[00:10:20] legoktm: Don't use negative configuration names. :-) [00:10:28] And that seems very specialized. [00:11:20] I guess, but those log entries are pretty useless if RC patrol is enabled [00:12:08] I mean you could do something like... $wgUnloggedActions = array(); or something. [00:12:24] But performing an action without a log entry seems pretty against God to me. [00:12:30] Er [00:12:32] These are edits. [00:12:44] The edit by default is patrolled [00:12:54] Right... are you patrolling the edits, though? [00:13:06] You can just disable patrolling altogether. [00:13:07] No, these are auto-patrolled. [00:13:12] Patrolling is useful! [00:13:19] Auto-patrolling is still patrolling. [00:13:42] It's useful when you're making 500 edits a minute with a bot? [00:13:43] Susan: I think what we really need is a "patrolexempt" [00:13:57] If you're patrolling an edit, it should have a logged action. [00:14:03] Otherwise, there's no need to patrol it. [00:14:10] Yes it should, but not if it's patrolled by default. [00:14:13] You either have accountability or you don't... [00:14:14] anyone on here an admin for the Wikimedia github account? [00:14:34] ^demon|away probably. [00:14:48] If I patrol an edit that wasn't autopatrolled that should generate a log entry. But if a user makes an edit that gets autopatrolled, there's no need for a log entry on that. [00:15:09] How do you know the edit has been patrolled, then? [00:15:35] I don't remember this part of the code very well, but I believe patrolling _is_ the log action. [00:15:39] No its not [00:15:44] Where's it stored? [00:15:47] it's the rc_patrolled column [00:15:55] in the recentchanges table [00:15:57] So you can only patrol recent edits? [00:16:03] Yes. [00:16:10] But the log action is supposed to be permanent. [00:16:16] Yeah. [00:16:24] So past 30 days, you wouldn't know if an edit had been patrolled or auto patrolled. [00:16:49] Well if there's a patrol log entry for it, you know it was patrolled. [00:17:13] You can also check the user rights log to see if/when the user was given the autopatrol userright [00:18:22] Perhaps. [00:18:39] Is patrolling being used? [00:18:45] Yes [00:18:46] I'm not sure why you wouldn't just disable it. [00:18:52] Who/what is using it? [00:18:55] I use it and I know a few others who do :P [00:19:00] On all edits? [00:19:10] I doubt the wiki is patrolling that many edits per day... [00:19:17] Yeah, its not that hard if you use Krinkle's tool [00:19:37] Well we normally have ~500 epm [00:19:45] So that's ~490 log actions per minute. [00:28:41] Ryan_Lane: do you have admin access to the Wikimedia github account? [00:28:55] yes, but only so that I can remove people [00:29:23] could you add me to it? [00:29:50] or know someone who could? [00:30:08] apergos: what's the latest on your query? try running from scratch on a slave with slave SQL thread stopped? [02:52:14] Susan: well if patroling is causing issues, just bug Reedy to turn it off :p then ask the community if they want it back on and how they want it setup [02:54:51] uh [02:54:52] no [02:54:55] we need it on [02:57:04] need or want? [02:59:25] need [03:05:15] The way Kathie Lee needed Regis. [04:29:13] jeremyb_: don't know the latest, I was in bed at 3:30 am so... [04:29:17] (when you replied) [04:30:00] good time to be in bed [04:34:04] I'm pretty sure that if there s no other activity on the db it will run at regular speed [04:34:23] the only other thing going on on db73, aside form a very occasional other query, was replication [04:34:39] so that was enough to make it unhappy, and progressively more unhappy over time [04:41:51] is someone able to explain to me the criteria for small <-> medium <-> large for WMF's wikis in dblist? [04:42:06] http://noc.wikimedia.org/conf/ [04:45:34] apergos: well i wanted to be absolutely sure replication was the cause... [04:46:21] sDrewth: some might be for dump priority... apergos would know. idk exactly though. you should check the history of those files [04:46:28] No. [04:46:37] No? [04:46:45] No to them being for dumps. [04:46:48] They're related to https://bugzilla.wikimedia.org/show_bug.cgi?id=39667 [04:47:48] https://gerrit.wikimedia.org/r/#/c/33694/7/makeSizeDBLists.php [04:48:10] thx mzm [04:48:38] https://bugzilla.wikimedia.org/show_bug.cgi?id=46098 [04:49:47] ewww, why do people use side by side? [04:50:35] It's there? [04:50:44] unified > side by side [04:50:55] sDrewth: So a small wiki is a wiki with less than 10,000 pages. [04:50:56] what kind of breakage do you mean for trailing new line? [04:51:12] A medium wiki is a wiki with less than 1,000,000 pages but more than 10,000 pages. [04:51:15] Large wikis are the rest. [04:51:22] k, thx [04:51:23] jeremyb_: It breaks shit, as I recall. [04:51:25] nice [04:51:35] Susan: tell me more [04:52:29] I don't remember. [04:52:34] I thought it was a rule. [04:54:08] and Susan that explains to me which tests are killed, and why [04:54:13] so bonus :-) [04:54:18] no, the dblists' definitions of small medium and large are separate from my lists [04:54:31] jeremyb_: Maybe I'm crazy. [04:54:41] Susan: ^^^^ DUH! [04:54:49] * sDrewth runs [04:58:45] jeremyb_: it's related to replication *or* to the specific db activity (level, type) that is happening on master, ie. I expect though can't prove that if this query ran onn master I'd see the same behavior [04:58:58] not going to test it out though :-P [05:01:18] https://wikitech.wikimedia.org/wiki/Server_admin_log/Archive_16 looks blank. [05:02:08] Oh, here we go. [05:02:12] jeremyb_: > 21:48 RoanKattouw: Removed empty line from the end of all.dblist, broke account creation ^^ [05:02:37] https://wikitech.wikimedia.org/wiki/Server_admin_log/Archive_15#February_25 [05:02:40] I think that was it. [05:02:55] wtf [05:03:35] I think that was about \n\n, though. [05:04:09] probably [05:06:00] $ for dblist in small.dblist medium.dblist large.dblist all.dblist; do printf '%-15s' "$dblist: "; tail -c 4 "$dblist" | xxd; done [05:06:04] small.dblist: 0000000: 6172 790a ary. [05:06:06] medium.dblist: 0000000: 7263 650a rce. [05:06:09] large.dblist: 0000000: 6e61 7279 nary [05:06:11] Susan [05:06:13] all.dblist: 0000000: 6172 790a ary. [05:06:22] that's HEAD. so should be what's live now [05:07:55] I'm not sure what your paste is demonstrating. [05:08:10] well 0a is the same as \n [05:08:25] so 3 of those 4 have a trailing \n [05:08:33] 1 is missing [05:08:45] Scandalous. [05:09:12] anyway, you may want to update your gerrit comment [05:10:01] Not really. [05:11:16] * jeremyb_ wonders why [07:38:36] anyone happen to know why the "add image to this article" on mobile doesn't show up for all articles that lack an image, e.g. https://en.m.wikipedia.org/wiki/Coastline_%28sculpture%29 [07:39:36] edsu: errr, i thought you're in my timezone but now i'm not sure [07:39:49] jeremyb_: well :) [07:39:51] anyway, maybe try #wikimedia-mobile or try a more normal hour :) [07:40:31] does https://wikipedia-irc.herokuapp.com/ work for you? i've been waiting for some hours :P [07:40:35] i did try wikimedia-mobile ; but will try at another time [07:41:12] so i see [07:41:26] you waited 10 mins before repeating here :) [07:41:47] jeremyb_: seems to be working for me [07:41:57] it's been 7 hours since any other human besides you spoke in #-mobile [07:42:16] after waiting how long? [07:42:24] jeremyb_: 10 seconds [07:42:46] errmmm, not happening [07:42:53] jeremyb_: i think it hates you [07:43:14] i'm iceweasel 19.0.x fwiw [07:43:25] checked console. no reqs were waiting [07:45:20] * edsu shrugs [07:46:06] jeremyb_: i really didn't have much to do with the app, despite being listed on the paper ; twitter.com/tomayac would be interested to hear from you though [07:46:29] huh, ok [07:46:35] jeremyb_: does it work in another browser? [07:46:56] i suppose i could try chromium. but that's dangerous (freezes up my machine) [07:47:02] i could try my phone [07:49:43] firefox and chrome both work on phone. iceweasel don't work on laptop [08:03:43] jeremyb_: doesn't seem to work in Aurora either [08:05:56] jeremyb_: or Firefox 20.0 for that matter [08:06:44] jeremyb_: doesn't seem to be a general problem with socket.io since http://wikistream.inkdroid.org/ seems to work fine [08:06:59] * edsu tickets [08:11:59] jeremyb_: thanks for letting me know: https://github.com/tomayac/wikipedia-irc/issues/3 [08:12:18] jeremyb_: hopefully tomayac will be more helpful than me :) [08:32:42] edsu: but it *does* work on 20 on my phone... [08:33:02] jeremyb_: odd [08:33:14] jeremyb_: feel free to add to that ticket ... [08:33:26] yeah... when not asleep [09:04:45] Hi [09:04:58] Is there a blamefinder tool? [09:05:12] I'm needing to track down some very specifc edits I made [09:27:24] Qcoder00: just go to the history page and click revisions history search? [09:27:40] Yes but that shows everything [09:27:49] It doesn't go down into very specfic additions [09:29:52] Qcoder00: there are tools.. I think you need to search for blametool [09:30:35] If and since full histories are kept, people are going to make blametools. [09:32:21] Blametool only deals with single articles [09:32:32] I'm needing an article list [09:33:08] Namely I need to find all articles containing {{cite web|author= which User:Sfan00_IMG has edited [09:33:58] Whats more these edits were ALL done using Reflinks (so there should be a recognisable edit summary) [09:36:38] do you use reflinks for anything else, other than the edits your trying to find now? [09:36:43] 'k .. totally rocket scienced me [09:37:58] p858snake|l: Not over the 1 hour period on the 22nd April concerned [09:38:46] So [[Special:Contributions]] isn't usable why? [09:39:06] because it lists ALL the edits... [09:39:16] Including the ones that aren't immediatly relevant [09:39:34] just copy and paste and remove the illrelevant ones? [09:40:07] Hah [09:40:31] I am effectivly needing to do a grep for a certain patterns in my edit history [09:40:39] (ie the pages in my edit history) [09:40:54] There doesn't seem to be a tool that will grep an entire page range [09:42:22] just use a larger limit on the contribs page... unless you did more than 500 edits in that hour it should be fine... [09:43:23] It's around 500 or so [09:43:36] but I don't like having to check EVRY single diff to find the specfic ones [10:36:40] Qcoder00: how many pages? [10:37:00] Less than 250 or so, I've checked the others [10:37:45] Earliest is http://en.wikipedia.org/w/index.php?title=Portal_2&oldid=551600003 latest - http://en.wikipedia.org/w/index.php?title=Ranjit_Sinha&oldid=551604497 [10:39:20] what makes them similar? [10:40:14] All the edits will contain the addition of a string of the form {{cite web|author= [10:40:52] and an edit summary containing - Filling in (.)*) references using [10:41:11] and were on the revision ID's noted edited by User:Sfan00_IMG [10:41:24] *Within the range of revid's noted [10:41:35] Qcoder00: drop me an email with the details, Im about to head out, I should be able to whip something up [10:41:57] Betacommand: PM me your e-mail? [10:42:08] @toolserver.org [11:46:42] apergos: wot, the logging tables was only 200 MB big? [11:46:49] no [11:47:16] 307 but it got stopped in the middle apparently when the db was upgraded to mariadb [11:47:28] I'm rerunning i [11:47:30] t [11:47:49] the export of the logs? i think it would be OK to skip it this time [11:48:07] it won't tak 4 days this time [11:48:23] I think. I'm testing out a theory [11:49:17] cool [11:49:53] ? [12:02:34] just to confirm our current understanding: autopatrolled edits are also logged in the logs, and there is no need for that. if we didn't have that, about 90% of those log messages would just go away. is this correct? [12:02:54] * legoktm thinks so [12:03:09] autopatrolled edits are logged on all projects [12:03:33] the difference with wikidata is that on the other projects, only new page creation are subject to patrolling (and autopatrolling) [12:03:39] on wikidata it seems that all the edits are [12:04:10] so step one might be to limit it to just new pages [12:04:46] step two if you wanted to go further might be to not log bot autopatrols to the log (but this would be a change in core, prolly needs discussiion) [12:09:18] er, there are plenty of projects that use RCPatrol [12:09:34] its just that none of those have the volume of edits that wikidata does [12:17:53] ah srry [12:18:04] I'd only checked en wp (which doesn't) and a couple other randoms [12:18:44] apergos: You should've checked 841 public projects [12:18:54] clearly [12:19:41] still seems to me that bot autopatrol could generally not be logged [12:19:47] log the edits and be done with it [12:19:53] across the projects I mean [12:21:12] so my theory (which isn't very interesting really) is that if I run this job in batches of 2 million pages I'll escap the slowdown [12:21:27] I'll have the answer to that in about half an hour [12:22:03] We should point out to wikidatawiki that this is why we can't have nice things [12:22:16] awww [12:22:31] because they edit too much? :-D [12:23:10] reminds me of the bad old days when we had to tell people "uh could you not bulk upload all that stuf right now? we're running low on capacity/space" [12:23:18] :D [12:24:38] so if this theory pans out we are going to want a batchsize argument to export.php that can be passed in there, and would let it do its miain queries in chunks of x at a time for certain types of export [12:25:13] yes I could just call the export routine multiple times and have a bunch of little files that then have to be recombined stripping off the siteinfo and mediawiki tags but meh [12:25:19] much rather do it the "right" way [12:25:28] Reedy: or just a better antispam system to kill of page "patrolling" [12:26:34] Patrolling and FR and stuff just confuse me [12:28:01] fr amuses me [12:28:07] Reedy: burn them all! [12:28:15] er *cough*pending changes*cough* [12:28:30] * p858snake|l rejects apergos's revision >.> [12:28:37] :-D [12:29:07] * Reedy goes to add DAS IST VERBOTEN to site notice on wikidatawiki [12:29:43] meh it's much easier to just run my query on master in eqiad.. that'll hold em up for awhile [12:29:54] WFM [12:30:13] "Nope, you're edting too fast, you're causing the rep lag" [12:30:23] :) [12:31:13] my bot now no longer edits faster than 120epm :( [12:31:46] and i have a safeguard that blocks it if it hits 300+epm [12:31:55] that is a goooood idea [12:32:06] and no fair giving it to five of your friends to run it [12:32:58] Either way, that is pretty impressive [12:33:41] Reedy: impressive that legoktm was able to screw up the replag? [12:56:21] oh yeah my theory is correct, batches work like clockwork, I'm already at 22 million and will finish up here in a few more minutes [12:56:38] this is me being slow and running them by hand from the command line [13:17:06] Haha [13:17:12] Still faster than waiting for the old one [13:25:01] yes, tons faster [13:25:16] 4 days faster in fact (plus however mny days it would have taken but was interrupted) [13:30:33] Lets all go patrol some pages in the Thread namespace on mw wiki! [13:31:24] sounds deadly dull [13:32:19] i forgot i needed to add {{sarc}} at the end of that [13:34:56] or was that {{snark}} [13:35:35] we're past 28 million, wonder how many million it's racked up in the meantime [13:38:20] 29538611 [13:38:23] that was the last one [13:39:10] lol [13:40:21] for now! [13:55:19] is anyone here competent with .js ? [13:55:48] Dragonfly6-7: What do you need? [13:56:10] I have some .js issues involving customized tools that seem to have been broken by the skin transition [13:56:25] hold on, let me get the URL [13:57:22] http://en.wikipedia.org/wiki/User:Mr.Z-man/patrollinks.js [13:57:27] this was written for me [13:57:33] Dragonfly6-7: if no one is around, you can always post on the tech page and someone will generally help [13:57:34] important maintenance tool [13:57:44] p858snake|l - I know. But I asked here first. [14:00:02] it worked fine on Classic (most of the time), and I imported it into Vector and it's not working. [14:00:25] which might, or might not, be the result of an incompatibility [14:00:58] Dragonfly6-7: What is it supposed to do and where (on which page... give me an URI at best) [14:02:12] hold on [14:02:35] http://en.wikipedia.org/w/index.php?title=Special:NewPages&dir=prev&namespace=1&hidepatrolled=1&hidebots=1 [14:02:37] there [14:02:47] each unpatrolled page in the list should have a little "click to patrol" link [14:03:05] for the ones that are obviously good, I just click the link and they're patrolled without my even having to open the page [14:03:46] saving a few seconds each time, which adds up quite quickly [14:04:34] Dragonfly6-7: Where exactly does it break? The links turn up for me [14:06:55] try clicking one [14:08:16] does it work for you? [14:08:17] Dragonfly6-7: Here's your problem: http://toolserver.org/~alexz/util/rcid.php [14:09:00] * hoo can't even ssh into the toolserver atm [14:09:52] hoo - as in, the .js references a toolserver address which is now dead? [14:10:12] Dragonfly6-7: Exactly... might be just a temporary TS outage, tough [14:10:22] okay [14:10:23] thanks [14:11:19] in principle, do you think you'd be able to write an equivalent tool for me? [14:11:31] I sure would [14:11:39] Seems temporary, btw "Wikimedia Toolserver | Known issues: SSH, CRON, user.store" [14:11:46] okay, good [14:11:55] coincidence. [14:12:56] one of my current projects is "nag people into helping me replicate the appearance of Classic, in Vector" [14:13:31] would you have a few minutes to look at a list of features I need and/or want? [14:13:57] (I've been careful to separate the two) [14:16:19] Dragonfly6-7: I don't htink I got time for that... dunno what that TS tool does anyway [14:16:38] seems superfluous to me (w/ having had a look at it) [14:17:25] yeah, its a "temporary" outage [14:17:40] hoo - too many pronouns, not enough antecedents. *What* seems superfluuous to you? [14:18:31] Dragonfly6-7: Using a toolserver tool in that script ;) [14:19:56] ah [14:20:32] and "I don't think I got time for that" -- 'that' = a) rewriting the TS tool, b) looking at my list of needs and wants? [14:25:50] * Dragonfly6-7 waves arms [15:34:25] okay, confirmed - just a toolserver glitch, and now it's working fine. [15:41:17] 439mb now fyi [15:41:22] if anyone was keeping track [15:59:57] sumanah: the new mac mini in the conf room is foiling us... one second [16:00:19] ah ok [16:00:54] https://www.mediawiki.org/w/index.php?title=Extension:Score&oldid=676548&diff=cur extension:score is considered to be stable now right? [16:04:39] Probably [16:04:54] Though deployed doesn't always mean stable [16:05:16] well i approved it :P [16:05:44] shouldn't be far off [16:06:08] hello hello [16:06:18] can I nag someone into looking at a script for me [16:06:49] Dragonfly6-7: I feel that if you learned js/css, you'd probably get it fixed faster. [16:10:35] Dragonfly6-7: or sleeping with a developer may be another option [16:10:45] or even threatening too sleep with them [16:29:09] AaronSchulz: Hi, could you comment on https://bugzilla.wikimedia.org/show_bug.cgi?id=36085#c6 ? [16:46:58] [[Tech]]; PiRSquared17; /* Classic skin and CSS */ +; https://meta.wikimedia.org/w/index.php?diff=5415680&oldid=5410032&rcid=4099776 [17:26:46] ok yay I finally gt why the history list gets longer. how slow was I to get that. sheesh [17:27:01] and I see why there's going to be contention over locks etc [17:27:28] but I still don't see why the query gets slower and slower over time. that's the missing bit [17:28:50] actually I take tht back, not sure about history,. sure about contention for locks though. [17:44:43] greg-g: Looking at the calendar, it seems there's an open spot from 12:30-1:00 PDT. Can I take that to quickly backport gerrit change 59508 to 1.22wmf2? (I have other obligations this week during the Lightning Deploy timeslot) [17:46:26] anomie: yeah, sounds ok,add it to the calendar if you can, please [17:49:11] greg-g: Added to the wikitech page. I don't seem to have access to the deployments google calendar. [17:50:03] anomie: thanks, I'll get that. [17:51:10] Does anyone know where the upload.wikimedia.org favicon is? [17:55:15] uh [17:55:27] well it's gonna either be in swift or bits [17:55:41] if we even have one that is [17:56:23] found the history list thing finally: every edit => user edit count is updated -> update to the row means we have to put it in th udo log cause the logging dup does a join against the user table [17:56:29] piece o crap [17:57:21] really should have the stuff that doesn't change often in one place (user id, name, email) and the stuff that changes all the time somewhere else, meh [17:57:51] Isarra: what is it you're looking for exactly? [17:58:00] Isarra: you can't change it via Gerrit, if that's what you're asking [17:58:18] but I know how to change it [17:59:42] Can you change it to be the one used by foundationwiki and whatnot? [18:00:21] are we talking about favicon.ico? [18:00:23] or a png? [18:00:28] favicon.ico [18:00:35] looks the same to me? [18:00:52] It's not. [18:00:54] yea, http://upload.wikimedia.org/favicon.ico [18:01:18] The upload one is the old broken one. [18:01:27] okay [18:01:30] Should be the same as http://wikimediafoundation.org/favicon.ico [18:03:02] greg-g, I mentioned https://bugzilla.wikimedia.org/show_bug.cgi?id=47457 yesterday (some User page don - a number of users now state that disabling WikiLove helps. [18:03:12] meh. I didn't want to hit Enter yet. [18:03:42] again: greg-g, I mentioned https://bugzilla.wikimedia.org/show_bug.cgi?id=47457 yesterday (some User page don't load in Firefox) - a number of users now state that disabling WikiLove helps. So sounds like platformeng territory to me? :) [18:04:37] Yay, I'm not the only one who sometimes hits enter instead of '... [18:06:19] Isarra: fixed [18:06:23] Isarra: thanks [18:06:29] paravoid: Thank you! [18:06:46] what was broken on the old one? [18:06:50] purely out of curiosity :) [18:07:17] andre__: huh, wikilove... [18:07:30] paravoid: high res, I think [18:07:45] as in, it wasn't high enough res for those fancy new macintoshes ;) [18:07:56] Not sure, but something about the way it was saved made it show up rather poorly in some browsers. Wasn't the best render even when it did show up as intended, though, and also didn't have a high-res version. [18:08:16] what's the high-res version? [18:08:19] I doubt an ico can be very much high-res [18:08:20] they're embedded somehow [18:08:32] It's saved in the file. Basically it has different layers for different resolutions. [18:08:41] Browser or whatever picks out the correct one. [18:08:43] okay [18:08:52] greg-g has the simple answers, Isarra has the complete ones ;) [18:09:13] greg-g used to be the new guy, now he's not anymore :-) [18:09:21] greg-g: see my summary in https://bugzilla.wikimedia.org/show_bug.cgi?id=47457#c8 [18:09:23] :) [18:09:31] Well, more complete. All I really know is making layers works and saving that fixed the other problems too. [18:09:39] Isarra: btw, upload's index page sucks [18:09:50] greg-g: I'd love to have somebody to look at that (as it gets quite some coverage in VP), and it doesn't feel like ops territory anymore :) [18:09:54] s/to// [18:10:00] if you want to fix this and/or track down the right people, I'd happily put it up :-) [18:10:07] Bahahah, it totally does. XD [18:10:41] Would something more like the 404 page elsewhere work? [18:11:06] andre__: what Fx versions are affected, that first one, User:Bidgee, worked for me :/ [18:11:06] I guess? [18:11:07] :) [18:11:15] ...a 404 page that probably wouldn't hurt to move over as well... [18:11:29] It could just... dump people on Commons or something, maybe. [18:11:31] upload's error pages also suck, but that's less trivial to fix [18:11:49] greg-g: whou, interesting. I run FF18 here, but it also happens with FF20 according to VP [18:11:54] Aye... [18:11:55] and I can reproduce the problem :-/ [18:12:02] andre__: huh, I'm on Fx 20..... alright [18:12:31] User:Moogsi loads quickly for me as well [18:13:06] andre__: but yeah, let's escalate this one [18:13:13] er, I'll escalate it ;) [18:13:27] or whatever, I'll poke the right people [18:13:32] Isarra: I'll be waiting ;) [18:13:45] greg-g, thanks! [18:15:26] http://en.wikipedia.org/wiki/Favicon#HTML5_recommendation_for_icons_in_multiple_sizes [18:16:06] Apple mobile devices running iOS 5 and above ignore the HTML5 recommendation and use the proprietary apple-touch-icon method detailed above. The Google Chrome web browser however, will select the closest matching size from those provided in the HTML headers to create 128×128 pixel application icons when the user chooses the Create application shortcuts... from the "Tools" menu. [18:16:08] "Apple mobile devices running iOS 5 and above ignore the HTML5 recommendation and use the proprietary apple-touch-icon method detailed above." [18:16:12] heh, yeah [18:16:49] [18:16:53] from foundationwiki [18:17:06] foundation.wikipedia.org doesn't exist [18:17:09] lol :) [18:18:25] :P [18:18:46] heh [18:21:51] "ignore the HTML5 recommendation" [18:21:55] mentioned in https://bugzilla.wikimedia.org/show_bug.cgi?id=27911 [18:22:00] lol, sure. [18:22:03] more than 2 years ago :) [18:23:14] brion: this ticket looks very mobile related [18:23:20] *hint* *hint* [18:23:53] <^demon|away> Even if it wasn't, the apple touch icon code was his fault anyway ;-) [18:30:42] RoanKattouw: kaldari: heya, this bug seems like it might be related to WikiLove (as best we can determine right now) since your name is on extensions:WikiLove, I'm pinging you ;) https://bugzilla.wikimedia.org/show_bug.cgi?id=47457#c8 Thoughts? [18:31:07] * RoanKattouw washes his hands [18:31:23]