[00:25:51] TrevorParscal - are you there? [00:25:58] yes [00:26:01] how's it going? [00:28:02] TrevorParscal: i'm not seeing the jsMsg popup for ajax watch/unwatch in vector. still works in monobook... [00:31:49] Hi! [00:31:57] I wanted to talk about template call stuff again. [00:32:08] brion: popup? [00:32:31] oh [00:32:32] yeah [00:32:34] hmm [00:32:36] hld [00:33:06] tx [00:33:12] howdy YaronK [00:33:12] it IS changing the dom [00:33:18] but it's not showing the div [00:33:20] Hi brion. [00:33:25] the display:none is stickin [00:33:29] looking into it [00:33:50] TrevorParscal: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/56902 ? [00:37:15] ok, the div is part of the HTML output in Vector and not in Monobook [00:37:30] the #mw-js-message div [00:37:49] so, the code that creates it automatically just skips over it [00:38:09] when the automatically created version got changed, they didn't take into account the possiblity that it already existed [00:39:04] bah [00:39:21] so I can either remove the inclusion of this div in the HTML output of Vector.php - resulting in the div being created and thus treated the same as Monobook, or change the JS to handle pre-existing divs of that ID [00:39:48] the former seems to make for a simple fix, the latter would fix other skins that still include the div [00:40:07] i say just revert r56902 [00:41:01] note - the div was added because this JS function didn't used to add the div automatically.... [00:41:04] r44025 [00:41:23] interesting [00:41:40] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/56924 <- is this all resolved at this point? [00:41:41] r44025 is where it appeared, and it looks like it's intended to fix Modern [00:42:10] brion: that's my refactor [00:42:33] er [00:42:40] that's not mine - yes [00:42:53] :) [00:43:04] I refactored that code in r57030 [00:43:11] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/57030 [00:44:15] so I think that the situation with r56902 removing what r44025 added seems like reverting r56902 doesn't break anything, and keeping it does [00:44:29] so I would second your motion to dismiss r56902 :) [00:44:50] i quadruple that motion. [00:45:07] yay [00:48:39] evil reverted :D [00:55:02] brion: FYI - there is one blocker bug reported https://bugzilla.wikimedia.org/show_bug.cgi?id=20669 [00:55:39] but it cannot be reproduced by us, so i don't think it will be a show-stopper [00:55:48] will send out confirmation to calcey tonight [00:55:55] high YaronK! [00:56:10] Hi! [00:56:27] It looks like everyone's there. [00:58:08] nkomura: did we get any useful feedback from calcey this time? [00:58:52] brion: yes, they have been reporting good bugs for the last two weeks [00:59:04] yay [00:59:22] they just ran another round against prototypes [00:59:46] and they will do another round against deployment branch tonight [01:01:06] i wouldn't bother with that, as deployment will be updated tomorrow [01:02:35] ok [01:10:37] TrevorParscal - do you have time to talk now? [01:12:32] YaronK: howdy [01:14:07] i scared him away [01:58:39] Oh, I got logged off somehow - only realized it now. [02:00:08] TrevorParscal - are you still there? :) [02:01:25] yes [02:01:27] sadly [02:01:31] waiting for my wife to get me [02:01:33] :( [02:03:05] Well, your misfortune is my good luck. :) [02:03:32] What do you think about all the template-call stuff, based on the email discussions, etc.? [02:09:31] well [02:10:01] I think we are pretty much getting allot of noise from people who either don't understand the problem, proposed solution, or what reality looks and feels like [02:10:09] gtg [02:10:19] talk tomorow [15:27:07] need to nab a trevor when he appears [16:34:18] hey nimish_g, is trevor in yet? [16:35:19] werdna: not yet [16:35:26] when's he usually in? [17:17:50] hey nkomura [17:18:01] hey [17:27:54] Morning everyone [17:29:17] good morning RoanKattouw [17:29:27] nice summary of bug status, thanks! [17:29:54] brion: Permission to merge r56208 and r56949 (both OKed on CodeReview) to wmf-deployment? [17:30:00] No problem [17:30:14] I said it was "quick" and I intended it to be, but it took me like an hour :P [17:30:30] RoanKattouw: feel free, but note i'll be merging everything today [17:30:37] so probably no need :) [17:31:06] Ah OK [17:31:24] Where are you with reviewing things? Are you gonna go as far up as r57100? [17:32:43] i'm at r57020 right now [17:34:17] :O nice [17:34:39] Dude you should leave us more often ;) [17:36:13] hah [17:38:44] hrm [17:39:06] active tab is showing orange-red in vector, it looks like a broken link [17:39:15] Needs a revert? [17:39:32] Oh dammit Adam's changes [17:39:41] That's unfinished stuff, we should branch it [17:39:53] *brion waves at adam_miller_away :) [17:39:57] ok [17:40:12] Ah why don't you revert all of Adam's changes and followups (1 or 2 by Trevor) and I'll branch them [17:40:26] *RoanKattouw joins the waving [17:41:17] Or better yet, /I'll/ revert them if you don't feel like sorting out dependent revs [17:41:21] yay :D [17:41:21] thx [17:50:48] adam's work is for acai enhancements and still in discussion [17:51:03] so they are not tied to today's deployment [17:53:31] howdy [17:54:31] hey TrevorParscal [17:54:46] saw the little one :D [17:54:52] :)) [17:54:59] nkomura: Yup, that's why I'm gonna take it out of trunk and move it to a branch [17:55:25] TrevorParscal, adam_miller_away: FYI, you guys' watch icon + color scheme work is being taken out of trunk and branched, so Brion can deploy trunk [17:55:46] the watch icon is fine; it's got a control switch [17:55:54] the color changes i'm not so hot on [17:56:23] Hm [17:56:36] They were introduced in the same rev, though [17:56:48] So I'm not gonna bother taking out the color changes but leaving in the watch icon ATM [17:57:16] whichever way's easier for you guys to handle right now [17:59:09] TrevorParscal: poke poke, you around? [17:59:22] i am [17:59:25] aawesome [17:59:35] one issue before a test deployment of LQT: IE6 [17:59:53] I want to disable the fancy google wave look for IE6 -- is there an easy way to make those lines in the CSS invisible to IE6 somehow? [18:00:18] *RoanKattouw suggested child selectors earlier [18:01:17] crap... login broken on my local instance again [18:01:23] ... [18:01:27] other than that everything's looking awesome for a test deployment [18:01:28] this worries me because i don't know why it's doing that :P [18:01:33] I've got a tech blog post ready to go :) [18:01:38] TrevorParscal: Can you grab your IE install and take a look at bug 20900 (especially the last comment)? [18:01:38] *brion slaps self [18:01:39] ok i see it [18:01:53] i turned off memcached at some point which breaks centralauth :D [18:02:05] brion: I'm thinking a deployment for all talk pages on a labs wiki, and only for pages with the magic word that turns it on on testwiki. [18:02:36] werdna: sounds good [18:02:40] whee [18:02:51] should we stick it on existing en.labs or create a dedicated lab wiki? [18:03:13] I dunno, I was thinking dedicated, but I think it's ready enough to stick it on en labs [18:03:21] although that's not the one where the FR feedback is going, right? [18:03:41] would be fun if we could capture all the FR feedback with it -- would give users some experience using it [18:03:47] brion: we have an en.labs wiki? [18:03:55] en.labs.wikimedia.org [18:04:01] werdna: best way to do what you are talking about - making something work one way for some brwosers and another way for others [18:04:09] is to have a non-js fallback [18:04:37] I already have that -- I'm talking about the non-JS stuff [18:04:51] Mostly, I think, the toolbar's appearance, rather than its functionality [18:04:58] it does crazy stuff on IE6 apparently [18:04:58] then blacklist a couple browsers [18:05:14] yeah, but how? [18:05:24] sorry, my lines were disconnected [18:05:25] um [18:05:34] This is CSS stuff, not JS stuff if I understand correctly [18:05:38] well, if you are wanting to get IE6 to use different CSS [18:05:47] then as Roan said, you use the child selector [18:05:59] let me explain [18:06:23] #myid > .someclass {} [18:06:29] is ignored by IE6 [18:07:04] so you can do #myid .someclass { .. ie6 stuff ... } #myid > .someclass { ... override ie6 stuff for everything else ... } [18:07:13] that's done throughout vector [18:08:01] It seems to be standard to put /* OVERRIDDEN BY COMPLIANT BROWSERS */ above the former statement [18:08:02] it avoids needing multiple style-sheets, which helps keep the CSS more maintainable (related styles are grouped together) [18:08:14] RoanKattouw: yes, excellent point [18:08:26] see http://svn.wikimedia.org/viewvc/mediawiki/trunk/phase3/skins/vector/main-ltr.css?view=markup [18:08:30] Well that's what I see in vector/main-ltr.css anyway [18:08:55] yes, "div.vectorTabs ul li" for instance [18:09:16] it solves IE6 having no support for inline-block, but strange handling of block [18:10:03] werdna: have you done any RTL testing yet? [18:10:59] TrevorParscal: ewwww, but okay [18:11:07] TrevorParscal: no, haven't sorted RTL [18:13:13] need to file a bug about that [18:13:24] werdna: we use CSS Janus to build our RTL stylesheets [18:14:05] hmm [18:14:15] where needed, you can add @noflip before a rule [18:14:19] and it won't flip it [18:14:23] it's very cool [18:17:53] RoanKattouw: I uploaded PDF file for https://bugzilla.wikimedia.org/show_bug.cgi?id=20669 [18:18:15] RoanKattouw: did you fix the deployment wiki? [18:18:27] it was broken last night, and I don't have root access anymore [18:18:28] Yeah I think so [18:18:33] Oh yeah I fixed that one [18:18:35] i need root access [18:18:39] Aye you do [18:18:42] I'll make you a sudoer [18:18:46] What's your username? [18:18:54] yes, and it asks for root password still [18:19:00] not just mine [18:19:01] Hm doesn't for me [18:19:05] my username is tparscal [18:19:19] i tried to sudo svn up in wikis/d/ [18:19:27] and it asked for my password, then the root password [18:19:51] good thing you got to it though [18:19:55] # Developers [18:19:57] tparscal ALL=(ALL) ALL [18:19:58] catrope ALL=(ALL) ALL [18:20:04] It was a conflict marker breaking it BTW [18:20:46] It says you have the right [18:21:10] Oh wait [18:21:15] There's a different reason [18:21:21] ? [18:21:26] It won't let you svn up because I checked out d/ using my private key [18:21:33] ah [18:21:36] So it's asking for the passphrase for root@svn.wikimedia.org [18:21:38] well that's interesting [18:21:43] Yeah [18:21:51] so maybe we could fix that [18:21:51] I never thought anyone else would touch it [18:21:59] cause ryan would also have the same issue right? [18:22:01] Yeah, I should make it an anonymous http checkout again [18:22:03] Yes [18:22:10] There's a workaround: sudo su catrope [18:22:21] :) [18:22:23] what's with this FF 2.0 bug? [18:23:13] Hm [18:23:19] Maybe anons have NTOC disabled on that wiki [18:25:09] Hm [18:25:18] No wait [18:25:22] It's there for me now [18:25:24] WTH [18:26:07] OK not a config issue, WFM in Firefox 3, don't have Firefox 2. If you can't reproduce on FF 2.0, tell Calcey to clear cache [18:26:43] I can't repoduce on FF 2.0.0.20 [18:26:51] thats the FF2 I have [18:26:55] running on XP [18:27:00] Right [18:27:25] Hm wait, it's a section edit [18:28:56] Nope, still WFM on FF 3; did you try visiting the exact URL with FF2? http://prototype.wikimedia.org/en-wp/index.php?title=Victoria_University_of_Manchester&action=edit§ion=1 [18:39:59] adam_miller_away, TrevorParscal: I moved the watch icon and color scheme stuff to /branches/usability/phase3 and switched sandbox.5 to use that [18:40:25] RoanKattouw: t [18:40:30] the watch stuff has a switch [18:40:34] you don't have to branch it [18:40:38] please don't branch it [18:40:46] I installed the switch for a reason :) [18:40:52] I had to, it was in one rev with the color change [18:41:02] I'll try to reintroduce the watch icon on trunk [18:41:03] how is that possible? [18:41:18] The rev in which Adam introduced it, he introduced both [18:41:35] really? [18:41:37] So I could've reverted part of that rev, if it weren't for the fact that both features change the same file [18:41:38] what rev # [18:41:40] r56924 [18:41:47] I'm gonna try and pull them apart now [18:42:21] wow [18:42:23] ok sorry [18:42:27] No worries [18:42:27] I'm still catching up I guess [18:42:38] svn: Cannot reverse-merge a range from a path's own future history; try updating first [18:42:38] yes, split them up please [18:42:44] "future history" ? [18:43:08] That's either humor or overly complex wording :P [18:47:10] OK I put the watch icon back in trunk and updated prototype, please test [18:47:29] In particular the CSS [18:52:17] RoanKattouw: do you mean to test deployment? [18:52:38] If by deployment you mean prototype.wikimedia.org/deployment then yes [18:52:49] k [19:16:48] nimish_g: What does ExperimentalSkins do? [19:34:02] its for Adam to use [19:34:27] so he can add in some CSS that's tweaks vector (testing out color scheme changes) [19:34:32] without messing with head [19:34:33] *RoanKattouw suggests [[MediaWiki:Vector.css]] for that purpose [19:36:09] *TrevorParscal doesn't really dissagree [19:36:10] hmm [19:36:14] why didn't I think of that! [19:36:16] oh! [19:36:17] I know why [19:36:26] because we have the same DB for all sandboxes [19:36:37] It'd be easy to decouple s-5 [19:36:43] so if you do it there not only does it no longer represent the en-wp stuff [19:37:12] (custom stuff we need to be aware of on en-wp) but it also would affect all sandboxes [19:37:25] Well, nimish_g wrote that in like 4 minutes [19:37:31] Heh :P [19:37:41] and it will keep our CSS work in the SVN - code review tools are good [19:38:17] Does that mean I can kill the branch I've created once Adam has moved over his CSS? [19:38:52] yes [19:38:56] that was our hope [19:39:04] OK will do [19:42:34] you can kill it now if you want [19:43:24] OK killing [21:45:35] brion: Are we gonna do Babaco deployment today or is it being postponed till tomorrow? [21:45:50] today with general updates [21:48:43] OK [21:48:50] Is it gonna be turned on today as well? [21:54:43] how are we doing bug-wise? [21:55:34] See my usability@wikimedia post earlier today [21:55:46] yeah - looked intimidating :) [21:55:53] There's still a few open bugs I need you to hunt down, and a bunch that require team-wide discussion [21:55:59] ok [21:56:39] For you: bug 20874 bug 20867 bug 20895 bug 20900 [22:03:32] Babaco is being generally deployed? [22:03:37] that's awesome, congratulations! [22:07:57] If it really happens that means we've made schedule [22:08:28] (Although originally Babaco deployment was scheduled for like Sept 9th, but that quickly became unrealistic after the long aftermath of the Acai deployment) [22:14:08] aawesome [22:15:07] brion: are we deploying clicktracking with this release? [22:15:16] *RoanKattouw waits for Naoko [22:15:37] werdna: the new stuff is going to be opt-in only for whatever reason [22:15:45] so it's not that exciting a deployment initially i guess :D [22:15:52] brion: :( [22:15:54] get some more feedback first on things... [22:15:58] RoanKattouw: i'm back [22:16:00] Oooh nice link icons TrevorParscal [22:16:02] but it'll be opt-out eventually right? [22:16:02] Whee [22:16:10] nimish_g: we'll turn it on after other stuff, need to make sure things work :D [22:16:14] For some definition of eventually [22:16:14] RoanKattouw: can you plese help me with the tagging of http://www.mediawiki.org/wiki/Special:Code/MediaWiki/r57160 [22:16:34] Tagged as babacofix; is that what you wanted [22:16:37] I was out while the tagging procedures changed... [22:16:37] ok [22:16:48] thank you [23:06:31] adam_miller_away: you really away? [23:13:41] TrevorParscal: I really was [23:13:52] i was eating dinner [23:13:58] :) [23:14:16] adam_miller: you have interesting email [23:15:26] we need to update the sandboxes though [23:15:27] well, yes I do [23:15:33] or at least the one you are working with [23:15:34] :) [23:15:50] sorry about the confusion on how to do something so simple [23:16:11] there have been some changes made to the output page object and how skins use it which threw us for a loop [23:16:51] so the extra stylesheet links will come AFTER the link to main-ltr.css? [23:18:17] or is it still putting the extra's before? [23:23:26] AFTER [23:23:31] awesome [23:23:32] thanks [23:23:36] that was the whole issue I solved by doing this [23:23:37] n/p [23:24:05] yeah, like I said, things changed with the skin that caused our assumptions that extensions styles would come after the skin styles to be very very wrong [23:29:20] TrevorParscal - are you there? [23:30:45] indeed, I am [23:30:58] Hello! Do you have time to talk about templates? [23:31:06] i.e., TTT? [23:31:37] sure [23:31:42] where are we at? [23:32:44] Well, there are a few questions - most notably, the template-calls-within-template-calls issue, and the format issue. [23:33:34] For the format, specifically, whether there should be different formats for multi-language vs. non-multi-language wikis. [23:33:48] Format for defining the template specs, that is. [23:35:33] I just wanted to know your thoughts on any of these. [23:39:47] Test [23:48:47] well - translation happens in the tags [23:48:56] so any wiki supports UI languages [23:48:59] not just commons [23:49:13] remember, this is UI i18n not content i18n [23:49:32] Well... not really. [23:49:56] I mean, it's content that generates a UI. [23:50:48] It's the equivalent of customized language values - every Wikipedia has some, hardcoded in its own language. [23:52:42] what's the issue here? [23:53:28] The issue is that, if internationalization isn't required, the format could be a lot simpler. [23:54:10] Simple enough that it could be added to the template definition itself, and/or done in a format other than XML, like YAML. [23:54:16] then it's optional [23:54:28] Text [23:54:29] or [23:54:30] svn is giving me way too much crap on the merge [23:54:36] i'm afraid it won't be as quick as hoped :( [23:54:43] Text [23:54:58] brion: i see [23:54:59] :( [23:55:02] eta? days? [23:55:31] tomorrow prolly for core [23:55:45] so let's see if we can go ahead and whip in just the UsabilityInitiative bits [23:55:47] \o/ [23:56:18] :) woo hoo [23:56:38] YaronK: making the msg tags optional is sort of silly through [23:56:51] Why is that? [23:56:59] because now you have removed the explicit explaination of what language the message is [23:57:19] but i guess, it sort of mediawiki-esque [23:57:28] that's how MediaWiki:MessageName works [23:57:44] Yeah, I mean, if it's on a single-language wiki, it'll be obvious. [23:57:48] if you want to do another lang you do MediaWiki:MessageName/ja [23:58:04] Oh, I didn't know that one. [23:58:24] so - yeah, just allow text elements to either have a string in them, or children each contining strings for specific languages [23:59:21] http://prototype.wikimedia.org/en.wikipedia.org/MediaWiki:Preferences/ja [23:59:22] Alright, so I guess it'll remain in XML. [23:59:33] thats' the japanese label for "My Preferences" [23:59:40] Right. [23:59:52] There were various suggestions to use another format, like YAML.