[07:27:46] @trusted [07:27:46] I trust: petan!.*@wikimedia/Petrb (2admin), .*@mediawiki/Yuvipandan (2admin), .*@wikipedia/preilly (2admin), [07:27:58] @rss-on [07:27:58] Rss feed has been enabled on channel [07:33:57] @configure style-rss=[bugs] $bugzilla_status $title (priority: $bugzilla_priority severity: $bugzilla_severity) $link [07:33:57] Value [bugs] $bugzilla_status $title (priority: $bugzilla_priority severity: $bugzilla_severity) $link was stored into style-rss to config [07:34:11] @rss+ bz https://bugzilla.wikimedia.org/buglist.cgi?chfieldfrom=-1h&chfieldto=Now&list_id=152259&product=Wikimedia Mobile&query_format=advanced&title=Bug List&ctype=atom [07:34:12] Item was inserted to feed [07:34:33] @rss- bz [07:34:33] Item was removed from db [07:35:57] @rss+ bz https://bugzilla.wikimedia.org/buglist.cgi?chfieldfrom=-10h&chfieldto=Now&list_id=152259&product=Wikimedia+Mobile&query_format=advanced&title=Bug+List&ctype=atom [07:35:57] Item was inserted to feed [07:42:20] @rss+ mobile1 https://bugzilla.wikimedia.org/buglist.cgi?bug_status=UNCONFIRMED&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&chfieldfrom=-2h&chfieldto=Now&list_id=152262&product=Wikipedia+App&query_format=advanced&resolution=---&resolution=LATER&resolution=DUPLICATE&title=Bug+List&ctype=atom [07:42:20] Item was inserted to feed [07:43:58] @rss+ extension https://bugzilla.wikimedia.org/buglist.cgi?list_id=152263&resolution=---&resolution=LATER&resolution=DUPLICATE&chfieldto=Now&query_format=advanced&chfieldfrom=-20h&component=MobileFrontend&product=MediaWiki+extensions [07:43:58] Item was inserted to feed [07:44:40] Unable to parse the feed from https://bugzilla.wikimedia.org/buglist.cgi?list_id=152263&resolution=---&resolution=LATER&resolution=DUPLICATE&chfieldto=Now&query_format=advanced&chfieldfrom=-20h&component=MobileFrontend&product=MediaWiki+extensions this url is probably not a valid rss, the feed will be disabled, until you re-enable it by typing @rss+ extension [07:45:34] @rss- extension [07:45:34] Item was removed from db [07:45:45] @rss+ extension https://bugzilla.wikimedia.org/buglist.cgi?chfieldfrom=-20h&chfieldto=Now&component=MobileFrontend&list_id=152263&product=MediaWiki%20extensions&query_format=advanced&resolution=---&resolution=LATER&resolution=DUPLICATE&title=Bug%20List&ctype=atom [07:45:45] Item was inserted to feed [09:43:37] are there instructions somewhere for users how to provide a stacktrace when the Wikipedia App crashes on iOS? [09:44:14] [bugs] NEW [Bug 41027] App is crashing in landscape mode on iOS 6.0 (priority: Unprioritized severity: normal) https://bugzilla.wikimedia.org/show_bug.cgi?id=41027 [11:00:36] are there instructions somewhere for users how to provide a stacktrace when the Wikipedia App crashes on iOS? [11:06:34] [WLMMobile] siebrand pushed 1 new commit to master: http://git.io/YnkFNQ [11:06:34] [WLMMobile/master] Localisation updates from http://translatewiki.net. - Siebrand Mazeland [11:06:52] [WikipediaMobile] siebrand pushed 1 new commit to master: http://git.io/7n4dbQ [11:06:52] [WikipediaMobile/master] Localisation updates from http://translatewiki.net. - Siebrand Mazeland [11:06:53] Project WLMMobile - Nightly builds build #244: SUCCESS in 8.8 sec: https://integration.mediawiki.org/ci/job/WLMMobile%20-%20Nightly%20builds/244/ [11:06:54] s.mazeland: Localisation updates from http://translatewiki.net. [11:07:04] Project WikipediaMobile - Nightly builds build #417: SUCCESS in 10 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/417/ [11:07:04] s.mazeland: Localisation updates from http://translatewiki.net. [11:07:05] Project WiktionaryMobile - Nightly builds build #104: SUCCESS in 13 sec: https://integration.mediawiki.org/ci/job/WiktionaryMobile%20-%20Nightly%20builds/104/ [14:04:18] are there instructions somewhere for users how to provide a stacktrace when the Wikipedia App crashes on iOS? [14:26:54] I don't think so [14:27:09] And from what Tomasz said before, Apple won't tell you about it until theres X occurances [14:36:51] hmm, I see. Was wondering which instructions to provide to users so they can provide useful info in bug reports. [15:10:33] New patchset: Reedy; "Fix test related paths" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28030 [15:11:07] New patchset: Reedy; "Fix test related paths" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28030 [15:11:34] Change merged: Reedy; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28030 [15:16:41] andre__: Ask again when the SF guys are in the office [15:16:43] couple of hours from now [15:42:58] [bugs] NEW [Bug 41039] Cannot select search results Opera Mobile (priority: Unprioritized severity: normal) https://bugzilla.wikimedia.org/show_bug.cgi?id=41039 [17:01:12] http://www.androidpolice.com/2012/10/15/google-play-has-a-brand-new-developer-console-completely-overhauled-interface-new-statistics-and-more/ [17:20:30] ewwww, mobile-testing has very little local drive space [17:20:44] * yuvipanda looks around [17:20:58] * MaxSem hides from yuvipanda [17:21:02] * awjr waves [17:21:05] hi yuvipanda [17:21:18] MaxSem: yeah :( [17:21:18] greetings all [17:21:39] hi tfinc [17:21:41] * tfinc waves from the HTML5DevConf [17:21:45] awjr: need anything from us ? [17:21:58] tfinc the only thing i could think of i just sent an email about [17:22:09] yuvipanda: i've been busy testing your commons app [17:22:16] yuvipanda: you have some newly filed issues [17:22:31] my favorite is that i can't dismiss any of your status bar notifications for completed upload [17:22:37] awjr: responded [17:23:03] ty [17:23:22] http://www.androidpolice.com/2012/10/15/google-play-has-a-brand-new-developer-console-completely-overhauled-interface-new-statistics-and-more/ [17:24:21] Reedy: i've been trying that out for the last couple of weeks [17:24:28] ahh [17:25:15] Reedy: what i really want is for them to flip the bit so that i can respond to comments [17:25:22] its only available to their "top developers" [17:26:27] tfinc: I fixed a good part of them :P [17:27:19] tfinc: did you see the screenshot I tweeted to you? [17:28:57] yuvipanda: not yet. let me pull twitter up [17:29:01] mmm, directory listing is atrocitously slow. is the whole instance borked? [17:29:04] awjr, ^^ [17:29:26] or it's due to Gluster? [17:29:29] MaxSem: not that i could tell, the only thing that was obviously totally f'd to me was mysql [17:29:31] yuvipanda: not seeing it. link please [17:29:36] MaxSem: could well be due to gluster [17:33:49] okay, I've recreated the DB, but main page takes 25 seconds to load - it's unacceptable [17:34:19] MaxSem: perhaps memcache just needs to be warmed [17:34:44] nope, that's on the *second* view, first was smth like 40s [17:35:04] oh [17:35:10] tfinc:https://test.wikipedia.org/wiki/File:Tee.png [17:35:19] tfinc: ^ [17:35:27] MaxSem: no good :( [17:35:34] * tfinc wonders what a tfinc: url handler would do :) [17:35:53] MaxSem: is that generally the case for other pages too? [17:35:54] yuvipanda: what about just calling it commons ? [17:36:05] yuvipanda: awesome to see the account integration [17:36:12] tfinc: a *lot* of people confuse it with Creative Commons [17:36:17] tfinc: it would probably take you climbing and then open a beer [17:36:31] like, about everyone who asked me what I was working on at that hackathon last weekend [17:37:17] oh .. next presentation is starting … bbl [17:37:59] maxsem@mobile-testing:/var/www/w/languages$ time ls [17:38:07] real 0m0.596s [17:38:38] multiplied by the number of files MW executes on pageview... [17:39:02] wow [17:41:02] MaxSem: would it be better if we used a much smaller subset of enwiki articles and moved the db off of gluster? [17:41:22] MW is on gluster too [17:41:29] oh [17:42:11] root@mobile-testing:/var# rm -rf www [17:42:11] root@mobile-testing:/var# mkdir www [17:42:11] root@mobile-testing:/var# chmod 777 www [17:50:12] so who's not at the conference this week? [17:50:18] me [17:50:30] Maryana: basically anybody on the mobile team who is not in SF [17:50:43] ah. so brion and jon are out? [17:50:48] Maryana: correct [17:50:56] boo-urns [17:51:01] heh [17:51:11] Maryana: what do you need? [17:51:16] watchlist ain't gonna code itself! [17:51:24] :) [17:51:28] truth. [17:52:06] i'm really excited to get it up on beta so we can put it in front of some wikipedians [17:52:12] and they can tell us the 50 things we did wrong with it [17:52:53] i am too - i dont think anyone's started on the watchlist changes yet though [17:53:45] no, though design has been working overtime on it and it's pretty much ready to go [17:53:53] hopefully later this week [17:53:56] * Maryana crosses fingers [17:54:39] awesome. brion has been working on ResourceLoader stuff which i imagine will occupy most of his time for the rest of the iteration, although jon is not currently working on anything, so i suspect he'll be taking the watchlist when he's back from the conf [17:55:12] Maryana: can you get the design assets attached to the card? [17:55:19] will do [17:55:32] hey, when did jon deploy that language selector button to beta? [17:55:48] Maryana: i think that went out last week [17:56:05] unless i am thinking of something different [17:56:33] uh, i believe so. the lang. selector no longer appears in the ribbon [17:56:46] it's a big gray button above the footer now [17:57:10] that is in beta, right now? [17:57:10] which is sort of what i was going for in the mingle card. i just didn't know it would happen in between deployment windows... [17:57:13] yep [17:58:22] huh, and it doesn't appear to be localized. [17:58:32] nope [17:58:32] i dont know how that got to production [17:58:42] i dont think it should have. [18:00:28] it looks like most htings we would deploy today were deployed [18:00:48] also, it looks like the spacing/font changes messed up table displays a little [18:01:09] let me see if ic an figure out who did this [18:03:22] looks like reedy and/or preilly [18:03:41] Me, likely [18:03:53] * awjr shakes fist [18:04:14] Reedy a bunch of MF changes were pushed out to production that had not been tested [18:04:22] bleh [18:04:32] https://gerrit.wikimedia.org/r/#/c/28030/ [18:04:32] according to the server admin log it looks like it was likely either you or preilly [18:04:38] Most likely me earlier [18:04:50] I was bringing that revision in to try and reduce the spam on the apache logs [18:05:18] https://gerrit.wikimedia.org/r/#/c/28033/ [18:05:29] ahha [18:06:28] Just pushing out a revert of that now... [18:08:29] awjr: ok, back to whatever revision it was on previously [18:09:13] thanks Reedy [18:09:16] Maryana: ^ [18:09:22] thanks reedy! [18:09:38] Reedy: i wonder if we should set up a separate deployment branch for MF [18:09:52] since there a particular considerations for MF deployments that don't typically apply to other extensions [18:10:10] We could do... [18:10:12] namely device testing and varnish cache wonk [18:11:37] I can cherry pick this revision onto whatever production is running for the moment.. [18:12:39] Reedy: let's do that for now [18:12:56] Reedy: we'll probably sync produciton to master today or tomorrow depending on today's testing [18:13:19] at least now we know what issues to look for when we test [18:13:20] heh [18:13:25] :p [18:18:08] grr 100 (0.96 pages/sec 0.96 revs/sec) [18:22:32] reedy@fenari:/home/wikipedia/common/php-1.21wmf1$ git submodule update extensions/MobileFrontend [18:22:32] fatal: reference is not a tree: cb20aa5ac42d91e2c0c20340ef69a81ee4bb888e [18:22:32] Unable to checkout 'cb20aa5ac42d91e2c0c20340ef69a81ee4bb888e' in submodule path 'extensions/MobileFrontend' [18:22:34] Fai [18:22:35] l [18:23:09] * Reedy reverts to save his sanity [18:32:11] Reedy: what happened? ^ [18:32:40] the cherry pick apparently didn't transpose very well [18:32:48] :( [18:32:52] I'm just used to being lazy, update an extension to master and deploy as such ;) [18:33:25] I think this is actually properly documented somewhere.. [18:33:34] heh ok, so is the current state whatever was in production before your recent push, w/o the fix? [18:33:55] yup [18:35:56] I guess I need to do something more like http://wikitech.wikimedia.org/view/How_to_deploy_code#Case_1b:_extension_changes [18:36:51] yeah we've done that before for cherry-pick changes [19:03:14] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/27935 [19:03:55] New review: awjrichards; "Was the updated css compiled and submitted elsewhere?" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/27935 [19:14:00] okay, the import is running http://mobile-testing.wmflabs.org/w/index.php/Special:AllPages [19:14:16] but the performance is disastrous [19:36:12] MaxSem: does it seem worse than before? i recall it taking roughly forever when pages were previously imported [19:36:46] yes, it seems 3-4 times slower [19:40:11] :( [20:09:59] MaxSem: i just started putting together the changelog for today's deployment and then remembered you're handling it today; but here it is: https://www.mediawiki.org/wiki/Extension:MobileFrontend/Deployments/2012-10-15%2616 [20:10:28] cool, thanks [20:10:51] ugh, lots of frontend changes [20:12:55] awjr: we did a deployment? [20:12:55] http://en.m.wikipedia.org/w/index.php?title=Main_Page&mobileaction=toggle_view_mobile seems to be broken [20:13:29] search in non-beta now shows a button with a label it shouldn't and footer doesn't expand/collapse [20:13:41] damn it [20:13:49] jdlrobson: probably fallout from Reedy's deployment [20:13:59] surprise surprise debug=true fixes it [20:14:11] we really need to get this shit sorted [20:14:24] guessing files need touching again [20:14:28] jdlrobson: im on it [20:14:32] thanks awjr [20:14:43] jdlrobson: see my email about deployment branch that i sent this a [20:14:43] m [20:14:48] sure [20:21:11] jdlrobson: when you have a sec, can you dbl check? things look ok now [20:21:35] * jdlrobson looks [20:21:43] still getting old cached version [20:21:47] guess we have to wait a bit [20:22:09] yeah, i just saw a cached version too, but some pages look fine, let's check again in 5 mins [20:24:12] i have to force refresh [20:24:28] I still think half our bugs are down to our caching mechanisms [20:25:15] hard refreshing on http://en.m.wikipedia.org/wiki/Kaiser_class_battleship and still old css [20:25:47] jdlrobson, FYI: debug=true enables RL debug mode. it can't serve outdated stuff - simply because it serves everything uncacheable [20:26:32] New review: Jdlrobson; "The output CSS is exactly the same so doesn't change. Its just a case of merging variable names." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/27935 [20:26:54] MaxSem: I know that - that's what I'm saying [20:27:12] we're serving old css/javascript to people who are not appending debug=true suggesting something wrong with our caching [20:29:31] it simmply does its job by caching for 28 days everything it hasn't been told otherwise [20:29:50] the problem we've been running into is a resource loader bug [20:30:07] ? [20:30:08] sure.. we need to sort this though - it's killing us imo - mobile is changing more dramatically then anywhere else on the site [20:30:23] if the html markup changes it can throw a javascript error [20:30:33] (for example) [20:31:10] yeah, hence why im starting to feel like having a more controlled deployment branch would help us with a situation like this, where someone off the team pushed from master [20:31:21] we should plan this stuff in a way that edge cases where new html/old resources or vice versa don't cause anything catastrophic [20:34:17] MaxSem: that seems like an unnecessary overhead and will lead to lots of TODO: comments and code bloat [20:35:45] in such case, we will have the ops flush varnish even after RL is fixed. and when our cluster will not be able to handle the load from empty cache... oh well [20:36:01] the easy solution is to fix the RL bug and make it so we don't have to worry about flushing the varnish cache anymore. [20:36:09] cause RL will be pretty good at serving new resources to old HTML [20:37:07] mmm great http://www.youtube.com/watch?v=RjlKjeJKjVs [20:38:28] jdlrobson & awjr, have you committed everything for today's deployment? [20:38:52] MaxSem: we still have a deployment tomorrow yes? (I want to get some css changes in this week) [20:39:03] but yes I have committed everything I needed to [20:44:50] MaxSem: yes [20:48:39] ugh Notice: fwrite(): send of 1117 bytes failed with errno=11 Resource temporarily unavailable in /var/www/w/includes/objectcache/MemcachedClient.php on line 1104 [20:49:48] jdlrobson: issue now appears to be resolved to me [20:51:22] awjr: WFM [20:51:36] jdlrobson: it was a combination of the RL bug and mobile varnish cache needing to be cleared [20:51:55] awjr: can we get this written up so anyone deploying MFE knows this [20:52:03] this is causing me lots of pain [20:52:22] yes, but writing it up does not mean that other non-mobile deployers will be aware of it [20:52:39] which is what happened today. [20:52:41] sure - but at least that will save the mobile team causing this problem [20:52:57] is there anyway we could automate it? [20:54:07] jdlrobson: the RL bug needs to be fixed regardless, and getting RL fully mobile-friendly should make it so we don't need to flush mobile varnish cache [20:54:25] jdlrobson, we could have a dedicated deployment branch (and make sure that Reedy's scripts know about it). this way, general deployments will not result in new MF code being deployed [20:54:58] mmm [20:55:03] awjr, RL is not a silver bullet, new resources will need to work nicely with old HTML [20:55:08] Though, you shouldn't merge untested code ;) [20:55:30] then we need 50 testers [20:55:44] and an actual production-like environment in which we can test [20:55:48] * Reedy grins [20:56:12] :p [20:59:22] awjr, so what's the battle plan for today? you wanted to deploy just to testwiki, however leaving stuff undeployed to cluster is dangerous [21:00:05] so gotta fly… sorry [21:00:19] MaxSem: let's start with deploying to testwiki and see it how it goes; if things look ok then we should be fine to push to production [21:00:38] MaxSem: we'll have Maryana and jcmish to help QA/test [21:00:54] yup [21:01:02] indeed :) [21:01:37] sehr good [21:04:22] okay, /me prepares commits [21:07:28] question awjr and MaxSem when we say production does that really mean production? Or beta or simple? [21:08:05] jcmish: good question. when i say production, i mean anything running live on the production cluster (beta or otherwise) [21:08:18] production = WMF cluster [21:08:19] to differentiate between beta and non-beta, i usually say beta or non-beta :) [21:08:32] it would be nice if we had a better name for non-beta... [21:08:40] hehe cool I'm learning all of these [21:08:49] whether testwiki is 100% production or just 99.9999% is a philosophical question [21:08:57] lol [21:09:03] hahahhaha [21:13:56] +100 [21:14:07] non-beta is a weird way to talk about the mobile site [21:14:41] it's also non-desktop [21:14:50] and non-unicorn :) [21:15:03] pro and anti unicorn :D [21:15:07] the latter is disputable! [21:15:30] everyone knows unicorn denotes wikimedia labs [21:15:34] yet another test environment! [21:16:46] one day it will all make sense :) [21:16:49] test and test2 should have properly been named "staging" or "pre-production" or some such [21:17:09] haha [21:17:14] that would be sensible [21:18:02] pfft, test and test2 is much more clear and understandable [21:18:18] mind the idiots! [21:18:21] ...like me [21:18:24] hahahahha [21:39:13] MaxSem: fyi i've created some documentation around MF-specific deployment stuff: http://wikitech.wikimedia.org/view/MobileFrontend/Deployment [21:41:02] MaxSem: wait that's not right, noe sec [21:41:05] *one sec [21:41:24] awjr, there's fatalmonitor [21:41:37] ok yeah it's right [21:41:45] MaxSem: oo - add it to the doc [21:42:36] done [21:58:05] New patchset: Brion VIBBER; "initial stab at touchy for gesture detection in MF" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28158 [22:06:22] awjr, jcmish, Maryana: new changes are live on https://test.m.wikipedia.org/wiki/Main_Page - please test [22:06:50] MaxSem will do [22:07:08] Maryana can you do the searches? [22:07:15] I can look at the icons and make sure they're all working etc [22:07:26] on it [22:08:03] thanks! [22:09:55] footer expansion appears to be broken [22:10:16] (both in ios and android) [22:10:18] also, should 'read in another language' appear if there are no other languages available? [22:10:34] no [22:10:43] Change abandoned: Brion VIBBER; "This doesn't work as-is, need to rethink use of this lib or similar. :)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28158 [22:11:02] what is supposed to appear in the ribbon now that the language selector has moveD? [22:11:22] just watchlist star for logged in, though i think we'll probably want to move it out of there [22:11:30] darn you awjr! [22:11:34] getting a JS error in chrome too [22:11:42] you took away my joy :D i was gonna say that about the footer :D [22:11:43] jcmish what'd i do?! [22:11:46] this is odd. do a search for "sand", find result "\sandbox". normal page is "http://test.wikipedia.org/wiki/%5CSandbox", mobile page is https://test.m.wikipedia.org/wiki//Sandbox (and broken) [22:11:55] jcmish: heh feel free to say it again if you like :) [22:12:11] no no :) [22:12:31] chrismcmahon, works for me: https://test.m.wikipedia.org/wiki/%5CSandbox [22:13:19] MaxSem, chrismcmahon: i am seeing the ribbon not working and the footer expanded on that page; there is either JS broken or ResourceLoader is loading old resources [22:13:25] MaxSem: you clicked to that URL from search results for 'sand'? [22:13:31] yup [22:13:45] OH [22:14:02] MaxSem: I don't. Using Chromium on Linux, checking FF/OSX just to see [22:14:04] actually, the search result for \sandbox sends the user to: http://test.m.wikipedia.org/wiki//Sandbox (not url escaped) [22:14:09] right [22:14:48] it's a bit tricky testing search on test.wiki - i wish we had a list of articles we knew to search for that at least sort of replicate wikipedia functionality [22:14:58] there's san francisco [22:15:11] what else..? [22:15:23] i think the /sandbox case is a MF bug but probably not a show-stopper [22:15:24] 'unicorns' used to work, but it's small now and has some Russian [22:15:36] is the obama one always there Maryana? [22:15:38] * Maryana eyes maxsem suspiciously [22:16:00] * MaxSem прикидывается шлангом [22:16:02] oh odd, the 'sand' search in FF gets the correct result [22:16:20] the ribbon looks broken with nothing in it when it expands [22:16:26] maxsem: haha [22:16:41] the slash is rendered the other way, .../wiki/\Sandbox" ha [22:17:54] the page for "golden-crowned sparrow" has a lot of stuff on testwiki just fyi [22:17:56] barack obama - yep, another good 'un. all looks good on the search front on iOS for beta and non-beta [22:18:26] it's just slow [22:18:38] Maryana I squirreled that away last time :D [22:29:28] MaxSem: any idea what's going on with the footer? is it the RL issue or is something actually busted? [22:29:46] K so other than the things that have been brought up. I think everything looks good (at least to my untrained eye) [22:30:11] footer WFM - caching? [22:30:17] i haven't tested anything on android or opera mini, but i figure maxsem will get the latter (he always does) [22:30:18] MaxSem: could be [22:30:38] I test on desktop Opera [22:33:35] hmm yeah some pages ahve broken footer + language links section header (with no languages) but some do not [22:35:42] watchlisting works just fine, it seems [22:36:00] hmm but debug=true does not fix the footer/language issues [22:36:04] s/but/bug [22:36:11] der [22:36:14] nm [22:36:21] bug debug? [22:36:26] :) [22:36:35] * awjr needs more coffee [22:37:32] awjr, can you post examples here? [22:37:49] sure [22:37:52] home page and Unicorn [22:37:56] are broken [22:38:05] it appears they are getting the same js as everyhting else [22:39:22] unicorn looks ok to me.. [22:39:58] i am seeing 'Read in another language' at the bottom of the articles and the footer will not expand [22:40:10] beta or non-beta? [22:40:12] on ios6, desktop chrome, android 4.1 stock browser [22:40:13] beta [22:40:37] is there another unicorn article? [22:40:55] not seeing read in another language button and the footer expands [22:40:58] http://test.m.wikipedia.org/wiki/Unicorns [22:41:00] on ios6 chrome [22:41:08] hmm, home's broken for me too [22:41:15] home's broken, yeah [22:41:45] ReferenceError: Undefined variable: actionMenuButton [22:41:56] TypeError: Cannot convert '$this.attr('id')' to object [22:42:01] ah indeed [22:42:07] aaand now it's broken [22:46:15] New patchset: Jdlrobson; "add id's to the language section so that javascript doesn't break" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28169 [22:46:51] jon is apparently reading our minds. [22:47:34] a wild jdlrobson appears! [22:52:00] New patchset: Jdlrobson; "add buttons for styling and languageSelector class to language selection button" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28170 [22:54:03] review, anyone? [22:54:20] i have been checking those do not appear to be fixes to the problem [22:54:29] :( [22:54:30] i sent jon an email about the issue we're seeing and am trying to understand how this works [22:54:37] but i am no js master :( [22:55:27] our deployment window is about to end, correct? [22:55:54] no [22:55:57] we have until 5 [22:55:58] New patchset: Jdlrobson; "die quietly when no other languages for an article" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28171 [22:56:04] ah, cool [22:56:13] ^ that might fix the problem [22:56:37] jdlrobson needs to stop reading beat poetry [22:57:09] it's making his patch titles really depressing [22:57:33] lol [22:57:38] running around corner to get the boy [22:57:50] back in a bit [22:57:58] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28171 [22:58:04] MaxSem: ^ [22:58:34] i havent been able to replicate the broken footer expansion locally, but that patchset fixes the JS error for me [23:00:34] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28169 [23:01:09] MaxSem: can you get those two patchsets up on test? [23:04:33] we're live [23:04:42] test please [23:04:57] * Maryana looks [23:05:05] yeah that looks better [23:05:16] success! [23:11:43] * awjr wants sticky nav [23:12:48] ruhroh [23:12:51] found a bug [23:13:12] the new dynamic loading of images in sections does not appear to honor disabled images [23:14:14] sux [23:15:52] meeting, brb [23:15:53] that's the only outstanding issue i've been able to find [23:16:12] are they that good? [23:16:48] eh? [23:16:58] you said it was outstanding [23:17:04] AHHAHAHAHA [23:17:06] funny. [23:17:26] sooo... [23:17:46] MaxSem: wondering whether or not to push to production? [23:17:55] yup [23:18:17] Maryana: we're waiting on you to signoff before pushing to production - is that image loading issue a show stopper? [23:21:19] i hope she went to a short meeting. [23:23:41] ...otherwise, I'll be reverting soon [23:24:19] back [23:24:39] * awjr waves [23:24:52] howdy guys :) [23:25:15] jcmish found an issue where when images are disabled, they still get loaded in dynamically loaded sections [23:25:27] waiting on Maryana to make the call whether or not to push to prod but it looks like she ran into a mtg [23:25:40] you have time to find more bugs:) [23:25:55] i want to go work on login! [23:26:20] MaxSem: let's give her a few more minutes otherwise just revert and we'll try again tomorrow. [23:26:57] hehe make myself useful testing [23:30:17] okay, I declare that with 30 minutes left before the end of the window I'm not going to deploy [23:30:43] make it so [23:32:49] hmmm [23:33:02] looks like before me we ran on https://gerrit.wikimedia.org/r/#/c/28030/ [23:34:09] er actually MaxSem im not sure - i think Reedy cherry-picked that change into prod [23:34:20] MaxSem: did you only update code in 1.21wmf2? [23:34:24] Nope, it failed and I reverted [23:35:08] It looks like that isn't causing any noise in production, so it's not urgent [23:35:24] hi, is anyone here who knows of recent changes to the mobile version that might affect screen-reader software? a visually-impaired user reported that there is no longer any possibility for him to expand article sections. [23:36:00] pill, when this started? [23:37:16] Reedy, so when creating a new branch you simply use master MF? [23:37:30] MaxSem if you only merged code to 1.21wmf2 you probably don't really need to revert as the only wikis running that branch are test, test2 and mediawiki [23:37:53] MaxSem: by default, yeah [23:37:54] mhm [23:38:37] make-wmf-branch should allow for specifying revisions/hashes or whatever [23:38:37] recent weeks, apparently it still worked on october 3. [23:39:03] awjr, would still be fun cuz when someone runs scap they might deploy these changes to mw.org [23:39:52] oh, yeah i guess MF is enabled on mw.org, just no auto redirect [23:40:16] so I'll have revert wmf2 to wmf1 [23:44:04] MaxSem: while you're taking care of that I can mail out about today's deployment [23:45:11] cheers, that would be great [23:45:27] he's using a nokia n82 with talks and wrote that he had already enquired the producer of that screenreader about the issue. apparently it was figured out that only the latest smartphones with an experimental version of the software can still handle wikipedia content (on m.wikipedia.org). [23:46:10] have there been any significant changes recently? [23:46:19] we deploy every week [23:46:42] is screenreader support tested in some way? [23:47:59] no [23:48:04] sadly [23:48:25] could you connect us with that user directly? [23:48:27] ok. hmm, could it make a difference if he switched to the beta version? [23:48:31] pill we do not regularly test screenreaders - also, symbian (the operating system running on n82) is not a part of our regular testing regimen [23:50:20] bah, sorry guys - back [23:50:35] hi Maryana, Max made the call to postpone deployment til tomorrow [23:50:44] since we were waiting on the goahead from you [23:51:08] sorry bout that :( did you fix the image loading issue? [23:51:26] nope not yet [23:51:53] Maryana: do you regularly meet at 4pm on Mondays or was that a one-off mtg? [23:52:14] yeah, weekly product mtg w/howie and steven. should probably change the time on that [23:52:43] oh, and i see that on the engineering calendar in google calendar the window is only 3-4, while on the deployment page it's listed at 3-5 [23:53:02] yeah, i had it as 3-4 on my cal, too [23:53:06] lame alert [23:53:09] i'll email rachel about it [23:53:13] okay, https://test.m.wikipedia.org/wiki/Unicorns has been rolled back [23:53:32] \o/ [23:53:39] good thing there's another deployment window [23:53:40] :) [23:54:53] MaxSem: he has emailed to info-de, i can give you the email address, though i'm not sure if he speaks english. shall i ask? [23:55:39] pill, can you ask him to email mobile-l? [23:56:09] in english? [23:57:06] hmm [23:57:08] pill if he can, otherwise german is fine, we can find someone to translate :) [23:58:57] ok, i will ask him. thanks for your help. [23:59:20] mobile-I@lists.wikimedia.org, i guess? [23:59:23] wee ... [23:59:51] awjr: thanks for the deployment update