[00:14:52] the sue banner looks awesome. [00:15:15] and i gotta drop props to kaldari on the dynamic thermometer. [00:15:35] what's the betting i can't view them.. [00:16:23] no idea. [00:18:42] Reedy: http://meta.wikimedia.org/w/index.php?title=Special:CentralNotice&method=listNoticeDetail¬ice=Fundraiser-en-US [00:19:39] ahh [00:20:18] you'll need to go here to get the letters: http://wikimediafoundation.org/w/index.php?title=Special:PrefixIndex/WMF [00:28:32] are you getting UK chapter ones or something? [00:29:42] i only ever see jimmy [00:36:13] yeah, we haven't got any of the editor appeals [00:36:20] just plain old Jimmy [00:36:52] I bet Liam's itching to have his ugly mug up in Australia :() [00:36:54] :)* [00:37:12] well, sue's banner is awesome. [00:37:25] yeah, it is [00:38:33] *aude reads on twitter "Wikipedia finally uses sex instead of Jimbo to solicit donations. The result: I donate. You should too (cos #Wikipedia rocks). #keepitfree" [00:39:21] *werdna wonders what that refers to [00:39:45] *aude doesn't know but they weren't seeing jimbo any more [00:40:35] Mayo was on a few days ago [00:41:28] we've had a bunch of different banners... kartika, now sue [00:44:20] i don't remember seeing kartika. [00:45:03] oh, yeah i do. [00:45:24] *aude is on the wikipedia social media team [00:45:59] all sorts of interesting feedback [00:46:05] via twitter [00:46:06] where's the branch point for 1.17? [00:46:14] anyway... [00:46:20] aude: I really really want the good banners in Australia :( [00:46:38] sounds like it's up to the chapter to put different ones up? [00:46:42] *aude doesn't understand [00:46:43] werdna, http://www.mediawiki.org/wiki/Special:Code/MediaWiki/77974 [00:46:58] Reedy: thanks [00:47:48] ugh: "wikipedia supporters strike out iwth ddos attack, why didint we listen to the personal appeal from jimmy whales" [00:50:19] I really cannot get into the work groove today. Going to do some code review, I think [00:58:15] werdna, qhttp://www.mediawiki.org/w/index.php?title=Special:Code/MediaWiki/status/new&dir=prev [00:58:18] 1240 to go ;) [00:59:43] I'm working on it :) [18:54:52] hi RoanKattouw [18:54:57] how's it going? [18:55:12] having fun yet? [18:55:18] heh [18:55:33] did you see me replies to the sessionId thing? [18:55:36] I just remembered I hadn't read Twitter/identi.ca in three weeks, and I got bored of CR temporarily :D [18:55:38] Yes [18:55:43] Well one [18:55:48] I think there are some interesting points, i will probably do another rev soon [18:55:55] but I need more input [18:56:12] Oh and I'm getting more frustrated with my laptop by the day [18:56:35] Maybe I should buy a new one earlier than I anticipated and get one with longer battery life in time for my next trip [18:56:35] I was kind of ammused by the "OMG are we even allowed to do this?" sort of tone that... what's his name had - I'm thinking "uh, well we already do.. hmmm" [18:56:56] dude, you should be running the best hardware available [18:56:58] really [18:57:08] People seem to get nervous about the word cookie every time [18:57:15] But sometimes they have valid points [18:57:44] Dude when I bought this, it came with 1GB RAM. I upgraded it to 2GB back when I used VirtualBox a lot [18:57:51] But still... [18:58:32] woah, cookies! [18:58:44] Om nom nom [18:59:01] do the cookies only get set if the page is running articlefeedback, clicktracking, etc. [18:59:07] or on all pages for anons? [18:59:17] ha ha - we could rename the $.cookie plugin $.unicorn [18:59:32] maybe people would be more happy to see unicorns on wikipedia than cookies [18:59:39] hah! [19:00:06] damn wireless [19:00:18] Roan probably answered already, but aude: it's only if it's used [19:00:24] no need to set cookies we don't need [19:00:47] The implementation is on-demand [19:00:53] as it should be [19:01:01] that's okay, i suppose [19:01:08] accessing mediaWiki.user.sessionId() will set a cookie if it's not already there [19:01:09] So only when some code asks what the current session ID is do we check for the cookie and create it if not present [19:01:22] Except for registered users, we'll just use their username as sessionID [19:01:53] right [19:53:56] RoanKattouw_away: Is there more to be done in your opinion on mw.util ? I have implemented all ideas I initially had. Now what :) ? [20:10:19] hey folks, i had a query from within google... is there anything we can do in the way of sponsoring a new server or something to bring the wikipedia dumps back? [20:11:41] an excellent question [20:12:03] (i presume it's a more complex problem than just "the server died"... but thought i'd ask anyway) [20:12:32] what's the actual cause of the outage, if you don't mind me asking? [20:14:03] Skud: download.wikimedia.org has a link to a page describing in detail the errors and progress of repair [20:14:16] i'm not sure why it wasn't just swapped to another machine weeks ago though [20:15:45] is ariel the person to talk to? [20:19:33] and, i've forgotten ariel's nick. what is it again? [20:30:36] ariel usually going as apergos atm [20:30:43] not in this channel just now, but elsewhere [20:31:12] Skud: in general looks like the story is "bought a single piece of hardware, it broke, nobody got around to setting up alternate plans other than wait until it's fixed/replaced" [20:31:28] so far nobody's entirely sure why we only bought one [20:40:16] TrevorParscal: RE: https://bugzilla.wikimedia.org/show_bug.cgi?id=25486 - by define I didn't mean defining the message content (ie. int:foo is 'bar') but defining which messages are used (ie. needed-msgs: { 'recentchanges', 'pagehistory', 'talkpagelinktxt' }) [20:40:57] Krinkle: that's in interesting idea [20:41:32] what we need is a way to say "load these messages" [20:41:47] wrapping that into a using call would be great too [20:42:06] so I could do mw.loader.using( 'jquery.ui', function(){ /* do stuff */ }, { 'message-key', 'message-key', 'message-key' }); [20:42:19] but also mw.loader.register() should support using messages. [20:42:42] or, instead of a string or array of string for the module, it could also take an object [20:43:04] which can contain 'modules': [list of modules] as well as other things like 'messages': [list of message keys] [20:43:19] yeah, more verbose and understandable [20:43:52] two paramters (or three including the error-callback) [20:44:29] but yeah - it should be used sparingly, like if you have a module that you've written in an extension or core, and you know you will need some messages, it's much faster to register them on the server performance wise, and you get the added bonus of invalidation based on message updates [20:44:32] TrevorParscal: but that would be a breaking change, should be keep the string-way/array-way as backward compat and short-hand easy of use ? [20:44:45] TrevorParscal: Sure, but I mean for user scripts and gadgets. [20:44:46] there really should be an error callback if there isn't one... [20:44:51] There is [20:44:58] Krinkle: yes - for userscripts, it would be awesome [20:45:00] currenty its (dependencies, success, error) [20:45:15] I think we can keep the old ways around [20:45:32] that could become (using, success, error) where using can be a module-stringname, module-array or object with modules and messages [20:45:37] it's easy to know if something is a string or an array ($.isArray) or an object [20:45:50] yup [20:45:56] TrevorParscal: Am I correct in the assumption that loader.load() does not make a server request untill go() ? [20:46:01] and then we can add more features later in the object calling style [20:46:07] yes [20:46:08] ie. one can queue things ? [20:46:11] Nice [20:46:16] and after go it does immediately [20:46:28] request batching is turned off when go() is called [20:46:44] might be interesting to make it possible to do stop() as well [20:47:07] go() can only be used once ? ie. loader.load()'s after go() make server request directly ? [20:47:10] which could temporarily allow you to build up a request then do go again [20:47:34] right now it's only being used once, and it would (should) have no effect if used again [20:47:49] but if we had a stop method that could change [20:47:57] something to add when we need it perhaps [20:47:58] Hm.. I think that's a little illogical, I think load() should be a queue at all times and go() simply empties the queue(), so if one needs only one thing, one does load() and go() [20:48:04] or just add in now for completeness [20:48:38] the problem there is that you might forget to call go() and wonder what's going on [20:48:49] and it adds lots of go() calls that would otherwise not be needed [20:48:49] I dont consider that a problem. [20:48:59] it does save resources [20:49:09] and the stop(), add things, go() model is more exotic [20:49:19] that way gadgets and big user scripts can load several things (userscripts are read after the first go() ) and do a go() [20:49:41] so without stop() after go() things are loaded directly. [20:49:43] we should support it, but not make it the default behavior, because it's a more advanced functionality [20:49:47] yes [20:49:49] k [20:49:55] I'll open up a ticket to remember [20:50:01] cool [20:50:45] by the way - just wanted to say I'm really impressed with your contributions to our JS code - we are really glad to have you around [20:51:16] if there's ever anything you need help with, I'm always happy to do whatever I can [20:51:39] Thanks, that's good to hear! [20:55:56] TrevorParscal: The stop(), should I put the bug in 1.17 or 1.18-svn ? [20:57:08] 1.18 [20:57:35] without a specific use we shouldn't block 1.17 with it [20:58:01] but if it gets done before we deploy, it could always be backported pretty easily I think [20:58:34] Yeah. I'm looking at that javascript deprecation page of RL @ mw.org, I see they haven't all been ported yet. [20:58:47] That is, I guess, a to-fix bug for 1.17, right ? [20:59:24] I mean, they work without being ported, but I think it's a good goal to want to have a replacement for each mw.legacy module. [21:01:53] yeah, some of the functions too - as you are aware I'm sure - are not really meant for public use, they are functions used by other functions, or functions that just run on load [21:02:16] so, much of that work is just moving the code into a document ready closure [21:02:23] modernizing the dom manipulation [21:02:26] etc. [21:02:51] Yup [21:03:17] TrevorParscal: I think I'm going to add the new tablesorter this weekend, but what I dont quite get is what it has to do with bug 4740. [21:03:27] http://www.mediawiki.org/wiki/ResourceLoader/Status#Tasks [21:03:36] Table sorting: replace old sortable table code with http://tablesorter.com - requires resolving bug 4740 [21:04:47] oh [21:05:24] well, since tables don't put th elements in a thead element, existing table sorting code, such as tablesorter doesn't work [21:06:11] so, we either modify tablesorter to work with or without thead, tbody, tfoot - or we fix the parser to generate tables with thead, tbody, tfoot (not sure that last one has any syntax for it, but you get what I mean) [21:06:26] or we can just jQuery-plugin-ify the existing code [21:06:34] I don't think the existing code is evil [21:06:37] it just needs some love [21:07:57] Oh, the jquery.tablesorter plugin requires the TH to be in a Ah, that explains it. [21:08:22] I see mediawiki doesn't do that. Not sure it should by default either (could break other stuff) [21:08:31] there has been some discussion about that. [21:09:01] afaik, putting th inside thead will break nothing except css rules that were incorrectly written. [21:09:02] But I guess changing jquery.tablesorter to if (no-thead){ get first th and it's TR [21:09:17] now you're just talking crazy. [21:09:33] Yeah, I just realized that [21:09:35] :P [21:10:08] So we change the parser to output the first table row, if it cointains "!" or to wrap it in a [21:12:06] TrevorParscal: True, the current tablesorter by joostdevalk in wikibits isn't evil, but it does leave things to be wished for which jq.tablesorter has addressed [21:12:20] otoh, the ont from wikibits is very lightweight [21:12:41] which I like a lot [21:12:44] and we know it works [21:12:45] :) [21:17:31] dude, neil showed up [21:17:40] i guess that means he's feeling better... [21:24:24] wb flipzagging [21:24:30] I hear you're back in the office [21:24:56] RoanKattouw: yes [21:25:00] Krinkle: About mw.util: I'll look at that later, I'm on 1.17 full-time now [21:25:20] RoanKattouw: thanks [21:28:03] brion, Skud: What went on with the dumps server is we only had one box, no copies of the data AFAIK (no servers with comparable storage capacity) and then the RAID controller/backplane/whatever on the box broke so we can't access the data. Or something. (Disclaimer: this is just an uninvolved bystander's understanding of the situation) [22:02:41] werdna, about? [23:09:31] Reedy: about. [23:09:59] "MEDIA STUDIES STUDENTS. Stop protesting. Under the new 23k threshold, you will never have to pay back your fees. /via @PegnosePete" [23:10:17] Is that targetted at the brits? Or are you protesting too? :P [23:11:41] brits [23:11:56] It's just funny [23:11:59] *werdna waves pdhanda [23:12:17] sup werdna [23:12:29] nm, busy as usual [23:12:49] when are you visiting next? [23:12:55] werdna, it is. Oh, we want something, so, I don't know, let's see, riot? Yeah, that's a way to show the government [23:14:13] pdhanda: I get in Jan 14th :) [23:14:24] Reedy: Seems to be the British Way[TM] [23:14:40] They should be standing in a queue then, complaining about the weather [23:14:50] well, that's what kettling is for [23:14:59] the Brits couldn't bear to riot without a queue [23:15:02] so they added one [23:15:06] We're having student protests here too, much the same reason [23:15:49] hmm, flights to Paris these days are $125 cheaper than flights to London [23:15:56] sounds like a good case for flying to Paris and taking Eurostar [23:16:31] From? [23:16:36] sydney [23:16:39] Yeah Eurostar Paris-London is, what, EUR 35 youth rate? [23:16:47] eur 49 return youth [23:16:51] Right [23:16:55] Isn't that almost 100 AUD? [23:17:01] yeah [23:17:04] wait, no [23:17:09] it's 4/3 these days [23:17:13] Huh [23:17:14] so about $65 [23:17:23] Is there such a large difference between AUD and NZD then? [23:17:32] ~20% I think [23:17:40] 30% [23:17:41] werdna: cool! [23:17:50] Cause last I checked EUR:NZD was like 2:1 [23:17:52] pdhanda: I have a full three weeks too [23:17:56] But admittedly that was years ago [23:18:04] 1 Euro = 1.7686847 New Zealand dollars [23:18:10] 1 Euro = 1.34443087 Australian dollars [23:18:13] Huh [23:18:37] The NZD rate was stuck on 1 NZD = 0.49-0.51 EUR for like two years prior to the financial crisis [23:18:46] Oh well [23:19:08] I know this sort of question is perhaps a lost cause on you, but why were you watching the nzd/eur exchange rate? [23:19:11] :p [23:19:13] haha [23:19:21] I have an uncle down there [23:19:26] Moved like 20 years ago [23:19:41] So one day I was asked to look up the rate cause we have a small amount of NZ money [23:20:13] And... well I don't think I have to tell you the rest of the story [23:20:35] :) indeed not [23:30:34] RoanKattouw: incidentally, what did you think of that revision that stops the link colour betraying page existence? [23:30:37] I believe you looked it over? [23:30:53] I did [23:31:13] But it needs to be tested in unconventional skins like standard and modern [23:31:42] <^demon> Skins need to be checked in general. [23:31:46] <^demon> I think some of them are broken atm. [23:31:56] Well I'm talking about this specific fix [23:32:14] But yeah someone needs to go and verify all of them [23:35:34] <^demon> standard is broken. [23:35:54] Surprise surprise [23:36:01] <^demon> Myskin and simple seem to be ok [23:36:01] TrevorParscal: Hey what about that Monobook-to-RL thing? [23:36:26] i'm thinking we should back it out for now [23:36:32] <^demon> Cologneblue is also broken [23:36:40] did it get ported to RL? [23:36:44] Alright, I'll back it out tomorrow [23:36:47] cool [23:36:48] <^demon> Standard and cologneblue are broken in the same way, the left navbar is getting pushed below the content. [23:37:04] Are they both non-SkinTemplate skins? [23:37:27] <^demon> Yep, both extend Skin [23:38:01] <^demon> Nostalgia seems to be ok [23:38:43] <^demon> Yep, just cologneblue and standard it seems [23:39:23] Wondering whether our RL work added stuff to SkinTemplate [23:40:26] Hm, doesn't look suspicious [23:40:55] ^demon: Do you see any warnings in the load.php output? [23:41:07] <^demon> Lemme check [23:41:26] *RoanKattouw goes and checks for himself because he can [23:41:38] You should see any warnings in /* */ comments [23:42:30] <^demon> No, but the two load.php URIs are serving blank content. [23:42:54] <^demon> load.php?debug=true&lang=en&modules=mediawiki.legacy.oldshared|mediawiki.legacy.shared|mediawiki.legacy.wikiprintable&only=styles&skin=cologneblue [23:43:08] <^demon> I have E_ALL | E_STRICT [23:43:12] *^demon checks the error_log [23:43:50] Hmm [23:43:54] ul.gallery, li.gallerybox { [23:43:57] display: inline-block; [23:43:58] zoom: 1; [23:44:00] *display: inline; [23:44:02] } [23:44:04] In shared.css [23:44:35] Unrelated, sure, fixing [23:45:50] ^demon: & ? [23:46:11] Do the URLs work with s/&/&/ ? [23:46:45] <^demon> copy+pasting it from the HTML: [23:46:50] <^demon> [23:47:11] Yeah the browser should do & -> & conversion there [23:47:18] <^demon> Ah there we go. [23:47:31] <^demon> If I remove the amp;s it works [23:48:08] I suggest using Firebug or LiveHTTPHeaders to see which URL the browser hits [23:48:12] Should be the one without amp; [23:48:43] <^demon> I'm checking in chrome's dev tools now [23:49:52] <^demon> Seems to be fine, looks like a problem with chrome copy+pasting the URLs [23:51:34] UGH [23:51:43] The positioning for #quickbar is in the user-generated CSS [23:52:30] We have some kind of equivalent for that in RL [23:53:58] *^demon pukes [23:54:02] <^demon> I hate that quickbar pref. [23:54:05] <^demon> We should kill it [23:56:19] Ugh there's some horrible horrible handling of it [23:56:31] I'm watching a movie at 1am so I'm not gonna pike todaq [23:56:35] *poke today [23:57:30] <^demon> I wonder how many people have the quickbar pref enabled. [23:57:55] Let's find out [23:57:57] <^demon> Minus the people who have it turned on with a skin it doesn't work on [23:58:12] <^demon> Like if you have quickbar and vector, you're obviously not using the former anymore [23:59:52] <^demon> It only works on cologneblue and standard. [23:59:53] Ugh