] I would like to start with documentation [16:41:49] ah, do you have an account on mediawiki.org? [16:42:13] oh , no [16:42:15] (also: https://www.mediawiki.org/wiki/How_to_become_a_MediaWiki_hacker ) [16:42:27] I'll create one :) [16:46:31] I'm good with RoR , Css , Js and html . Is there a tiny bug or a small project that I could start off with ? to get the hang of how it works :) [16:48:09] YuviPanda: on a more interesting note, who should be poked about https://translatewiki.net/wiki/Thread:Support/About_MediaWiki:Shorturl-desc/en [16:48:11] (03CR) 10Ori.livneh: [C: 031] "This looks good; reviewing dependencies now." [core] - 10https://gerrit.wikimedia.org/r/82828 (owner: 10Bartosz Dziewoński) [16:48:43] pallavi: hrm, yes! [16:48:59] Nemo_bis: well, I think that message should be removed, since it was converted to an unlisted special page sometime ago [16:49:03] Nemo_bis: want me to do that now? [16:49:25] YuviPanda: unlisted? O_o that's the desc in special:version [16:49:27] pallavi: https://en.wikipedia.org/wiki/User:Ori.livneh/common.js needs to be implemented as a module in core, it's on my todo list of things to do [16:49:53] ori-l: checking :) [16:49:55] pallavi: i feel a bit selfish dumping that on you, but on the other hand it's pretty small and self-contained as a task [16:49:59] Nemo_bis: oh, hmm. Thought it was desc. [16:50:05] for the special pages [16:50:15] Nemo_bis: I also don't remember writing the original text [16:50:23] Nemo_bis: anyway, I agree with the poster's rewording. [16:50:24] pallavi: it's a debug tool; gives a breakdown of loaded modules on the page [16:50:41] YuviPanda: fantastic, submit it? :) or I submit and you speedy merge? ;) [16:50:49] Nemo_bis: I like the second option ;) [16:51:34] Nemo_bis: poke me here with patchset, I'll +2. Thanks! :) [16:51:57] yeah, rhetoric question, I'm already commiting :P [16:52:06] heh :) [16:53:02] pallavi: sample output: http://i.imgur.com/kG1LioY.png ; it could also be rendered on the page itself (if activated) [16:53:20] (03PS1) 10Nemo bis: Clarify description: the short URL *are* redirects, for/to normal pages [extensions/ShortUrl] - 10https://gerrit.wikimedia.org/r/87591 [16:53:40] YuviPanda: ^ [16:54:00] * YuviPanda lets the patch rot for 4 months [16:54:01] (03PS2) 10Nemo bis: Clarify description: the short URL *are* redirects, for/to normal pages [extensions/ShortUrl] - 10https://gerrit.wikimedia.org/r/87591 [16:54:16] grr [16:54:18] pallavi: the steps for getting in would be: get a mediawiki developer account and set up the relevant tools (https://www.mediawiki.org/wiki/How_to_become_a_MediaWiki_hacker has info) [16:54:19] (03CR) 10Yuvipanda: [C: 032] Clarify description: the short URLs *are* redirects, for/to normal pages [extensions/ShortUrl] - 10https://gerrit.wikimedia.org/r/87591 (owner: 10Nemo bis) [16:54:22] Nemo_bis: :) [16:54:23] (03Merged) 10jenkins-bot: Clarify description: the short URLs *are* redirects, for/to normal pages [extensions/ShortUrl] - 10https://gerrit.wikimedia.org/r/87591 (owner: 10Nemo bis) [16:54:53] ;) [16:54:54] pallavi: then to add the file to the resources/mediawiki folder and to declare it in resources/Resources.php [16:54:57] ori-l: Thank you :) I just created an account now [16:56:10] OPW mentors around? [16:57:11] pallavi: I was a GSoC mentor this year, and marktraceur was an OPW mentor last time, I believe [16:57:25] Yup [16:57:32] pallavi: Are you asking about the next round? [16:57:39] marktraceur: Yup [16:57:45] (03PS1) 10Krinkle: mw.Title: Add missing dependency on jquery.byteLength [core] - 10https://gerrit.wikimedia.org/r/87593 [16:57:51] Then I'm not your guy, 'cause I don't do OPW anymore [16:58:11] marktraceur: do you know who does? [16:58:23] qg dammit [16:58:37] He's coming [16:58:37] pallavi: yeah, qgil is the person to ask, but he's not around at the moment [16:58:38] marktraceur: Oh , do you know who's going to be mentoring this time ? [16:58:52] (03PS2) 10Krinkle: mw.Title: Add missing dependency on jquery.byteLength [core] - 10https://gerrit.wikimedia.org/r/87593 [16:59:08] ok , anyway , I'll get started :) [16:59:12] marktraceur: your ing summoned him :) [16:59:18] qgil: pallavi is asking who's mentoring for OPW this round [16:59:26] Or something [16:59:38] pallavi: qgil is the person making the whole thing happen [16:59:57] (03CR) 10Ori.livneh: [C: 032] mw.Title: Add missing dependency on jquery.byteLength [core] - 10https://gerrit.wikimedia.org/r/87593 (owner: 10Krinkle) [16:59:58] (03CR) 10Legoktm: [C: 032] Remove pointless override teardown method [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87526 (owner: 10Addshore) [16:59:59] hi pallavi - we haven't started with the new round yet but you can of course start contributing and defining a proposal. How can I help? [17:00:06] marktraceur, thank you for the ping :) [17:00:16] My pleasure [17:00:31] qgil: I wanted to start with documentation [17:00:41] (03PS3) 10Krinkle: mw.Title: Add missing dependency on jquery.byteLength [core] - 10https://gerrit.wikimedia.org/r/87593 [17:00:53] (03CR) 10Krinkle: "(longer hash so that Gerrit gets the link)" [core] - 10https://gerrit.wikimedia.org/r/87593 (owner: 10Krinkle) [17:01:08] qgil: I have an account [17:01:33] (03CR) 10Legoktm: [C: 032] Use covers tags in tests [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87525 (owner: 10Addshore) [17:01:40] qgil: ori-l already suggested a nice, *small*, concrete and will-be-deployed-if-done project to pallavi to get started on [17:01:55] (03Merged) 10jenkins-bot: Remove pointless override teardown method [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87526 (owner: 10Addshore) [17:02:24] But I wanted to start with documentation [17:02:41] (03CR) 10Krinkle: [C: 032] "Per Ori (sorry)." [core] - 10https://gerrit.wikimedia.org/r/87593 (owner: 10Krinkle) [17:02:48] (03CR) 10Catrope: [C: 032] mw.Title: Add missing dependency on jquery.byteLength [core] - 10https://gerrit.wikimedia.org/r/87593 (owner: 10Krinkle) [17:03:09] (03Merged) 10jenkins-bot: Use covers tags in tests [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87525 (owner: 10Addshore) [17:03:42] whooo, brion! [17:04:42] addshore: thanks :D [17:04:53] how do I access machines with no public IP? for example FQDN is i-0000079d.pmtpa.wmflabs [17:05:25] (03Abandoned) 10Ori.livneh: Use '//'-style comments in LESS doc blocks [core] - 10https://gerrit.wikimedia.org/r/87536 (owner: 10Ori.livneh) [17:07:24] (03Merged) 10jenkins-bot: mw.Title: Add missing dependency on jquery.byteLength [core] - 10https://gerrit.wikimedia.org/r/87593 (owner: 10Krinkle) [17:07:28] yhager: https://wikitech.wikimedia.org/wiki/Help:Access [17:07:45] yhager: short version: you need to go through an intermediate host [17:07:51] long version: that page :) [17:08:03] ori-l: thanks. that's what I was looking for. [17:12:21] legoktm: np :) was just walking through the code today :) [17:14:31] pallavi, just checking: are you aware of https://www.mediawiki.org/wiki/Annoying_little_bugs ? [17:15:40] (03PS11) 10Liangent: (bug 35378) Support multiple collations at the same time [core] - 10https://gerrit.wikimedia.org/r/87288 [17:16:28] (03CR) 10Liangent: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/87288 (owner: 10Liangent) [17:16:38] good to know ori-l YuviPanda - which project is this? [17:17:13] pallavi, https://bugzilla.wikimedia.org/show_bug.cgi?id=35234 is related to documentation, although it requires coding skills [17:20:08] pallavi, there is also https://bugzilla.wikimedia.org/show_bug.cgi?id=47486 - which is indeed about documentation. You would probably need to investigate / ask about the topic but it would be a good exercise and once you would start others probably will help you. [17:20:50] (03CR) 10LuisVilla: [C: 04-1] "(3 comments)" [qa/browsertests] - 10https://gerrit.wikimedia.org/r/86355 (owner: 10Yuvipanda) [17:20:51] pallavi, in addition to this we have PLENTY of wiki pages at http://mediawiki.org that need some (or a lot of) love. [17:21:00] Hello [17:21:18] Hi, natuur12 wants to know whether it is possible to block users from using the thankyou feature in Echo [17:21:20] (jdlrobson: spagewmf: StevenW: YuviPanda: superm401:) [17:21:25] (03CR) 10MarkTraceur: "Rillke, FYI https://gerrit.wikimedia.org/r/85926" [core] - 10https://gerrit.wikimedia.org/r/87487 (owner: 10MarkTraceur) [17:21:38] Krinkle: is thank trolling occuring? [17:21:41] We got a user who is abusing the thank you function [17:21:43] ^ MaxSem [17:21:55] I think hiding with css is the only option? [17:22:07] A list: https://nl.wikipedia.org/w/index.php?title=Speciaal%3ALogboeken&type=thanks&user=Bart_Versieck&page=&year=&month=-1&tagfilter=&hide_patrol_log=1&hide_thanks_log=1 [17:22:08] natuur12, block him [17:22:23] jdlrobson: thanktrolling is real! [17:22:29] natuur12: yeh block him for thank trolling [17:22:31] warned ya! [17:23:01] Ofcourse, any public API can and will be abused. Especially thankyou because its effect quite visible to users. [17:23:09] Another block for him? :P https://nl.wikipedia.org/w/index.php?title=Speciaal:Logboeken/block&page=Gebruiker%3ABart+Versieck [17:23:09] pallavi, see https://www.mediawiki.org/wiki/Category:Cleanup [17:23:12] I could block him but that should not be the only option and is not covered by our guidelines [17:23:23] for spammers as well (high likelyhood you'll go to this user page and click whatever link is there) [17:23:31] could we update the guidelines to state thank trolling is an offense worthy of this? [17:23:48] This is about nl.wikipedia admin guidelines [17:23:52] pallavi, ok, now you have several links to look at. You can look at them and choose one. :) Let me know what you decide! [17:23:58] also out of interest to what nature is he abusing the thank button [17:24:18] jdlrobson: isn't it easier to instate a stricter ratelimit [17:24:21] natuur12: Yeah, I think in this case it would make sense to update nl.wiki guidelines. This is a new feauture, it doesn't cover it becuase it didn't exist yet. [17:24:34] is he thank nuking (ie. flooding a user with notifications) or is he just thanking bad edits? [17:24:34] qgil: Im setting up the development environment and thank you so much , sure I'll come back :) [17:24:49] jdlrobson: thank nuking [17:25:01] for the same edit or different edits? [17:25:10] different edits [17:25:11] i assume the latter (otherwise that sounds like a bug in echo) [17:25:16] He thanked me for eating all the stroopwafels. [17:25:33] He is thanking all users on all pages [17:25:35] jdlrobson: that bug used to exist but it was fixed iirc [17:25:41] He even thanks you for reverting him and so on... [17:26:24] quite the thank troll... [17:26:33] i think we should apply the same rules to editing to thanking [17:26:48] which would suggest blocking him [17:26:59] i don't think we should invent functionality for this specific purpose [17:27:06] There must be another solution [17:27:09] Nemo_bis> jdlrobson: isn't it easier to instate a stricter ratelimit [17:27:13] Some users told him to stop, but he didn't listen. [17:27:13] 'i was blocked for thanking too many people!' [17:27:20] ^d: is this a reasonable patch for FormatJsonTest? With the patch applied, both hhvm and php pass. http://p.defau.lt/?1kZGVrBtMrxysu4Ee5HqUA [17:27:38] WHY WONT YOU STOP THANKING! THINK OF THE CHILDREN! [17:27:52] a rate limit would work although i worry it might stop some genuine thanking? [17:28:01] on enwiki someone with a block log that long are routinely indeffed (unless they're G....) [17:28:23] 'sorry yu have thanked too much. wait for 5 minutes and try again' [17:28:44] We have a rule about disabling the "mark as controlled"-function on our Wikipedia. [17:28:54] jdlrobson: set it to 20 per day and, say, 5 per hour, or whatever email uses [17:29:23] *5 for newbies and 20 for normal users per day [17:29:39] 0 for thank you trolls [17:30:06] but maybe i want to say thank you more than 20 times a day.. [17:30:14] Oh, thank god he has a cleaned talk page… https://nl.wikipedia.org/wiki/Overleg_gebruiker:Bart_Versieck [17:30:22] <^d> yhager: Looks reasonable. Let me test it out locally and if all passes we'll push to git. [17:30:25] could we report users who are thank trolling and use that to cap thanking? [17:30:29] Warning: It will freeze your browser. :P [17:30:36] 20 per day is a lot [17:30:44] ^d: okay, thx [17:30:52] yeh but 1 thank per hour seems reasonable [17:31:15] 20 per day allows 20 in a minute :) [17:31:17] what if i'm collaborating with another polite user we might want to send more than 20 thanks between us [17:31:25] O_o [17:31:35] 20 per day AND 1 per 1 hour [17:31:36] Nemo_bis: we could rate cap them per 10 minutes [17:31:37] make it 100 if you wish, just start somewhere [17:31:46] I agree with Jurgen [17:31:56] i mean if i'm doing the software equivalent of THANKS! THANKS! THANKS! THANKS! THANKS! THANKS! THANKS! THANKS! THANKS! that's obviously thank trolling [17:31:58] A thank you should be special [17:32:09] +1 on natuur12 [17:32:15] but if I do THANKS! THANKS! that seems reasoanble [17:32:17] COMBAT THE THANKS DEFLATION [17:32:27] * inflation [17:32:59] i could imagine a system where thanks per day are limited [17:33:06] >_< [17:33:08] WMF is the FED of thanking, an aggressive quantitative easing is ongoing [17:33:12] i get to say thank you ten times a day for example [17:33:24] and get told I have 9 thank yous left [17:33:33] now we only need to figure out how to earn 80 G$/y like them with this [17:33:34] who are we to say how many thanks are reasonable in a day? [17:33:36] then they become more magical and meaningful and we remove the ability to thank troll [17:33:52] awjr: who are we to say how many edits are reasonable in a hour? :) but we do [17:33:57] same for email, move etc. [17:34:13] but thank yous are different [17:34:36] are we able to detect the difference between a sarcastic thank and a genuine thank? [17:34:42] something like Abusefilter [17:34:55] lol [17:35:01] if username == Bart Versieck - sarcastic thank [17:35:07] jdlrobson: AI [17:35:10] bayesian thanking [17:35:12] yeh we could define rules for thanking [17:35:14] if username != Bart Versieck - genuine thank [17:35:19] bayesian thanking! [17:35:21] Maybe limit per namespace? [17:35:22] Jurgen: that would be a good start :) [17:35:33] (03CR) 10Krinkle: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/87487 (owner: 10MarkTraceur) [17:35:43] if page_blanking and THANK = sarcastic thank [17:35:53] That you can thank more in the normal namespace then like a user talk page. [17:36:45] sjoerdisdruk: Jurgen Nemo_bis has someone raised a bug about this to make sure we work out a better solution for this? [17:36:51] jdlrobson: I don't like the idea of policing thanking, an inthrisically ephemeral action, but just make it go through abusefilter if you wish [17:37:06] legoktm: how's labs-vagrant working out? :) [17:37:06] I think at the moment though, all we could really do is block the user :-( [17:37:34] jdlrobson: sjoerdisdruk Jurgen Nemo_bis 'thank' userright, granted by default? [17:37:40] you can take it away if they troll.. [17:37:41] we could add some JavaScript to Common.js I guess [17:37:43] i would also be against any kind of thank filtering, no need to second guess users [17:37:54] it could use local storage / cookie to track how many clicks to thank buttons occurred [17:37:56] CSS is enough jdlrobson [17:38:00] if particular users dont want a thank button, they can add some css to hide the thank button [17:38:01] and then destroy the thank button [17:38:04] Oh. [17:38:11] there's always triggering captcha for thanks [17:38:13] sjoerdisdruk: if you want to stop everyone from thanking yes you could do that [17:38:20] yeh captcha wouldn't be a bad idea [17:38:21] Or a setting. [17:38:27] captcha really sounds over the top [17:38:32] like, you hit a certain threshold of thanks and then a captcha is triggered [17:38:32] I think that is a good idea YuviPanda [17:38:41] And a thanks-blacklist. [17:38:43] perhaps with a threshhold [17:38:43] it's also not too hard. [17:38:49] sjoerdisdruk: but i think the right to thank should be one of the 10 amendments of wikipedians [17:38:51] "I don't want to get thanks from:" [17:38:58] here you go https://bugzilla.wikimedia.org/show_bug.cgi?id=54983 [17:39:06] now this works in css: .mw-thanks-thank-link { display: none !important; } [17:39:20] But then you have to lock the common.css page [17:39:21] sjoerdisdruk: but what if a thank troll changes their ways and wants to thank you to make peace? [17:39:37] It's Bart Versieck. [17:40:15] ebernhardson: some of this discussion seems to have heavy elements of humor, mostly because mobile team trolled awjr a month ago with a deep 1 hour discussiona bout the thankyoupocalypse and how we were planning on using AI methods to combat it [17:40:21] He gets blocked all the time because of editing in someone else's messages. [17:40:47] sjoerdisdruk: why is he not blocked now? [17:40:50] He even trolls using the thank mark button.. [17:41:07] the mark button* [17:41:26] i need to go into a meeting- can someone summarize the ideas that came out of this? [17:41:32] and update that bug [17:41:40] Because we don't have rules about it, but now: [17:41:45] jdlrobson, we don't take blokking a person lightly [17:41:50] sjoerdisdruk: thoughts on the userrights/ [17:41:56] natuur12: sure but this guy seems like a big troll [17:42:00] 'thank' user right [17:42:08] Yeah? [17:42:19] En indeed, it is not covered by our guidelines and it should not be nescacarry te develop guidelines for this [17:42:29] He is, but he cannot help it. [17:42:33] sjoerdisdruk: add a 'thank' user right, if someone thank trolls, remove the right from them [17:42:40] You could do that. [17:43:04] we don't do it for anything [17:43:27] sjoerdisdruk: I don't know your bugzilla email so I couldn't add you to cc [17:43:35] sjoerddebruin@me.com [17:43:42] :) [17:43:51] * YuviPanda tries to pronounce sjoerddebruin [17:44:03] :P [17:44:29] sjoerddebruin, natuur12, Jurgen, if you get local consensus for a ratelimit nobody can stop it ;) [17:44:35] (03CR) 10Brion VIBBER: [C: 032] "Confirmed that this works as designed when Extension:CLDR is loaded; when it's not you get the autonym again, but the tooltip is more cons" [core] - 10https://gerrit.wikimedia.org/r/35871 (owner: 10Wizardist) [17:45:01] (Though it may be broken right now; but it shouldn't be too hard to fix in that case.) [17:45:06] Nemo_bis: we have this now: https://nl.wikipedia.org/wiki/Wikipedia:Notificaties#Kan_de_bedank-functie_ook_weggenomen_worden_indien_er_misbruik_van_gemaakt_wordt.3F [17:45:30] People can request, a little discussion and we can block it. [17:45:53] you add that to the abusers' CSS? [17:45:55] (03PS1) 10Chad: Strip excess newlines from formatting test [core] - 10https://gerrit.wikimedia.org/r/87601 [17:46:07] Yeah if we get consensus. [17:46:11] And then lock it. [17:46:12] sigh [17:46:17] yes Nemo_bis [17:46:27] don't they just use the API then? ;) [17:46:27] <^d> yhager: https://gerrit.wikimedia.org/r/87601 is the patch :) [17:46:52] We also did that with the mark as controlled-button with some user. [17:47:13] (03PS12) 10Liangent: (bug 35378) Support multiple collations at the same time [core] - 10https://gerrit.wikimedia.org/r/87288 [17:47:14] ^d: thanks! [17:47:38] YuviPanda: http://www.youtube.com/watch?v=jniKtV-ny5o ;) [17:49:07] Here. Learn some Dutch: http://www.youtube.com/watch?v=3ak6AmAn_cY [17:49:29] hehe :D [17:49:34] (03Merged) 10jenkins-bot: Mouseover explanations for interlanguage links in native language [core] - 10https://gerrit.wikimedia.org/r/35871 (owner: 10Wizardist) [17:49:51] (03CR) 10Brion VIBBER: "So do we need further work on this? Open questions:" [core] - 10https://gerrit.wikimedia.org/r/87165 (owner: 10Jdlrobson) [17:50:31] (03PS13) 10Liangent: (bug 35378) Support multiple collations at the same time [core] - 10https://gerrit.wikimedia.org/r/87288 [17:50:34] <^d> YuviPanda: Want a one-line patch to review? :) [17:50:37] (03CR) 10Liangent: "PS12 & 13 is an attempt to add back previous-collation support." [core] - 10https://gerrit.wikimedia.org/r/87288 (owner: 10Liangent) [17:50:42] ^d: sure? [17:50:42] <^d> s/1/2/ [17:50:43] Dutch is the worst language ever [17:50:56] <^d> YuviPanda: https://gerrit.wikimedia.org/r/#/c/87601/ [17:51:52] (03CR) 10Yuvipanda: [C: 032] "Looks harmless" [core] - 10https://gerrit.wikimedia.org/r/87601 (owner: 10Chad) [17:52:10] ^d: merged, assuming that the tests do still pass (I didn't run them) [17:52:32] Yeah for sure [17:52:52] <^d> YuviPanda: I just ran them against hhvm master, zend 5.3.20-something and zend 5.5.4 [17:52:53] <^d> :) [17:53:03] ^d: good enough :) [17:53:26] (03CR) 10Brion VIBBER: [C: 032] LESS embeddable(): Use lessc::toBool [core] - 10https://gerrit.wikimedia.org/r/87493 (owner: 10Ori.livneh) [17:54:36] fyi Nemo_bis sjoerddebruin even if you add to the user's CSS they can still disable css styling in their browser to see the button :) [17:54:46] (03Abandoned) 10EBernhardson (WMF): Convert from SASS to LESS [extensions/Flow] - 10https://gerrit.wikimedia.org/r/87488 (owner: 10EBernhardson (WMF)) [17:54:48] He's not so smart jdlrobson [17:55:34] (03PS5) 10Jdlrobson: Rewrite collapsibleNav as LESS file [core] - 10https://gerrit.wikimedia.org/r/86691 [17:56:37] <^d> yhager: My next goal was trying to get the dba and xdiff extensions working in ext_zend_compat. Tried dba on my fork, still not working :) [17:59:36] ^d: how are you testing these? [17:59:39] (03PS11) 10Jdlrobson: Turn screen.css into screen.less [core] - 10https://gerrit.wikimedia.org/r/85920 [17:59:51] superm401: ping [18:00:00] <^d> yhager: Testing dba? Well step 1 would be getting a clean compile :) [18:00:36] ^d: oh, okay :). Want me to step in thre, or should I continue with phpunit? [18:01:11] <^d> Well if we can knock out dba and/or xdiff, that'll take care of a ton of the skipped tests. [18:01:55] (03Merged) 10jenkins-bot: Strip excess newlines from formatting test [core] - 10https://gerrit.wikimedia.org/r/87601 (owner: 10Chad) [18:04:34] ^d: okay, since you're already on dba, let me try xdiff then [18:05:58] Nemo_bis: I don't get e-mails from thanks [18:07:31] (03CR) 10Brion VIBBER: [C: 032] "Seems to work! RTL & embedded icons work as expected." [core] - 10https://gerrit.wikimedia.org/r/86691 (owner: 10Jdlrobson) [18:07:41] sjoerddebruin: good for you :) I did iirc at some point [18:07:45] but I may misremembering [18:07:59] I never changed the setting for that... [18:08:27] brion: ping? [18:08:31] YuviPanda: yo [18:08:48] brion: i was thinking of ways to make the app super fast [18:08:59] brion: one thing that mobile web enjoys but we don't is... serverside caching [18:09:02] brion: api isn't cached [18:09:10] mmmm [18:09:15] brion: if we switch to the parsoid / content storage api thing, we'll get parsed content [18:09:16] err [18:09:18] *cached content [18:09:22] ooooh [18:09:30] brion: with proper frontend caches! [18:09:32] brion: and invalidation! [18:09:37] brion: other thing - SPDY! [18:09:39] \o/ [18:09:52] brion: however, the minute we hit the mediawiki API, things will slow down [18:09:57] brion: because the API isn't cached [18:10:00] well [18:10:04] brion: frontendcached [18:10:10] in theory we can use cacheable parameters, but things won't get invalidated if we do that :) [18:10:29] that brings us back to enhanced content api..... [18:10:49] brion: assuming we will get to use the content API for the bulk of what we want to do... [18:11:06] well then let's help plan that api >:D [18:11:06] brion: it still leaves all the other things - search, interlanguage, interwiki, 'last-modified-user!', etc [18:11:13] apiv2! [18:11:29] brion: content api is being plotted about in #mediawiki-parsoid :) [18:12:47] (03CR) 10Krinkle: "Does the @embed actually work like that? Afaik LESS might drop it or move it around before CSSMin gets to it. Should perhaps use the embed" [core] - 10https://gerrit.wikimedia.org/r/86691 (owner: 10Jdlrobson) [18:12:53] ori-l: ^ [18:13:23] brion: https://bugzilla.wikimedia.org/show_bug.cgi?id=54986 [18:13:52] (03CR) 10Brion VIBBER: "Krinkle: yes this got fixed upstream and merged to our copy in core." [core] - 10https://gerrit.wikimedia.org/r/86691 (owner: 10Jdlrobson) [18:14:58] (03CR) 10Umherirrender: "What about the action=purge of the api?" [core] - 10https://gerrit.wikimedia.org/r/87269 (owner: 10Reedy) [18:15:04] brion: btw, you bypassed jenkins on that merge. It got merged by you with the V+2 of the last patchset-test instead of by jenkins with the V+2 of the merge-test [18:15:17] Krinkle: why would jenkins fail that way? [18:15:31] It didn't, you did ;-) [18:15:33] (03PS2) 10Jdlrobson: Introduce .transition mixin [core] - 10https://gerrit.wikimedia.org/r/87165 [18:15:46] Krinkle: if a new patch set comes in that needs to invalidate the verified +2 surely [18:15:52] that seems like a failure in either jenkins or gerrit [18:16:00] It does invalidate that [18:16:07] then why would gerrit merge it? [18:16:14] you did [18:16:27] (03CR) 10Jdlrobson: "Ori's upstream patch sorted this out (Ib6bc76736d460081dc8f48aee09ca32066dae296)" [core] - 10https://gerrit.wikimedia.org/r/86691 (owner: 10Jdlrobson) [18:16:29] you clicked "Publish & Submit" instead of "Publish" [18:16:30] i pushed +2, which starts gate-and-submit jobs [18:16:33] submit is gerrit lingo for merge [18:16:37] ah then gerrit is fucked up [18:16:38] * ori-l files a bug against the brion component in bugzilla [18:16:44] cause normally it refuses to merge until +2 verified is present [18:17:00] (03CR) 10Jdlrobson: [C: 04-1] "whoops something weird happens - should have a dependency!" [core] - 10https://gerrit.wikimedia.org/r/87165 (owner: 10Jdlrobson) [18:17:11] brion: I admit Gerrit UI sucks but you told it to merge now by clicking on Publish & Submit instead of Publish on the CR+2 form [18:17:28] Krinkle: whenever i use it, it refuses to submit if there's no verified +2 [18:17:32] so…. something failed this time. [18:17:43] brion: There is V+2 because the last patchset passed the test. [18:17:50] Krinkle: i thought you said it cleared the V+2? [18:17:52] Krinkle: anyhow, what Brion said on the patch is right -- the improvement to comment-handling fixes all outstanding issues with /* @annotations */ AFAIK, including making it once again possible to have @noflip apply to individual rules rather than entire blocks. [18:17:53] no, there is nothing out of the ordinary here [18:18:17] brion: It clears the V+2 (as it should) if there is a new patch set, yes. Becuase then it starts a new jenkins job to test that new patch set [18:18:30] ok so there's two possibilities here [18:18:35] 1) it did clear the V+2 [18:18:43] 2) it didn't clear the V+2, so it submitted [18:18:44] or a third [18:18:51] no [18:18:57] 3) it did clear the V+2 but REACHED BACK IN TIME to find the V+2 that had been cleared [18:19:02] nobody cleared anything. [18:19:15] "Krinkle: brion: It clears the V+2 (as it should) if there is a new patch set, yes. Becuase then it starts a new jenkins job to test that new patch set" [18:19:17] a patchset was submitted, jenkins starts testing it and reported V+2, great [18:19:19] <^d> yhager: Oh, what was the problem with the "Creating default object from empty value" again? Curious if it's on MW or HHVM's end we need a fix. [18:19:30] Krinkle: oh so i *didn't* bypass the tests? [18:19:35] it ran new tests and set a new V+2? [18:19:39] in that case what the hell are we talking about :D [18:19:49] LESS @embeds! [18:19:59] and how wonderfully they work [18:20:10] brion: There is 3 types of pipelines in Jenkins. check (for unauthorised users), test (for whitelisted users) and submit. [18:20:21] brion: on each patchset of that change, it ran the test pipeline [18:20:25] that sounds bad. it should run the same tests for everybody, surely [18:20:44] brion: anyway, that's besides the point [18:20:46] ok so you have just confirmed that it ran tests... [18:20:49] i'm really confused [18:20:54] brion: point is, don;t click on "Publish& Submit" click on "Publish" [18:20:58] are there different tests it's supposed to run on submit, that it doesn't if you push "submit"? [18:21:00] cause that sounds WRONG [18:21:03] just remember that, the rest is just everything sucking. [18:21:26] if you click Submit (which is gerrit language for "Merge now"), it merges, becaus you have that extra user right. [18:21:35] (e.g. to bypass it in case Jenkins is down) [18:21:56] what you should be doing is just CR+2 and then Publish that vote, then jenkins comes along, tests it and merges it for you. [18:22:14] Krinkle: ok i've filed https://bugzilla.wikimedia.org/show_bug.cgi?id=54987 [18:22:20] <^d> Ideally we'd disable manual submission so we don't get confused. [18:22:20] which is what everyone else does and you do most of the time, just pointing it out as a gentle reminder, nothing new. [18:22:21] (03PS12) 10Jdlrobson: Turn screen.css into screen.less [core] - 10https://gerrit.wikimedia.org/r/85920 [18:22:22] <^d> But jenkisn. [18:22:22] please add comments there if my analysis seems incorrect :) [18:22:22] (03PS3) 10Jdlrobson: Introduce .transition mixin [core] - 10https://gerrit.wikimedia.org/r/87165 [18:22:50] ^d: indeed, people clearly don't understand how sucky gerrit is and we should just disable that because it's only misused (by accident) [18:23:15] (03CR) 10Jdlrobson: "Brion I added a generic arguments parameter. This seems the most flexible, covers all use cases and allows us to manage vendor prefixes ea" [core] - 10https://gerrit.wikimedia.org/r/87165 (owner: 10Jdlrobson) [18:23:27] jdlrobson: sounds good, lemme look it over [18:23:39] <^d> Krinkle: I wouldn't mind trying to disable it on core. [18:23:43] brion: the screen.less file needs to be merged first thoguh [18:23:50] <^d> Do we think jenkins is ok enough these days on core tests? [18:24:26] what's the difference between "publish and submit" and "CR+2 V+2"? I thought there was one [18:25:06] (03CR) 10Brion VIBBER: [C: 032] Introduce .transition mixin [core] - 10https://gerrit.wikimedia.org/r/87165 (owner: 10Jdlrobson) [18:25:12] ok lemme poke that dep [18:25:27] <^d> Nemo_bis: Publish and submit means "publish my comments and go ahead and merge." Since we want jenkins to do the merging, people should almost never click it. [18:25:29] <^d> Just publish. [18:25:43] ^d: sounds good. I think the emergency case is overrated. If something is that important, it probably isn't a regular patch in gerit anyway, but a private commit on the cluster. [18:25:46] <^d> Historically, jenkins hasn't been reliable enough so I've left it on to overrule jenkins. [18:26:01] <^d> But if we think jenkins is ok enough I'd be willing to change it on core. [18:26:05] it should be fine to not be able to merge for a few hours if it really comes to it, shouldn't block any real work. [18:26:21] I know what that button does but I don't know where the V+2 comes from [18:26:39] <^d> V+2 comes from jenkins. [18:26:42] the one which is regularly used for the repos where it's not jenkins +2'ing [18:26:44] V+1 and V+2 are jenkins votes. [18:26:50] CR+2 is code review approval [18:27:01] <^d> Krinkle: You know my one regret about gerrit? I wish I hadn't made mediawiki/core and just put the code in mediawiki [18:27:02] liars :P [18:27:17] Publish = Saving that score (e.g. adding the CR+2 vote, or if you're jenkins, adding the V+2 vote) [18:27:23] Submit = Try to merge right now [18:27:35] (03PS1) 10Aaron Schulz: Added an expiry field to upload stash table [core] - 10https://gerrit.wikimedia.org/r/87607 [18:27:39] (which will reject if there is no CR+2 and V+2 at that point) [18:27:44] 2013-10-01 14.02 < grrrit-wm> (CR) Dzahn: [V: 2] "just a border color / CSS" [wikimedia/bugzilla/modifications] - https://gerrit.wikimedia.org/r/86799 (owner: TTO) [18:28:02] is that another button or the same [18:28:03] <^d> Um, this is bad. [18:28:07] (03CR) 10Aaron Schulz: [C: 04-2] "WIP" [core] - 10https://gerrit.wikimedia.org/r/87607 (owner: 10Aaron Schulz) [18:28:08] <^d> https://gerrit.wikimedia.org/r/#/admin/projects/mediawiki,access - why can't I see the acl? [18:28:11] ^d: at least we didn't go with mediawiki/core.git on github, but wikimedia/mediawiki-core.git [18:28:13] (03Abandoned) 10Physikerwelt: Adding svgtex to run MathJax on sererside [extensions/Math] - 10https://gerrit.wikimedia.org/r/87309 (owner: 10Physikerwelt) [18:28:21] (03Abandoned) 10Physikerwelt: Try to introduce some error handling [extensions/Math] - 10https://gerrit.wikimedia.org/r/87429 (owner: 10Physikerwelt) [18:28:40] (03CR) 10Brion VIBBER: [C: 032] "Looking good now." [core] - 10https://gerrit.wikimedia.org/r/85920 (owner: 10Jdlrobson) [18:28:41] ^d: I can see it, but can't change it. gerrit's interface is borked, I always use meta/config commit's from cli [18:28:42] 2013-09-23 20.57 < grrrit-wm> (CR) Jeroen De Dauw: [V: 2] Add parsing option to the runners [extensions/ParserHooks] - https://gerrit.wikimedia.org/r/85705 (owner: Jeroen De Dauw) [18:28:47] etc. etc. [18:28:57] there's plenty, not everyone has jenkins doing the job [18:29:04] <^d> Krinkle: Hard refresh and it showed. [18:29:06] <^d> Weird. [18:29:24] <^d> Meh, I always hate screwing with this acl...so messy [18:29:26] ^d: https://gist.github.com/Krinkle/a031027bbbe4597214bf [18:29:27] <^d> I'll do it after lunch. [18:29:59] (03PS27) 10Jdlrobson: Beta: Apply mobile typography lessons to Vector on desktop [core] - 10https://gerrit.wikimedia.org/r/79948 [18:30:40] (03CR) 10jenkins-bot: [V: 04-1] Added an expiry field to upload stash table [core] - 10https://gerrit.wikimedia.org/r/87607 (owner: 10Aaron Schulz) [18:30:46] ^d: https://gerrit.wikimedia.org/r/#/q/owner:krinkle+ref:refs/meta/config,n,z [18:31:36] <^d> Do all things in mediawiki/* have jenkins jobs to merge them? [18:31:43] <^d> Don't wanna prevent people from merging :) [18:31:49] Ha, no. Not all extensions. [18:32:05] and I imagine tools/ might not, though they should have basic lint jobs [18:32:15] <^d> We should at the very least have some kind of linting on all projects by default. [18:32:19] Yep [18:32:31] apps/ don't [18:32:38] brion: how's the iOS app migration going? [18:33:01] YuviPanda: it's moved over but i have to sync the latest updates and make sure monte's up on gerrit :) [18:33:05] (03CR) 10Umherirrender: [C: 031] "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/87187 (owner: 10Anomie) [18:33:06] YuviPanda: oh we should set up the bot [18:33:17] <^d> YuviPanda: Lucky for you apps/* isn't in mediawiki/* :) [18:33:22] brion: I should bring the bot back up... [18:34:00] ^d: oh we ended up with a stray apps/win8 repo -- is that just a side effect of having apps/win8/wikipedia or can that be deleted safely? [18:34:35] (03Merged) 10jenkins-bot: Introduce .transition mixin [core] - 10https://gerrit.wikimedia.org/r/87165 (owner: 10Jdlrobson) [18:34:36] (03Merged) 10jenkins-bot: Turn screen.css into screen.less [core] - 10https://gerrit.wikimedia.org/r/85920 (owner: 10Jdlrobson) [18:34:56] <^d> brion: Yeah totally, remind me after lunch [18:36:49] ^d: heh, you reminded brion-lunch of lunch ;) [18:38:58] (03CR) 10Umherirrender: "What about a default entry in $wgRateLimits in DefaultSettings.php? For 'linkpurge' this was done." [core] - 10https://gerrit.wikimedia.org/r/86007 (owner: 10Aaron Schulz) [18:42:04] (03PS4) 10Aaron Schulz: Support rate-limiting thumbnail generation [core] - 10https://gerrit.wikimedia.org/r/86007 [18:43:48] (03CR) 10Physikerwelt: "(4 comments)" [core] - 10https://gerrit.wikimedia.org/r/87479 (owner: 10Physikerwelt) [18:44:22] (03PS2) 10Physikerwelt: Mark Math-specific functions in core as deprecated [core] - 10https://gerrit.wikimedia.org/r/87479 [18:44:33] csteipp: https://gerrit.wikimedia.org/r/#/c/86007/ [18:45:12] (03CR) 10Umherirrender: [C: 031] API: Add output of wgLogo to meta=siteinfo [core] - 10https://gerrit.wikimedia.org/r/85846 (owner: 10Anomie) [18:45:35] (03PD2) 10Addshore: Remove use of deprecated LoadBalancer pickRandom [core] - 10https://gerrit.wikimedia.org/r/87608 [18:46:48] (03CR) 10Aaron Schulz: [C: 032] Remove use of deprecated LoadBalancer pickRandom [core] - 10https://gerrit.wikimedia.org/r/87608 (owner: 10Addshore) [18:47:23] (03CR) 10Physikerwelt: "(1 comment)" [extensions/Math] - 10https://gerrit.wikimedia.org/r/85801 (owner: 10Physikerwelt) [18:47:25] (03CR) 10CSteipp: [C: 031] "I like the concept, and this seems like a good way to do it. Untested, but otherwise looks good." [core] - 10https://gerrit.wikimedia.org/r/86007 (owner: 10Aaron Schulz) [18:48:50] (03PS2) 10Aaron Schulz: Added an expiry field to upload stash table [core] - 10https://gerrit.wikimedia.org/r/87607 [18:49:12] (03CR) 10Umherirrender: "Sounds like a follow up to Ia85a97434c14adcdaafc0802cbe0530bfa57a435" [core] - 10https://gerrit.wikimedia.org/r/85420 (owner: 10TTO) [18:50:40] (03CR) 10Brian Wolff: "Note, that's only true if that variable is set. If not it falls back to http timeout I believe." [core] - 10https://gerrit.wikimedia.org/r/85420 (owner: 10TTO) [18:50:41] (03Merged) 10jenkins-bot: Remove use of deprecated LoadBalancer pickRandom [core] - 10https://gerrit.wikimedia.org/r/87608 (owner: 10Addshore) [18:51:56] i'm here <-- sam [18:52:46] (03CR) 10jenkins-bot: [V: 04-1] Added an expiry field to upload stash table [core] - 10https://gerrit.wikimedia.org/r/87607 (owner: 10Aaron Schulz) [18:57:42] (03CR) 10Liangent: "(1 comment)" [extensions/Math] - 10https://gerrit.wikimedia.org/r/85801 (owner: 10Physikerwelt) [19:07:12] (03PS2) 10Addshore: Uncomment tag_summary, valid_tag tables for mysql [core] - 10https://gerrit.wikimedia.org/r/80367 [19:07:51] (03PS3) 10Addshore: Split change_tag sql patch [core] - 10https://gerrit.wikimedia.org/r/80367 [19:14:39] (03PS1) 10Umherirrender: Use always context for messages on ImagePage [core] - 10https://gerrit.wikimedia.org/r/87610 [19:15:12] (03PS2) 10Umherirrender: Use always context for messages on ImagePage [core] - 10https://gerrit.wikimedia.org/r/87610 [19:15:42] (03PS3) 10Umherirrender: Use always context for messages on ImagePage [core] - 10https://gerrit.wikimedia.org/r/86008 [19:20:10] YuviNoPower: just fine, i just need to figure out how to hack it up to get a wikifarm working [19:20:54] hmm, okay [19:20:54] consider submitting a vagrant patch if you get that to work :) [19:30:11] (03CR) 10Brion VIBBER: [C: 04-1] "Need to update docs/export-0.8.xsd with the new field?" [core] - 10https://gerrit.wikimedia.org/r/85846 (owner: 10Anomie) [19:31:41] (03PS1) 10Physikerwelt: Adding svgtex as tex to mathml and svg renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 [19:31:55] (03CR) 10jenkins-bot: [V: 04-1] Adding svgtex as tex to mathml and svg renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [19:32:07] (03CR) 10Chad: "This info wouldn't go in that though, this is just for the API's siteinfo..." [core] - 10https://gerrit.wikimedia.org/r/85846 (owner: 10Anomie) [19:33:24] AaronSchulz: ping re: https://gerrit.wikimedia.org/86883/ [19:33:32] (03CR) 10Brion VIBBER: "Bah... I hate it when different things have the same name." [core] - 10https://gerrit.wikimedia.org/r/85846 (owner: 10Anomie) [19:35:15] (03CR) 10Aaron Schulz: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/86883 (owner: 10BryanDavis) [19:35:22] huh, I had a draft there [19:35:26] (03PS2) 10Physikerwelt: Adding svgtex as tex to mathml and svg renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 [19:35:33] ahh, the same gerrit UI quirk [19:35:38] (03CR) 10jenkins-bot: [V: 04-1] Adding svgtex as tex to mathml and svg renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [19:36:02] (03CR) 10Brion VIBBER: [C: 032] "does what it says on the tin" [core] - 10https://gerrit.wikimedia.org/r/85846 (owner: 10Anomie) [19:36:07] (03PS1) 10Chad: Remove placeholder tests [core] - 10https://gerrit.wikimedia.org/r/87612 [19:36:44] (03PS3) 10Physikerwelt: Adding svgtex as tex to mathml and svg renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 [19:37:04] AaronSchulz: so "good enough" or more tweaking needed? [19:37:35] (03CR) 10Chad: [C: 032] Move maintenance scripts out of top level [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/87578 (owner: 10Chad) [19:38:13] ori-l: thoughts on varnish by default on vagrant? [19:38:54] (03CR) 10jenkins-bot: [V: 04-1] Adding svgtex as tex to mathml and svg renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [19:39:08] (03Merged) 10jenkins-bot: API: Add output of wgLogo to meta=siteinfo [core] - 10https://gerrit.wikimedia.org/r/85846 (owner: 10Anomie) [19:39:16] (03Merged) 10jenkins-bot: Move maintenance scripts out of top level [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/87578 (owner: 10Chad) [19:41:31] omg, the TimedText ui needs love. It just took me a minute to disable the subtitles that were enabled by default and made the video hard to watch (as the video player was almost entirely covered by text) [19:41:50] (03PS1) 10Legoktm: Add a config flag, $wgAllowGlobalMessaging, to allow global messaging [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/87613 [19:42:18] (03PS1) 10Chad: Also mention Elastica [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/87614 [19:42:26] (03CR) 10Chad: [C: 032] Also mention Elastica [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/87614 (owner: 10Chad) [19:42:28] (03Merged) 10jenkins-bot: Also mention Elastica [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/87614 (owner: 10Chad) [19:47:17] YuviNoPower: maybe; would have to think about the implications of that and test it out. Submit a patch? [19:47:25] ori-l: yeah, working on it now [19:48:54] (03CR) 10Brion VIBBER: [C: 04-1] "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/86007 (owner: 10Aaron Schulz) [19:53:02] (03CR) 10Nikerabbit: "The removal of () from require calls has not sneaked in here it seems." [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/87614 (owner: 10Chad) [19:53:36] Nikerabbit: you really hate parentheses don't you ;) [19:55:25] (03CR) 10Chad: "Because it doesn't matter." [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/87614 (owner: 10Chad) [19:55:44] require/require_once is a statement, not a function [19:56:55] Using parens there is equivalent to `$foo = ("some string");` [19:57:15] <^d> Right. Which means it doesn't matter. [19:57:49] ^d: You're right, but it's icky [19:58:24] <^d> Meh, plenty of other things to worry about. [19:58:52] <^d> Anyway, I filed a media bug ;-) [19:59:33] Awesome. There's plenty of them still in the wild I fear. [20:00:16] <^d> I don't think the Tiff handler works on hhvm :\ [20:00:34] ^d: Does the tiff handler work anywhere ;) [20:00:53] (03PS1) 10Anomie: Add necessary functions to $wgRedactedFunctionArguments [extensions/OAuth] - 10https://gerrit.wikimedia.org/r/87617 [20:00:55] (03PS4) 10Physikerwelt: Adding svgtex as tex to mathml and svg renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 [20:00:56] <^d> All the unit tests pass for me with zend, which is something :) [20:01:14] (03CR) 10jenkins-bot: [V: 04-1] Adding svgtex as tex to mathml and svg renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [20:01:36] I still haven't made time to get an hhvm vm running [20:02:07] If only I could just `vagrant enable-role hhvm` to test such things [20:02:15] (03CR) 10Nikerabbit: "It doesn't matter functionality wise. As a coding convention it does matter." [extensions/CirrusSearch] - 10https://gerrit.wikimedia.org/r/87614 (owner: 10Chad) [20:02:17] there's a bug for that :D [20:02:38] vagrant delegate-bug bd808 [20:02:39] bd808: brion I'm working on a varnish role :) [20:02:43] \o/ [20:03:16] <^d> If someone would be willing to sit here and help me with vagrant I'd be more than happy to do an hhvm & elastic role. [20:03:30] (03CR) 10Aaron Schulz: "(2 comments)" [core] - 10https://gerrit.wikimedia.org/r/86883 (owner: 10BryanDavis) [20:03:37] ^d: sold; Monday? [20:03:53] bd808: I noticed another issue looking at that query now [20:04:05] I'll just try to amend it into something stupid that works [20:04:09] <^d> ori-l: Sounds good, anytime other than 2-3 [20:04:34] bd808: you don't have any uncommitted changes for that do you? [20:04:56] AaronSchulz: nope. Hack away [20:05:53] AaronSchulz: asserting minTime <= maxTime seems like a good idea too [20:06:22] * bd808 makes a note to pay more attention to details [20:06:31] AaronSchulz: does https://gerrit.wikimedia.org/r/#/c/87030/ look good? (It's the profiler output one) [20:08:29] csteipp: tyler mentioned something yesterday about a plan to deploy a content security (presumably not the unsafe allow inline scripts one) on the site -- do you have a plan/timeline for that? [20:09:02] CSP? I'd love to [20:09:08] ori-l: too many things to do :) [20:09:19] np [20:09:30] No direct timeline. Jon robson had an rfc about styles, which would be one step [20:09:32] I don't want to send bd808 on other goose hunt [20:09:47] * bd808 appreciates that [20:10:01] So definitely not near term, but on my wishlist [20:11:12] (03CR) 10BryanDavis: "(2 comments)" [core] - 10https://gerrit.wikimedia.org/r/86883 (owner: 10BryanDavis) [20:13:12] (03CR) 10Bartosz Dziewoński: [C: 031] "Looks sane, didn't test." [core] - 10https://gerrit.wikimedia.org/r/80367 (owner: 10Addshore) [20:15:05] (03PS5) 10Physikerwelt: Adding svgtex as tex to mathml and svg renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 [20:25:06] (03CR) 10Physikerwelt: "(1 comment)" [extensions/Math] - 10https://gerrit.wikimedia.org/r/85801 (owner: 10Physikerwelt) [20:29:18] (03PS6) 10Physikerwelt: Adding svgtex as tex to mathml and svg renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 [20:33:34] (03CR) 10Anomie: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/87187 (owner: 10Anomie) [20:36:05] (03PS1) 10Yuvipanda: [HEAVY WIP] Add varnish role [vagrant] - 10https://gerrit.wikimedia.org/r/87623 [20:36:12] ori-l: bd808 ^ [20:36:24] beginnings of varnish integration [20:36:38] [HEAVY NOT REVIEWING YET] [20:36:45] hehe [20:36:50] * YuviNoPower accidentally merges it [20:40:56] (03CR) 10Nemo bis: [C: 04-1] ""Proposed code was updated; please update patch."" [extensions/WikimediaMessages] - 10https://gerrit.wikimedia.org/r/76709 (owner: 10Amire80) [20:42:26] (03CR) 10BryanDavis: "(3 comments)" [vagrant] - 10https://gerrit.wikimedia.org/r/87623 (owner: 10Yuvipanda) [20:45:09] ori-l, BTW, I'm not necessarily opposed to separating extensions and roles. [20:45:22] I just haven't had a chance to think about it yet, which is why I haven't replied. [20:46:46] bd808: copypaste from mediawiki.org fail [20:47:05] bd808: also,why 80? I was going to try keep internal and external ports the same [20:47:51] YuviNoPower: copypasta tastes bad. Also 80 is "normal" and easier to reason about when debugging apache problems. [20:48:08] bd808: it was originally on 8080, port'd out to 8080 [20:48:10] no? [20:48:13] hmm [20:48:14] no [20:49:02] YuviNoPower: I think it was on 80 + @forwarded_port [20:49:17] xhmm, yerah [20:49:32] bd808: so apcahe on 80, varnish on whatever, then port forward varnish to 8080 and apache to 8081? [20:49:47] bd808: or just don't have separate access to apache at all? [20:50:23] I's expose both. Maybe leave apache on 8080 and put varnish on 8081. Least surprise [20:50:32] *I'd [20:50:46] bd808: either way, if we're gonna expose both we need to figure a way to unset $wgServer [20:51:18] YuviNoPower: agreed. Ori pointed out that my hack for that was a hack and I haven't gotten back to it [20:51:35] bd808: true. [20:51:40] bd808: i've killed the whitespace now [20:51:42] (03PS2) 10Yuvipanda: [HEAVY WIP] Add varnish role [vagrant] - 10https://gerrit.wikimedia.org/r/87623 [20:52:28] Reedy: Got a minute to answer some api caching questions? [20:56:20] (03PS3) 10Yuvipanda: [HEAVY WIP] Add varnish role [vagrant] - 10https://gerrit.wikimedia.org/r/87623 [20:58:25] bd808: deleting $wgServer makes them all work fine [20:59:29] YuviNoPower: Have you tried messing with the `install` command to see if it's possible to not set $wgServer in the first place? [20:59:34] RoanKattouw: pin [20:59:47] * bd808 will look into it later if not [21:00:09] bd808: I had a patch that sortof did that, but couldn't vagrant destroy / up at that time (limited bandwidth) [21:00:16] bd808: and I'm going to sleep now, so please do poke at it! [21:00:29] bd808: plus look at the varnish too. I'm varnishing for the first time... [21:00:40] goodnight sleepy lowpower panda [21:01:22] bd808: night [21:01:28] bd808: power came back though, about 30mins ago :) [21:08:01] AaronSchulz: saw the changes you made to UploadStash.php does anything else need to be done? i see jenkins is reporting several errors on the surrounding code [21:09:33] dan-nl: sqliteupdater needs to be changed [21:10:03] csteipp: regarding csp again -- I just read your reply -- but on the plan; do you think we'll implement csp before we implement varnish esi? [21:10:31] mwalker: What's the timeline for esi? [21:10:39] fuzzy future :) [21:10:54] fuzzy future == fuzzy future? [21:11:08] but it's about the same time that we make the final transition from squid cache to varnish cache [21:11:16] I'd say > 9 months on CSP [21:11:51] But if you're planning to implement something that would require an insane CSP policy, I'd recommend against it [21:12:16] So avoid inline scripts, inline styles. [21:12:37] I would like to; for reasons of limiting the number of round trips and limiting the latency of a feature; do an inline script in the head of content pages [21:12:48] which will be able to be replaced once we roll out esi [21:13:43] We have a lot of inline scripts in mediawiki that would have to be fixed.. so yeah, do it for now, and we'll fix it when I go on that rampage. [21:14:17] AaronSchulz: looks like the test db it's using needs the additional field in the table … how do you add that? or is that not the issue? [21:14:38] ok; for reference this is what I'm talking about -- https://www.mediawiki.org/wiki/Requests_for_comment/Caching_Overhaul_-_Frontend_Proxy [21:14:57] one of your favourite extensions I'm sure :p [21:16:13] AaronSchulz: do all of those too long lines in MysqlUpdater.php need to be addressed? looks like part of the issue haste do with all of those array entries and the whitespace in there … i could clean that up if you'd like. [21:17:43] Krinkle: pong, what's up? [21:18:04] RoanKattouw: 2 things, templatedata api and visualeditor api, with regards to caching [21:18:18] RoanKattouw: I just learned that we have getCacheMode and setCacheMode, and they're set to private. [21:18:40] how dat does that impact? does it mean there is no squid caching either for anon GET requests? [21:18:45] (e.g. templatedata query) [21:18:57] how deep* [21:19:13] ^d: is there an hhvm instance I can/should use? [21:19:19] dan-nl: they shouldn't matter [21:19:37] anyway updating sqliteupdater is just like the mysql one, you also don't need to add a patch file [21:19:47] it will fallback to the mysql one [21:19:51] I can understand it being the default, because of revision delete and oversight etc. it's hard to determine generically whether something contains anything private. so I;m tryign to determine it and set it to public (or anon-public-user-private) if possible. [21:20:09] but I dont' really understand how those are used and how it impacts in wmf config with anon cachign in front of apaches [21:20:40] <^d> yhager: Not particularly. You can create new instances in labs if you want. [21:20:54] (03CR) 10Mattflaschen: "makeConfigDeprecateScript looks good." [core] - 10https://gerrit.wikimedia.org/r/76497 (owner: 10Bartosz Dziewoński) [21:21:00] ^d: I'm not quite clear on when/how instances should be used. How can I create one? [21:21:49] <^d> So if you go to https://wikitech.wikimedia.org/wiki/Special:NovaInstance on wikitech, you should be able to see projects you're a member of. [21:22:01] <^d> You should be able to create new instances in those projects. [21:22:14] <^d> There's also docs on wikitech for ssh'ing to your instance(s) [21:22:26] AaronSchulz: k, i'll leave things be. thanks for putting the change set up. it looks straight forward enough, i'll modify the code i have to start using the us_expiry value [21:23:18] ^d: I've figured out ssh'ing.. now for instances creation (baby steps). Thanks! [21:23:28] <^d> You're welcome [21:24:32] Krinkle: can i have 5 minutes of your time? https://gerrit.wikimedia.org/r/#/c/76497/ "mw.config: Deprecate wgMonthNames and wgMonthNamesShort" wants your input :) [21:28:48] mwalker: helps! [21:28:52] mwalker: non-urgent [21:29:16] what's up greg-g [21:29:20] (03PS1) 10Nemo bis: Remove unused or incognito mwe-upwiz-campaigns-editing [extensions/UploadWizard] - 10https://gerrit.wikimedia.org/r/87628 [21:29:39] MatmaRex: ok [21:29:49] mwalker: see [[wikitech:Deployments]], see "ongoing" table. see no Thursday deploys. Edit. See: [21:30:08] {{#invoke:Deployment_schedule|row [21:30:08] |when={{#invoke:Deployment_schedule|dateOf|Thu}} 11:00 PDT [21:30:08] |who=Sam / Greg [21:30:08] |what=MediaWiki deploy window, currently following the [[mw:MediaWiki 1.22/Roadmap#Schedule for the deployments|1.22 schedule)]] [21:30:11] * group2 (all Wikipedias) gets old wmf [21:30:13] * group0 (test/test2/testwikidata/loginwiki/mediawiki/) gets new wmf [21:30:16] }}| [21:30:18] eg: "Thu" [21:30:37] Krinkle: grep core for examples of how setCacheMode is used for modules that return data that's entirely public [21:30:55] RoanKattouw: done that already [21:30:58] Also, API responses are always uncached by default, you have to pass &maxage= and/or &smaxage= to get caching headers [21:30:59] mwalker: but, I have to bow out right now, no rush, it is working correctly when set with an explicit date, eg: 2013-10-07 [21:31:05] What a private cache mode does is disobey those [21:31:13] hurrrm [21:31:13] I see it is not used a lot, only never. [21:31:25] It isn't never used [21:31:25] I'll take a poke at it in a bit [21:31:35] Also, ApiQueryBase sets a public cache mode I think? [21:31:41] And that's inherited by a crapload of modules [21:31:41] typo [21:31:45] No [21:32:11] RoanKattouw: Both base and querybase default to private, the only module that sets it to public is open search [21:32:26] and the only modules using anon-public-user-private are expandtemplates and parse [21:32:27] mwalker: yeah, no problem, this is the first I've noticed it (well, awjr was like "there's no Thursday deploy, what u smoking?") [21:32:34] That's ... weird [21:32:37] That sounds broken [21:32:46] Yes [21:32:53] hence my confusion [21:33:13] I blame Tim-away , he built this cache mode stuff :) [21:33:30] greg-g: FTFY ? [21:33:48] greg-g: I just saved with no changes and now the rows show up [21:33:53] huh... [21:34:06] bd808: yessir [21:34:14] RoanKattouw: wait, so even with cache:public, it will bypass anon squid/varnish by default, it needs (s)maxage set in the req? [21:34:14] mwalker: crises averted [21:34:38] bd808: thank you kindly [21:34:45] "have you tried kicking it?" [21:35:01] have you tried turning it off an on again [21:35:21] lua man, black art [21:35:28] s/lua/mediawiki/ [21:35:35] * greg-g sets self up for joke [21:36:16] fine, don't laugh [21:36:32] Krinkle: Yes [21:36:38] * bd808 tells greg-g to get off his lawn [21:36:41] Needs to be in the query string [21:36:52] * greg-g steals bd808 cane [21:37:00] alright, back online later all [21:38:28] (03CR) 10Krinkle: [C: 04-1] "The makeConfigDeprecateScript doesn't look right to me." [core] - 10https://gerrit.wikimedia.org/r/76497 (owner: 10Bartosz Dziewoński) [21:38:56] RoanKattouw: Hm.. I can't help but wonder how come the cluster hasn't crashed from overload then [21:39:46] almost nothing is setting 'public' or 'anon-public-user-private', and even then, I don't think we're using maxage anywhere in API calls, both in core (e.g. search suggestions) and in extensions using the API. [21:40:09] opensearch clients do use it [21:40:14] But most other API consumers do not [21:40:15] I suppose it's fine since they're relatively cheap apache requests (As far as an apache hit can be cheap), since the data itself is cached very well. [21:40:37] Well the thing is too, on-demand purging for API results is pretty hard [21:40:57] So caching has to be for a set time, which brings all sorts of suckage with it [21:41:05] And so that's why the client has to opt into caching by setting a query param [21:41:17] thanks Krinkle [21:41:31] sorry I don't have better news :) [21:42:30] (03CR) 10Mattflaschen: [C: 04-1] "Good point. If we do this, it should be removed from OutputPage::getJSVars. I missed that before." [core] - 10https://gerrit.wikimedia.org/r/76497 (owner: 10Bartosz Dziewoński) [21:44:51] Krinkle: superm401: so, in short, let's abandon that for now? [21:45:17] MatmaRex, well, per my first comment, I agree we should only deprecate it if we're actually going to remove it sometime. [21:45:27] (i don't mind abandoning, i was not sure about doing that myself) [21:45:48] superm401: i think that generally we want to clean up some stuff from mw.config [21:45:53] I don't like having things in deprecated status forever, unless it's really unavoidable (API can not be changed). [21:46:04] in particular things that are not config at all, like these two :) [21:46:20] MatmaRex, well, there's lots of stuff in mw.config that's not config. [21:46:28] It would be nice to separate it, but it's not really hurting anything. [21:47:28] So my top-of-the-head thoughts are, let's only do this if we commit to removing it by X date. [21:49:36] superm401: yeah… i don't think we can do that in this case (yet) [21:49:56] superm401: (but if you want to merge some deprecation with a set removal date, then https://gerrit.wikimedia.org/r/#/q/topic:htmlform,n,z is waiting ;) ) [21:50:16] Sorry, pretty busy on GettingStarted [21:50:47] alright [21:52:12] (03CR) 10Mattflaschen: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/76497 (owner: 10Bartosz Dziewoński) [21:55:38] unreviewed commits below 700 O_o [21:59:31] MatmaRex: Well, mw.config not being config is a sunken boat. [21:59:49] we essentially have mw.serverconfig and mw.pageproperties together at last in mw.config [22:00:07] this is something that belongs in neither [22:00:51] but they're a bit too fragile imho to deprecate individually. It needs a lot of traction throughout to weed them out everywhere. Something I'd rather do together with something else. [22:01:29] yeah, understood (and i think i agree) [22:01:33] I mean, we're still dealing with broken scripts from before 2008 that haven't been fixed to work with 1.16 and earlier, not to mention 1.17 and ResourceLoader [22:02:01] MatmaRex: on that note, did I ask you to help with the resource walker? [22:02:09] https://meta.wikimedia.org/wiki/User:Krinkle/Le_Tour_de_Wik%C3%AD/2011_Resource_Walker [22:02:26] aww where is http://krinkle-tools.grizzdesign.nl/CountLogs.php [22:02:30] (03Abandoned) 10Bartosz Dziewoński: mw.config: Deprecate wgMonthNames and wgMonthNamesShort [core] - 10https://gerrit.wikimedia.org/r/76497 (owner: 10Bartosz Dziewoński) [22:02:47] (found in bookmarks while looking for the URL pasted just above) [22:03:07] I don't remember, but that domain is pretty old. I imagine I just deleted it when it wasn't working [22:03:10] Krinkle: i think you did, but i didn't ;) [22:04:07] for example: https://ab.wikipedia.org/wiki/%D0%A6%D0%B0%D1%81%D1%82%D3%99%D0%B8:%D0%92%D0%BA%D0%BB%D0%B0%D0%B4/Krinkle [22:04:31] https://als.wikipedia.org/wiki/Spezial:Byytreeg/Krinkle [22:05:25] it's probably one of the hardest things to do, as it requires knowledge and interest in a wide range of things about mediawiki, but even if you don't feel comfortable doing an entire wiki, anything would help. [22:05:29] I think you'll do fine :) [22:05:59] heh. i only have admin rights on pl.wp ;) [22:06:11] Things can change [22:06:43] and i don't really have an excess of free time now, i started university and i'm still adjusting :) [22:06:53] basically the purpose is to establish a base line where we can say: all site scripts and gadgets are in a state that is compatible and non-deprecated post MW 1.17 and either maintained by local community, deleted as part of the resource walker, or included from a place where it is maintained. [22:07:33] well, that those last 2 points are not the case in reality, but would be nice :) [22:08:15] Krinkle: on a related note, do we have some browser tests or something that visit main pages of various and look for output in the console? that sounds like it would be useful here [22:08:23] (along with more liberal usage of mw.log.deprecate) [22:08:35] of various wikis* [22:08:48] No, but it is on my long list of things I want to happen. [22:09:05] (03PS10) 10Aaron Schulz: Add script to purge changed pages from cache [core] - 10https://gerrit.wikimedia.org/r/86883 (owner: 10BryanDavis) [22:09:11] could be combined with ori-l's efforts to measure performance [22:09:29] (03PS11) 10Aaron Schulz: Add script to purge changed pages from cache [core] - 10https://gerrit.wikimedia.org/r/86883 (owner: 10BryanDavis) [22:09:33] it does a routine check (not yet on all wikis, but that could be expanded) to see how large the default module load is. [22:09:57] should be relatively trivial to extend that phantomjs run to look in the console for exceptions or deprecation notices of any kind [22:10:08] as well as 404 errors [22:10:15] (well, any HTTP error) [22:14:04] anybody knows why "git review" would suddenly start saying "git: 'review' is not a git command. See 'git --help'."? [22:14:23] I am on a mac machine, git and git review used to work just fine [22:14:48] I have tried reinstalling git-review following instructions here http://www.mediawiki.org/wiki/Gerrit/git-review#Mac_OS_X [22:14:57] zeljkof: is the script in your PATH? [22:15:04] lemme see... [22:15:05] zeljkof: `which git-review` [22:15:20] ok, not [22:15:23] i don't think git itself cares where the script is, as long as it can see it [22:15:33] zeljkof: echo $PATH; [22:15:34] (and can access it) [22:16:24] so I should add /Library/Python/2.7/site-packages to PATH then, right? [22:19:28] AaronSchulz: That's quite a change. [22:20:05] zeljkof: I'm on Mac, I have git-review in /usr/local/bin/git-review [22:20:36] I don't think site-packages is a bin directory, is it? It's an import directory for Python (like PATH is to bash) [22:20:50] indeed [22:21:47] zeljkof: What does your PATH look like at the moment? [22:22:02] Krinkle: hm, git-review is not in /usr/local/bin/ for me [22:23:07] "/Users/z/.rvm/gems/ruby-2.0.0-p247/bin:/Users/z/.rvm/gems/ruby-2.0.0-p247@global/bin:/Users/z/.rvm/rubies/ruby-2.0.0-p247/bin:/Users/z/.rvm/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/opt/X11/bin" [22:23:29] my brain is sleeping already [22:23:56] zeljkof: `which pip` and `which easy_install` [22:24:23] Krinkle: /usr/local/bin/pip and /usr/bin/easy_install [22:25:08] Krinkle: and if it matters, /usr/bin/git, git version 1.8.3.4 (Apple Git-47) [22:25:35] zeljkof: yeah,. [22:25:41] (03CR) 10BryanDavis: "(4 comments)" [core] - 10https://gerrit.wikimedia.org/r/86883 (owner: 10BryanDavis) [22:25:52] zeljkof: Do you have a package manager of some kind? homebrew, macports etc. I recommend homebrew [22:26:01] yes, homebrew [22:26:08] good [22:26:26] me too, so we shoudl be able to figure this out [22:26:27] brb [22:26:27] Krinkle: "Your system is ready to brew." [22:26:34] wee [22:26:51] (03CR) 10Aaron Schulz: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/86883 (owner: 10BryanDavis) [22:27:06] bd808: you can add those sanity checks...it's your patch again :) [22:27:48] zeljkof: brew install git [22:27:52] as a start :) [22:28:12] <^d> I love how you have to have git installed to use homebrew. [22:28:14] Krinkle: I am pretty sure I had git installed via home-brew before [22:28:25] <^d> But you should always update to brew's git because osx git is lame and outdated. [22:28:26] (03PS7) 10GWicke: Adding svgtex as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [22:28:35] (03CR) 10jenkins-bot: [V: 04-1] Adding svgtex as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [22:29:32] git is now at /usr/local/bin/git, but version still says "git version 1.8.3.4 (Apple Git-47)" [22:29:40] open a new shell [22:29:47] new tab close old tab [22:29:50] zeljkof: btw, https://gist.github.com/Krinkle/bd882618430761c36bf7 [22:30:15] you're missing /usr/local/sbin [22:30:26] probably not related, but seems appropiate [22:34:06] Krinkle: I do not even have that folder [22:34:10] "-bash: cd: /usr/local/sbin: No such file or directory" [22:34:15] zeljkof: so, with that new shell, if you install easy_install / pip / git-review, and then restart shell again, does it still not work? [22:34:22] zeljkof: where is homebrew installed then? [22:34:35] Krinkle: /usr/local/bin/brew [22:34:46] do you have /usr/local/.git and /usr/local/Formulas ? [22:34:48] Krinkle: let me try installing git-review again [22:34:54] (03CR) 10Physikerwelt: "(1 comment)" [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [22:35:22] that bin/brew is bin's brew, not where homebrew lives, Homebrew manages /usr/local, primarily through /usr/local/Cellar [22:35:25] Krinkle: I have /usr/local/.git but not /usr/local/Formulas [22:35:53] zeljkof: ah, they renamed it. I assume you have /usr/local/Library/Formula/ [22:36:09] Krinkle: yes [22:36:11] k [22:37:55] AaronSchulz: will do. [22:38:28] Krinkle: git review still does not work :( [22:38:30] brb [22:38:50] k [22:39:14] can you do a `find` and see if you can figure out (if you don't know already) where it ended up being installed? [22:40:33] (03PS8) 10GWicke: Adding svgtex as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [22:40:43] (03CR) 10jenkins-bot: [V: 04-1] Adding svgtex as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [22:46:05] (03PS1) 10BryanDavis: [WIP] Add script to purge changed files from cache [core] - 10https://gerrit.wikimedia.org/r/87635 [22:46:53] Krinkle: thanks a lot for helping me, I am too tired now, I give up [22:47:01] (03CR) 10BryanDavis: [C: 04-2] "Work in progress. Not ready to merge at all." [core] - 10https://gerrit.wikimedia.org/r/87635 (owner: 10BryanDavis) [22:47:03] will try tomorrow or next week [22:50:17] good night [22:51:53] (03PS9) 10GWicke: Adding svgtex as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [22:51:59] (03PS5) 10Aaron Schulz: Support rate-limiting thumbnail generation [core] - 10https://gerrit.wikimedia.org/r/86007 [22:52:07] (03CR) 10jenkins-bot: [V: 04-1] Adding svgtex as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [22:53:32] (03CR) 10BryanDavis: "(2 comments)" [core] - 10https://gerrit.wikimedia.org/r/87635 (owner: 10BryanDavis) [22:55:30] AaronSchulz: Could you look at https://gerrit.wikimedia.org/r/87635 and see if I'm on the right track? [22:56:10] when I get the chance, sure [22:56:33] Krinkle: For QUnit tests, is there an equivalent to @dataProvider in phpunit or do I need to roll my own? [22:57:01] I'm thinking just a for loop through an array, calling the same test with different arguments [22:57:15] AaronSchulz: thanks [22:58:07] (03CR) 10GWicke: "(1 comment)" [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [22:59:06] marktraceur: basically either $.each( dataSet, -> QUnit.test( .., cases.length, -> (assert) { [22:59:07] or [22:59:34] Right right [22:59:43] marktraceur: basically either QUnit.test( .., cases.length, (assert) { $.each( cases, -> [22:59:57] I'd recommend the latter if you can. [23:00:16] Krinkle: I actually want the separate tests, because it gives me context for the messages [23:00:34] it depends on how many assertions for each data set [23:00:37] yeah, should be fine [23:00:38] Oh, but...hm [23:00:52] Krinkle: I see that the rest of the tests in mw.Title are very broadly scoped into big buckets [23:01:15] So me adding 8 separate tests doubles the size of the results page for the module [23:01:19] Or, nearly so [23:01:38] (03CR) 10BryanDavis: [C: 031] Support rate-limiting thumbnail generation [core] - 10https://gerrit.wikimedia.org/r/86007 (owner: 10Aaron Schulz) [23:01:43] indeed, you shoudn't need separate tests [23:01:54] I usually follow the design of 1 QUnit.test per method [23:02:13] or general feature (e.g. "Transformation", "Case-sensivity") [23:02:25] and then the rest per method (constructor, newFromText, getExtension, getUrl) [23:02:41] 'kay done [23:02:46] but if you need your teardown/setup to run, separate tests is fine [23:02:59] (03PS3) 10MarkTraceur: Add file-link "parser" to mw.Title from commons [core] - 10https://gerrit.wikimedia.org/r/87487 [23:03:04] that is, if you find yourself implementing teardown/setup, go for separate tests ;) [23:03:12] No, it doesn't even have teardown/setup, so it's cool [23:03:13] (03PS1) 10Bsitu: Add style to new topic [extensions/Flow] - 10https://gerrit.wikimedia.org/r/87637 [23:03:29] (03CR) 10jenkins-bot: [V: 04-1] Add file-link "parser" to mw.Title from commons [core] - 10https://gerrit.wikimedia.org/r/87487 (owner: 10MarkTraceur) [23:03:30] One Big Test is fine [23:03:39] Oh, god damn it Jenkins [23:04:29] marktraceur: "Empty string was magically turned into an mw.Title" that doesnt' make sense as the assertion name [23:04:56] the assert message is not the "error message for if this fails" [23:05:12] (03CR) 10MZMcBride: "Does this have an associated bug report? The current commit message describes what is being done, but offers no rationale as to why it's b" [core] - 10https://gerrit.wikimedia.org/r/86007 (owner: 10Aaron Schulz) [23:05:45] Oh [23:05:46] Hm [23:05:50] I'll fix it [23:06:09] Krinkle: I guess that's a phpunitism that I carried over [23:06:30] marktraceur: doesn't make sense to me to do there either. [23:06:43] Krinkle: You only get that message if the test fails [23:06:57] Or I guess if you expand the successful tests. [23:06:58] maybe from the cli runner, but even then, so what. [23:06:59] * marktraceur learns [23:10:51] (03PS4) 10MarkTraceur: Add file-link "parser" to mw.Title from commons [core] - 10https://gerrit.wikimedia.org/r/87487 [23:16:23] Marg, still need to change it [23:25:57] (03PS12) 10BryanDavis: Add script to purge changed pages from cache [core] - 10https://gerrit.wikimedia.org/r/86883 [23:26:45] (03CR) 10BryanDavis: "(3 comments)" [core] - 10https://gerrit.wikimedia.org/r/86883 (owner: 10BryanDavis) [23:27:38] (03PS2) 10Bsitu: Add style to new topic [extensions/Flow] - 10https://gerrit.wikimedia.org/r/87637 [23:31:57] (03PS10) 10GWicke: Adding svgtex as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [23:32:33] (03CR) 10jenkins-bot: [V: 04-1] Adding svgtex as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [23:41:33] (03CR) 10BryanDavis: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/86883 (owner: 10BryanDavis) [23:42:08] (03PS11) 10GWicke: Adding mathoid as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [23:42:17] (03CR) 10jenkins-bot: [V: 04-1] Adding mathoid as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [23:44:13] (03PS12) 10GWicke: Adding mathoid as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [23:44:22] (03CR) 10jenkins-bot: [V: 04-1] Adding mathoid as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [23:45:24] (03CR) 10GWicke: [C: 031] Adding mathoid as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [23:48:01] (03PS13) 10GWicke: Adding mathoid as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [23:48:11] (03CR) 10jenkins-bot: [V: 04-1] Adding mathoid as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [23:50:16] (03PS14) 10GWicke: Adding mathoid as tex to MathML and SVG renderer [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [23:54:19] Krinkle: yt? [23:54:22] * ebernhardson sighs ... the bug where vagrant sees corrupted files after i move them around is really anoying [23:54:34] ebernhardson: update-guest-additions [23:54:49] ahh, i shouldn't have been ignoring that so much :) ok [23:58:10] ori-l: sort of [23:58:35] is it possible to prevent a jqxhr from resolving until i decide it's ready? [23:58:52] short of wrapping it in another deferred [23:58:57] (03CR) 10Physikerwelt: "(2 comments)" [extensions/Math] - 10https://gerrit.wikimedia.org/r/87611 (owner: 10Physikerwelt) [23:59:49] for some infuriating reason resource timing information for XHR is available 1 tick after readyState == 4