[13:30:21] hexmode: hello! I trust your weekend was good? Ready to have our phone chat anytime [13:30:55] hexmode: (except not in the next 15 min) [13:42:56] hexmode: back [14:32:34] hexmode: ping [15:05:26] if mdale is around anyone, i figured out the reason for one of the bugs he was seeing. [15:56:58] sumanah: sorry, was exhausted. I'm awake and here now :P [16:01:04] hexmode: so, want to have a short phone call now? or around 2pm local? [16:09:34] sumanah: network connection problems and me being exausted and sleeping till 11 mean I'm only now ready to go if you want to talk [16:09:50] hexmode: let's wait till 2, [16:09:56] that way you can catch up a bit first. [16:09:59] :) [16:10:17] np... write my weekly report, too. [17:49:47] RoanKattouw: howdy [17:49:52] Morning [17:57:54] jorm: Do the -1 to 100 scrums still happen every morning at 10:15? I totally forgot about that today [18:03:41] we haven't had any late. [18:03:43] lately. [18:15:47] *hexmode looks for sumanah [18:16:01] hello :) [18:16:20] hashar: ! [18:16:37] I managed to fix a bug you assigned to me 2 weeks ago [18:16:44] (merging two scripts in maintenance/ ) [18:16:46] :) [18:17:00] oo... fixing bugs is dangerous [18:17:13] it only encourages me to assign more to you :P [18:19:02] :p [18:19:14] could you possibly update the revision report at http://www.mediawiki.org/wiki/MediaWiki_roadmap/1.18/Revision_report ? [18:19:23] one day we will have to ask reedy to code this in CodeReview [18:24:14] RoanKattouw_away: is the protocol-relative stuff in good enough shape to push to all sites? [18:41:26] robla: do you know who is assigned to code review the mw swift extension? [18:41:57] binasher: there's no specific assignments. why do you ask? [18:42:59] I suspects it needs review before any type of deployment [18:43:31] robla: we are discussing deploying in the ops meeting and have been wondering if its being reviewed as commits occur [18:44:13] Nope [18:44:37] If something is live on the cluster, it'll be reviewed [18:44:45] If it's not, it doesn't usually get done, till someone reviews the whole thing [18:44:48] binasher: I haven't been involved in that. rnelson should be lining up the review, and woosters should be making sure he does that [18:45:22] ok, thanks. we'll make sure it happens before deployment [18:45:47] robla: I will be point to carry on that conversation [18:46:35] (along with russ) [18:52:19] Ryan_Lane: I am gonna deploy an update to the protocol relative handling code tonight [18:52:30] ok [18:52:37] Ryan_Lane: Then I wanna run it on test a few more days before we roll it out to all wikis [18:52:43] sounds good [18:52:52] the date was the 18th anyway ;) [18:53:19] I hadn't heard about that, but OK [18:53:27] that was in the email, wasn't it? [18:53:31] or was it today? [18:54:06] ah. week of 15th [18:57:31] Meh [18:57:42] The toolsever is killing my AFT stats queries because they're too slow [19:01:25] RoanKattouw, use the enwiki slave that the interns are using :P [19:02:04] Hmm, yeah, I guess I should, huh? [19:02:26] They can't be giving it that much load [19:02:42] Do they let them have their own tables on that box? [19:02:56] Pass. binasher would probably know [19:04:00] RoanKattouw: yeah - all of the interns have their own db on there that they can write to [19:05:01] Sweet [19:05:15] Maybe we can give K4-713 one of those so she can move the AFT stats production there? [19:07:02] *K4-713 is interested [19:10:39] certainly - email me details [19:12:21] I will write an e-mail to the two of you later [19:12:26] When I'm not doing something else [19:12:51] It's db42, right? [19:13:38] db42 is fairly loaded at times though - the interns like to run queries that write gigs of data to temp tables [19:13:39] yep [19:15:54] I don't really care, I just need a weekly cron job [19:16:04] Which is heavy but not /that/ heavy [19:16:19] And I don't care if it's slow, as long as there's no query killer killing it after half an hour [19:16:40] It normally doesn't run for that long, this Monday was the first time that happened [20:21:27] Grah, I want 1.18 [20:21:38] Merging things to 1.17wmf1 is frustrating if every single rev conflicts somehow [20:22:23] <^demon> +1 [22:26:28] ^demon: you around? [22:26:36] <^demon> Yep, what's up? [22:26:58] I've been asked to figure out a review strategy for SwiftMedia [22:27:20] have you been keeping up on it? [22:27:38] (that's the thing that Russ has been working on) [22:27:43] <^demon> Causally, haven't done any in-depth reviews. [22:28:11] <^demon> Casually, even [22:28:31] my understanding is that Russ thinks its pretty close to being ready to deploy [22:28:46] As one of the first steps, I suggest contacting Bryan Tongh Minh and asking him what his availability looks like [22:28:58] There seems to be quite a lot of commented unused code and stuff lying about in it [22:29:16] He wrote most of the current file-related code. I'm not saying he should review the entire thing, but where the two things tie in together his expertise should be useful [22:29:43] <^demon> He wrote a lot of the Upload code. [22:30:28] <^demon> FileRepo is mostly Tim's doing. [22:30:38] <^demon> But both Bryan and I are pretty familiar with it. [22:31:23] <^demon> robla: Russ' core changes are almost completely reviewed (I only see 1 outstanding rev). The ext needs a top-to-bottom review most likely. [22:31:35] there's quite a few entries in his extensions TODO [22:32:47] *robla looks over at CodeReview [22:34:22] does Russ have a non-obvious IRC nick? [22:34:33] <^demon> nelson, iirc. [22:35:03] sounds right [22:35:15] I'm not seeing him online now [22:35:28] he's rather sporadically online usually [22:35:49] maplebed: something else to find out from Russ: does this extension require MediaWiki 1.18? [22:36:06] <^demon> I think so. [22:36:07] I'd guess so with his changes to core [22:36:11] unless we start backporting [22:36:12] <^demon> That ^ [22:36:14] *Reedy barfs [22:36:17] <^demon> No more backporting, plz. [22:36:20] nope [22:36:44] robla: I might be able to answer that... [22:36:49] well, we *may* even need to backport trunk stuff to 1.18 [22:36:58] <^demon> All of his changes are pre-branch [22:37:00] <^demon> So we should be fine [22:37:02] #^ [22:37:28] <^demon> Yep, last core commit was July 19th, which was before I rebranched. [22:37:50] OK [22:37:57] So we could block Swift deployment on 1.18, right? [22:38:05] Speaking of which I really need to start reviewing more [22:38:16] Speaking of which, robla, which tag do I use to exclude stuff from the revision report? [22:38:27] 560 new [22:38:40] *robla looks up that tag for Roan [22:39:01] RoanKattouw, reviewing Timos stuff would be great [22:39:21] most of it ist stuff ^demon and I can't really review [22:39:42] Yeah [22:39:47] I have some tabs open [22:39:51] And a million other tasks, as you konw [22:39:58] yup [22:39:59] I should do some deployment scheduling tomorrow [22:40:21] RoanKattouw: all of the following are excluded: '1.18revert', '1.18ok' and 'nodeploy' [22:40:21] robin has the most news... [22:40:42] robla: So what do I do with a rev that has /already/ been reverted on 1.18? I tagged them 1.18reverted for now [22:40:47] <^demon> I never used the word "new" as a noun until CR. [22:40:50] <^demon> "You have 4 news" [22:41:13] robla: Also, I tagged some directionality revs by robin for Nikerabbit , but I don't think anyone told him he'd been assigned revs to review :D [22:42:42] RoanKattouw: hrm....I guess I can add "1.18reverted", and leave "1.18revert" in the report, if we want to be that granular [22:42:59] robla: the MW installation on the swift cluster has a HISTORY file that statrs with 1.17. [22:43:28] robla: Well 1.18revert is an action tag, just like the merge tags. They're removed when the action has been taken [22:43:43] <^demon> maplebed: Perhaps, but he's made changes to core MW that aren't in 1.17, so I imagine his work depends on those. [22:43:53] Specifically, a lot of DieBuche's and Timo's commits are excluded because they concern AJAXCategories, which was backed out [22:44:28] Which isn't obviously evident from looking at the new list :) [22:45:40] Nope, which is why I tagged them and just asked robla to blacklist that tag [22:45:50] BTW, in better news [22:45:56] I cleared out the 1.17wmf1 merge queue today [22:46:02] Merged and deployed everything save two revs [22:46:03] hexmode: you around? [22:46:17] robla: yep [22:46:19] Ryan_Lane: That means most prot rel URL issues (all of the known ones, I believe) are now fixed on the cluster [22:46:40] Status Number of revisions [22:46:40] new 560 [22:46:40] fixme 125 [22:46:57] 538/124 [22:47:04] hexmode: read everything starting at "RoanKattouw: Speaking of which, robla, which tag do I use to exclude stuff from the revision report?" [22:47:41] (about 10 min ago) [22:49:13] robla: russ says "svn head, but yes, 1.17." is that enough of an answer to your question? [22:49:15] Hmm [22:49:31] Can we get maxsems code review popups to work on non CR pages? [22:49:41] or is that unconditional load on every page? [22:51:31] RoanKattouw: \o/ [22:51:31] thanks [22:51:31] maplebed: I can't figure that out for sure. He's saying that his extension (in trunk) will run on top of an unmodified 1.17? [22:51:51] "I fetched phase3/trunk back in March and have been working with that ever since." [22:52:10] Ryan_Lane: In fact, there are just two open bugs. The e-mail notif bug is open just because I haven't confirmed the fix yet (will do shortly), the other one is the external link search bug [22:52:17] Which isn't really that much of a big deal in the short term [22:52:23] also, nearly all the changes are contained within the extension. [22:52:28] cool [22:52:31] ("nearly") [22:52:38] <^demon|away> robla: I agree, that doesn't really make sense. The trunk he pulled in march was already 1.18alpha [22:52:40] Yes, surely he made *some* changes to core [22:52:43] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/author/nelson [22:52:44] I'm sure we'll get a flood of new bugs when we enable it site wide [22:52:52] Probably [22:52:58] But that's why we deploy stuff, right? [22:53:03] heh [22:53:11] Wikipedia isn't really an encyclopedia, it's where we test our code [22:53:15] lol [22:53:18] at least 16 of his revisions touched core [22:53:18] RoanKattouw: can we exclude 1.18revert as well as 1.18reverted? [22:53:27] I wonder how we should roll it out [22:53:41] half != some [22:53:44] hexmode: Yes, but 1.18revert should be listed separately on the report because it requires action. 1.18reverted doesn't need to be displayed at all [22:53:55] RoanKattouw: any sane idea on order of wikis to deploy it to? [22:53:57] RoanKattouw: gotcha [22:54:12] Ryan_Lane: Start with meta and commons, which already have the https DNS thingy? [22:54:14] hexmode: I'd just say s/1.18revert/1.18reverted/ in the script [22:54:20] yeah. sounds like a good idea [22:54:39] central auth is going to give us headaches [22:54:45] log-in-globally... [22:54:47] I guess the script just misunderstands what the 1.18revert tag means. It means please-revert-me, not I've-been-reverted [22:55:01] ? [22:55:04] robla: well, that is a quick first step, but I think putting "these need reversion" in their own section makes sense [22:55:06] Wouldn't that already be broken on test? [22:55:19] let's see :) [22:55:20] hexmode: that will happen kinda automatically [22:55:38] anything tagged as 1.18revert will still be in a line on the report [22:56:00] robla: ah, k [22:56:01] For extra clarity in case someone didn't get it yet: 1.18revert means "I need to be reverted, please revert me" and 1.18reverted means "I have been reverted, ignore me" [22:56:01] seems to work fine [22:56:11] I wonder if it pulls the images from secure.wikimedia.org [22:56:24] The former should be a separate line on the report. The latter could be but need not be, as long as its revs don't pollute the untagged list [22:56:38] Ryan_Lane: Hmm let me firebug that [22:57:04] it doesn't [22:57:22] Nope, http [22:57:27] which is going to cause mix-content issues, but otherwise shouldn't be a problem [22:57:32] Actually, that sucks ass [22:57:38] You think you're logging in securely [22:57:45] but you definitely are not [22:57:46] And you're really logging in insecurely to like 12 domains [22:57:53] RoanKattouw: robla: report updated [22:58:00] Thanks, [22:58:02] *RoanKattouw looks [22:59:33] 166 new untagged [22:59:36] *RoanKattouw loves jQuery + CSS# [22:59:38] *CSS3 [22:59:43] >>> $('#sortable_table_id_0 tr:nth-child(14) td:nth-child(2) a').length [22:59:45] 166 [23:01:03] Or a full list of counts per tag with $('#sortable_table_id_0 tr').each(function() { console.log($(this).find('td:first').text() + ' ' + $(this).find('td:last a').length); }); [23:02:47] http://i.imgur.com/zvjhw.jpg [23:03:24] RoanKattouw: does it look right? [23:03:57] Ryan_Lane: Could you do me a favor and edit http://test.wikipedia.org/wiki/User_talk:Catrope?action=edit ? Just add anything, I need to test the e-mail notif [23:04:21] Ryan_Lane: In fact, do it over HTTPS while you're at it [23:04:30] Devilish [23:05:42] done [23:05:44] Hmm, BTW, my virtual attendance for the NOLA hackathon isn't looking too bad [23:05:46] Thansk [23:05:51] cool [23:05:54] I have the GRE thing at 8:30am on Saturday in Hamburg [23:06:03] So Friday is shot because I have to drive out there and go to sleep early [23:06:13] But the thing ends at noon, so I can drive back here before you guys wake up [23:06:15] Reedy: where can I get one? [23:06:23] No idea [23:06:35] I should definitely be up and running by noon CDT on Saturday [23:06:56] RoanKattouw: I studied like a dog for the math part of the GRE and did great [23:07:09] felt more confident on the english portion [23:07:13] Oh I already took the general GRE [23:07:17] didn't do so great [23:07:18] Math: 800, English: 550 [23:07:30] The English part was rife with very obscure worsd [23:07:38] same [23:07:47] not my score [23:07:51] you did better [23:07:52] But same idea? [23:07:57] yes [23:08:06] Yeah in the math part I got one question wrong at least, I'm surprised I still got 800 [23:08:07] English makes no sense at the best of times [23:08:12] But I guess if that's the only one I got wrong... [23:08:21] Anyway, this is the GRE CS Subject Test [23:08:29] ah... [23:08:33] Which is why I have to do it concurrently with the NOLA hackathon in Hamburg [23:08:44] I was taking it 3-4 years after getting my BS [23:08:45] Because subject tests are only offered 3 times a year, and are not offered in my country [23:08:53] so that is my excuse [23:08:54] Nearest locations were Hamburg, Brussels and Munich [23:09:35] IIRC, I got a 770 on math [23:09:40] *hexmode sighs [23:10:04] *RoanKattouw sees Ryan's barnstar (thanks) and wonders whether the enotif got stuck in the job queue [23:10:29] could have [23:10:32] hexmode: I get more fun: I had to take TOEFL too [23:11:05] Only [23:11:14] So soon I will have taken three graduate admissions tests, in three different cities [23:11:15] Only ~64 jobs on enwiki, so queue is low [23:11:21] Reedy: This is *test* wiki [23:11:23] Feck [23:11:25] Yeah [23:11:27] 28152 [23:11:28] There shouldn't really be any jobs there [23:11:29] Why so high? [23:11:30] *hexmode googles TOEFL ... clicks wikipedia 3rd hit [23:11:31] But the job runner rotates afaik [23:11:35] yeah [23:11:47] hexmode: Test Of English as a Foreign Language [23:12:03] Required for students who didn't have an English-language bachelor's education [23:12:14] RoanKattouw: too late, enwiki already told me ;) [23:12:24] Which, in practice, is almost anyone who did undergrad in a non-English-speaking country [23:12:31] RoanKattouw, the job queue is full of enotifyNotify stuff [23:12:52] RoanKattouw: since I don't know: where are you going to school? [23:13:11] hexmode: Some place in the Bay Area. I'm gonna apply to Berkeley and Stanford at least [23:13:23] :) [23:13:24] RoanKattouw: good luck [23:13:26] Hmm, a lot of enotfiy jobs about 24 hours old [23:13:27] Thanks :) [23:13:34] *vvv is applying there for undergrad [23:13:37] Oh, cool [23:13:47] I thought you said you didn't like SF? :P [23:13:58] hey, I dunno if you should go to stanford... you might leave us to start google or something [23:13:59] Yep, but I like Stanford quite a lot :) [23:14:02] Right [23:14:08] And it is not so close to SF [23:14:13] Yeah, that's a minus for me [23:14:45] For now, at least, I prefer Berkeley because they seem to give you more freedom in picking courses, seem to offer more courses I like, and it's less of a hassle to get to [23:14:58] It's closer and connected to the city by a rail system that's actually decent [23:15:07] *RoanKattouw glares at Caltrain [23:16:07] Ryan_Lane: Could you give me another barnstar on officewiki via https? [23:16:13] sure [23:16:23] via wikilove? [23:16:25] or manually? [23:16:28] Caltrain is not decent? [23:16:50] RoanKattouw, it seems on testwiki, there are 25k jobs from the last hour or so :| [23:17:07] Bah [23:17:12] What about office? [23:17:21] vvv: If it runs hourly between peak times, no [23:17:29] 1 [23:17:54] And SF-Stanford takes more than 3x as long as SF-Berkeley, although that's partly a function of distance [23:18:18] Also, Caltrain only has two stops in SF, compared to BART's 8 [23:18:32] And one is right near the office [23:18:36] But maybe changing trains at Millbrae is decent, I wouldn't know about that [23:18:53] The 4th&King depot is like a 15-20min walk from the office, yeah [23:19:04] And in 2017 it's gonna be a 3-stop MUNI ride [23:19:19] With the Central Subway [23:19:19] I still remember when I first were in SF, I did only two things: [23:19:25] 1) Visited WMF office; [23:19:38] 2) Went to the nearest Friday's in the airport [23:20:32] One of the first thing that surprised me in US is fast food coverage [23:21:02] In Russia, if there is an internation ff network, then it is somewhere near the center of Moscow => ~near me [23:21:19] The Wikipedia page User talk:Catrope has been changed on 15 August 2011 [23:21:21] by Ryan lane, see http://test.wikipedia.org/wiki/User_talk:Catrope for [23:21:22] SWEET [23:21:24] So that works [23:21:38] \o/ [23:21:43] one more bug smashed [23:21:45] Yay [23:21:52] I think I won the BZ weekly report last week [23:21:55] RoanKattouw, I'm sure there are far too many enotifNotify jobs for testwiki [23:21:57] heh [23:21:59] I'll probably make top 5 again this week [23:22:07] I'll likely never even show up on that ;) [23:22:17] I remember my first week in the office [23:22:25] And seeing the report after flying home [23:22:33] Trevor and I completely took it over [23:22:47] This was before we had Reedy and hexmode being so prolific [23:22:52] So people getting >10 was uncommon [23:23:11] lolol [23:23:11] And it said Roan: 26, Trevor: 14, next person: <10 [23:23:20] Actually, I can go find it [23:23:37] I'm on 401 for the last 180 days [23:23:40] hexmode is on 240 [23:23:40] Win [23:23:52] Oh here goes [23:23:55] roan.kattouw [AT] gmail.com 23 [23:23:56] tparscal [AT] wikimedia.org 17 [23:24:04] Bugs RESOLVED : 107 [23:24:15] So just under 40% of all bugs closed that week was us [23:24:30] Fatal error: Call to a member function getNamespace() on a non-object in /home/wikipedia/common/php-1.17/includes/job/JobQueue.php on line 309 [23:24:31] Nasty [23:25:02] for UTC, I add 7 hours if I'm in PDT? [23:25:05] I hate time zones [23:25:13] Yes [23:25:16] thanks [23:25:16] PDT = UTC - 7 [23:25:20] *Ryan_Lane nods [23:25:24] I hate that too [23:25:30] I'm usually much closer to UTC [23:25:58] I'm often in random time zones [23:26:18] True [23:26:53] I can do conversions between CEST and PDT and between CEST and UTC, but when I have to convert between PDT and UTC it gets harder [23:27:21] When converting between CEST and e.g. EDT or CDT I frequently find myself converting to PDT first, overshooting my target, then working back from there [23:28:02] test has no jobqueue now [23:28:51] I just get insanely confused :) [23:29:13] I only really need UTC when I'm posting times to events