[17:42:31] nkomura: Done [17:42:41] thanks [18:17:56] adam_miller: RoanKattouw: good morning [18:18:06] good morning sir [18:18:21] Good morning [18:18:25] I was pretty badly misquoted on village pump... [18:18:27] *RoanKattouw looks at bug 22435, just came in [18:18:31] TrevorParscal: What's that? [18:18:52] brb [18:20:29] TrevorParscal: bug 22435 is a case of dirty diffs, I'll prioritize it [18:20:41] Oh wait it's at blocker :D can't beat taht [18:21:16] yes [18:23:12] nkomura: Seems like the staff meeting is starting late from what I hear. Can you ping me when I should call in? [18:23:34] it's done already [18:23:39] so we can start in 7 minutes [18:24:20] Oh OK [18:24:29] we should be on time [18:25:52] RoanKattouw: you working on the & dissapearance thing? [18:26:01] No I'm still catching up on e-mail and stuff [18:26:16] I've been mostly out of touch for the weekend and apparently I chose the wrong one :( [18:26:20] Or it chose me :) [18:26:25] ha ha [18:26:56] have you seen the watchlist notice on en.wp ? "some bugs were introduced by the usability team..." [18:33:23] hey adam_miller and RoanKattouw [18:33:30] will you call the conf number? [18:33:33] sure [18:33:39] i can't call two lines [18:51:54] nimish_g: Show Trevor http://pastebin.com/m58d07ec8 [18:53:15] http://pastebin.com/m58d07ec8 [18:55:18] TrevorParscal: Better patch http://pastebin.com/m4125bec2 [19:24:02] nkomura: Please test the newline eating bug on the deployment prototype to see if it's fixed there [19:43:44] selenium could be good for this kind of thing [19:44:09] we have a contractor who is working on a test suit for mediawiki. i'll poke them to put it into core. [19:44:20] Daniel_WMDE: That's why Ryan_Lane set up a Selenium server like last week [19:44:37] :P [19:45:26] Ryan_Lane, Roan-meeting: check out the selenium stuff in http://svn.wikimedia.org/viewvc/mediawiki/trunk/extensions/PagedTiffHandler/ [19:45:54] *Roan-meeting is not involved with this Selenium thing at all at this point [19:46:03] markus (mglaser) is happy about feedback/encouragement :) [19:47:23] nkomura: OK just to clear this up: is it OK if I change "Try Beta" to "Leave Beta" for partially opted-in people like right now? [19:47:56] Roan-meeting: hey [19:48:13] ok [19:48:16] good meeting [19:48:42] TrevorParscal: nkomura: OK just to clear this up: is it OK if I change "Try Beta" to "Leave Beta" for partially opted-in people like right now? [19:48:43] whoa, you can copy and paste the wikieditor INTO the wikieditor [19:48:48] http://etherpad.com/SNZzKGbmRD [19:48:48] which is kinda awesome looking [19:48:50] TrevorParscal: AKA http://pastebin.com/m4125bec2 [19:49:04] Cool [19:49:31] RoanKattouw: are templates still haunted? [19:49:39] RoanKattouw: do you mean before the full fix we discussed [19:49:50] just use etherpad for a while [19:50:05] we are going to setup an etherpad instance on one of our own servers soon [19:50:07] it's open source now [19:51:09] nimish_g: Yes, laptop battery empty on train, no chance to code :( [19:51:22] nimish_g: Once this release nightmare finally blows over, that's the first thing on my list [19:51:24] TrevorParscal: let me know how it goes, i want one too. or, actually - perhaps wmde can just use yours? [19:51:24] Purple person, name youself! [19:51:31] nkomura: Yes, like a temp fix right now [19:51:40] Daniel_WMDE: we shall see how loaded our server gets [19:51:44] i need to get some food [19:51:51] TrevorParscal: Purple person is me [19:51:54] :P [19:52:05] let's keep using the etherpad instance today [19:52:08] see how we like it [19:52:16] I think this will help use keep things on track today [19:52:20] Yeah [19:52:20] brbh [19:52:21] It's awesome [19:52:43] nkomura: Take a peek at http://etherpad.com/SNZzKGbmRD [19:53:15] TrevorParscal: a different idea on the copy/paste issue would be to capture the text prior to and after pasting and do a diff of the two strings to identify the bits that need reformatted [19:53:30] nkomura: Please approve/disapprove the items marked as (QUICK FIX) [19:53:33] adam_miller: yes, that was Roan/my original idea [19:53:44] but i think it's possible t [19:53:54] ah, sorry if i missed that [19:53:55] that nimish's idea will work / be reliable / be easy to do [19:53:59] no worries [19:54:19] TrevorParscal: That was your idea actually, I'm more in favor of the class tagging thing [19:54:23] RoanKattouw: i'm reviewing it right now [19:54:34] Nimish is typing out some options on the Etherpad [19:54:57] RoanKattouw: i am in favor of that too, I'm saying before that idea, that was my naive implementation concept [19:55:01] i'm going to get food [19:55:16] mee too! [19:55:17] brb [20:00:59] !log catrope synchronized php-1.5/wmf-config/CommonSettings.php 'Track NTOC and dialogs as well in PrefStats' [20:02:16] *RoanKattouw <3 Etherpad [20:04:56] wow, that got big quick [20:05:05] i'm back from scavenging for food [20:15:54] RoanKattouw: let me know when you stage the "Leave Beta" change on prototypes and test [20:16:02] Will do [20:25:46] *RoanKattouw just closed some tabs to go from 65 to 11 [20:27:19] yay :) [20:27:35] nkomura: Staged "Leave Beta" change on deployment prototype [20:27:56] *RoanKattouw makes a config change so it'll work properly [20:29:38] RoanKattouw: hey [20:29:47] let's stage it first [20:29:54] I meant a config change on deployment-en [20:29:59] oh good [20:30:19] I'd forgotten to actually add usenavigabletoc and usebetatoolbar-cgd to $wgOptInPrefs there; this already happened on the cluster [20:32:15] TrevorParscal: When people enable a boolean preference that's off by default, then disable it again, it leaves *no traces* in the user_properties table. We have prefstats for our new prefs now though and Naoko says it's not problem to potentially screw the few people it missed [20:32:50] nkomura: I tested this by messing around with my experimental preferences and looking at the try/leave beta link, and it seems to work for me [20:34:50] can i test it on deployment now? [20:34:53] Yes [20:35:00] let me check [20:35:11] TrevorParscal, nimish_g, adam_miller: What are you guys working on? Can you indicate this in the EtherPad please? [20:35:20] could i get a link to deployment [20:36:08] http://prototype.wikimedia.org/d-en/index.php?title=Special:Preferences [20:36:55] hi [20:37:24] RoanKattouw: that's not a link to deployment... [20:37:42] hi werdna [20:39:51] RoanKattouw: see any problems with being this specific with my tab fix for safari? http://pastebin.com/m6d8b474f [20:43:25] Looks fine to me [20:43:48] Except that it doesn't respect tabindex at all; best make it Safari-specific I tihnk [20:47:56] RoanKattouw: syntax highlighting is back on [20:48:06] we do not want to make it visible [20:48:17] nkomura: Turning that off on deployment now [20:48:23] Not an issue on the cluster [20:48:43] Done [20:49:11] still testing [20:52:53] TrevorParscal, nimish_g, adam_miller: What are you guys currently working on? Can you indicate this in the EtherPad please? [20:54:12] opting in looks good [20:54:15] testing opting out [20:57:15] opting out looks good too [20:57:53] RoanKattouw: TrevorParscal and nimish_g are in the brown bag session [20:58:16] Ah [21:02:58] was there code change for opt-in/opt-out? [21:03:20] What do you mean? [21:03:39] I did change the code on the deployment prototype to accomplish this, obviously [21:04:28] trying to see if code review is needed [21:04:39] as it is more than config change [21:04:51] Changes @ http://pastebin.com/m25b13837 [21:04:54] Rather trivial [21:06:35] k, let's have it reviewed by trevor as soon as this meeting is over [21:06:43] here in the office i mean [21:07:39] Yeah [21:11:46] adam_miller: Apparently Chrome inserts a literal tab character when tab is pressed, we'll want to fix that as well [21:12:46] i'm writing some code to look for the next tabindex and focus on that [21:13:00] so it can be used for all browsers [21:13:12] Cool [21:13:30] Take care taht there can also be elements with the same tabindex that are after the iframe in the document order, that's also legal [21:13:59] So has a tab order of foo-bar-whee-baz [21:14:19] yep, got it [21:19:40] RoanKattouw: skin system rewrite? [21:19:42] ugh [21:19:53] seriously, that's not a very good proposal [21:20:07] I think we could do better than that... [21:20:08] It's supposed to be back-compatible [21:20:16] And I haven't looked at any of the details at all [21:21:40] ok [21:21:53] interesting, you have to reconnect to etherpad after a while of inactivity [21:22:36] RoanKattouw: what was that trick to grab wikieditor's context from the console? [21:22:58] var context=$j('textarea').data('wikiEditor-context') [21:23:04] context.fn.getContents(); [21:23:12] thanks [21:29:54] RoanKattouw: how many of these revs need review? http://www.mediawiki.org/wiki/Special:Code/MediaWiki/author/catrope [21:30:04] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/61958 and up? [21:30:13] dang you IRC client [21:30:21] 61958 and up? [21:30:22] Define need review? [21:30:39] You mean like revs that are already deployed don't need review? [21:30:40] well, Naoko was hoping I could take a look at some of the revs we are proposing to push [21:30:48] you pushed them already? [21:30:56] hmm [21:31:03] Everything in that list was deployed on Thursday, look at the rev timestamps [21:31:13] oh [21:31:15] yes [21:31:24] But she wanted you to review http://pastebin.com/m25b13837 [21:32:06] ah! [21:32:12] i thought that was allot of revs [21:32:18] sorry, I don't know what day it is aparently [21:32:26] ok, this code looks good [21:32:46] did you run it anywhere public yet? [21:34:03] Deployment prototype, Naoko and I both tested it [21:36:22] TrevorParscal, nkomura: So I wanted to push this fix as soon as you're both OK with it [21:37:49] i'm ok with it [21:38:55] *RoanKattouw looks at nkomura [21:49:20] hi [21:49:23] i'm back [21:49:42] hi nkomura [21:50:59] nkomura: So you OK with me pushing that change? [21:51:49] so the code review checked out? [21:52:02] TrevorParscal said it looked good [21:52:14] k [21:52:16] *RoanKattouw points to scrollback, like 8 lines up [21:53:36] have you all tried from the user interface too? [21:54:22] I have, I think you have as well [21:54:35] i'm testing one more time [22:00:22] it looks good to me [22:00:31] let's deploy [22:00:47] OK here we go [22:01:05] wheeee [22:02:44] just committed a tabindex fix for safari/chrome yo [22:02:55]

catrope * r62140 /trunk/extensions/UsabilityInitiative/OptIn/SpecialOptIn.php: OptIn: Show "Leave Beta" link if a user has ANY feature enabled (rather than only displaying it if they have ALL of them). Also don't consider the skin preference for this. [22:03:09] Cool I'll look at it in a sec [22:03:12] r62139 [22:04:08] adam_miller: Rather than slicing the list, why not just iterate through it? [22:04:39] adam_miller: Also a brief comment explaining why we're hijacking the tab key would be nice [22:04:54] probably a good idea [22:05:03] And we should probably check that Shift and Ctrl aren't pressed, because Ctrl+Tab and Ctrl+Shift+Tab have other effects in Firefox [22:06:58] OptIn fix deployed [22:07:13] *nkomura goes to look [22:19:20] looks good [22:19:39] i pointed out a user case for users who chose Vector without beta features [22:20:03] if this user chose to opt-in and opt-out s/he loses Vector [22:20:08] we do not need to worry about it now [22:20:14] but something to keep in minda [22:20:15] mind [22:21:48] nkomura: anyone who figured out how to set it once can figure out how to set it again [22:22:44] yeah, but it is not a nice experience to be overridden though [22:25:01] Yeah that's a good one [22:25:41] Lots of design flaws in the OptIn system; I have some ideas for a redesign but we're only gonna use it for a few more weeks so I'll do that redesign if and when it's used again in the future [22:27:46] how do other bug fixes look so far? [22:31:27] I'm still poking at the

s getting eaten [22:31:33] Setting up an escaping system for that but it's tricky [22:31:56] has tabbing bug been resolved? [22:32:10] Adam has a fix for it I think [22:32:25] But I had some comments on that [22:34:59] adam_miller: are these the fix? [22:35:00] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/62139 [22:35:07] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/62142 [22:35:56] for safari and chrome, yes...i'm adding exclusions for modifier keys now though [22:36:00] so there will be one more commit [22:36:24] k [22:36:39] how about Firefox? [22:39:01] RoanKattouw: so you could reproduce the problem reported with Bug 22401 right? [22:39:31] No [22:39:33] Haven't looked at it [22:39:43] Read my comment on that bug [22:39:51] k [22:40:25] adam_miller: Use .size() instead of .length [22:40:57] And I recommend .attr('id') == 'foo' over .is('#' + foo), the latter pretty much has to be slower [22:42:05] RoanKattouw: are they not deprecating .size() ? [22:42:35] Huh yeah you're right [22:42:44] I was kinda paranoid here but the jQuery docs say .length will work [22:42:49] .length should work [22:42:58] adam_miller: So ignore my .size() vs. .length comment [22:48:05] TrevorParscal: could you reproduce this bug? [22:48:06] https://bugzilla.wikimedia.org/show_bug.cgi?id=22401 [22:48:16] disappearing line breaks? [22:48:47] i can reproduce it easily on my ubuntu easily [22:50:32] this one is annoying me quite a bit :) [22:52:35] catrope * r62143 /trunk/extensions/UsabilityInitiative/ (4 files in 3 dirs): UsabilityInitiative: (bug 22435) Stuff like

and in the original wikitext gets eaten. Using a proper escaping scheme for this now [22:53:39] \o/ [22:54:14] TrevorParscal, nkomura: If we can get this speedy-reviewed, I can speedy-deploy it [22:55:30] have you cross-browser tested? [22:55:55] the problem is only existent with FF3 AFAIK [22:56:02] so we don't want to break other browsers [22:56:07] No, this is a different bug [22:56:22] I do need to cross-browser test though, will do in a sec [23:00:29] works for me [23:00:38] nkomura: works for me on FF [23:00:40] 3.6 [23:00:42] coming over [23:06:24] three line breaks are translated into two [23:06:39] ok, i'm starting to see it [23:07:19] it happens only when you have

[23:07:35] not just
by itself [23:08:54] but I can only reproduce on en.wikipedia [23:08:57] my localh [23:09:01] localhost is working [23:09:37] *guillom plays with his jumping cursor [23:09:43] (twice today already) [23:10:15] nimish_g: what is the bug id for the jumping cursor? [23:10:23] *werdna pokes guillom's cursor in a surprising way [23:10:29] that's why it's jumping [23:11:03] hm, I'll check [23:11:23] RoanKattouw: whatever newline handling code that hasn't been deployed yet is solving the collapsing issue [23:11:45] I can repoduce https://bugzilla.wikimedia.org/show_bug.cgi?id=22401 on deployment, but not trunk [23:12:59] Interesting [23:13:19] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/62143 [23:13:27] I think that fixed it [23:14:20] Wait that fixed the *newline* issues? [23:14:33] I'm fixing up r62143 now BTW, it's incomplete on IE7 [23:14:33] yes [23:14:35] verifying [23:14:50] hmm [23:14:57] no, even before that things are working [23:16:13] WOOT I can repeatedly reproduce the jumping cursor bug! [23:16:47] what can I do to help you guys fix it? [23:16:57] TrevorParscal: What probably happened is that you deployed your "fix" reverting your earlier fix but didn't bump your style version [23:17:01] s/your/the [23:17:07] ooooooh [23:17:09] hmmm [23:17:19] The Varnish server running bits in Tampa requires that $wgStyleVersion be raised [23:17:34] So you deployed a rev that broke stuff, then luckily didn't push it out properly [23:18:00] well, I'm trying to see how far back I have to svn up to be able to reproduce on my localhost [23:18:18] I think it's the rev where you kill the Firefox-specific code [23:19:02] is there any way I can somehow record what's happening in the iFrame, so that I can record what's happening when my cursor jumps back? [23:19:23] now that I've found a way to reproduce it [23:19:33] nkomura: I can't find it or it's not explicitly labelled. I'll create a new explicit one for it [23:19:46] thanks [23:20:06] guillom: could you just list steps to replicate what you're seeing? [23:20:43] *TrevorParscal was doing svn up -r on the wrong dir [23:20:45] adam_miller, yes, I could, but since it's happening on a private wiki I should probably try to reproduce it elsewhere first; gimme a minute [23:22:08] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/62041 [23:22:15] that rev is clearly not deployed [23:22:20] and yes, it did include style bumps [23:22:28] can we verify that it is or is not? [23:22:29] It is deployed [23:22:39] But it needs $wgStyleVersion to be bumped as well, and that didn't happen [23:22:48] hmmm [23:22:50] So it'll be deployed for Europe only [23:22:50] ok [23:22:57] Because of the crazy way that bits works currently [23:23:11] then indeed, you are correct [23:23:16] can we bump it? [23:23:41] Yeah but you said you could reproduce on deployment but not on production. Are you sure this won't actually *break* stuff rather than fix stuff? [23:24:00] (I'll do it anyway if you want, just pointing this out) [23:24:15] I can only repoduce this problem pre r62041 on my localhost [23:24:39] OK bumping then [23:26:04] (BTW I'm really bumping the appendix to $wgStyleVersion) [23:26:08] Done [23:27:12] shall i check now or does it take time to propagate? [23:27:28] looks like that did it [23:28:14] ssssslow [23:28:44] yay [23:28:55] line breaks is working [23:29:00] guillom: will you try? [23:29:12] ok [23:29:14] Cool [23:29:32] I'm writing how to reproduce the jumping cursor bug right now [23:29:46] can anyone verify with IE [23:29:53] nkomura: So what happened is Trevor's Friday afternoon fix was good, but because of recent infrastructure changes something else needed to be done to push it out to non-European users and Andrew didn't know that [23:29:55] i don't have my evil computer with me today [23:30:17] RoanKattouw: we should get that fixed [23:30:28] inconsistent display is bad when it comes to testing [23:30:31] Yes [23:30:51] I think it's a feature though (more aggressive caching), but then it should be deployed at knams as well [23:30:57] I'll send a private-l e-mail about this [23:31:36] bits.wikimedia.org - cool idea with a rushed and poorly tested execution [23:32:28] Not really [23:32:45] two bugs down! [23:32:47] What other issues were there apart from this (fairly obscure) one? [23:33:11] RoanKattouw: how much longer will you be available [23:33:23] ? [23:33:44] it'd be good if we can push out the fix for tabbing problem [23:33:49] the jumping cursor issue seems to be a combination of cut/paste and header issue [23:33:53] RoanKattouw: michael had something break on him, it's been really latent for allot of us, and Tomasz mentioned other things, he may be able to provide more info on request [23:33:57] almost done with the description [23:34:42] nkomura: I've also now fixed the

eating problem on IE as well [23:35:37] I'll be around for a while [23:35:42] nkomura: did we ever reproduce this? https://bugzilla.wikimedia.org/show_bug.cgi?id=22375 [23:35:48] I can't [23:38:22] TrevorParscal: that was on prototype [23:38:29] so it was before we went live [23:38:38] you can close it [23:38:49] nkomura: Also that's not necessarily representative, use the deployment prototypes [23:39:00] RoanKattouw: is there bug case for

eating bug by IE [23:39:41] nkomura: Just type

or into an article [23:39:47] And try to save it [23:39:58] i don't have IE with me today [23:40:02] On production it'll currently produce diffs like https://bugzilla.wikimedia.org/show_bug.cgi?id=22435 [23:40:04] This is on all browsers [23:40:16] (diff linked on bug report) [23:40:47] TrevorParscal: will you review RoanKattouw's code [23:40:51] ? [23:41:07] TrevorParscal: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/62413 , followup linked [23:41:08] RoanKattouw: cross-browser testing completed? [23:41:16] Firefox and IE, best I can do [23:41:32] i can do the rest [23:41:34] ok, I sent an e-mail on how to reproduce the jumping cursor issue [23:41:53] OK I'll have another look at Adam's tabbing fix, IIRC it was mostly OK [23:41:57] i'm fine with proceeding to deploy, if cross-browser testing and code review are completed [23:42:09] RoanKattouw: what was your test case? [23:42:31] loking at https://bugzilla.wikimedia.org/show_bug.cgi?id=22435 [23:42:59] TrevorParscal: Literal

or in the article --> dirty diffs on all browsers [23:43:45] Marked Adam's revs as OK [23:44:21] testing in chrome [23:47:13] looks good in opera, chrome, safari, and firefox [23:47:20] and you tested in IE [23:47:27] Yes [23:47:44] Could you test Adam's tab fix while you're at it? (If this isn't done already) [23:52:55] parutron, ? [23:52:59] hi [23:53:04] hi guillom [23:53:11] you're not using firefox on ubuntu, are you? [23:53:21] nope. firefox on a mac. [23:53:26] okay :) thanks [23:53:51] i wrote having no idea what the extent of the jumping cursor issue is....! [23:54:09] not sure if it helps, but whatever. at least now i know what you have been doing last week! [23:54:10] ;) [23:54:21] I have Firefox on Ubuntu, I'll try to reproduce in a sec [23:54:47] parutron, h??h?? ; apparently it happens only on firefox/ubuntu ; basically, the cursor jumps a few characters back [23:54:56] straaaaaaaange. [23:55:06] RoanKattouw: so, the old toolbar disappearing / encapsulation bridge thing... [23:55:14] Yes... [23:55:25] where are you on that? [23:55:32] Nowhere just yet [23:55:34] not / poking / done / etc [23:55:35] ok [23:55:41] I'm going to investigate it [23:56:19] Should be rather easy, figure out which function the old toolbar uses (skins/common/edit.js probably) and hijack it somehow [23:56:57] nimish_g: what's the bug ID for the jumping cursor? [23:57:23] i'll be in the meeting in 4min [23:57:40] i delegate deployment decision of patches to TrevorParscal [23:57:58] 20922 [23:57:59] *TrevorParscal is drunk with power and will deploy JS2 [23:58:00] guillom: Reproduced cursor moving successfully, moves back one char for me [23:58:10] \o/ [23:58:16] *RoanKattouw rats TrevorParscal out to Tim [23:58:24] in some other cases it jumps back a few characters [23:58:49] but it may be cumulative [23:58:54] wait nkomura [23:59:07] 22440 is the one I created, the other one was the original bug I thought was related but doesn't seem to be