' ) [07:39:18] .addClass( 'mlb-overlay' ); from libhtboxinterface.js, I find that the black backdrop is gone. Should that mean, $overlay is the object for the backdrop ? [07:39:27] Yeah. [07:39:40] Wait, what? Removing it? [07:43:14] marktraceur: nope, I was trying it for a start. To find out which class the backdrop belonged to [07:44:00] marktraceur: but the thing is, when I give another function, say the click.close() to the $overlay. Nothing happens when I click on the background [07:46:47] tonythomas01: No, because there are other elements in front of the overlay [07:49:54] tonythomas01: Didn't I explain that this.$imageDiv was the thing to listen for? [07:53:06] (03PS2) 10Adamw: (bug 56849) Pass true oldid for atomic page save [core] - 10https://gerrit.wikimedia.org/r/94584 [07:56:32] (03CR) 10jenkins-bot: [V: 04-1] (bug 56849) Pass true oldid for atomic page save [core] - 10https://gerrit.wikimedia.org/r/94584 (owner: 10Adamw) [08:12:30] marktraceur: Earlier, when I gave the click functionality on the imageDiv. The thing just disappeared even on clicking the image [08:12:50] legoktm, hi [08:13:07] now let me try. - > .click( function() {removeClass('overly')) ) [08:13:10] yurik: does the summary I wrote for https://www.mediawiki.org/wiki/Architecture_Summit_2014/Configuration#Overview look ok to you? [08:13:20] (since one of them is your rfc( [08:13:21] )* [08:13:23] tonythomas01: Then check if the image is the target of the event and ignore the click then [08:13:32] tonythomas01: That will do...nothing [08:13:59] marktraceur: that will do. Let me try coding it [08:14:04] legoktm, looks good to me :) [08:14:20] btw, i don't intend mine to be "the complete solution" [08:14:34] more like a convenient stopgap [08:14:44] tonythomas01: You shouldn't ask only me for help - ask a question and hilight me, someone else may be able to help [08:15:20] marktraceur: Yes. from next time [08:15:42] I'll hold you to it :) [08:16:03] yurik: do you want to mention that in your RfC then? my goal is to have a config interface that is well designed and will last rather than a temporary one [08:18:04] legoktm, not sure -- i am not yet convinced that having just one solution is good for all configuration needs - json style is good for "blob" configuration style where its used in all of the projects in a cluster, whereas config system might be better for about:config (firefox) style of settings [08:18:28] ok, makes sense [08:18:29] like a key-value approach [08:19:00] legoktm, btw, you mention that uploadwizard has something like a json config? [08:19:03] i can't seem to fintd it [08:19:08] yeah, for campaigns [08:19:11] let me find one [08:19:55] yurik: https://commons.wikimedia.org/w/index.php?title=Campaign:wlm-ca&action=edit and https://commons.wikimedia.org/wiki/Campaign:wlm-ca [08:20:27] yuvipanda wrote it [08:20:28] uuu, me like! thanks :) [08:20:38] this is exactly in the same style [08:20:52] I'm pretty sure he took the code from EventLogging :P [08:21:37] i will add that link to my rfc [08:21:40] he who? [08:21:47] yuvipanda [08:23:49] legoktm, thx a lot for the info! [08:24:16] np, I'm looking forward to meeting you at the summit [08:24:24] sounds good :) [09:03:56] (03CR) 10Ori.livneh: [C: 032] "Nice" [extensions/GettingStarted] - 10https://gerrit.wikimedia.org/r/103737 (owner: 10Mattflaschen) [09:04:00] (03Merged) 10jenkins-bot: Client-side wrapper for GettingStartedGetPages API [extensions/GettingStarted] - 10https://gerrit.wikimedia.org/r/103737 (owner: 10Mattflaschen) [09:10:18] (03CR) 10Amire80: "This broke GENDER support in flow-moderation-confirmation-* messages. mw.user is needed for gender support in the imperative that speaks t" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/100953 (owner: 10Werdna) [09:43:43] (03PS1) 10Umherirrender: treat true as empty string, skip false in xml format [core] - 10https://gerrit.wikimedia.org/r/108315 [10:00:40] (03PS3) 10Legoktm: Allow posting in all talk namespaces by default [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/94698 [10:01:05] (03CR) 10Legoktm: "Bump?" [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/94698 (owner: 10Legoktm) [10:05:13] (03CR) 10Reedy: [C: 04-1] "I'm not sure if this is actually going to help." [extensions/WikimediaMaintenance] - 10https://gerrit.wikimedia.org/r/108293 (owner: 10Legoktm) [10:06:12] Reedy: at what point in the process is the wiki's config added to $wgConf? [10:08:59] (03CR) 10Legoktm: "Will the wiki be in $wgConf->getLocalDatabases()? The only other thing that needs to be set is $wgCanonicalServer, and for any new Wikiped" [extensions/WikimediaMaintenance] - 10https://gerrit.wikimedia.org/r/108293 (owner: 10Legoktm) [10:13:58] (03PS7) 10Adamw: Basic state machinery [extensions/Workflow] - 10https://gerrit.wikimedia.org/r/107307 [10:15:18] (03PS1) 10Umherirrender: Remove outdated fixme/todo in CoreParserFunctions.php [core] - 10https://gerrit.wikimedia.org/r/108316 [10:27:05] (03CR) 10Reedy: "I don't think so. It will be the all.dblist from before mwscript is/was run" [extensions/WikimediaMaintenance] - 10https://gerrit.wikimedia.org/r/108293 (owner: 10Legoktm) [11:02:49] (03PS2) 10Legoktm: Clear MassMessage cache when creating a new wiki [extensions/WikimediaMaintenance] - 10https://gerrit.wikimedia.org/r/108293 [12:34:48] (03CR) 10Frédéric Wang: "Any update on this?" [extensions/Math] - 10https://gerrit.wikimedia.org/r/105187 (owner: 10Physikerwelt) [12:43:26] (03CR) 10Frédéric Wang: [C: 031] "I don't know about MediaWiki caching and options but this looks good to me. The changeset is small, so that does not seem something which " [extensions/Math] - 10https://gerrit.wikimedia.org/r/104991 (owner: 10Physikerwelt) [12:44:03] (03PS1) 10Gerrit Patch Uploader: Go to top button for pages [core] - 10https://gerrit.wikimedia.org/r/108323 [12:44:05] (03CR) 10Gerrit Patch Uploader: "This commit was uploaded using the Gerrit Patch Uploader [1]." [core] - 10https://gerrit.wikimedia.org/r/108323 (owner: 10Gerrit Patch Uploader) [12:45:55] (03CR) 10jenkins-bot: [V: 04-1] Go to top button for pages [core] - 10https://gerrit.wikimedia.org/r/108323 (owner: 10Gerrit Patch Uploader) [13:56:12] (03PS1) 10Whym: (bug 53132) properly get timestamp position in DiscussionParser [extensions/Echo] - 10https://gerrit.wikimedia.org/r/108329 [13:56:15] (03CR) 10jenkins-bot: [V: 04-1] (bug 53132) properly get timestamp position in DiscussionParser [extensions/Echo] - 10https://gerrit.wikimedia.org/r/108329 (owner: 10Whym) [14:13:11] (03CR) 10Hoo man: "Mostly style things, looks good in general... didn't look at the Special page changes." (0310 comments) [core] - 10https://gerrit.wikimedia.org/r/106517 (owner: 10Daniel Kinzler) [14:26:24] (03PS1) 10Dereckson: Improve UTF-8 and links support [extensions/TweetANew] - 10https://gerrit.wikimedia.org/r/108331 [14:31:19] (03CR) 10Dereckson: [C: 04-1] "I18n issue." (031 comment) [extensions/TweetANew] - 10https://gerrit.wikimedia.org/r/108331 (owner: 10Dereckson) [14:34:17] https://gerrit.wikimedia.org/r/#/c/108329/ Any idea to solve the error from jenkins-bot? [14:35:08] I'm pretty sure I rebased it. [15:22:40] (03CR) 10Martineznovo: "This change doesn't make the process run in the background, and may causes other problems. See bug 60210, bug 58719, and bug 60208" (031 comment) [core] - 10https://gerrit.wikimedia.org/r/59797 (owner: 10Aaron Schulz) [15:26:11] (03PS6) 10EBernhardson: Hack together apache -> hhvm [vagrant] - 10https://gerrit.wikimedia.org/r/105834 [15:30:07] (03PS1) 10Umherirrender: Add ParserOutput::addJsConfigVars [core] - 10https://gerrit.wikimedia.org/r/108333 [15:40:48] (03PS2) 10Umherirrender: Add ParserOutput::addJsConfigVars [core] - 10https://gerrit.wikimedia.org/r/108333 [15:41:34] (03CR) 10Umherirrender: "Patch Set 2: Removed trailing space" [core] - 10https://gerrit.wikimedia.org/r/108333 (owner: 10Umherirrender) [15:57:15] (03CR) 10Martineznovo: "Filed bug 60231 about the synchronous problem." [core] - 10https://gerrit.wikimedia.org/r/59797 (owner: 10Aaron Schulz) [17:11:51] (03CR) 10Parent5446: [C: 032] Remove outdated fixme/todo in CoreParserFunctions.php [core] - 10https://gerrit.wikimedia.org/r/108316 (owner: 10Umherirrender) [17:14:47] (03Merged) 10jenkins-bot: Remove outdated fixme/todo in CoreParserFunctions.php [core] - 10https://gerrit.wikimedia.org/r/108316 (owner: 10Umherirrender) [17:55:09] (03PS7) 10EBernhardson: Hack together apache -> hhvm [vagrant] - 10https://gerrit.wikimedia.org/r/105834 [18:29:11] anomie: ping [18:29:35] anomie: https://en.wikipedia.org/wiki/Wikipedia_talk:Lua#.22Hacking.22_into_Wikidata_to_get_properties_of_miscellaneous_articles [18:30:32] > No worst come to worst Anomie will take your toys away :) —TheDJ (talk • contribs) 11:28, 13 January 2014 (UTC) [18:31:04] lol [18:31:27] hoo: What about it? The hard part is deciding what exactly to do about mw.message's non-plain modes. [18:32:09] There's a reason we don't support that yet (although it's like a two-line change) [18:34:53] well, I guess we can't do anything about that right now... but if someone creates a module and wants to use that widespread I'm going to do something about that ... [18:35:07] they wont [18:36:35] legoktm: Ok... going to comment on that also, I guess [18:43:41] (03CR) 10Bartosz Dziewoński: "recheck" [extensions/Echo] - 10https://gerrit.wikimedia.org/r/108329 (owner: 10Whym) [18:46:13] (03CR) 10Bartosz Dziewoński: "DiscussionParser seems to have quite an extensive test suite (tests/DiscussionParserTest.php), you might want to add/adjust tests there." [extensions/Echo] - 10https://gerrit.wikimedia.org/r/108329 (owner: 10Whym) [18:48:02] (03CR) 10Brian Wolff: [C: 04-1] "I'm concerned that this may cause some performance problems, especially when first deployed." (032 comments) [extensions/Math] - 10https://gerrit.wikimedia.org/r/104991 (owner: 10Physikerwelt) [18:55:03] Blocked loading mixed active content "http://stats.grok.se/jsonp/en/latest30/foo?callback=pageviewsCallback&_=1390157596276" [18:55:11] ^^^ How do I work around that? [18:55:38] stats.grok.se doesn't have an https version.. [18:55:55] Is there some way I can force Wikipedia into the HTTP version? [18:56:17] http://jsfiddle.net/DZaE2/1/ is what I'm trying to get [18:56:53] !log faidon updated /a/common to {{Gerrit|I48a026eb8}}: Revert wmgRC2UDPAddress to ekrem [18:59:05] !log faidon synchronized wmf-config/InitialiseSettings.php 'switch wmgRC2UDPAddress back to ekrem' [18:59:23] Technical_13: Well there is a preference I guess [18:59:39] techman224: use a better browser that allows you to toggle that [18:59:42] Technical_13: ^ [18:59:51] (Opera 12 does, dunno about others) [19:00:08] For now I've enabled active mixed content in about:config on firefox 26 [19:00:23] The only issue is others won't be able to run this script... [19:00:42] Is there any chance of getting stats.grok.se moved to labs or something? [19:01:01] You would have to ask its maintainer [19:01:36] But the raw stats data is public, so there's nothing stopping someone from putting something on labs that does the same thing [19:04:06] Is the data accessable through the api? [19:05:17] or only directly from the database through labs? [19:05:53] (03CR) 10Physikerwelt: "@Brian Wolff: I think this would ideed effect all pages, even if they don't contain math at all. However, I see no alternative beside leav" (032 comments) [extensions/Math] - 10https://gerrit.wikimedia.org/r/104991 (owner: 10Physikerwelt) [19:06:44] (03CR) 10Bartosz Dziewoński: "No tests? No in-code examples? :(" (034 comments) [core] - 10https://gerrit.wikimedia.org/r/108045 (owner: 10Jack Phoenix) [19:06:58] Technical_13: neither, its accessible via text files you can download from the dumps server [19:07:12] Technical_13: http://dumps.wikimedia.org/other/pagecounts-raw/ [19:07:40] grok.se just post processes those files into a more useful form [19:08:19] (03CR) 10Bartosz Dziewoński: New class for parsing MediaWiki:Sidebar-like messages into modern, nested navigation menus. (032 comments) [core] - 10https://gerrit.wikimedia.org/r/108045 (owner: 10Jack Phoenix) [19:08:39] There's some other people who also use those files. I have a bot on wikinews that updates n:template:Popular_articles based on the raw page counts dumps [19:10:51] (03CR) 10Anomie: Remove call to deprecated ParserOptions::getMath (032 comments) [extensions/Math] - 10https://gerrit.wikimedia.org/r/104991 (owner: 10Physikerwelt) [19:16:46] (03CR) 10Brian Wolff: Remove call to deprecated ParserOptions::getMath (031 comment) [extensions/Math] - 10https://gerrit.wikimedia.org/r/104991 (owner: 10Physikerwelt) [19:39:51] (03CR) 10Anomie: [C: 04-1] "The change to not display the warning when no text is given looks good at first glance." [core] - 10https://gerrit.wikimedia.org/r/108307 (owner: 10Nullzero) [19:53:25] anyone around who has experience reading json from the api with javascript? [19:53:28] ping me, pls. [20:03:25] !ask | jorm [20:03:26] jorm: Please feel free to ask your question: if anybody who knows the answer is around, they will surely reply. Don't ask for help or for attention before actually asking your question, that's just a waste of time – both yours and everybody else's. :) [20:03:56] loading json data via javascript into an object. [20:04:20] the api data for language links is like this: [20:04:20] { [20:04:21] "lang": "af", [20:04:21] "url": "http://af.wikipedia.org/wiki/Kat", [20:04:21] "*": "Kat" [20:04:21] }, [20:04:45] accessing the * entry fails spectacularly. [20:04:55] use ['*'] instead of .* [20:04:55] Access it via [] operator [20:05:10] there's a bug somewhere about this being silly [20:05:14] thanks for clarifying, MatmaRex. [20:05:21] MatmaRex: https://bugzilla.wikimedia.org/show_bug.cgi?id=31629 [20:05:49] TL;DR version: "Because XML". [20:06:01] (I really did search bugzilla before filing that svg bug, just using the wrong words I guess.) [20:06:11] ragesoss: :) [20:06:20] ragesoss: bugzilla search sucks, don't worry about that :) [20:06:43] bugzilla search is not really a search [20:06:46] it's a query engine [20:06:51] yeah, oops, i had .['*'] [20:06:51] (and it excels at being one) [20:07:30] MatmaRex: Unfortunately I think we need a good search more than a query system [20:08:02] a good search would definitely be nice, but searching is hard [20:08:13] and i will fight tooth and nail against anyone trying to take my query engine :P [20:08:35] hey Nemo_bis, I was thinking of ideas for a bot, and wondered if a tempaltedata bot might work? [20:09:00] like a bot that would scrape the parameters of a template and generate basic template data to 1. help people using visualeditor to at least have a skeleton templatedata [20:09:20] 2. givep technicals in the back a headstart on making templatedata [20:17:20] MatmaRex: I don't look for bugs, bugs find me [20:17:32] (03PS1) 10Umherirrender: Expand documentation around $wgAPI*Modules [core] - 10https://gerrit.wikimedia.org/r/108411 [20:18:19] * bawolff glares at TimedMediaHandler angrily [20:19:00] heh [20:19:14] Appearently its css is written in such a way, that it has to be loaded by addModuleStyles (to prevent flash unstyled content) and addModules (the css is not specific enough, and the change in load order causes things to break) [20:19:32] I think, or perhaps its just haunted [20:20:00] who thought "*" for a key was a good idea? [20:20:02] (03PS2) 10Umherirrender: Expand documentation around $wgAPI*Modules [core] - 10https://gerrit.wikimedia.org/r/108411 [20:20:46] (03CR) 10Umherirrender: "Patch set 2: Removed trailing space" [core] - 10https://gerrit.wikimedia.org/r/108411 (owner: 10Umherirrender) [20:22:07] Probably Yuri. [20:22:32] You could look if you're really interested. Though the answer may be hiding in an extension. [20:22:44] who thought xml was a good idea for anything ever? [20:23:14] in 1998, xml was a *godsend* [20:23:31] it was the best idea, ever, at that point in time. [20:23:42] the concept of divorcing data from display logic... [20:24:24] the concept is fine, the choice of implementing format... [20:24:37] well, we didn't have much else to work with. [20:25:02] it had to be machine readable *and* human parseable/editable [20:25:32] and you have one single CLOB to hold everything that is infinitely extensible. [20:25:54] the data is structured but you can modify the structure *ad hoc* without having to add/remove rows/tables from a database. [20:26:02] and you can run diffs on it. [20:26:50] there was lots to love. [20:27:10] wading through what our API spits out in JSON, i'm kind of appalled. [20:27:29] but that's mostly because pages themselves are just CLOBS. Completely, terribly, unstructured CLOBS. [20:27:55] in the first iteration of the winter prototype, i manually marked out sections into thier own bits. [20:28:22] now that i've got it reading the JSON, i can't do that with any ease - I have to get the sections[] object and then do fucking *byte math* in order to mark sections out. [20:28:26] (03PS1) 10Umherirrender: Split documentation for $wgRCLinkLimits/$wgRCLinkDays [core] - 10https://gerrit.wikimedia.org/r/108466 [20:28:30] that's frankly retarded. [20:30:11] What? you're doing byte math? [20:30:27] if i want to find section starts and ends, i do. [20:30:39] i feel like a baby in a hardcore 1337 club [20:30:45] (which is true) [20:31:24] consider: [20:31:25] { [20:31:25] "toclevel": 1, [20:31:25] "level": "2", [20:31:25] "line": "Nomenclature and etymology", [20:31:25] "number": "1", [20:31:25] "index": "1", [20:31:25] "fromtitle": "Cat", [20:31:26] "byteoffset": 6363, [20:31:26] "anchor": "Nomenclature_and_etymology" [20:31:27] }, [20:31:35] that byteoffset thing, that's the marker the section starts. [20:31:55] so i find that, and then i find the byteoffset for the NEXT section, and that's the section size. [20:32:02] jorm: I think there's an option to retrieve only a single section [20:32:07] sure there is. [20:32:15] but i can't do that - [20:32:26] i need to pull full text, and then mark sections, as a single request. [20:32:48] otherwise, we're doing shite like pulling a list of all sections, and then pulling each section in its own request, and assembling them. [20:32:57] which is a nightmare when done async. [20:33:11] because i'll get them in this order: 2, 5, 4, 7, 3, 1, 6 [20:33:28] hmm, that's probably an offset in utf-8 bytes, but js encodes strings as utf-16. That will probably add to the nightmere [20:34:07] yup. [20:34:19] you want to see the problem i'm trying to solve? [20:34:33] sure, I'm kind of curious at this point [20:34:47] go here: http://unicorn.wmflabs.org/winter/ [20:34:52] and hover over any section edit link. [20:34:59] (not the top link) [20:35:13] you'll notice that the *targeted* section highlights. so you know what you'll be editing. [20:35:35] that only works in that version of the prototype because i hand-massaged the HTML to mark sections. [20:35:36] [random aside]: I really hope that's a real article [20:35:50] it is [20:35:55] now that the prototype is reading real data via json, that feature gotstago bye-bye [20:35:59] it is. it's my favorite article. [20:36:11] jorm: hmmmm [20:36:27] i hope that's a "hmmm, i might have an idea" [20:36:35] yeahhh maybe [20:36:52] the section are delimited in HTML by [20:36:54] that's cool, would help users see which section the edit link goes to which people always complain is confusing [20:37:00] when i saw that deme, i thought you were just using that [20:37:05] demo* [20:37:20] (03CR) 10Nullzero: "@Anomie: Can you look at http://en.wikipedia.org/w/api.php?action=parse&text={{Project:Sandbox}}? It gives a warning "No 'title' or 'conte" [core] - 10https://gerrit.wikimedia.org/r/108307 (owner: 10Nullzero) [20:37:23] hrm. [20:37:29] it's more complicated than it looks at first glance because these s can be nasted in all kinds of elements [20:37:34] nested* [20:37:49] you're suggesting that i tag the link to the hN, and then on hover say "next children", until next ? [20:37:53] but you could definitely mark sections using just that and walking the DOM tree [20:37:55] that gets really goofy with nesting. [20:38:07] yeah, it's not very pleasant [20:38:30] eventually (sooner, rather than later, hopefully), these demo things are going to be in git/gerrit [20:38:36] and you'll easily be able to hit the code. [20:38:49] jorm: this exists: https://developer.mozilla.org/en-US/docs/Web/API/document.createTreeWalker [20:38:56] i'm not sure about what exactly it does [20:38:59] but it might help [20:39:04] probably more heavyweight than i need for a demo. [20:39:07] (i'm also not sure about browser compatibility) [20:39:19] ideally, we'll jsut make a thing in core that just adds the right

tags around sections. [20:39:31] and then i won't have to worry about trickery [20:39:43] What's a section? :-) [20:40:07] jorm: that sounds impossible [20:40:10] due to the nesting [20:40:34] and doing that by walking the tree should be doable enough [20:40:34] well, i got it figured when i manually did it. [20:40:43] there's probably a jquery module for it [20:40:55] or extension, or however are these things called [20:41:04] library [20:41:14] i'm trying to avoid adding any [20:41:20] winter will require jquery.waypoints [20:41:25] but other than that i'm pure. [20:42:36] (03PS1) 10Umherirrender: Remove outdated comment from Preferences.php [core] - 10https://gerrit.wikimedia.org/r/108468 [20:44:17] jorm: it would be like 50 lines of code [20:46:32] * MatmaRex starts coding [20:50:17] (03PS1) 10Raimond Spekking: Module namespace translations for Serbian Latin and Croatian [extensions/Scribunto] - 10https://gerrit.wikimedia.org/r/108469 [20:51:58] (03CR) 10Anomie: "Yes, using wikitext when no title or model is given is for BC. And yes, that example link should probably be fixed." [core] - 10https://gerrit.wikimedia.org/r/108307 (owner: 10Nullzero) [20:53:31] GEOFBOT: perhaps an AWB plugin would work better, or you could improve that tool which is already available and advertise it on some community pages [20:54:20] i.e. http://tools.wikimedia.pl/~mlazowik/templatedata/ [20:59:34] jorm: okay, it'd take me an hour or two to devise something that's both robust and fits what you need to do [20:59:47] and i don't feel like spending that time right now [21:00:04] but it's definitely possible and feasible [21:02:02] hrm. [21:02:27] let me get my shit in a place where it's uploadable/publishable/deployable, and then i can point you at the (hacky) source. [21:02:28] (it's 10 pm here and i've got classes tomorrow, eh) [21:02:41] it's a holiday! [21:02:47] wait, maybe not for you. [21:02:47] heh. [21:02:56] yeah [21:03:20] (03PS8) 10Ori.livneh: Run MediaWiki on Apache + fastcgi + HHVM [vagrant] - 10https://gerrit.wikimedia.org/r/105834 (owner: 10EBernhardson) [21:03:22] (03PS1) 10Ori.livneh: Add apache::env [vagrant] - 10https://gerrit.wikimedia.org/r/108470 [21:03:24] and then i need to sort out some things before i hop on a plane on wednesday [21:04:07] where are you heading? [21:04:23] the architecture summit? :D [21:04:31] (03CR) 10Ori.livneh: [C: 032] Add apache::env [vagrant] - 10https://gerrit.wikimedia.org/r/108470 (owner: 10Ori.livneh) [21:04:33] (03Merged) 10jenkins-bot: Add apache::env [vagrant] - 10https://gerrit.wikimedia.org/r/108470 (owner: 10Ori.livneh) [21:04:39] oh! [21:04:53] ebernhardson: updated your patch, check it out [21:05:39] it works suuuper nicely now. Special:Version says php; vagrant enable-role hhvm && vagrant provision ; refresh Special:Version and presto, hhvm. [21:07:00] i'm going to rebuild a vm from scratch to test, will merge then [21:07:17] (03CR) 10Bartosz Dziewoński: [C: 032] Remove outdated comment from Preferences.php [core] - 10https://gerrit.wikimedia.org/r/108468 (owner: 10Umherirrender) [21:10:18] (03Merged) 10jenkins-bot: Remove outdated comment from Preferences.php [core] - 10https://gerrit.wikimedia.org/r/108468 (owner: 10Umherirrender) [21:18:11] Nemo_bis: i think a bot might help the most immediately, because a bunch of templates that need data would a skeleton version of it [21:18:25] the tool would help more with tempalte authors [21:28:47] GEOFBOT: well but where would you save the skeletons? [21:28:52] talk page? [21:29:13] o is there a way to mark some templatedata as draft/sandbox [21:29:26] on the doc page but with a tempalte or something that would put it in a category for checking maybe [21:29:32] like how bot-marked move-to-commons works [21:30:28] they don't delete NowCommons automatically [21:31:07] I don't know how much harm wrong templatedata can do [21:31:29] surely you need consensus so perhaps you may start testing on a smaller wiki and then propose in a big oneù [21:31:47] ok [21:31:47] (03CR) 10Hoo man: "Wasn't this the change which made us think about a JavaScript hook thing, which later became mw.hook... any plans on using that here, if t" [core] - 10https://gerrit.wikimedia.org/r/53029 (owner: 10Bartosz Dziewoński) [21:34:01] (03PS9) 10Ori.livneh: Run MediaWiki on Apache + fastcgi + HHVM [vagrant] - 10https://gerrit.wikimedia.org/r/105834 (owner: 10EBernhardson) [21:35:51] (03CR) 10Ori.livneh: [C: 032] "So. Awesome." [vagrant] - 10https://gerrit.wikimedia.org/r/105834 (owner: 10EBernhardson) [21:35:54] (03Merged) 10jenkins-bot: Run MediaWiki on Apache + fastcgi + HHVM [vagrant] - 10https://gerrit.wikimedia.org/r/105834 (owner: 10EBernhardson) [21:39:06] ori: looks like path_translated - document root is how they decided to detect mod_fastcgi https://github.com/facebook/hhvm/commit/0e20e402d928011d903b3472a5eb9d9300747f74 [21:39:23] looking into if there is a better way, unfortunatly php-fpm is a mess i dont yet see how they do it [21:39:51] ebernhardson: http://lists.wikimedia.org/pipermail/wikitech-l/2014-January/074089.html [21:40:59] wooooo, this is cool [21:41:27] excellent [21:42:31] btw, tangentially related.. mw-v with NFS is ridiculously faster than virtualbox shared folders [21:42:49] going to bite the bullet and have a config that automatically uses NFS where available (osx and linux hosts) [21:45:37] its funny how NFS ends up being better than all the things invented since :) [21:45:59] heh [21:46:04] * ori runs off [21:46:15] have a good rest of your weekend, etc., and thanks again! [21:46:27] really glad to see this happen [21:46:44] one good step twords getting it tested in dev stuff [21:46:52] i already found a few errors in flow :) [21:47:22] (03CR) 10Brian Wolff: "I think this may have broken things locally for me (Dialogs are now sized very weirdly), but everything looks find on commons, so I'm a li" [extensions/MwEmbedSupport] - 10https://gerrit.wikimedia.org/r/88943 (owner: 10Mattflaschen) [21:48:32] man, NFS. [21:48:37] ori: I looked into the heart of the javascript of TimedMediaHandler, but all I saw was madness [21:48:51] is that michael dale's code? [21:48:59] yes [21:49:21] bawolff: i am really running off, but <3 <3 <3 if you remove those things from [21:50:33] Its not even working for me locally at the moment, and the commit I thinks at fault was from 2 months ago, but it didn't show up for me until today, and I'm just very confused [22:13:04] (03CR) 10Brian Wolff: "Sorry nevermind. Please ignore me." [extensions/MwEmbedSupport] - 10https://gerrit.wikimedia.org/r/88943 (owner: 10Mattflaschen) [22:55:42] ugh TMH [22:59:04] bawolff: i was thinking of picking up that patch to make TMH load only when it's needed.. has anything changed in the last few weeks that would affect it? [22:59:16] is anything that you're currently doing related? [22:59:29] no nothings changed [22:59:40] Its not really related to what I'm currently doing [23:00:03] plenty of TMH ughs for everyone [23:00:59] the other day I merged https://gerrit.wikimedia.org/r/#/c/92052/ , it was pointed out that could cause a flash of unstyled content, so I went to fix that part, and no TMH is for some reason displaying dialog boxes with the wrong width [23:03:22] The more I investigate, the less things seem to work [23:03:38] stop investigating! :P [23:03:42] you're breaking the site! :P [23:03:51] j/k [23:04:39] wtf. Ok now (only thing I did was switch to an earlier commit in the branch), and videos are unclickable [23:04:47] I'm not kidding about this code being haunted [23:04:55] caching? [23:05:51] maybe, but they're only unclickable on previews, and none of the in between commits had anything to do with that [23:06:42] (03PS2) 10Ori.livneh: Add AUTHORS.txt [vagrant] - 10https://gerrit.wikimedia.org/r/107527 [23:07:22] Now all js on my local wiki is gone, but no errors in the js console... [23:10:12] ok, that one wasn't TMH's fault. mediawikis debug comment block somehow didn't get terminated and swalloed up all the