[13:11:50] can someone make me admin on test2? [13:13:46] hexmode, what's your username there? [13:14:01] guillom: MarkAHershberger [13:14:03] eww, the wiki seems a bit borked [13:14:10] Special pages, at least [13:14:24] yeah, I got a blank page on login [13:14:51] reload seems fine [13:14:57] hexmode, done [13:15:01] lol @ the first line: https://test2.wikipedia.org/wiki/Special:Log/rights [13:15:19] tyvm, guillom [13:15:37] So now all users are called "Male user" or "Female user"? [13:15:39] Wow. [13:16:20] guillom: I think that was an old test [13:16:51] I'm getting blank pages intermittently there, though [13:16:51] It seems it's how all user links are formatted now. [13:17:17] guillom: right, but I think that is just a test2 config thing [13:17:24] ah, ok [13:18:38] I don't think this is a good test... or, if it is, it is revealing some problems [13:18:51] *hexmode gets cranky when he sees blank pages [13:19:16] CSS was all borked on the special pages I used to give you admin rights [13:21:18] guillom: I can't give admin rights to user:saibo, could you? [13:21:26] *hexmode has to step out now [13:21:31] sure [13:22:09] if I can get past the random blank pages, of course [13:22:32] done [13:22:59] wait, no, it didn't get through [13:23:40] I hope someone goes through the error log to see where these blank pages are coming from before we deploy this to more sites. [13:24:22] Oh, of course, 1.19 deployment [13:24:52] Yeah, looking at test2.wm.o, seeing random blank pages, and broken (absent) CSS in special pages, stuff like that [13:25:03] hmm [13:25:09] I'll have a look at the error logs [13:25:16] Special:Userrights has no CSS, for example [13:25:23] wtf, fenari rebooted? [13:25:33] ah, kernel upgrade [13:26:47] Hmm, this one looks suspicious 3 Fatal error: Class 'IP' not found in /usr/local/apache/common-local/php-1.19/includes/WebRequest.php on line 167 [13:37:11] I see all sorts of disturbing errors in the Apache logs [13:37:20] Maybe it's APC cache corruption or something [13:43:04] WTF [13:43:11] Oh, I guess Tim disabled wmerrors, didn't he [13:43:24] That would be why fatal.log hasn't been written to since Jan 31 [13:49:22] guillom: Well, I think I may have uncovered the cause of the missing CSS. I'm getting a 503 Service Unavailable from bits.wm.o for one of RL's JS requests [13:49:38] RoanKattouw, ah, good, I guess? [13:50:07] WEll [13:50:08] It worries me, but at least I found /something/ [13:54:14] RoanKattouw: fenari was restarted because I killed NFS1 [13:54:22] It was then restrarted again for kernel updates [13:54:23] Ah, right [13:54:34] Do you have any idea what's up with all these bogus missing class fatals? [13:54:39] (See /home/catrope/119apache [13:54:45] I've seen some of them [13:54:55] Tim suggested it was apc cache pollution or something [13:55:02] And also, why the hell is bits giving me 503 Service Unavailable responses for some but not all of my JS/CSS reqs [13:55:11] I think he thought we had this last time [13:55:20] I like the error 500 every n requests [13:55:57] Did we increase the APC cache size appropriately for 1.18 and 1.19 to coexist? [13:56:10] Nope [13:56:39] Well, we did for the 1.17/1.18 duality, I think we just kept that one [13:56:49] mmm [13:57:03] But maybe we need even more now [13:57:24] It's possible, what's it set to now? [13:57:31] nfi [13:57:52] There is some way that you can see how full the APC cache is, I think [13:57:58] But I've never worked with that kind of stuff [13:58:19] There's an extension for that [13:58:48] I'm sure there's a simpler way though [13:59:03] Found it [13:59:17] We have a script for this, will PM you the URL [14:00:18] I see on the first load, over 1000 files for 1.18, 240 for 1.19 [14:02:00] This is one benefit of test being only on srv193 [14:02:12] True [14:02:17] Is test on 1.19 too? [14:02:30] nope [14:03:17] No reason we can't do it [14:04:29] *Reedy pushes [14:04:30] Yeah, I say let's do it [14:04:48] Can easily enough move it back if it's needed [14:05:21] done [14:07:33] Well the el random error 500s don't seem to be happening [14:07:42] Which is good [14:07:46] hmm [14:07:48] mw.loader::execute> Exception thrown by user.options: mw.user is undefined [14:07:55] mw.loader::execute> Exception thrown by mediawiki.page.startup: mw.util is undefined [14:07:58] Those aren't good things [14:08:19] Resource interpreted as Font but transferred with MIME type application/x-woff: "https://test.wikipedia.org/w/extensions-1.19/WebFonts/fonts/Hebr/MiriamCLM-Book.woff?version=0.105&20111213". [14:08:26] hi RoanKattouw, glad you are feeling better. [14:08:59] My stomach and I are still negotiating, but we're getting somewhere [14:09:07] haha [14:09:19] application/x-font-woff [14:09:28] I slept a lot last night and I'm not really sick anymore, but I've barely drunk or eaten anything in 36 hours so there's that to deal with now [14:09:33] Oh, is that the correct MIME type? [14:09:39] Apparently [14:09:41] For chrome at least [14:09:43] I'll fix it [14:09:55] Might aswell fix it in svn [14:10:36] Hmm, JS errors were transient, and I understand why now [14:10:46] Computers suck [14:10:46] Dependency map changed, takes 5 mins to fully propagate [14:10:53] We re-established this last night [14:12:40] RoanKattouw: shall I tell you of Traditional Indian(-American) remedies? [14:12:47] not to JS, but to nausea. [14:12:54] Although maybe JS is causing your nausea. [14:13:02] hehe [14:13:06] No, the nausea is over now [14:13:25] It's just that my digestive system is ... reinitializing [14:13:34] Right [14:13:43] Essentially, it rebooted [14:13:48] yeah [14:14:05] like Linux on an airplane seatback entertainment system [14:14:16] Reedy: What happened last night re computers suck, specifically? [14:14:25] *RoanKattouw wasn't around for the 1.19 deploy obviously [14:14:31] Using NFS made the kernel segfault [14:14:33] Repeatedly [14:14:37] ouch [14:15:04] Hmm, what else happened [14:15:42] That meant most servers didn't have all the php-1.19 files, and in many cases had bad permissions [14:16:20] OK [14:16:31] Are we sure the "not all servers have all files" issue is fully fixed yet? [14:16:36] Maybe APC is caching the nonexistence of files? [14:16:42] It should be [14:16:46] I deleted all the 1.19 files [14:16:55] and then resync'd them all at a much slower rate [14:16:59] OK [14:17:13] So that was the -F30 thing [14:17:14] ? [14:17:21] Yeah, using -F5 instead [14:18:01] Pediapress decided to change their default image size from 1200px to 1199px [14:18:12] Because they were getting some thumbnail issues related to swift [14:18:19] And didn't bother to even tell us [14:18:26] *facepalm* [14:18:48] I'm sure this caused nasty thumb infrastructure issues? [14:18:53] So that'd caused a lot more load on the image scalers [14:19:02] How did you guess!? [15:02:39] unable to create an account on test2? Error: ERR_ZERO_SIZE_OBJECT, errno [No Error] at Tue, 14 Feb 2012 14:59:14 GMT [15:02:53] Try again [15:03:10] test2 is flaky [15:03:18] test.wikipedia.org is also now on php-1.19 [15:05:01] second time worked thx Reedy [17:20:12] Getting something more than an error 500 would be nice for test2 [17:25:29] RoanKattouw: getting 503's again... [17:25:45] or sometimes [17:25:46] meh [17:25:58] and white pages [17:26:00] this is so fun [17:26:06] Let's get Tim to poke at the APC issue first [17:26:35] More than half of fatalmonitor is APC corruption bogus right now [17:26:52] Yeah [17:27:11] A few obvious errors (though, non fatal) have been cleaned up [17:27:13] so it's progress [18:14:03] *sumanah gets ready for 20% meeting [18:15:47] brion, you wfh today? [18:15:54] yep [18:16:28] need me to call in for anything? [18:16:43] we're just doing IRC for this...coordinating 20% time today [18:16:54] RoanKattouw_away: feeling better? [18:17:10] Trevor said he'll be on in a bit [18:17:24] there are two big things for today: [18:17:34] spiff [18:17:35] 1. Getting test2 actually functional enough for real testing [18:17:45] 2. finishing up the last bits of code review [18:17:49] hi hexmode [18:18:36] yesterday, a bunch of us got distracted by the Swift stuff (including Tim) so that took our focus away from figuring this stuff out [18:18:45] Reedy: you around? [18:18:53] sumanah: hey [18:19:39] so...to repro the 500 problem: http://test2.wikipedia.org/wiki/Special:Random <- go here a few times until you hit it [18:20:18] maybe a graceful apache restart is in order? [18:20:20] *sumanah waits for lag to resolve, can't hear anything except by control-r on the log [18:20:55] hello [18:21:31] on the code review front: http://www.mediawiki.org/wiki/MediaWiki_1.19/Revision_report [18:21:31] any logs from that server ? [18:21:44] i can confirm on the client end [18:21:49] but there's no output visible [18:21:51] robla: I'm getting just blank pages no 500s [18:22:00] my blank pages are 500s [18:22:11] we're getting random errors like-a so: [18:22:21] Feb 14 18:21:19 10.0.11.28 apache2[17932]: PHP Fatal error: Allowed memory size of 125829120 bytes exhausted (tried to allocate 523800 bytes) in /usr/local/apache/common-local/php-1.19/extensions/LiquidThreads/LiquidThreads.php on line 14 [18:23:45] yeah, firebug shows they are 500s [18:23:55] I was just asking Aaron in real life how to correlate the exact 500 to the error in the log [18:24:01] brion: you can't do anything w/o server access, there are only ~20 CR revs left [18:24:11] whee [18:24:16] :) [18:25:18] brion and TrevorParscal: how do you want to divvy up the work? [18:26:02] is santosh around? [18:26:12] sorry, santhosh [18:26:30] I don't see him :( [18:26:42] was hoping he could make a quick fix for r110745 [18:27:11] hello? [18:27:21] is this thing on? [18:27:41] TrevorParscal: I'll ping him when I see him, but how about CR? [18:29:32] *hexmode runs CR report one more time hoping for disappearing revs [18:29:44] robla: yo [18:30:03] *robla throws something at TrevorParscal irl [18:30:15] heh [18:31:09] hexmode: could you tag the remaining "miscextension" revs with their actual extesnion? [18:31:25] yep, makes a lot of sense [18:31:27] :) [18:31:42] TrevorParscal: brion: how do you want to divide up the remaining review work? [18:31:50] AaronSchulz will work on the 500 stuff, I think [18:32:24] guess i'll look over things on https://www.mediawiki.org/wiki/MediaWiki_1.19/Revision_report ??? ? [18:32:37] Rob and I are doing the frontend stuff [18:32:40] atm [18:32:43] robla: am now [18:33:05] just want to make sure you guys don't both burrow in on the same ones [18:33:18] so...Trevor just called dibs on the "frontend" tag? [18:33:51] i'll take the categorytree one for now [18:34:01] any others to volunteer me? [18:34:34] hexmode: maybe start tagging names as people volunteer [18:34:38] make sense? [18:35:41] yeah, "stick a pin in it" -- so I don't forget [18:37:23] we've got a couple for aft (not aftv5) [18:37:29] anyone want those? [18:38:13] *werdna waves [18:38:20] wassup people? [18:38:46] TrevorParscal: will there be a feature meeting today? [18:38:54] um... [18:38:56] lolz [18:38:58] yes! [18:39:23] sorry to wake you so rudely ;) [18:39:32] eep, features team meeting [18:50:28] TrevorParscal: http://etherpad.wikimedia.org/FeaturesTeam2012-W07 [18:52:47] gwicke: just sent out email - thank you for doing that! [18:53:11] TrevorParscal: sure! [18:53:56] Hello [18:56:35] hi Inez! [18:57:44] uga chaga [18:58:13] And now I see a dancing baby. [18:59:53] nooooo!!! [19:02:53] re: code review - this needs fixing -> http://www.mediawiki.org/wiki/Special:Code/MediaWiki/109808#c30954 [19:03:34] re: features team meeting - WELCOME! [19:07:29] lots of progress on the parser front: the parser is now known as Parsoid.. [19:07:38] apart from that- slow and steady mostly [19:07:53] general template and -argument support in attributes now works [19:08:08] main page gets closer to the original [19:08:31] next up: microdata representation of templates and -arguments [19:08:40] we should use more code names. [19:08:50] I like "Parsoid". It sounds like a robot. [19:09:06] sounds like something green you don't want to eat [19:09:16] we are already the first hit on google for it [19:09:20] A robot made of parsely. [19:09:21] surprisingly [19:10:09] Audrey Tang will help us a bit [19:10:34] that's it from me mostly [19:11:20] so, I will take the mic then [19:11:26] mmmmm Parsoid [19:11:43] her talk page is despressing. [19:11:45] for those of you who don't know, the visual editor team is in 3 parts atm [19:12:19] it's like a triforce of visual editor engineering power http://upload.wikimedia.org/wikipedia/commons/thumb/8/86/Triforce.svg/210px-Triforce.svg.png [19:12:35] at the top you have gabriel and roan doing the parser and data model stuff [19:13:07] triforce? or triad? http://venturefans.org/vbwiki/The_Order_of_the_Triad [19:13:22] at the bottom left is Inez and Christian working on a ContentEditable implementation of the display surface, the top-most layer of the visual editor stack [19:14:07] and at the bottom right you have Rob and me prototyping some "you can't do that, can you?" sort of things for the edit surface, the display layer we've been building so far [19:14:39] so the CE and ES teams met last week and shared ideas and whatnot [19:14:47] excellent [19:14:56] they are making good progress on a CE implementation that doesn't suck that bad [19:15:02] ah, finally I understand those abbreviations [19:15:10] how does visual editor tie into this? [19:15:28] with the mobile team fleshing out i'll probably poke more at what you guys are doing :) [19:15:46] brion: yay! [19:16:07] *gwicke hopes to draw brion back into parser stuff [19:16:08] sounds like some DOM goodness is happening [19:16:23] (it's all visual editor?) [19:16:23] rob is focusing on IME atm [19:16:23] making some progress there but FireFox is kicking him in the teeth right now [19:16:23] (i will let him tell his side of the story) [19:16:23] and I'm working in my laboratory on floating content [19:16:24] aka, supporting float left and right and clear and all that jazz [19:16:24] that's about where we are at [19:16:52] so you are working on two alternative approaches at the same time? [19:16:57] yes, goodness indeed [19:16:57] Rob? [19:16:57] anything to add? [19:16:57] yes [19:17:19] So far this week I have IME working mostly in Chrome, but then realized that this solution will not work at all in Firefox.. [19:17:44] IME = input methods? [19:17:45] Nikerabbit: then we will choose one, and recombine the team - crossing the streams Ghostbusters style [19:17:47] Yeah [19:17:49] (hence aforementioned teeth kicking) [19:18:00] what's IME? [19:18:20] I have some ideas about how to better this, like comparing text blocks.. Perhaps reworking general input [19:18:26] brion: we are building an HTML DOM, but it does not yet properly track expanded template contents etc [19:18:36] does it compare to Narayam? [19:19:00] IME is the native stuff on phone [19:19:01] s [19:19:15] kaldari: IMM = Input method (OS language modes) [19:19:15] Narayam is a JS emulation of IME, essentially [19:19:29] IME ^^ [19:19:39] are there any plans for text selection on mobiles? [19:19:57] some devices should be able to use Narayam, but native methods can be much better [19:20:10] Yes [19:20:19] But native methods don't work automatically on the edit surface [19:20:45] k [19:21:05] so no silver bullet in sight yet [19:21:25] Trevor is having wifi issues [19:21:44] mmm, they still haven't fixed wireless in the office? [19:21:52] "fixed" [19:22:30] "not-quite-fixed" [19:22:39] it has been a problem in almost all meetings I've had with SF folks [19:23:10] wireless on my mac os x laptops has been flawless; on other things???. flakier [19:23:25] maybe i'm just lucky on that machine ;) [19:23:46] *RoanKattouw just plugs in to wired Ethernet all the tiem [19:23:47] I remember my first visit, it stopped working few times, and didn't start working until I booted Windows in the between [19:23:59] yeah. i try to use wired as much as possible. [19:24:20] *gwicke had no problems with wireless in SF either, but alolita had all the time [19:24:37] alolita also has a thinkpad. [19:24:44] *Nikerabbit too [19:24:57] Ubuntu wireless hardly works with the lenovo [19:25:02] broadcom wireless? [19:25:34] *brion back in a bit [19:25:42] don't all thinkpads have intel's wireless [19:25:53] It's Eloquence. [19:26:22] my head is a big bowl of snot. [19:26:27] the agenda says editor retention would be next [19:26:32] jorm, sounds familiar [19:26:37] it's awesome. [19:26:42] Editor Retention. [19:26:50] okay. i can slip up on that. [19:27:08] Focusing on Article Creation Workflows. Hoping to have a phase one spring deployed on Thursday. [19:27:21] Was supposed to be yesterday but we've still got work to do. [19:27:54] *TrevorParscal_ emerges from the server room with an ethernet dongle [19:28:01] We were going to meet about doing an NPT sprint to start work on some back end stuff while we see how ACW handled itself on the outside. [19:28:16] but people are sick, etc. [19:28:42] AFT has been chugging forward. they're working (i think) on handling issues with oversight. [19:28:50] What's NPT? [19:28:57] New Page Triage. [19:29:15] An interface for new page patrol that doesn't pre-suppose guilt. [19:29:27] *gwicke wondered the same thing [19:29:33] are we aware of the interesection between ACW and NPT? [19:29:36] we should have more code words ;) [19:29:40] srsly. [19:29:51] I've been out of town for the past week, so I've been out of the loop on ArticleCreationWorkflow... [19:29:58] ACW (Article Creation Workflow) is an attempt to make the process of creating articles friendlier. [19:30:01] on the surface. [19:30:15] Should I work on the data collection part or has someone else taken that piece? [19:30:24] under the surface, it's mostly about reducing crap. it's supposed to subtly encourage people to NOT create articles. [19:30:35] i think that's a fine choice, kaldari. [19:30:38] cool [19:30:42] i believe diederik has some data for you? [19:30:46] yes [19:30:50] he shared it with us in our product meeting this morning. [19:31:18] anyways. NPT is where experienced editors curate the articles that get created. [19:31:29] but it's a really goofy process. [19:31:58] we've been in analysis for quite some time; it's time for a design iteration. then there will be much wailing and gnashing of teeth. [19:32:21] other members of the team may have other updates. [19:32:43] if not, then I18n would be next [19:33:00] yay [19:33:14] grab the mic ;) [19:33:16] Siebrand, Amir, Santhosn and Alolita have been in India [19:33:38] I don't know much about what fun I've been missing, but there is something in this pad about the hackathon: http://etherpad.wikimedia.org/123 [19:34:44] No i18ndeploy yesterday, waiting for 1.19 to be rolled out [19:35:13] new translation memory is progressing, already enabled on translatewiki.net, and I've started playing with the labs to test tm in a environment closer to wmf cluster [19:35:35] also blogged about the translation memory design [19:35:53] Some UI work Special:Translate: http://translatewiki.net/sandwiki/i.php?title=Special%3ATranslate&taction=translate&group=ext-translate-core&language=ru&task=untranslated [19:36:22] compare with: https://meta.wikimedia.org/w/index.php?title=Special:Translate&group=page-Wikimedia+Highlights%2C+December+2011&language=fi&task=view [19:36:51] pretty. [19:37:00] Gerard has been blogging as usual, and reaching out all the wikipedias and such [19:37:26] I will continue working on getting translation memory to wmf, probably after 1.19 [19:37:41] hoping to sync with rest of the team which is getting back today I think [19:37:43] nice work on the UI [19:38:37] and final words: I hope to stary playing with git soonish, to make sure we can handle it from i18n side [19:39:00] kaldari: are you still in the fundraising group? [19:39:06] no [19:39:10] k [19:39:16] I'm in Editor Engagement now [19:39:47] ah! [19:40:17] hmm- the order would be fundraising, resourceloader [19:40:36] MW 1.19 [19:40:59] and finally AFT [19:41:35] Fundraising up next, then? [19:41:46] K4-713: yep! [19:41:55] Okay, I'll take it then. [19:42:06] forgot multimedia, but nobody seems to be screaming [19:42:23] Basically, everything I have to report is more or less the same as last week. [19:42:49] I'm still working on the auditing tools for civicrm, and doing some code review for Jeremy as needed. [19:43:13] We're also reviewing a lot of resumes for new fundraiser devs. [19:43:31] This week: Even more of the same. [19:44:01] we don't really have a multimedia team right now. [19:44:01] We will be starting phone screenings and interviews this week for the new position. [19:45:04] i can go now if you are done K4-713 [19:45:13] Yeah, that's about all I've got. [19:45:28] I am wrapping up CiviCRM recurring payments [19:45:31] have I missed editor engagement? [19:45:52] I have been helping with interviewing candidates for fundraising [19:45:53] werdna: the agenda only lists editor retention [19:46:04] same thing [19:46:09] renamed [19:46:17] I am also working on LastModified and UnitTest [19:46:20] NikeRabbit: Do you know if PLURAL is working for javascript? [19:46:20] I am getting this output to pages: [19:46:21] Last updated {{PLURAL:23|23 day|23 days}} ago [19:46:22] Seibrand said, in December, that NeilK merged in 1.18-wmf and trunk [19:46:31] I'm double-parked in the meeting department [19:46:48] jpostlethwaite: in theory yes, we are working on integrating it with mw.msg/mw.message [19:47:08] it is not removing {{PLURAL... [19:47:11] what other meeting/ [19:47:13] ? [19:47:15] i can talk to you about this off line [19:47:33] analytics for acw [19:47:55] I gave a tutorial to OmniTI on the extension UnitTest for AFT last week [19:48:00] that is it for me [19:48:21] werdna? [19:48:28] yeah? [19:48:46] is it our turn? [19:48:49] i have one last thing, btw, unless we want to talk about any mobile stuff. [19:48:55] editor engagement already went. [19:48:59] werdna: just grab it.. [19:49:08] ok, off you go jorm [19:49:23] check out gwicke, running the meeting like a boss. [19:49:52] hehe [19:50:02] werdna: i just basically told everyone where we were with ACW. [19:50:15] sweet, glad you know where we're at :D [19:50:49] mobile has a couple design iterations running. mostly working on a project called Glaucus, which is a visual refresh and intended to be an intermediate step twoards Athena. [19:50:59] (hence the name "Glaucus," which is Athena's owl) [19:51:23] two other small mobile projects that heather is handling - full screen search stuff and then mobile beta opt-in interactions. [19:51:27] so that's that. [19:51:32] the last thing i have is this: http://www.zazzle.com/wikimedia_labs_shirt-235234255296411069 [19:51:38] Unicorn t-shirts. [19:51:55] We don't have budget to make them for everyone, so if you want one you gotta get it yourself. [19:52:15]