[00:05:04] robla: OK good luck with the deployment, I'm off to bed now. I'm glad I could fix three problems on a day I wasn't supposed to work :) [00:05:18] thank you RoanKattouw! [01:24:10] robla, I'm around in that I exist and am on the internet. Currently I'm doing web standards work (and some Gecko hacking) for Mozilla, and don't have any time for MediaWiki, if that's what you were asking, unless someone wants to ask something about code I wrote or something related to web standards. [01:26:01] hi AryehGregor - we were really looking for someone who could tell us if hewikisource is broken after we deploy 1.19, but we may not get to it today/tonight. thanks for getting back to me though! [01:26:27] robla, okay. If it's blatant brokenness that I could spot at a glance, I can look briefly if I'm available. [01:26:31] pgehres: when do you think we'll be ready to head out? [01:26:50] AryehGregor: cool, thanks! we'll ping you if we move forward [01:27:15] we're figuring out what to do about some out-of-memory problems right now on #wikimedia-tech [01:29:15] AryehGregor: how's mozilla going? I have some good friends from there [10:02:31] TimStarling: hi. I really like your comments "Despite the name, this MWDebug class doesn't handle debug logging." :-) [12:00:43] sumanah:will you please help me out in Gsoc ? [12:00:49] hi hber! [12:00:53] I'm happy to help. [12:01:07] So, you're interested in applying to GSoC, to work with MediaWiki? [12:02:15] sumanah:yes [12:02:22] hber: why MediaWiki specifically? [12:04:30] sumanah: mediawiki looks user friendly and I think it is developer friendly too though they are mutually exclusive [12:05:11] hber: well, we hope they are not mutually exclusive in our case! :-) [12:06:50] sumanah: so, I am in mediawiki [12:07:10] hber: what do you mean? do you mean that you have an interest in MediaWiki, or that you have already started working on MediaWiki? [12:08:10] hber: I'm curious -- what gives you the impression that MediaWiki is developer-friendly? I hope it is, but I'm wondering what made you think so. [12:08:13] sumanah: I started learning how todevelop extensions [12:08:17] great! [12:08:28] by the way, hber, how did you hear about us? [12:10:10] sumanah: I am following the tweets and blog posts of yuvi panda from last six months , I heard from him [12:10:25] oh cool! yuvipanda is an interesting and smart guy [12:12:05] hber: so, have you ever installed MediaWiki? [12:14:03] sumanah: I have installed mediawiki and started learning to develop extensions from the link "http://www.mediawiki.org/wiki/How_to_become_a_MediaWiki_hacker/Extension_Writing_Tutorial#Exercise_4:_Introducing_.24wgRequest" [12:14:36] sumanah: and I submitted the patch according to the procedure given in that link [12:14:43] Let me look. [12:15:15] hber: are you Bharath? [12:17:36] sumanah:yah? [12:17:56] hber: You tell me -- your "yah?" tells me you are not sure. [12:18:34] sumanah:sorry ,yes I am [12:18:53] hber: ok. Thanks. [12:19:15] (it's tough when all we have is text to communicate with -- little things like question marks take on more meaning....) [12:19:50] hber: so do you have an idea of what to try next? [12:20:08] hber: also, will you go to the Chennai hackathon? Chennai hackathon 17 Mar http://ur1.ca/82v84 [12:23:43] sumanah: I don't have any idea yet , and I heard about the hackathon today [12:24:20] hber: by the way, you can spread the word about MediaWiki and GSoC using these leaflets: https://www.mediawiki.org/wiki/Summer_of_Code_2012#Spread_the_word [12:24:38] hber: come into #mediawiki where there are more people [12:25:25] sumanah:may I know about any hackathon plans in Hyderabad? [12:25:41] hber: you could create an event if you want! [12:25:48] let's talk in #mediawiki [15:21:36] hashar: fancy breaking some wikis? :p [15:23:43] are you plotting 1.19 upgrades? *squint* [15:24:07] oh noes! [15:24:10] robla is up earlyt [15:24:41] I was like "oh no, is it Tuesday? am I about to miss a meeting?" because robla is up so early [15:24:51] (if it were, I would have a meeting with him in 6 min) [15:25:13] just checking in for a bit.....going to go talk to real estate agent and then be into work [15:26:45] Reedy: nothing spectacularly broken, I take it? [15:26:55] there seems to be a little javascript weirdness on meta [15:27:32] anyway, rushing off in a couple of minutes [15:27:42] I've not seen anything [15:27:54] enwikibooks and enwikiquote (iirc?) would give us a bit more coverage [15:28:46] alright....if you can scare up Amir, I'd say also do hewikisource. [15:28:59] make sure guillom or someone helps with edit notices [15:30:04] No sign of Amir at the moment [15:37:51] Boo, guillom isn't actually about [15:55:38] Reedy: sorry I did not catch your message earlier [15:55:46] "hashar: fancy breaking some wikis? :p" [15:55:50] heh, it's ok ;) [15:55:51] go ahead. I can break whatever you want [15:56:12] I got a bit less than 2 hours though [15:56:16] If Guillom was around, there's a few more we could 1.19 [15:56:34] RobH: [15:56:42] RobH: err sorry I mean Roan [16:03:28] PHP Warning: filemtime() [function.filemtime]: stat failed for /usr/local/apache/common-local/php-1.17/extensions/WikiEditor/modules/./images/toc/open.png in /usr/local/apache/common-local/php-1.18/includes/resourceloader/ResourceLoaderFileModule.php on line 380 [16:03:30] yeahhh [16:03:35] that is spamming the apache error log [16:04:01] not sure why it references a php-1.17 file [16:08:22] hashar: old cache entries that aren't supposed to be there any more [16:08:59] which cache is it? Cant we clean them up?? [16:09:34] we can [16:09:42] it's the resource loader database tables [16:09:46] Roan was going to do it at some point [16:09:50] But feel free to do it before! [16:10:00] i will just open a bug for him [16:10:17] cause I have no idea what cache it is nor how to flush it :D [16:10:30] truncate table? :D [16:10:46] shah I am not allowed to do that :-) [16:11:00] my insurer refuse to cover the TRUNCATE statement ! [16:11:05] s//filemtime/@filemtime/ [16:11:19] that happens each time we remove files [16:11:30] yeah [16:12:04] we could symlink 1.17 to 1.18, but that just feels icky [16:23:05] how do I authenticate with gerrit ? [16:23:44] OrenOf: with your Labs account [16:25:27] no certificate - just password ? [16:25:52] yup [16:27:24] what port ? [16:28:00] it crashes [16:28:16] it cannot get branches [16:30:22] git clone ssh://@gerrit.wikimedia.org:29418/some/path/to/repo.git [16:30:27] so port 29418 [16:30:33] you will want to install git-review too [16:30:39] using pip (Python package installer) [16:31:05] https://labsconsole.wikimedia.org/wiki/Git-review [16:31:36] hashar: can you add git-review details to the [[Git/Workflow]] document, please? [16:32:03] whenever I read Git/Worflow yes :-) [16:48:52] hashar: want to do frwikisource? [16:49:13] sure :-D [16:50:06] Best set sitenotice now and we can do it in a bit ;) [16:55:46] hashar: I do like the new diffs in CR [16:56:26] hexmode: your r93397 cause bug 34421 (duplicate headers in email) see https://www.mediawiki.org/wiki/Special:Code/MediaWiki/93397 / https://bugzilla.wikimedia.org/show_bug.cgi?id=34421 [16:56:52] I was afraid it was me :( [16:57:08] it is probably a trivial error [16:57:13] hashar: you're fixing? [16:57:18] you can talk about it to iAlex, I think he know that part of the code [16:57:55] well I suspected the bug to be either from iAlex or you so I just handled bissecting [16:58:03] might give it a shot later to night but not right now [16:58:09] oh, I can fix probably unless ialex is, thanks for tracking. [16:58:28] Reedy: I did a table for the diff to be able to insert comment bellow each line (ala gerrit ) [16:58:45] but, yeah, I've got some other stuff [16:59:02] Reedy: the commenting patch did not receive that much attention since we decided to move to gerrit and patch got reverted in january IIRC [16:59:16] hexmode: maybe tag it for platformeng? [16:59:27] will do :) [17:00:06] frwikisource is on 1.19wmf1 [17:00:37] that is fast migration! [17:00:52] do you have a step by step process to follow for wmf upgrades? [17:01:01] cause I might be able to help if you need [17:01:03] wikiversions.dat [17:01:06] find the dbname [17:01:11] change php-1.18 to php-1.19 [17:01:23] sync-wikiversions message here [17:01:48] eowiki, hewikisource and betawikiversity are left [17:01:52] can't do he without amir [17:02:13] so I can do eowiki ? :D [17:02:48] do you speak esperanto? ;) [17:03:05] yeah it is much like french with O and A added at the end of word [17:03:10] :D [17:03:19] jeo parla esperanto courramento [17:03:28] (I speak esperanto fluently) [17:03:52] brion does speak esperanto but he is hiding behind TWO underscores!! [17:04:09] is that to write the site notice? [17:05:06] nope, that needs doing manually [17:05:09] just did it for eowiki [17:08:30] see [17:08:36] I told you esperanto is not that hard! [17:09:40] I just got welcomed to frwikisource [17:09:42] wheee [17:11:59] Reedy: once sync-wikiersions is done, how do you get the database upgraded? [17:12:08] it's already been done [17:12:18] all changes are backwards compatible [17:15:52] hashar: going to do it? :D [17:17:40] for eowiki ? [17:17:45] sure. Lets try [17:18:04] !seen robla [17:18:04] --elephant-- I don't know anything about "seen". [17:18:08] hrm [17:18:24] he was there 2 hours ago [17:18:25] roughl [17:18:36] ya [17:18:38] I am not sure how long is his commute [17:18:53] istr one of the bots around here knows about !seen. no big deal. [17:20:22] eowiki php-1.19 [17:20:42] ya [17:20:44] yeah that file is not under subversion :D [17:21:08] $ sync-wikiversions 'moving eowiki to 1.19' [17:21:09] wikiversions.cdb successfully built. [17:21:09] Copying wikiversions dat file to apaches... [17:21:34] *hashar various apaches sending connection timed out [17:21:44] success http://eo.wikipedia.org/wiki/Speciala??o:Versio [17:22:02] so you have applied 1.19 schema changes to all wiki before the switches ? [17:22:40] Asher has been doing them all [17:22:45] adding rev_sha1 isn't quick ;) [17:24:54] thanks for training me :-?? [17:24:59] might be useful oneday [17:25:06] heh [17:25:14] it's all on the hetdeploy page on wikitech [17:26:55] ohhh [17:27:05] Orange (major ISP in France) is deploying deep packet inspection [17:27:23] they are monitoring all their customer content !! [17:29:34] Lovely... [17:32:13] the good internet is going to die :-( [17:56:49] So apparently inheritance in php doesn't work how you think it should [17:59:04] Reedy: is this an LSB rant? [18:13:18] 15 min stand-up in IRC in 2 min [18:16:38] hi folks! [18:16:54] *robla digs up the list of people to ping [18:17:26] hexmode: replag cleared up on s3 yet? [18:17:46] I just ran the rev report about 3 min ago [18:19:33] oh, I have a bug someone just emailed me about that looks pretty serious in a "this would be a good way for someone to put stress on the site" kind of way [18:19:34] today it's just preilly (who I believe is still out) and Aaron|away [18:19:52] https://bugzilla.wikimedia.org/show_bug.cgi?id=34451 [18:21:33] hexmode: is that new one? [18:21:55] robla: yep, just recorded it [18:22:18] hexmode: as in "was this new since 1.18"? [18:22:38] they use enwiki as an example, so I assume not [18:23:35] I'll see if the 1.19 wikis show the same problem [18:24:18] ok....let's make that a normal priority bug [18:24:19] ok, so, we've upgraded he.wikisource, fr.wikisource, eo.wikipedia, beta.wikiversity, en.wikiquote, and en.wikibooks [18:24:35] sumanah: is this today? [18:24:54] robla: I have just gone through to all of those sites' Special:Version pages and verified -- I presume it was today [18:25:10] "1.19wmf1 (r111194)" [18:25:14] funny ol' Reedy [18:25:41] robla: happened while you were at your appointment, I figure. [18:26:00] did those folks get banner notices? [18:26:14] I do not know. [18:26:43] I didn't see them on the sites when I checked [18:27:18] regardless, hexmode, are you checking all their Village Pumps? and who else is on the effort to look and listen for feedback from those communities? [18:27:43] so, I got some people to check village pumps for non-english wikis [18:28:04] next deployment: Tuesday, February 21 (to Feb 22), 23:00-03:00 UTC (3pm-7pm Pacific Time): MediaWiki 1.19 stage 2 deployment (commons) [18:28:05] and I'll be checking the english ones right after this [18:28:38] hexmode: could you list the people who are checking the non-English wikis? [18:30:03] robla: tomorrow is Tim, Nikerabbit, and awjr. [18:30:27] ? [18:30:33] Nikerabbit: 20% time [18:30:38] yeah http://wikitech.wikimedia.org/view/Server_admin_log [18:30:44] sumanah: the few people I've pinged are here: https://en.wikipedia.org/wiki/User:MarkAHershberger/Weekly_reports/2012-02-20 [18:30:49] robla: awjr is probably still going to be catching up from his trip; I don't know what Niklas and Tim should prep to do [18:30:55] brion__: ping? [18:31:04] sumanah: ^ server admin log shows when the changes happened [18:31:15] of course, duh. thanks Rob. [18:31:30] Reedy seems to be afk. I wonder who he handed off to [18:31:36] thanks hexmode. [18:31:44] hashar doesn't seem to be around either [18:31:53] dinner perhaps? [18:35:02] robla: ok, I know you have to scoot. I'll add a status to the 1.19/roadmap about today's rollout, if you don't mind [18:35:31] sumanah: fwiw I just got done talking to ironholds about how to focus on the task force [18:35:39] and steps going forward [18:36:19] *hexmode starts an email to phillipe and maggie as the first thing [18:37:09] so what's scheduled for today? (I'm looking at http://wikitech.wikimedia.org/view/Software_deployments) [18:37:23] no 1.19 rollouts :) [18:37:57] apergos: oh, rob h told me a bit more about letting people in _security [18:38:21] he said you can add them to the "invite not needed" list [18:38:32] so they can get in any time [18:38:36] sure [18:38:51] so, he did that for me :) [18:39:02] oh. I didn't realize you weren't! [18:39:03] :-( [18:39:32] but glad it's done [18:39:39] yeah, I thought you had set it up, evidently not. which is why I mentioned it. [18:39:44] :) [18:39:52] I did someone, I think Niklas [18:40:46] thanks sumanah [18:42:01] robla: done. [18:51:46] robla, sumanah: have any time today to chat about git and labs? [18:52:17] (this is really regarding account policy and access levels) [18:52:23] I do, over the next hour and then again in about 3 hours [18:53:15] Ryan_Lane: the next few hours look busy for him & for me, actually, with an open slot in about 90 min [18:54:09] posssssssibly. I'm still deep in 1.19 stuff, and have a lot of meetings today [18:57:03] *Ryan_Lane nods [18:57:07] it's no rush [18:57:15] we have a meeting with legal tomorrow to discuss some things [18:57:41] Ryan_Lane: do you already have, written down anywhere, "this is what I think the policy ought to be"? if I can read that, it'll make my life easier [18:57:48] I just wanted to make sure we're somewhat on the same page. [18:57:50] even if it's just indented bullet points [18:58:08] I do not. we'd discussed it some in the past. let me write something up really quick [18:58:25] I was going to write an email to engineering, but then decided it might be better to wait for legal [18:58:31] it's kind of chicken and egg [19:26:01] sumanah, robla: ok. I wrote an email to the engineering list. we probably don't need a meeting, I'd just like to know if this looks sane before I talk to legal about it (and other things) [19:27:24] thanks, Ryan_Lane [19:28:19] at some point we will need a unified glossary somewhere that explains: labsconsole, labs, project (Labs), project (git/gerrit), instances, bastion [19:28:47] in fact, I bet the lawyers will be happy if they can read such a thing tomorrow [19:29:16] kaldari: hope my list o' links email was helpful [19:29:20] morning K4-713 [19:29:33] yes [19:29:42] Good morning, werdna! [19:43:52] AaronSchulz: Hello :) Got a bug for you while you were sleeping https://bugzilla.wikimedia.org/show_bug.cgi?id=34440 [19:44:15] TrevorParscal: can we get some help on a resource loader issue? [19:44:22] https://bugzilla.wikimedia.org/show_bug.cgi?id=34450 [19:44:57] I think it's bad enough we should postpone any other deployments if we don't figure it out [19:46:58] robla: glad you pointed me that direction, then [19:48:04] back in 15 [19:49:03] *TrevorParscal talks to Roan about https://bugzilla.wikimedia.org/show_bug.cgi?id=34450 [20:09:54] hexmode: if you have a moment, could you double-check https://bugzilla.wikimedia.org/show_bug.cgi?id=33762 ? it says it's fixed in 1.19 but I don't think it is, I want to be sure I'm not mistaken. I believe http://meta.wikimedia.org/wiki/MediaWiki:Uncategorizedpages-summary should show something other than what it does. [20:11:44] chrismcmahon: I see it here: https://meta.wikimedia.org/wiki/Special:UncategorizedPages?uselang=qqx [20:13:13] hexmode: as I read 33762, MediaWiki:Uncategorizedpages-summary should exist but does not. [20:13:22] "MediaWiki:Uncategorizedpages-summary" [20:13:53] ah, I think I got it [20:14:10] yeah, compare enwiki with meta :) [20:14:18] got to go pick up my daughter [20:14:47] nope, still no https://meta.wikimedia.org/wiki/Special:UncategorizedPages-summary [20:35:21] hexmode: chrismcmahon: sumanah: I'd like chrismcmahon and hexmode to work together on whittling down our backlog of unprioritized bugs, and least catching all of them from this past week. Sound reasonable? [20:35:35] yes, sounds good to me [20:35:39] how many are there right now? [20:35:53] robla: I assume you mean, unprioritized bugs for MediaWiki core + WMF-deployed extensions, right? [20:36:01] sumanah: yes [20:36:06] hexmode stepped out for a bit I think [20:36:17] back [20:36:23] ohai :) [20:36:44] should be only 13 or so unprio bugs [20:37:05] chrismcmahon: that isn't talking about things tagged "enhancement" [20:37:14] which I don't prioritize [20:37:15] well, make sure you're looking at the whole list, not just a narrow query [20:37:26] it's too hard to get the query right without huge blind spots [20:37:36] robla: I've been doing that for a while now [20:37:42] robla: somewhat amusing to see that most (all?) of the fatals being reported seem to be 1.18 [20:38:05] "a while" = 4-6 weeks [20:38:42] Reedy: it's still quite possible that they are 1.19 deployment related, though, right? [20:39:43] chrismcmahon: oh, I also haven't been looking at the semantic mw extensions [20:40:27] *chrismcmahon fools with BZ searches [20:40:29] https://bugzilla.wikimedia.org/show_bug.cgi?id=34410 + https://bugzilla.wikimedia.org/show_bug.cgi?id=34417 are the only unprioritized non-enhancement bugs on WMF-deployed extensions [20:40:30] but those (!"enhancement" and !"*semantic*") are the only filters [20:40:53] sumanah: fixed [20:40:53] :p [20:41:08] chrismcmahon: [20:41:08] https://bugzilla.wikimedia.org/query.cgi?bug_severity=blocker&bug_severity=critical&bug_severity=major&bug_severity=normal&bug_severity=minor&bug_severity=trivial&bug_status=UNCONFIRMED&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&columnlist=priority%2Cassigned_to%2Cbug_status%2Cproduct%2Ccomponent%2Cchangeddate%2Cshort_short_desc&field0-0-0=component&field0-1-0=component&priority=Unprioritized&query_format=advanced&resolution=--- [20:41:14] robla: hexmode: chrismcmahon: I've shared a search in bugzilla that's also linked from https://www.mediawiki.org/wiki/User:Sumanah that is useful as a starter for these things [20:42:15] Reedy: clarify? [20:42:41] "fixed"? [20:43:00] it's now prioriti[sz]ed [20:43:12] sumanah: Last night I was thinking about starting a multi-lang volunteer task force to look at village pumps. Ironholds suggested pinging Phillippe and Maggie. Robla said I should ask you and Guillame to head up coordination. thoughts? [20:43:42] hexmode: yes, a multi-lang volunteer task force to report problems from Village Pumps is probably a good idea. [20:43:51] I suggest Guillaume to coordinate. [20:44:14] k, I'll talk to him and Phillippe and Maggie [20:44:21] Reedy: https://bugzilla.wikimedia.org/show_bug.cgi?id=34417 is not prioritized yet [20:45:19] sumanah: looks like a feature req to me: what priority do you think it should be? [20:45:47] hexmode: are you talking about Don't show the button on own userpage - https://bugzilla.wikimedia.org/34417 ? it's a moot point now, Reedy just prioritized it. [20:45:56] k [20:49:02] ok, robla, chrismcmahon, hexmode, I think there are currently 4 unprioritized non-enhancement BZ issues open for MediaWiki [20:50:00] I wouldn't mind a second set of eyes on the ones that have already been prioritized. I'd like to make sure we're not letting anything really important slip through the cracks [20:50:16] 34382, 34385, 34413, 34455 [20:50:47] chrismcmahon: if you get the chance to look at prioritized bugs, that would help... [20:50:57] 14 unprio overall [20:52:29] not sure I'm in a position quite yet to judge priority, but I'll certainly look. [20:52:51] also, I'm still not convinced that https://bugzilla.wikimedia.org/show_bug.cgi?id=33762 is actually fixed for 1.19 [20:52:54] robla: did brion mention this one to you? https://bugzilla.wikimedia.org/34385 [20:53:45] Hrm... are we responsible for Acawiki? I'll have to start looking at semantic mw (smw) [20:54:25] hexmode: SMW is not deployed on any WMF sites [20:54:50] (as far as I know) [20:55:13] sumanah: https://bugzilla.wikimedia.org/show_bug.cgi?id=34390 [20:55:27] It's on labs [20:56:12] labsconsole even [20:56:26] hexmode: http://acawiki.org/AcaWiki:Communications if you want to communicate with them [20:56:28] yes.... worry worry [20:56:39] AcaWiki is run by AcaWiki Inc., a 501(c)(3) non profit corporation incorporated in California, U.S.A. The project is grateful to have a seed grant from the Hewlett Foundation to get off the ground. AcaWiki was founded by Neeru Paharia. A small group of volunteers is working on the project, including Jodi Schneider, Mako Hill, and others. Also, Christopher Adams and Jon Phillips from Fabricatorz are growing the project actively and accelerating [20:56:40] development issues. (Feel free to add your name here!) [20:56:48] Reedy: how did Ryan_Lane get away with that? [20:56:51] http://acawiki.org/AcaWiki:FAQ#Who_is_running_this_project.3F [20:57:01] hexmode: it's not a large scale deployment [20:57:10] it's not working on a large dataset [20:57:20] hexmode: same as parserfunctions, it's fine until you go mad with it [20:57:31] What do you think enwiki would try if they had sm* available? [20:57:33] k, well, if you aren't freaking out about it, I won't [20:57:48] who knows [20:57:54] wait till they get LUA [20:58:12] hexmode: http://acawiki.org/Special:Version they're running MW 1.16.5 [20:59:06] ok, so, per robla's instructions, hexmode, chrismcmahon, checking prioritizations of the bugs that have come in this week is your current task. Right? [20:59:54] I thought robla said 1.19 and village pump patrol [21:00:09] robla: So about that 1.19 blocker [21:00:18] but if chrismcmahon can check my work, I'd be greatful [21:00:32] I think it's probably an infrastructural problem [21:00:39] hexmode: do you have a search link handy so we're on the same page? [21:00:52] yeah. labsconsole will never be a large enough wiki for SMW to cause performance problems [21:01:06] the one I sent above should work. 1s and I'll shorten it [21:01:45] chrismcmahon: http://hexm.de/ew [21:01:52] ew! [21:02:17] those are the currently unprioritized ones [21:02:21] *chrismcmahon tries to figure how to save a search in BZ [21:02:23] Hmph [21:02:25] 2012-02-16 21:01:14 mw27 enwikibooks: /wiki/Extraeditbuttons.js?action=raw&ctype=text%2Fjavascript Exception from line 897 of /usr/local/apache/common-local/php-1.19/includes/WebRequest.php: Invalid file extension found in the path info or query string. [21:02:29] That's a problem [21:02:47] ha, got it, or at least good enough [21:03:56] Oooooh [21:04:02] Some parts o f bug 34450 are IE7-specific [21:04:10] Thanks for that clarification chrismcmahon [21:04:36] robla: That blocker bug is an annoying one that I'll investigate, but I don't think it's a blocker [21:04:40] RoanKattouw: and the FF behavior looks definitely incorrect to me. [21:05:05] Well, you filed one ticket for three problems [21:05:12] So let me sit down at my desk and disentangle them [21:05:48] RoanKattouw: I actually questioned that at the time. Do you prefer that sort of report to be split out among different tickets? [21:06:06] If they're different problems, yes, unless you have reasons to believe they have the same root cause [21:06:10] When in doubt, file separately [21:06:22] RoanKattouw: in my mind it was one set of behaviors manifested differently in different browsers, but I'll remember that. [21:06:39] I personally don't mind duping tickets occasionally if that means I never have to split them [21:06:50] ack [21:06:56] Well, yeah, but arrow display is not related to tab order is not related to showing the Jump to: thing [21:07:53] But I'm glad you thought about it and I'm glad you're taking my recommendations :) [21:16:07] oh Bugzilla you so crazy: Request-URI Too Large [21:16:08] The requested URL's length exceeds the capacity limit for this server. [21:17:05] haha [21:18:03] chrismcmahon: <3 bugzilla [21:20:01] no unprioritized bugs :) [21:20:21] me starts CR FIXME nagging [21:20:33] how did we get revs with FIXMEs in them? [21:20:47] fwiw, I find 360 unresolved bugs since Jan 1 for MW/MW Extensions, any priority, any severity, not "semantic". [21:21:20] filter out enhancements [21:21:24] and the search is not saveable [21:21:26] ok [21:22:01] What do you mean "not savable"? Let me try.... [21:22:06] 225 bugs not 'enhancement' [21:22:32] "URI too large" [21:22:52] probably because I included all "Components" [21:24:33] heh, yeah, don't specify all of them, don't select any [21:24:43] and all will be found [21:26:24] https://bugzilla.wikimedia.org/buglist.cgi?chfieldto=Now&chfield=[Bug%20creation]&query_format=advanced&chfieldfrom=1-1-2012&list_id=90990&field0-0-0=component&type0-0-0=nowords&value0-0-0=semantic&resolution=---&product=MediaWiki&product=MediaWiki%20extensions [21:26:26] ah. https://bugzilla.wikimedia.org/buglist.cgi?cmdtype=runnamed&namedcmd=WM%20open%20bugs%20since%20Jan%201%202012&list_id=90989 [21:26:35] #380 bugs there [21:26:45] 225 in mine [21:27:40] yeah, I didn't keep enhancement out, so I believe you ;) [21:28:42] hrm... have to look at yours... I've got 241 [21:30:46] chrismcmahon: could you edit your search and then paste the url? I can't use un-shared named searches [21:32:00] one sec. [21:32:47] does this help? Creation date: (is greater than or equal to) 2012-01-01 [21:32:47] Severity: blocker, critical, major, normal, minor, trivial [21:32:47] Status: NEW, ASSIGNED, REOPENED [21:32:48] Priority: Highest, High, Normal, Low, Lowest, Unprioritized [21:32:48] Product: MediaWiki, MediaWiki extensions [21:32:48] Component: (does not contain the string) Semantic [21:33:08] hrm... yeah, let me see [21:33:24] Reedy: Can we MFT r111685 which is a minor OCD fix on the surface but might just possibly the the cause of the bug that's driving robla crazy right now? [21:33:28] !r 111685 | Reedy [21:33:28] --elephant-- Reedy: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/111685 [21:33:42] heh [21:33:45] yeah [21:33:59] heh [21:34:19] MFT? [21:34:49] merge from trunk (to a branch) [21:34:58] thx ;) [21:34:59] merge from trunk [21:35:42] it's basically cherrypicking a fix into our 1.19 branch for deployment [21:35:56] yep, just didn't know the acronym. [21:35:57] anyway: RoanKattouw: re whether 34450 is a blocker. it kinda depends [21:36:35] Chris, Erik and I have witnessed several strange intermittent javascript issues, mainly on meta [21:36:54] Most of them are probably infrastructural [21:36:54] *Reedy kicks formey [21:37:08] RoanKattouw: infrastructural how so? [21:37:15] I guess it so happens Chris filed a ticket with only the ones that have real causes [21:37:28] except that FF not tabbing to links or "Jump to" I'm pretty sure is a fairly heavy accessibility issue. [21:37:28] robla: Like some server somewhere randomly fails to give a proper response [21:37:45] Not a "real" bug but rather something weird happening in the cluster [21:37:48] RoanKattouw: except we're not seeing these in 1.18 [21:37:59] chrismcmahon: change "does not contain string" to "does not contain the word" [21:38:12] so...if 1.19 is more susceptable to "infrastructural" problems than 1.18, we need to fix that [21:38:14] hexmode: got it [21:38:23] chrismcmahon: Is the FF tabbing behavior a regression from 1.18? I'm extremely skeptical that that would be caused by an MW change [21:38:34] robla: It's probably not inherent to 1.19, more like the cluster itself being flak [21:38:36] y [21:38:38] But it's hard to say [21:38:42] that's a good question, looking. [21:38:46] *robla isn't buying it [21:39:03] We've seen weird shit before [21:39:29] But then again the issue Ian saw turned out to be a legit RL bug that was my fault [21:39:34] sure....we need to have a much deeper investigation before we chalk it up to cluster gremlins [21:40:07] Sure we need to be reasonably confident that it won't happen again [21:40:20] Done [21:40:30] But I maintain that we don't necessarily need to find the exact cause of each and every weird thing we see [21:40:36] Thanks Reedy [21:40:45] this is really really easy to repro for me and others [21:40:55] this isn't just a minor detail [21:40:57] I can't reproduce the IE7 thing [21:41:13] Well most of bug 34450 is a minor detail other than the a11y angle [21:41:16] so...I suppose we should break up the bug into several [21:41:25] So I'd love to hear what else you ran into that is a big deal [21:41:32] here's what I consider a blocker: edit bar dissappearing in Chrome [21:41:39] RoanKattouw: you're right, FF vs. enwiki also does not tab to links, Chrome does. (which surprises me, but not a regression) [21:41:41] Oh, yeah there was that too [21:41:53] chrismcmahon: Right, so I call that FF's fault [21:41:59] agreed [21:42:22] robla: In that case, yes, report bugs better. Split the report and provide developer console data [21:42:35] chrismcmahon: can you do that? [21:42:38] Even if there were no JS errors in your console, tell me that there weren't any [21:42:57] yep, look for disappearing edit bar on meta [21:43:09] Browser? [21:43:10] *robla sets about trying to repro again too [21:43:16] Does it disappear or never appear? [21:43:20] starting with Chrome [21:43:51] Edit bar WFM in FF [21:44:26] Hmm, now I have a JS error in IE [21:48:02] ok...got the edit bar problem [21:48:07] RoanKattouw: nothing on the console [21:48:19] In Chrome? [21:48:23] I have uncovered a JS error in IE [21:48:39] best way to repro in chrome: keep pulling up incognito windows and visit the edit url [21:48:42] http://meta.wikimedia.org/w/index.php?title=Main_Page&action=edit [21:48:57] duh...wait [21:49:00] eh? [21:49:03] You can't edit that page [21:49:12] So naturally there is no toolbar for a view source page [21:49:21] I'm using [[IRC]] as my test case [21:49:24] yeah, sorry....I'll keep trying to repro [21:54:14] morning TimStarling [21:54:20] morning [21:55:27] hey there Tim [21:55:38] *RoanKattouw throws stuff at IE [21:55:52] maplebed is writing up a plan of attack now on the swift stuff [21:55:57] *werdna throws stuff at mysql too [21:56:18] *robla throws stuff at ms5 for good measure [21:56:32] actually, I'll throw it at rewrite.py [21:56:32] robla: feel free to give Tim the tl;dr version! :) [21:56:36] I can reproduce this JS error reliably on meta, but not on localhost [21:56:49] And it's an error in the HTML parser so the debugger doesn't let me catch it [21:57:04] TimStarling, robla: just as a quick update, I have almost finished the module which adds some Unicode functions to Lua through ICU [21:57:23] oooooh, great to hear vvv! [21:58:26] *robla takes the swift conversation over to #wikimedia-tech [22:02:40] RoanKattouw: editing on meta is flat busted on IE7. get only an "IE cannot open" message attempting to edit my user page. [22:03:03] Yeah there's some issues here [22:03:12] I'm investigating but my Windows VM doesn't like me very much it seems [22:03:36] Also Erik reported another mw.user.options problem, I'm sick of that shit and I'm just gonna fix that properly now [22:04:30] even editing protected page kills IE7 [22:04:36] *chrismcmahon should have tried this earlier. [22:10:10] OK, I have a crazy idea [22:10:29] robla: I have the crazy idea that if I enable the experimental RL feature on meta, that'll *fix* the IE issue I'm seeing [22:10:45] TrevorParscal: ^ [22:10:46] Is there any wiki where we have this enabled at this point? [22:11:09] I do not feel like explaining why right now. It's stupid and I'll talk about it later [22:11:24] lolz [22:11:32] RoanKattouw: let's deploy to test2 first, since I don't think it's actually there yet [22:11:47] Let me see if the bug occurs on test2 too [22:12:29] *TrevorParscal holds back Roan from attacking meta [22:12:33] "it's not worth it man!" [22:13:14] No bug on test2 [22:14:30] robla: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/111427 [22:15:06] robla: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/111429 [22:15:10] RoanKattouw: I'd just like to see if we introduce *other* bugs on test2 [22:15:13] those two plus a few to clicktracking [22:15:52] OK [22:15:57] Well I'll enable it on test2 [22:16:06] And if it doesn't break anything there, I'll enable it on meta [22:16:08] Sound good? [22:17:16] RoanKattouw: yeah, make sure you coordinate with chrismcmahon on doing a very quick test pass [22:17:32] OK [22:19:02] RoanKattouw: my catastrophic case seems to be editing in IE7, so point me at an env when you like. [22:19:34] test2.wp.o [22:19:51] looking [22:24:59] chrismcmahon: Have you made test2 break yet? If not I'd like to make the same config change on meta [22:26:32] RoanKattouw: editing is working but I'm getting non-fatal "Access is denied" errors in IE7 on test2 [22:27:11] RoanKattouw: don't know what the access is though. [22:28:02] URL? [22:28:22] I wish I was as unlucky as you, I don't get that error in IE8-compat-mode-pretending-to-be-IE7 [22:28:57] URL is just the page url, User:Cmcmahon in this case [22:29:25] RoanKattouw: old leftover VM I never got rid of still has IE7 [22:30:12] OK, well it's working for me without the error [22:30:49] RoanKattouw: I reproed the no toolbar problem on test2 [22:30:55] *robla pulls up js console [22:31:18] *chrismcmahon checks a browser with decent js console [22:32:10] i have an idea (stand back) [22:32:26] no error on console [22:32:48] we could (probably) turn on the experimental system for logged in users only on meta [22:32:58] this would avoid populating anon read page cache [22:33:00] *robla futzes with prefs and retries [22:33:03] so it can be turned off instantly [22:33:11] if it explodes [22:33:12] I guess I had old toolbar on [22:33:12] I will hack meta to turn it on just for me [22:33:19] nice! [22:33:21] even better [22:33:57] no error in FF editing on test2, just seen in IE7. phooey. [22:39:43] Ryan_Lane: I have finally pushed the puppet linting script https://gerrit.wikimedia.org/r/2629 :) [22:39:59] hashar: Didn't Leslie already write one last month? (Asked her in the -operations channel) [22:40:30] there is the local-lint which use puppet internal validation [22:40:39] this one is different AFAIK [22:40:46] much like php -l versus CodeSniffer [22:41:09] this is for client side? [22:41:14] yes [22:41:17] ah ok [22:42:12] might become a Jenkins job one day :-D [22:43:26] brion__: around? [22:47:06] bed time for now. See you tommorrow [22:47:12] tomorrow [22:47:22] night, hashar \ [22:48:17] back in 15 or so [22:58:30] Reedy: I can haz deployment for https://www.mediawiki.org/wiki/Special:Code/MediaWiki/111695 and https://www.mediawiki.org/wiki/Special:Code/MediaWiki/111697 ? [22:59:06] js error on simple? https://simple.wikipedia.org/wiki/Wikipedia:Simple_talk#Issues_with_Navigation_Popups.3F [23:00:35] chrismcmahon|brb: When you're back, could you see if meta in IE7 is fixed? Please test both logged in and logged out. I gave your account a special setting when you're logged in [23:00:41] RoanKattouw: for what it's worth.... [23:00:54] it might make sense to deploy to all logged-in users on meta [23:01:14] hexmode: That bug report needs more details, right now it's just "it doesn't work" [23:01:22] robla: I know, but I'm lazy [23:01:30] I can add you if you live [23:01:30] noooooooooooooo [23:01:32] conflicts [23:01:32] *ike [23:01:36] rawr [23:01:38] RoanKattouw: I agree... on it [23:01:39] I'll do it myself then [23:01:46] Reedy: I'll do it myself then [23:01:52] RoanKattouw: it's alright [23:01:56] It's just svn being stupid [23:02:02] Oh OK [23:02:14] lemme glance over the merge diff before you deploy though [23:02:22] even if things get better, I'd prefer to leave it for just logged-in users for a while, per conversatoin I had earlier with TrevorParscal [23:02:43] right [23:02:57] Wouldn't you first want it tested though? :) [23:03:07] r111699 [23:03:14] !r 111699 [23:03:14] --elephant-- http://www.mediawiki.org/wiki/Special:Code/MediaWiki/111699 [23:03:30] RoanKattouw: sure....yes, testing is good :) [23:03:45] - })() [23:03:47] + }() ), [23:03:48] That's a problem [23:04:00] That line should add a comma and nothing else [23:04:24] *RoanKattouw fixes [23:05:57] !r 111700 | Reedy [23:05:57] --elephant-- Reedy: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/111700 [23:06:16] *RoanKattouw deploys [23:07:03] i svn up'd the rl php files [23:07:14] Thanks [23:09:10] rawr [23:09:17] set-group-write takes forever [23:11:39] *RoanKattouw wonders why the hell the edit toolbar is screwed up in IE [23:11:51] It seems to just not be styled or something [23:12:19] RoanKattouw: the simplewiki problem WFM... asking for browser info [23:18:43] OK, toolbar is WFM in meta [23:18:58] Not in debug mode but that's sort of expected due to the stylesheets limitation bug in IE I guess [23:19:51] actually, when did secure.w.o go away? [23:20:07] *hexmode looks at at the log [23:20:10] I have no idea [23:20:13] i think it started having issues today [23:20:16] It wasn't supposed to have gone away yet was it? [23:20:18] I'm sure Ariel was looking at something [23:20:19] oh, issues [23:21:16] OK, so the experimental loading stuff along with what Reedy deployed fixes the issue for me [23:21:33] If chrismcmahon can confirm that, I'll enable experimental loading for all logged-in users on all 1.19 wikis [23:21:37] *chrismcmahon checks meta for IE7 [23:21:38] robla: --^^ [23:22:01] let's do it [23:22:45] Compare logged-in vs logged-out, too [23:23:12] I get functional JS with a JS error when logged out and no JS error when logged in [23:23:28] (this is because I've configured meta to behave differently if the user is called Catrope or Cmcmahon) [23:24:03] Reedy: RoanKattouw: k, so the simplewiki directly refers to secure.w.o -- I'm sure that is what they problem is [23:24:10] hexmode: Probably [23:24:31] I can't reproduce the "Jump to is visible in IE7" issue BTW [23:24:38] Can anyone else still? [23:24:48] RoanKattouw: edit on meta IE7 succeeds when logged in, fails completely when not logged in. [23:25:00] tour of village pumps pays off on first hit [23:25:10] *hexmode doesn't have IE7 [23:25:13] (and I still have the non-fatal access error on edit when logged in) [23:26:03] That's weird [23:26:14] Does the non-fatal access error go away when you do a hard refresh? [23:26:21] going to see how you're supposed to load scripts from other wikis. they're doing it wrong [23:26:37] There's not really a nice way to do it [23:26:54] importScript? [23:26:56] cleared cached, edit still fails when not logged in meta/IE7 (this VM is really slow, too) [23:26:58] But for the least bad way, see the HotCat gadget on enwiki [23:27:19] Yes, but does the permission denied error go away after a cache purge? [23:27:22] *hexmode lols at "least bad" [23:27:25] (when logged-in) [23:27:29] RoanKattouw: checking... [23:29:44] RoanKattouw: I'm getting the Access error all over, not just editing. wish IE7 told me more about it. [23:30:11] Hmm, of course IE7 doesn't have the debugger [23:30:20] I saw the issue in IE8's compat mode once [23:30:29] e.g. shift-reload on my user page gives the error [23:30:32] And changed something that I thought would fix it [23:31:48] do not get error loading main page, but do when I navigate to my user page. [23:32:17] at least in a new IE7 process [23:34:56] hmm, ok [23:35:03] What if you navigate to pages on meta you hadn't visited before? [23:37:59] navigated to /Meetup, which this browser has never seen, got "Access denied" error [23:38:50] Where did that error appear? JS error? [23:40:39] hard to tell with IE, I just see the modal box saying "Problems... Line 51/Char 3/Error: Access is denied/Code: 0/URL: http://meta.wikimedia.org/wiki/Meetup" [23:40:43] hexmode: are you about? [23:41:01] werdna: hey [23:41:21] hexmode: I wanted to ask about the deployment-prep stuff you have going [23:41:27] chrismcmahon: Do you get the same error if you add ?debug=true ? What's the file & line number in that case? [23:41:40] werdna: ask away! [23:42:13] I've just set up a new prototype server for editor engagement (mosty to avoid treading on too many toes over here) [23:42:30] but I understand you guys have put a fair bit of effort into replicating the production setup on deployment-prep [23:43:01] yes... still needs squid work, but yes [23:43:26] werdna: petan is a good person to ask about that [23:43:27] I'm wondering 1) If I can get access to deployment-prep to see what you have set up; 2) if it makes sense for editor engagement to have its own wiki/prototype within your setup; and 3) if not, if what you have is easy to replicate in the editor-engagement project [23:43:30] chrismcmahon: Specifically, http://meta.wikimedia.org/wiki/Meetup?debug=true [23:43:56] werdna: I can give you access and I think it would be ok to use it [23:44:05] RoanKattouw: yes, exactly the same error when ?debug=true [23:44:09] OK [23:44:11] Line number? [23:44:15] okay, should I ask petan first? [23:44:15] (from the debug=true error) [23:44:20] werdna: I want to start tracking svn head again though [23:44:29] Line: 51/Char:3 [23:44:39] werdna: yes just to make sure we're on the same page :) [23:44:46] okay [23:44:50] what're his usual work hours? [23:44:56] CET [23:45:22] so it's 1am petan time, I should ask tomorrow morning :) [23:45:59] hey [23:46:01] ooooooooooooooooooooooooooooh [23:46:09] Line 51: [23:46:10] [23:46:15] werdna: yes [23:46:18] oooooh :) [23:46:21] it's the frigging htc file [23:46:27] *TrevorParscal hides [23:46:35] petan: to "ask tomorrow"? [23:46:42] no problem [23:46:58] *werdna isn't sure what is replying to what [23:47:12] no idea what you want to ask about [23:47:15] OK that's probably harmless then [23:47:25] if it's going to take a long time, we can handle it tommorow if u want [23:47:27] I'm gonna enable experimental async loading for all logged-in users now [23:47:29] RoanKattouw: what's harmless? [23:47:51] The "access denied" error that seems to be coming from something related to the htc file [23:47:57] werdna: you only need shell there? [23:47:58] ah [23:48:01] It's harmless compared to the other crap we've been dealing with today [23:48:03] petan: oyment-prep to see what you have set up; 2) if it makes sense for editor engagement to have its own wiki/prototype within your setup; and 3) if not, if what you have is easy to replicate in the editor-engagement project [23:48:07] oops [23:48:10] stop talking :p [23:48:18] petan: 23:43 < werdna> I'm wondering 1) If I can get access to deployment-prep to see what you have set up; 2) if it makes sense for editor engagement to have its own wiki/prototype within your setup; and 3) if not, if what you have is easy to replicate in the editor-engagement project [23:48:29] RoanKattouw: just realize, that HTC file is what enables hover in older IE and that means the vector tools sub-menu [23:48:31] RoanKattouw: harmless, agreed, doesn't seem to interfere with actions. [23:48:33] I know [23:48:37] werdna: you have access to labs? [23:49:03] *TrevorParscal is satisfied having confirmed he's not the only one who knows that [23:49:22] petan: I don't much care what level of access it is, the objective is to have a place where we can install extensions that editor engagement are working on so we can test them on a production-like environment and invite public testing [23:50:18] ok, problem is that I don't really know what all kinds of extensions we are supposed to have there, as far as I know, the purpose is to have identical environment to production to test only new releases of stuff we have installed on prod [23:50:28] however I see no problem in testing new stuff there too [23:51:05] request account on labs and I give you shell access there with root, so that you can use it as you need, there is nothing to break [23:51:26] *TrevorParscal heads out to catch the last bart train that allows bikes [23:51:51] petan: well, ideally we'd have somewhere where what we do doesn't conflict with what you do [23:51:55] people already use that site and registered there, so it makes a lot of sense to extend the purpose of testing cluster [23:52:01] i.e. an extra wiki [23:52:08] that's no problem [23:52:26] we have 400+ wikis there, creating one more isn't really hard [23:52:58] however as I understand it, you only need to install some extension there, which is probably ok [23:53:11] nod. [23:53:18] customizations of mediawiki core might be problem [23:53:18] my labs account is werdna [23:53:24] k [23:53:29] um, we have a rev or two to merge into trunk [23:53:34] they're all reviewed [23:53:35] but yeah [23:53:40] we don't use trunk [23:53:44] 1.19wmf1 [23:53:46] well yeah [23:53:52] to merge *from* trunk :) [23:54:00] ok [23:54:13] added [23:54:27] so that might be an issue.. [23:54:34] (the merging from trunk) [23:54:50] so there are several instances, you can open http://nagios.wmflabs.org/nagios3/ all deployment instances are listed there [23:55:09] there are 3 web servers, so changing config of apache needs to replicated to all of them and one squid [23:55:21] all files are stored on nfs and mounted to /usr/local/apache [23:55:33] mediawiki is configured just as on prod [23:55:49] so config is somewhere in /usr/local/apache/common-local/wmf-config [23:56:00] are they all prefixed with deployment- [23:56:10] there is some info on [23:56:11] http://labs.wikimedia.beta.wmflabs.org/ [23:56:23] however if you break the site, the info is hard to access :) [23:56:40] :) [23:56:53] yes they are all prefixed you have access all except backup one [23:57:23] in case you accidentaly rm -rf all instances, let me know and I restore them from -backup ;) [23:57:29] heh [23:57:32] I'll try to avoid doing that [23:57:37] heh [23:57:38] ok [23:58:18] test.wikipedia is probably best for you atm, having extra extension there is no problem [23:58:25] but if you need I can create another wiki just for you [23:58:44] what box should I ssh to? [23:58:55] dbdump is perfect for maintenance and such [23:59:01] because loading it doesn't matter [23:59:20] okay, and to clarify, I have two core patches. [23:59:22] what can I do with those? [23:59:23] you can start maintenance scripts there and change all configs [23:59:35] mediawiki is in common/live