[06:50:22] http://commons.wikimedia.org/wiki/File:Domas-the-Giant-has-a-Posse.png [17:09:56] Bah, x2003 is taken again [17:10:08] alolita: I'm in x2004 [17:14:23] https://secure.wikimedia.org/wikipedia/mediawiki/wiki/Extension:ArticleFeedback/Ramp-up_plan [17:25:33] API's should not try to do anything to $wgOut, am i correct? [17:25:58] Not usually, no [17:26:09] What were you gonna be doing to $wgOut then? [17:27:06] actually i was trying to segregate the account creation coding from Special:UserLogin into APISignup [17:28:25] wanted the verification stuff to go into the API [17:28:38] one method for verifying each field [17:29:17] the current methods in Special:UserLogin could then continue to do changes to $wgOut [17:29:47] Yeah it's probably mixing validation/DB/etc logic with output/UI stuff, right? [17:30:00] Like if ( something is wrong ) { show error message } [17:30:17] As opposed to if ( something is wrong ) { return an error that a caller could do something with } [17:30:24] Then if the caller is the UI, it can show the same old error message [17:30:30] And if the caller is the API, it can do its own thing [17:31:28] RoanKattouw: howdy [17:31:39] UserLogin returns error codes [17:31:41] Morning [17:32:07] which are then mapped to messages in the UI and shown as-is in the API [17:32:14] That's nice [17:32:30] RoanKattouw: http://svn.wikimedia.org/viewvc/mediawiki/trunk/parsers/wikidom/ front-end black magic (see the surface demo) [17:33:06] *RoanKattouw runs svn up ~/mediawiki/trunk and goes to the bathroom [17:33:13] It might still be running when I get back [17:33:59] <^demon> Platonides: That desperately needs refactoring out of SpecialPage [17:34:49] Platonides: i am somewhat unclear of the interaction between SpecialPage and the API [17:35:15] Platonides: could you give me a rough idea? [17:35:47] akshayagarwal, let me look at it [17:36:16] Platonides: ok [17:38:02] akshayagarwal, the login work is done in LoginForm::authenticateUserData [17:38:10] which returns a status code [17:38:26] the UI maps that to messages in processLogin() [17:38:45] API login creates a login form, calls authenticateUserData() and deals with the return [17:39:28] if you want to create new accounts [17:39:39] you would call instead addNewAccount() or addNewAccountMailPassword() [17:40:35] TrevorParscal: Oh, the selection stuff is pretty sweet. Is that what the demo is about? (I.e. select stuff on the left, mirrored on the right) [17:40:55] yeah, well the word wrap and text selection are completely synthetic [17:40:59] look at the dom [17:41:39] Oh wow you're right [17:41:43] Even the selection markers are synthetic [17:41:44] Platonides: i have noted down all your points, i ll go through the code again & check my understanding [17:41:57] Feature request: double-click selection interaction ;) [17:42:06] Platonides: thanks :) [17:42:38] you are welcome, akshayagarwal :) [17:42:52] I have no idea who you are, btw [17:43:00] RoanKattouw: it's the beginnings of the visual editor [17:43:03] TrevorParscal: What's the blue selection, an absolutely positioned element behind the text or somethign? [17:43:13] yes [17:43:17] This looks pretty sweet [17:43:20] :) [17:43:26] Oh, -range, I see [17:43:35] Platonides: i am a GSoC student working on Signup related stuff [17:43:45] And it creates one -range diff per line, then keeps them [17:44:26] TrevorParscal: The long word doesn't /quite/ work for me though. I see "WrappingBetw \n MultipleLines" [17:44:32] so 'een' gets cut off [17:45:16] yeah, there's some bugs still [17:45:28] Might be browser-specific, I don't know [17:45:41] it is, you are on Firefox i would guess? [17:46:01] Yes, FF4.0.1 [17:46:05] the black magic of divining pixel positions of each character's boundaries without rendering each character in it's own span is a little tricky [17:46:08] You'd be testing on Chrome presumably [17:46:29] yeah, but I also check in firefox - it's also dependent somewhat on your operating system [17:46:39] since the text rendering for firefox is delegated [17:46:40] Yeah especially font-wise [17:46:45] Exactly [17:47:02] Typical Linux chars are 13x8 I think [17:47:09] In Mac OS, I /think/ it's 14x10 [17:47:18] kerning is what causes the issues [17:47:19] *RoanKattouw consults the hacky JS code that uses these numbers [17:47:29] Oh gosh that's right [17:47:31] so I need to use a combination of character measurements and word measurements [17:47:35] that will correct it [17:47:38] What I just said only applies to monospace [17:47:59] Grmpfs, nl.wiki has a UTCLiveLock that doesn't indicate UTC but their local timezone. Blaming me for messing up the "UTC time" when I replaced it with a call to the one hosted on mediawiki.org [17:48:01] Oops [17:48:12] haha [17:48:44] ohwell, mediawiki.org introduces " Gadget:LocalLiveClock :-) [17:50:48] RoanKattouw: The autoEllipsis testcase is failing in IE6. (the one that checks what happends when one more character is added) [17:51:09] Ugh [17:51:18] Well, it's IE6 [17:51:19] I'm not sure whether it's the test case is too strict or whether it doesn't work properly in IE6 at all (perhaps the testcase should add 3 characters ?) [17:51:40] Will you accept "screw IE6" as a response? :) [17:51:59] Can we get erik to fudge the squid report cards? [17:52:00] Well, the testcase needs to not fail. Otherwise the red column in test swarm is useless and can't be used to know when something does brake badly [17:52:07] Hmm I guess [17:52:17] Krinkle: the autoEllipsis in IE6 will likely never work [17:52:17] I have no idea why it's failing. I don't have IE6 handy [17:52:23] let me look [17:52:32] And I have many, MANY other things to do that are more important than fixing IE6 issues :) [17:52:35] http://toolserver.org/~krinkle/testswarm/?state=runresults&run_id=497&client_id=1202 [17:53:37] ok, team meeting [17:55:04] http://etherpad.wikimedia.org/FeaturesTeam20110607 [17:55:07] RoanKattouw, http://eiximenis.wikimedia.org/FeaturesTeam20110607 [17:55:13] ah, timo beat me to it [18:09:00] Errr, that was quick [18:09:04] not that I'm complaining [18:09:31] but considering the amount of stuff in the pad we didn't talk about, should we just assume nobody's working on those? [18:10:42] Or they're not around [18:10:48] Ah, right. [18:10:49] (UploadWizard, TMH) [18:15:28] TrevorParscal: Looks like pref-section editing/beta is locally created by WikiEditor (i18n for the message as well). Any gotcha's integrating it into core ? Would make it easier for ArticleFeedback to place it's preference. As well as for WikiLove (which I think is currently not working properly either on wikis without wikieditor/vector) [18:15:56] 'section' => 'editing/labs', [18:16:02] well - I think the AFT thing shouldn't be in editing [18:16:33] Appearance / Advanced Options [18:16:51] right, because it's about hiding/showing the gadget. [18:16:52] just under "Exclude me from feature experiments" [18:17:07] That one doesn't exist in core. [18:17:38] well, just somewhere in that section is good [18:17:44] sure [18:20:21] neilk_: you around? [18:20:36] *TrevorParscal looks for neil [18:22:38] TrevorParscal: I've looked at some other extensions and basically there's two ways: Make it a "show" preference that is prefdefaults to 'on', or make it a Hide preference that defaults to 'off'. The latter seems more common lately, I'm not sure if that's good tho. [18:23:08] What do you think ? [18:23:25] it should be shown by default, and not checked - so it should be a hide preference [18:45:23] Guys. I think I'm seeing the end of the universe with this: http://en.wikipedia.org/wiki/Web_page [18:46:39] hahaha [18:46:40] Nice [18:46:45] It only let's me click so far [18:46:47] It's broken [18:58:26] RoanKattouw: Hm.. do you know a good hook/place to undo addModule actions ? ie. right before they're loaded ? [18:58:43] I'm looking to find a way to integrate the static index.html qunit thing into a SpecialPage [18:58:58] Ah, what do you want to do exactly? [19:00:01] After all addModules by core and extensions took place, undo them all and load A) addModule( array of moduls that have a test suite ), B) the qunit .test.js files [19:00:29] You would have to write support for that in OutputPage [19:00:32] ie. no gadgets, no user scripts, no extension modules etc. [19:00:45] Cause OutputPage manages the modules array [19:56:10] alolita: So how did you want to do this meeting with Ryan? Phone or IRC? [19:56:20] yes - [19:56:27] let me pull Ryan in [19:56:31] chat with you on the phone [19:56:57] *RoanKattouw calls into x2003 [19:58:53] RoanKattouw https://secure.wikimedia.org/wikipedia/mediawiki/wiki/Extension:ArticleFeedback/Ramp-up_plan [20:27:56] Stuff like this: wfErrorLog( $headers, 'udp://10.0.5.8:8420/vandalism' ); [20:40:19] "World IPv6 Day begins 24 hours from now." [20:41:36] ? Isn't it tomorrow? [20:43:19] Yes, I'm a little behind on my RSS reader [20:43:41] Ah [20:43:55] we are not joining, afaik. [20:44:28] Noope [20:45:07] No, we're not [20:53:03] we aren't joining because our database schema isn't updated to handle ipv6 addresses :( [20:53:27] otherwise we probably would have knocked out the remaining work last night [20:53:42] doing the schema updates was just too much ;) [20:54:11] and it begins in a few hours [23:56:19] hi TimStarling: what's left on the 1.17 release? [23:57:58] very little, I'm just going to double-check these backports that Roan and Chad did over the last 24 hours, then I guess I can do an RC