[00:00:25] yeah, i think that's it. i'll have a fix up in a moment. [00:00:41] thedj: fixing the CN problem... [00:02:35] ori-l: stall simply seems not to be defined for me [00:05:15] thedj: yes, it's not specified as a dependency by AFTv5 [00:05:33] mlitn: yt? [00:06:06] thedj, ori-l: CN problem should be fixed now [00:08:39] ori-l: E3 currently deploying? if not, can I sync two extensions (FundraiserLandingPage and ContributionTracking)? [00:09:18] pgehres: go ahead, we're done, just fixing an unrelated bug. [00:09:22] cool, thanks [00:14:21] pgehres: let me know when you're done; have a fixup for aftv5 to sync. [00:14:34] ori-l: just finished, thanks [00:14:35] thanks. [00:21:27] thedj: could you check now? [11:03:11] !log Created WikiLove tables on urwiki [11:03:24] Logged the message, Master [12:37:30] AaronSchulz: https://upload.wikimedia.org/wikipedia/commons/thumb/a/a4/Test-bug.svg/500px-Test-bug.svg.png [12:37:45] we have new squid error messages ? [12:41:16] so pretty [12:41:56] and not protocol-relative [12:42:18] chromium HTTPS users won't see the octopus /o\ [12:42:49] there's no https version ;) [12:43:24] hmm [12:44:04] fixed [12:44:19] indeed. how does that work ? :D [12:44:36] we had a failed disk in one of the backend squids [12:44:50] someone replaced that and reprovisioned the box [12:45:10] but we've changed the default distro for new installs to precise [12:45:35] and precise has squid3, which a) has that new error page, b) does not work with our configs at all [12:49:43] iow, a messup from our side, apologies. [13:29:33] Hey all. At an editathon, so can't track down very well (may have already been reported) [13:29:39] but in IE I'm getting an exception in http://bits.wikimedia.org/static-1.21wmf2/extensions/EventLogging/modules/ext.EventLogging.js [13:32:33] Line 115 ("if ( desc.enum && desc.enum.indexOf( val ) === -1 ) {" by my reckoning) [13:33:09] It seems to be cascading, taking out all JS for me [13:33:38] ori-l perhaps? [13:36:35] Or spagewmf maybe? [13:47:34] I don't know if that affects IE9 as well, but in any case I would extrapolate that 13% of users are no experiencing no JavaScript :/ [13:48:00] Anyone else got IE7/8/9 available to test [13:48:15] ? [13:49:40] Okay, IE9 seems to work on my other laptop [13:51:00] That's a relief [13:57:55] aww thedj[work] edit conflicting me on svgs [14:03:14] chrismcmahon: ping [14:03:29] hi Jarry1250 [14:03:54] Do we routinely test JS changes in IE8? [14:04:31] chrismcmahon: ^ [14:05:14] Jarry1250: do you have an example? I have a lot of versions of IE available. [14:05:45] chrismcmahon: Yes, well you neatly avoided my rant about how I think all JS is currently broken in IE8 :) [14:05:55] (based on 2 datapoints) [14:06:26] Jarry1250: define "all JS" for me? [14:08:43] chrismcmahon: Um, I get an error in ext.EventLogging.js which then cascades [14:08:58] taking out the editing toolbar, all my user JS [14:09:13] The "My sandbox" link [14:10:45] Jarry1250: this happens when you edit a page on enwiki? [14:10:59] chrismcmahon: Yup, in IE8, logged-in, not logged-in [14:11:06] Strcit mode or compatbility mode [14:11:40] But not in a Firefox install on the same hardware [14:20:19] chrismcmahon: WORKSFORYOU? [14:21:08] Jarry1250: agreed, I see the same thing, although I can't see the whole error text on my IE8. EventLogging went out late yesterday. I checked Bugzilla and don't see any bugs reported, I'll do that right now. [14:25:12] Nemo_bis: he nemo :D [14:25:42] hmm, i see you are categorizing the old broken images. in that case, you might want to hit my user contribs of today :D [14:27:00] thedj[work]: I think I opened the category before you started :p [14:27:10] but I will [14:28:17] thedj[work]: is https://commons.wikimedia.org/wiki/File:Test_only.svg ugly on purpose or broken? :) [14:30:34] broken. all 6 cases should draw the same [14:30:44] ok [14:32:34] chrismcmahon: sorry for the late response, at an editathon. I think the problem is "if ( desc.enum && desc.enum.indexOf( val ) === -1 ) {" [14:35:06] Nemo_bis: https://commons.wikimedia.org/wiki/File:20th-century-concert-dance-vectorise.svg that one, it also looks broken to me in Chrome... [14:35:06] think it might be just an authoring mistake... [14:36:57] thedj[work]: yes, same in FF [14:37:40] Jarry1250: I created https://bugzilla.wikimedia.org/show_bug.cgi?id=41428, thanks [14:38:00] Jarry1250: assigning that now [14:40:12] k, basically, i went trough all of the rsvg bugs and the entire category with buggy renderings. quite a few fixed, and some issues further pinpointed now that we had some of the more 'noisy' problems out of the way. [14:42:12] thedj[work]: I'd be interested to have a comprehensive-ish list of problems remaining in rsvg (tracking bug?) so I can update my svgcheck tool :) [14:43:00] thedj[work]: we seem to have several cars and chemicals fixed :) https://commons.wikimedia.org/wiki/Category:Pictures_formerly_showing_a_librsvg_bug [14:43:03] https://bugzilla.wikimedia.org/show_bug.cgi?id=8901 [14:43:26] Jarry1250: the font size issues are all gone, that's nice. [14:44:02] thedj[work]: It assumes pixels now? [14:45:12] (That was the issue before, wasn't it?) [14:45:15] chrismcmahon: Replied there [14:46:10] Jarry1250: yeah before it only understood 18px. Now also relative sizes work (although i have the feeling that in inheritence situations sometimes there area some small errors) [14:46:39] thedj[work]: Oh, good. No news on flowRoot ("black box syndrome") I take it? [14:46:47] still broken [14:46:55] as is x with multiple coordinates [14:47:29] thedj[work]: are you sure this is fixed? https://commons.wikimedia.org/wiki/File:Hierarchisches_Modell.svg [14:47:33] the latter affects the most images atm I think [14:48:08] Nemo_bis: i see the lines. (and didn't before my cache purge) [14:48:53] thedj[work]: I don't at 200px [14:49:16] incomplete purge? [14:49:54] maybe... 199 and 201 work [14:50:50] purging there is not working due to https://bugzilla.wikimedia.org/show_bug.cgi?id=41130 most likely then. [14:51:42] :/ purging a dozen more times didn't help [14:51:42] the 200px version will have to expire on age then :( But the root cause seems fixed to me if i try all the other sizes in that area. [14:52:05] Nemo_bis: no, there is no thumbnail on the system anymore, so it doesn't know what to purge from squid [14:52:20] annoying bug. [14:54:26] Jarry1250: oh, order of nested tspans is fixed as well btw. also an important problem we saw in the past. [14:58:34] thedj[work]: Ah yes, that was a particular odd one, and actually one that could have affected the development of translatesvg [15:00:21] chrismcmahon: i'm also seeing lots of reports of broken sorting, missing hotcats and all other kinds of scripts being broken on en.wp right now. [15:01:14] thedj[work]: scripts loaded from other wikis? [15:01:44] à la https://en.wikipedia.org/w/index.php?title=MediaWiki:Common.js&diff=prev&oldid=519826982 [15:03:13] Nemo_bis: no, that's not related. [15:04:29] chrismcmahon: Can we not rollback last night's change? Or just comment those lines? [15:07:11] Reedy: ? there seems to be an issue with EventLogging extension (possibly IE only) [19:26:22] http://blog.archive.org/2012/10/26/10000000000000000-bytes-archived/ ! [19:29:08] Yeah, there was a party at EFF last night. Some Wikimedians went (I couldn't make it, unfortunately), and I hear it was awesome [19:32:13] oh [19:38:24] aude: What was that other revision? [19:38:36] Reedy: the bug fix? [19:38:43] * aude brings up gerrit [19:38:44] yeah [19:39:40] ori-l: so....theoretically we could redeploy EventHandler at any time [19:39:56] https://gerrit.wikimedia.org/r/#/c/30113/ [19:40:09] ori-l: might be wise to wait until Monday though [19:40:09] wikipedians noticed that one [19:41:17] oh wait, scratch that [19:42:21] robla: sorry, was the "scratch that" directed at me or the other bug? [19:42:32] rather, to aude re: the other bug. [19:43:17] ori-l: I was thinking the fix was already +2 and in [19:44:13] * robla needs to get food [19:44:43] because wikipedians noticed, i'd say backport that one (since it'll be week and half until they get the next deployment) [19:45:02] otherwise, i think that's the last thing we'll need to backport for contenthandler stuff [21:53:15] what is the problem with this file: http://commons.wikimedia.org/wiki/File:Astronomische_Beobachtungen_an_der_k._k._Sternwarte_zu_Prag_4_79_2.jpg the error says invalid url [21:56:04] same with _97.jpg [21:56:27] http://upload.wikimedia.org/wikipedia/commons/8/8e/Astronomische_Beobachtungen_an_der_k._k._Sternwarte_zu_Prag_4_79_2.jpg says invalid url [21:56:42] It would seem the source file is missing [21:57:06] opens fine for me (29x29px jpg). [21:57:43] Hmm [21:57:47] You're still in the US, right? [21:58:22] yes [21:58:45] Based on enomils hostmask, this is potentially a esams squid issue [21:59:37] <- germany [22:00:00] apergos: paravoid: about? [22:03:33] Source file is definitely on /vol/originals [22:07:39] http://commons.wikimedia.org/wiki/File:Astronomische_Beobachtungen_an_der_k._k._Sternwarte_zu_Prag_4_97.jpg is a little bit stranger, thumbs seems okay, full size file not [22:09:53] zzz [22:10:14] there was a broken squid deployed earlier today, this looks similar to that problem this morning [22:11:09] "12:46 paravoid: stopping squid3(!) on sq82, box was reprovisioned with precise but we don't support squid3/squid on precise yet" [22:11:29] this too looks like squid3 html when i open the broken url [22:12:55] MISS from localhost, HIT from amssq57.esams. [23:14:22] Hello. [23:14:22] Bugzilla seems to be down. [23:14:22] https://bugzilla.wikimedia.org/show_bug.cgi?id=41437 isn't loading for me. [23:14:23]

PROBLEM - HTTP on kaulen is CRITICAL: CRITICAL - Socket timeout after 10 seconds [23:14:23] same for ssh [23:14:23] It's self evident in #wikimedia-operations [23:23:23] There are people who don't have nagios-wm ignored? [23:23:23] Yes [23:23:23] Well that's your problem [23:23:23] ;) [23:23:23] I like how Google Chorme suggests a bunch of things to do. [23:23:23] But none of them is "contact the Web site operator." [23:23:24] > Add Google Chrome as a permitted program in your firewall's or antivirus software's settings. If it is already a permitted program, try deleting it from the list of permitted programs and adding it again. [23:23:24] Useful. [23:23:25] Yep, nagios just brought up kaulen problems again [23:23:25] again suggests they were fixed ;) [23:23:25] Chrome * [23:32:33] gn8 folks