[00:07:48] New patchset: MaxSem; "Fix special pages spewing modules to Special:Specialpages" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63806 [00:07:53] jdlrobson, ^^ [00:08:38] setHeaders? [00:08:51] but it doesn't set any headers o_O [00:09:21] why is this cleaner? awjr any thoughts? [00:09:27] it sets some elements [00:09:34] looking jdlrobson [00:09:38] well, modules are also part of head [00:10:05] New patchset: Jdlrobson; "Fix special pages spewing modules to Special:Specialpages" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63806 [00:10:16] i just changed the commit message to give some more detail [00:10:27] MaxSem: well i'll leave awjr to decide [00:10:41] it will work just a case of whether this is the right way to do it [00:11:17] I prefer to have 1 function that must be called by every execute(), instead of 2 [00:15:11] [Commons-iOS] montehurd opened pull request #68: Over 500 uploads (master...over500Uploads) http://git.io/YL__qA [00:18:11] i think the approach is sound, jdlrobson and MaxSem and I agree with MaxSem [00:18:19] ok cool [00:18:25] MaxSem: go for it [00:18:45] it's one less thing for someone to worry about implementing a mobile special page, since you have to call setHeaders() anyway [00:20:58] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63806 [00:21:08] jdlrobson, MaxSem ^ [00:21:33] awjr: sweet [00:31:03] deployed [00:40:35] New patchset: Dr0ptp4kt; "Adding redirect query params to be used by MobileFrontend for footer." [mediawiki/extensions/ZeroRatedMobileAccess] (master) - https://gerrit.wikimedia.org/r/63811 [00:53:00] i just realized Special:Nearby on desktop probably should not have the add image CTA [00:53:04] jdlrobson: ^ [00:53:38] it doesn't awjr [00:53:45] yes it does [00:53:47] im looking at it :) [00:54:02] no i'm saying it does have the cta on the nearby page but doesn't on the article page [00:54:12] this is a feature to show how desktop should improve [00:54:21] right, but it's misleading [00:54:22] the desktop workflow is click edit button ;-) [00:54:26] lol [00:54:33] awjr: it's not.. you can still visit the page to edit an image :) [00:54:55] watch star isn't showing for some reason on nearby though :/ [00:55:17] New patchset: Dr0ptp4kt; "Adding intersitials to footer navigation for Wikipedia Zero users." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63814 [00:55:22] Maryana: thoughts? ^^ [01:03:14] well, here's where following the copy in the mingle card might have helped [01:03:14] grumble grumble [01:03:17] but i think it's fine [01:03:25] it actually functions like a work queue [01:03:33] Maryana: it would have been misleading regardless [01:03:39] "add image" is still misleading on desktop :) [01:03:45] no, it's not [01:03:50] also what about non logged in users? [01:03:55] you can add a photo from commons [01:03:58] anons can, too :) [01:04:06] you just paste in the markup [01:04:14] Maryana: heh markup [01:04:33] has anybody mentioned this on the enwiki village pump? [01:04:39] we should probably let folks know about this new feature [01:05:35] or we could wait for the blog post if you want to add in some fixes, jdlrobson [01:05:38] what do you think? [01:06:32] I think I need to go home right now - but I'd like to write that blogpost - worth a little note on village pump though as a heads up [01:07:21] it's an unlisted special page so it's not very discoverable, but there's been enough public discussion about it (particularly after the bug that was fixed a bit ago with nearby stuff leaking to Special:SpecialPages) that it's probably worth mentioning on the VP [01:07:48] if we're proactive in discussing it we'll probably be better off [01:08:09] yeah, true that [01:08:25] i always think 'what would CJ Craig do?' [01:09:07] woah… uh [01:09:15] there's some weird stuff going on in desktop nearby [01:09:25] scroll to the bottom of the page [01:09:29] are those… categories? [01:09:31] what is that stuff? [01:09:51] jdlrobson ^ [01:11:23] New review: awjrichards; "I haven't looked particularly closely at this yet, but is there a way to pull this off in Zero rathe..." [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/63814 [01:12:12] im not seeing anything weird, Maryana [01:12:21] try hitting search [01:12:30] wait for it.. wait for it [01:12:50] whoaaaa [01:13:02] that is f'd. [01:13:11] yeahhh [01:13:20] thats our suggested search styling [01:13:33] too much mobile leakage [01:13:37] if you scroll all the way up, there's a hovering clear search x! [01:13:39] :) [01:13:47] ok, let's, uh, fix this soonish [01:13:48] trippy! [01:14:04] Maryana: i'll file a bug [01:14:11] thanks awjr [01:14:32] i gotta run, but i'll leave a note about this on the enwiki vp and mention that we're working on fixing the bug [01:14:43] transparency :) [01:15:40] ok, ttyl, gang [01:15:46] good deployment mostly! [01:41:38] New review: Yurik; "Just to clarify -- apparently the links about license at the bottom of the page are rendered by the ..." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63814 [02:36:51] New patchset: Jdlrobson; "Bug 48490: Report correct mode and whether to disable images or not" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63822 [04:27:49] New patchset: Yurik; "Cache invalid X-CS IDs" [mediawiki/extensions/ZeroRatedMobileAccess] (master) - https://gerrit.wikimedia.org/r/63628 [15:51:02] New patchset: Jdlrobson; "Bug 48485: Stop mobile leaking into Special:Nearby on desktop" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63863 [16:06:33] wat? again? [16:06:53] MaxSem, will you be able to baby sit in a half an hour? [16:07:01] sure [16:07:05] thx [16:11:34] New review: MaxSem; "My comments now that I've had sleep." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63814 [16:12:50] New review: MaxSem; "Main concern: this should be done in Zero - MF just needs to provide a way to do that, e.g. with hooks." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63814 [16:18:31] Change merged: Jdlrobson; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63599 [16:21:30] ooooooh google cloud messaging (push notifications) now support dismissing the notification once for all your devices [16:21:33] that's..... awesome [16:22:48] all riiiiiight another platform-specific gaming middleware [16:23:43] hmm, supports ios and web too. interesting [16:27:58] are there any issues with storing an empty string in memcached? [16:28:10] (with a proper key of course) [16:28:39] hmm [16:28:52] it seems to work in VM [16:28:54] i think an empty string should store..... but i'm not 100% sure :D [16:29:15] hehe, well, lets assume it is, because it tests fine [16:29:25] :D [16:29:48] brion, is dr in the house? [16:30:09] (adam) [16:30:14] ooh, they're making an android-branded version of intellij with all the right goodies installed [16:30:23] i'm at the home office today [16:30:30] hehe [16:31:23] i'm watching a liveblog of the google i/o keynote from the comfort of my home while tfinc sits in a crowded room watching it live [16:31:26] mwahaha [16:32:00] MaxSem, could you +2 https://gerrit.wikimedia.org/r/#/c/63628/ since adam is not around? i want to include that in the release today [16:32:23] looking [16:32:45] do you already have +2 on deployish stuff btw? [16:33:03] MaxSem, what do you mean? [16:33:17] everything i want to deploy has already been merged [16:33:21] except that patch [16:33:24] mediawiki config and wmf branches [16:33:36] *+2 rights [16:33:50] i think so - last time i checked in config changes [16:33:54] and self +2 [16:34:08] wmf branches i'm not exactly sure - what repo are you takling about? [16:34:16] mw [16:34:45] * brion waves tfinc  [16:35:20] MaxSem, you mean core? [16:35:26] yup [16:35:29] MaxSem, yep [16:36:09] Yo [16:36:09] Greetings from the keynote [16:36:38] i'm watching a liveblog :) [16:36:49] Google play studio looks really intersting [16:36:49] Google android studio* [16:37:05] yeah, that'll be nicer than maintaining intellij manually i think [16:37:10] tfinc, we just have to deploy on glass, make sure you get a few [16:37:13] i really like the view preview stuff [16:38:00] Nice beta testing and staging support [16:38:06] ooooh [16:38:14] Built into play [16:38:32] * brion beams "copy this please" messages to apple [16:39:14] Yuck its connected to Google+ [16:39:14] I hope you can use other means [16:39:41] oh joy [16:39:47] that works so well for us with hangout [16:39:53] so well i have to use a second browser :P [16:43:01] yurik, is Zero from wmf4 much different from what we have in master? [16:46:12] Change merged: jenkins-bot; [mediawiki/extensions/ZeroRatedMobileAccess] (master) - https://gerrit.wikimedia.org/r/63628 [16:50:17] MaxSem, not sure what's in wmf4, master has had a significant refactoring work and a few minor fixes [16:51:48] MaxSem, ok, ready when you are [16:52:26] yurik, I have a standup in 8 minutes - we could start after it or start now and interrupt [16:52:57] after is fine [16:53:02] shuold be enough time [16:53:09] no need for settings changes, etc [17:08:22] whee, vp9 codec to roll out on youtube later this year [17:08:36] another format for us to throw into our mix i suppose :P [17:11:51] MaxSem, ping me when ready [17:13:18] yurik, ping [17:19:58] hehe google hangout audio fail :) [17:20:57] brion: i am not liked by the chrome box [17:21:00] mhurd, ok let's just go over what we're working on briefly :D [17:21:09] snds good [17:22:02] so the >500 uploads should be done, i just have to merge that and it looks like it should work :D [17:22:37] hopefully the >500 fix is cool. i was going to ask you what number it should request - still 500? 100? [17:22:53] let's keep it at 100 at a time and see how that does [17:23:02] sounds good [17:23:03] smaller batches keep the io smoother in theory [17:23:04] Maryana: can you please poke jon to get on irc? [17:23:18] mhurd, so the main ios thing sitting in the todo queue is 'show and edit categories on already-uploaded photos' [17:23:25] brion: ya [17:23:25] that's going to be a little scary :D [17:23:31] but i think you can handle it :D [17:23:35] brion: :) [17:23:45] brion: i will give it a go! [17:23:56] mhurd, why don't you start with fetching and displaying the categories [17:24:04] and we'll follow up with the editing part, which is trickier [17:24:07] brion: sounds good! [17:24:29] awjr: poked [17:24:32] thanks :) [17:24:42] you'll need to fetch the categories on a page from the mediawiki api, use the sandbox to get the right query [17:24:57] brion: woohoo! [17:25:02] and then tweak things to show the categories section with your new data [17:25:15] so that'll be some fun table-view stuff [17:25:39] brion: looking forward to more mediawiki api stuff! [17:26:16] actually editing them will be a laugh riot... ;) i'll write up some notes on how to go about it [17:26:26] hope you like xml [17:26:32] brion: lol [17:26:32] sorry awjr no idea why but for some reason my irc doesn't always connect automatically [17:26:34] or we can give up and use regexes ;) [17:26:35] whats up? [17:27:03] jdlrobson: nevermind, i just answered my own question :) [17:27:18] brion: yay regex! [17:27:49] brion: always had an irrational fascination with them [17:27:53] jdlrobson: see comments in https://gerrit.wikimedia.org/r/#/c/63822/1, some minor fixes would make jeroen happy [17:28:00] i'll take a look at the other patchset now [17:30:01] arg awjr looks like i committed too much :-S [17:30:12] SkinMobile.php changes were not meant to be there [17:30:13] eh? [17:30:42] heh, i was wondering about that, it seemed extraneous :p [17:31:32] [ 'wgImagesDisabled' ] / ['wgImagesDisabled'] whatever.. [17:31:53] yeah [17:32:05] personally, i often do [ 'foo' ] rather than ['foo'] anyway... [17:32:11] it's just esier to read. [17:32:27] anyway, is there a difference between 'universal' and 'common'? [17:32:29] jdlrobson: ^ [17:33:57] if not, can we standardize one one? [17:34:04] awjr: ohh awjr so [ 'foo' ] has been the convention of that function so far.. [17:34:12] i guess i was just being consistent with rest of function [17:34:22] awjr: i'll look at that commit after this one is done ;-) [17:34:26] kk [17:36:53] New patchset: Jdlrobson; "Bug 48490: Report correct mode and whether to disable images or not" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63822 [17:36:53] New patchset: Jdlrobson; "Code cleanup: [ '..' ] to ['']" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63883 [17:36:53] ok awjr ^ [17:37:00] first one is the important one :) [17:38:24] awjr: that ok? [17:38:50] looking jdlrobson [17:40:11] looks sane jdlrobson, testingn ow [17:42:20] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63822 [17:42:53] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63883 [17:44:03] sweet. ok next one [17:44:15] New review: awjrichards; "(1 comment)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63863 [17:45:30] New patchset: awjrichards; "Bug 48485: Stop mobile leaking into Special:Nearby on desktop" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63863 [17:45:55] awjr: agreed with your comment - let me do some cleanup [17:46:08] cool jdlrobson;im testing that patchset now [17:49:57] mhurd, notes for the editing portion later: https://www.mediawiki.org/wiki/Wikimedia_Apps/Commons/Category_editing [17:50:05] feel free to annotate that with questions and "what the hell"s :) [17:50:13] brion: awesome thanks!! [17:51:01] brion: playing in the api sandbox now using your live category mockup as a guide [17:51:08] whee :D [17:51:59] jdlrobson: im not seeing the watchlist star on nearby in alpha with that changeset [17:52:41] awjr: 1) you must be logged in 2) you must have a local copy of the article 3) it doesn't work when you load from cache 4) you must be in alpha [17:52:46] do you adhere to all the above? [17:52:52] oh right, im not in tucson anymore [17:53:04] i do except for 2 [17:53:22] New patchset: Jdlrobson; "Bug 48485: Stop mobile leaking into Special:Nearby on desktop" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63863 [17:53:24] also ^ [17:53:30] i standardised on .common [17:53:43] w00t [17:55:44] ok looks good, merging [17:56:07] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63863 [17:56:36] MaxSem: french nearby is broken :( [17:56:41] https://fr.m.wikipedia.org/wiki/Sp%C3%A9cial:Nearby < no geodata? [17:57:29] MaxSem: those fixes are now reviewed and merged. do you think you guys can get those out now or should we wait til this afternoon? [17:57:35] https://gerrit.wikimedia.org/r/#/c/63863/ and https://gerrit.wikimedia.org/r/#/c/63822/2 [17:57:46] French? it's not English - why do you care? :P [17:58:06] Fatal error: Cannot redeclare fn() (previously declared in /usr/local/apache/common-local/php-1.22wmf3/extensions/MobileFrontend/includes/skins/SkinMobileBase.php:162) in /usr/local/apache/common-local/php-1.22wmf3/extensions/MobileFrontend/includes/skins/SkinMobileBase.php on line 162 [17:58:12] LIVE ON WP [18:00:06] MaxSem: i'm learning french i'll have you know ;-) [18:00:13] MaxSem: ? where's that happening... [18:00:28] on wikipedia [18:00:28] which page? i don't see it.. [18:00:39] it's in logs [18:00:50] err… this got introduced by my patchset? [18:01:05] no [18:01:07] i haven't touched SkinMobileBase.. [18:01:24] i dont really understand that error, MaxSem [18:01:37] it's in mobilizeUrls [18:01:48] yeah, it's encapsulated [18:02:03] is this an error which showed up since last deployment ? [18:03:28] MaxSem: ? [18:03:42] yes [18:04:56] oh i see, if mobilizeUrls gets called more than once it will fatal [18:05:56] but why would it get called more than once [18:06:12] is there a stack trace MaxSem awjr ? [18:06:17] no [18:07:26] awjr: only place it gets used is to mobilize urls in language list [18:07:38] yeah i dont see why it would get called twice [18:07:45] nested functions in php are kinda weird [18:07:54] i haven't really used them [18:07:56] http://stackoverflow.com/questions/415969/what-are-php-nested-functions-for [18:07:59] is there a better way to write mobilizeUrl ? [18:08:08] well the easy thing would be to just foreach [18:08:29] maybe worth doing that to play it safe? [18:08:45] I think a closure is better in PHP than a nested function because AFAIR nested functions are defined in global scope anyway (!) [18:09:36] something like this https://gist.github.com/anonymous/3d218cdde7495cf7bac5 [18:09:43] oh yeah, that example on stackoverflow explains this [18:09:52] yeah [18:10:00] wait [18:10:23] we should make unit tests for our skins [18:10:39] you can do array_map( function( $url ) { ... }, $urls ); instead [18:10:46] yeah you could do that but... [18:10:51] and it won't leak into the global scope, right? [18:10:55] (because it has no name) [18:10:59] no, it's anon [18:11:00] yeah [18:11:25] oh shi... [18:11:29] <3 PHP [18:11:35] ;) [18:12:19] hehehehe [18:13:17] back to French stuffs [18:13:40] jdlrobson, wanna know how to tell if an extension is installed on a wiki? [18:13:47] Special:Version [18:13:55] brion: pulling categories nicely. now to display them… just below the title on the details page for now? [18:14:01] MaxSem: my concern is the nearby page shows up despite not being supported [18:14:03] it should 404 [18:14:25] mhurd, so the details page is a table view with two sections [18:14:33] the first section is static cells, that's the title/description etc [18:14:34] brion: ya [18:14:46] on completed images we crop off the section at the title cell [18:14:50] (nearby link doesn't show up in list on french site) [18:14:50] will you create a bug for every wiki that doesn't use {{#coordinates:}} ?:P [18:14:54] ohh wait it does.. [18:14:54] the second section lists the categories [18:15:06] MaxSem: well the api could return something better when no coordinates exist [18:15:07] so you want to modify the logic that hides the second section [18:15:12] jgonera: i think https://gist.github.com/anonymous/4bd6e9ebccdbcbd3228c would work too but i dont have languages set up locally so i can't test [18:15:12] brion: k [18:15:14] and the logic that fills up a list of categories to display [18:15:30] brion: hijack that? [18:15:35] yup [18:15:36] i dont know if there's really any benefit to that approach vs just doing a foreach [18:15:37] hehe [18:15:38] MaxSem: we can't deploy broken nearby pages to wikipedia's which have no coordinates [18:15:39] the problem is that there might be a few coordinates -from EXIF in images [18:15:52] the code for display is already in there for new images, you just have to take it over :D [18:15:55] so there's no way to tell [18:16:03] brion: awesomeness! [18:16:14] MaxSem: well we need to find a way.. it's unacceptable from a user perspective - plus it makes us look bad [18:16:22] unless you want to enable nearby on a opt-in basis [18:16:38] MaxSem: you mean by a global variable? [18:16:53] awjr, should work, fine, at least on PHP5.3+ [18:16:56] yup. like $wgMFNearby;) [18:16:57] do we support PHP older than 5.3? [18:16:58] $wgMFNearby should be false unless a project has geodata [18:17:03] jgonera, no [18:17:10] i don't know why it's enabled for all projects [18:17:26] and when do we plan to support only 5.4? I'd love to use the short array syntax ;) [18:17:27] is geodata enabled but not used on the tohers? [18:17:37] in 2-3 years [18:17:38] or is it not even enabled? [18:18:09] jgonera: what's the short array syntax? i havent seen that [18:18:17] brion, the former [18:18:23] awjr, [] [18:18:31] array( 'a' => 'b' ) can be simply [ 'a' => 'b' ] [18:18:39] oh, like $foo = ['bar','baz'] [18:18:40] yeah [18:18:43] yes [18:18:50] well maybe showing the feature will encourage those communities to make use of geodata :D [18:19:09] brion, the only real soltion is Wikidata [18:19:15] Change merged: Raimond Spekking; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63860 [18:19:19] 6 Warning: array_key_exists() expects parameter 2 to be array, null given in /usr/local/apache/common-local/php-1.22wmf3/extensions/ZeroRatedMobileAccess/include [18:19:19] s/PageRenderingHooks.php on line 564 [18:19:19] 6 Warning: array_key_exists() expects parameter 2 to be array, null given in /usr/local/apache/common-local/php-1.22wmf3/extensions/ZeroRatedMobileAccess/include [18:19:19] s/PageRenderingHooks.php on line 554 [18:19:37] jgonera: do you have languages set up locally? if so can you test one or both of those diffs for me? [18:19:50] dr0ptp4kt, yurik ^^ [18:20:05] awjr, I have only a single language wiki [18:20:17] jdlrobson: do you have multiple langs set up? [18:20:21] or MaxSem? [18:20:27] meanwhile... do we already have a fix for that fatal? [18:20:36] working on it MaxSem, need to test [18:20:36] awjr, I don't [18:20:40] Reedy, zero extension messages have changed for the meta [18:20:42] actually, is staging set up with multiple langs? [18:20:45] or beta? [18:20:50] MaxSem: ? [18:20:51] ^ [18:21:04] beta is [18:21:17] i didn't scap, and it really shouldn't affect anyone except meta zero admins [18:21:18] ok, let me submit a fix and then we can dbl check on betalabs before pushing [18:23:06] Reedy, oh, this is something else, sorry, checking [18:23:19] awjr: single language only i'm afraid.. [18:23:24] no worries [18:23:32] gah [18:23:33] git weirdness [18:24:23] git pull/fetch seems to be timing out for me... [18:25:10] http://www.php.net/manual/en/migration55.new-features.php oh my, PHP is finally getting some features other languages had for years! in like five years when everyone uses 5.5 I'll be able to do 'PHP'[0] == 'P'... [18:25:24] exciting [18:26:02] PHP parser must be o total mess if they need to "implement" something like this as a special case [18:26:07] o=a [18:26:40] let's not belittle other peoples' parsers [18:26:43] ours ain't so hot either ;) [18:26:55] haha [18:27:08] well wikitext never aspired to be a programming language ;) [18:27:32] {{{{{{{{{{{{{{{{{{{oh yeah?}}}}}}}}}}}}}}}}}} [18:27:32] yurik, i'll turn the footer rewriter into a Zero hook [18:27:50] ;) [18:27:51] New patchset: awjrichards; "Anonymize nested function to fix PHP fatal on multiple executions" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63892 [18:27:57] dr0ptp4kt, sounds good, make sure MaxSem is happy ;) [18:28:02] this is kinda cool http://www.php.net/manual/en/session.upload-progress.php [18:28:43] brion, i don't think all your braces match [18:28:56] that's the beauty of wikitext, they don't have to [18:29:02] good point [18:29:16] MaxSem: https://gerrit.wikimedia.org/r/63892 [18:30:43] is anyone else having issues with gerrit right now? [18:30:58] things are moving *very* slow for me [18:31:09] yurik, yes, i am having problems with gerrit [18:31:37] dr0ptp4kt, you are not alone [18:31:58] grrrrrrrrrrrrr [18:32:07] ruhroh [18:32:12] getting 500s from gerrit [18:32:56] google i/o keynote video crashed my linux laptop. :P [18:33:18] you probably should have been using chromeos [18:35:40] gah [18:35:48] gerrit is hose [18:35:48] d [18:36:03] oh [18:36:04] ^demon [18:36:04] 11:34 [18:36:04] !log restarting gerrit [18:38:18] New review: awjrichards; "janky jenkins?" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63892 [18:38:57] Change merged: MaxSem; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63892 [18:56:07] so i dunno how that fatal was getting triggered, but language links seem to be working fine on betalabs, MaxSem [18:56:21] okay [18:56:30] MaxSem: so i guess we have a handful of patchsets to deploy - should we coordinate with greg-g a time to push? [18:56:44] greg-g, we need to deploy a few MF fixes - when can we do this|? [19:01:34] damn it MaxSem [19:01:34] drwxr-xr-x 2 maxsem wikidev 4096 May 14 21:25 87 [19:01:35] drwxr-xr-x 2 maxsem wikidev 4096 May 15 00:29 8a [19:01:35] drwxr-xr-x 2 maxsem wikidev 4096 May 15 00:29 8b [19:01:35] drwxr-xr-x 2 maxsem wikidev 4096 May 14 21:05 93 [19:01:58] 55 top level folders you've no group write on :9 [19:02:16] fixing [19:02:27] done [19:02:49] ta [19:03:28] lunch…. [19:04:14] eh, what should my umask be set to? [19:05:17] right, 2 [19:26:44] reminder: story review is happening in 5 minutes in R66 [19:26:50] jdlrobson, jgonera [19:27:07] Maryana: yeh i just remembered but forgot to get food :/ [19:27:27] and praveena says we've gotta be out of there at 1:30 sharp, because the FDC is having their quarterly review in there [19:35:52] technical difficulties awjr and MaxSem - just getting setup [19:35:58] jdlrobson: figured, thanks [19:36:59] sigh [19:38:09] someone unplugged the hdmi cable [19:40:01] New patchset: Jdlrobson; "Redirect page diffs to Special:MobileDiff" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63907 [19:40:02] New patchset: Jdlrobson; "Alpha: Use diff page for history" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63908 [19:51:56] MaxSem, sorry to bug you again, can't figure out where the bug is coming from in prod [19:52:18] yurik, we're on a meeting [19:52:23] sorry [19:53:35] MaxSem, np, ping me later. I'm getting a warning in production... worysome :) [19:55:29] hah [19:55:35] not even a fatal [20:29:51] all right, deployment time! [20:30:27] am I right that we can just deploy master because everything merged so far we those fixes? [20:33:03] dr0ptp4kt: Is Maryana able to set up google hangout for us? [20:33:05] dr0ptp4kt, Maryana is there a hangout? [20:33:08] hehe [20:33:18] :) [20:33:22] yurik, Maryana is adding that now [20:33:28] link should be in the mail [20:34:16] awjr, jdlrobson: am I right that we can just deploy master because everything merged so far we those fixes? [20:36:49] MaxSem: looking [20:38:23] MaxSem: loks like it [20:38:25] *looks [20:38:29] i think it's safe to deploy master [20:38:35] all right [20:39:41] oh man i dug out my old iphone 3gs [20:39:44] THE PIXELS THEYRE SO BIG [20:40:09] brion: haha [20:40:14] technically it's the same density as the ipad mini, but those pixels just *feel* chunkier [20:40:38] lol [20:41:21] brion: hey quick question - when i do this: [20:41:24] format: 'json', [20:41:25] action: 'query', [20:41:26] prop: 'revisions', [20:41:28] rvprop: 'content', [20:41:29] rvparse: 1, [20:41:30] rvlimit: 1, [20:41:31] rvgeneratexml: 1, [20:41:42] i get a large description - not just the one the user entered [20:41:58] how would i get just the user descrip? [20:42:11] mhurd, ahhh now you're encountering the fun ;) [20:42:21] you have to kinda extract it from the page [20:42:23] oh i see now [20:42:27] there may be descriptions in multiple languages, also [20:42:31] in the mockup you're parsing [20:42:56] i'm on it :) [20:43:42] awjr, jdlrobson, jgonera - new code is live live on testwiki [20:45:41] groovy [20:45:53] jdlrobson: can you confirm your wgMFMode fix on testwiki? [20:46:51] Special:Nearby looks good [20:48:12] meh [20:48:17] dunno where he is, maybe getting food [20:48:23] MaxSem: i *think* things look ok [20:48:29] we're in a mtg w/zero folks [20:48:41] awjr ^ [20:49:04] roger, thanks Maryana - MaxSem is about to push out the bug fixes and the fix for the fatal error [20:50:21] sweet [20:50:47] one +1 for special:nearby on the enwiki vp from TheDJ :) [20:51:25] sorry in a zero meeting [20:55:04] alrighty, we're live [20:55:22] w00t - jdlrobson when you have a chance plz confirm that the wgMFMode issue is fixed in prod [20:56:00] oh wait… special:nearby doesn't seem to be working on desktop anymore :-/ [20:56:05] nm [20:56:08] i take that back [20:57:02] doh fatalmonitor doesnt work on tin :( [20:57:44] looks like that fatal is fixed tho, MaxSem [21:04:47] ok, lunch time [21:05:00] thanks for getting those fixes out, MaxSem [21:11:53] brion, what is the extension for captcha we use in production? [21:12:08] jgonera, ConfirmEdit using the 'FancyCaptcha' mode [21:13:00] thx [21:20:14] New patchset: JGonera; "Fix captcha field margin on create account page" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63966 [21:33:01] New patchset: Yurik; "TEMPORARY patch to help find the source of warnings in production" [mediawiki/extensions/ZeroRatedMobileAccess] (master) - https://gerrit.wikimedia.org/r/63969 [21:33:56] MaxSem, i got the patch ready, what are the depl steps? [21:34:26] we can poke at it tomorrrow while SF sleeps [21:34:42] MaxSem, how about i just push out one file? [21:35:01] ask greg-g [21:35:18] asking... [21:35:51] MaxSem, in general, the process should be fairly straightforward, right - update the git's module, and file-sync [21:36:30] MaxSem, grep is ok [21:36:34] grep [21:36:39] grr!!! Greg! [21:36:53] need to work on that muscle memory [21:36:57] lol @ macs [21:37:05] heh [21:38:01] MaxSem, dr0ptp4kt_ i'm about to +2 my own patch -- [21:38:06] https://gerrit.wikimedia.org/r/#/c/63969/ [21:38:10] yurik, have fun [21:38:20] dr0ptp4kt_, pls take a look just in case [21:38:28] New patchset: Dr0ptp4kt; "Adding intersitials to footer navigation for Wikipedia Zero users." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63814 [21:38:31] Change merged: jenkins-bot; [mediawiki/extensions/ZeroRatedMobileAccess] (master) - https://gerrit.wikimedia.org/r/63969 [21:38:52] MaxSem was quicker :) [21:39:30] New patchset: Dr0ptp4kt; "Adding redirect query params to be used by MobileFrontend for footer." [mediawiki/extensions/ZeroRatedMobileAccess] (master) - https://gerrit.wikimedia.org/r/63811 [21:56:17] New patchset: Yurik; "Revert "TEMPORARY patch to help find the source of warnings in production"" [mediawiki/extensions/ZeroRatedMobileAccess] (master) - https://gerrit.wikimedia.org/r/63977 [21:56:36] Change merged: Yurik; [mediawiki/extensions/ZeroRatedMobileAccess] (master) - https://gerrit.wikimedia.org/r/63977 [22:23:51] Maryana: i noticed you updated the a/c on https://mingle.corp.wikimedia.org/projects/mobile/cards/484 - think it/s ready to go? [22:24:35] yeah, jdlrobson and i did it together this morning [22:24:53] far out, can i move it to 'ready for dev' then, jdlrobson? [22:25:08] awjr: yup [22:25:27] cool [22:25:31] what about https://mingle.corp.wikimedia.org/projects/mobile/cards/495 Maryana, jdlrobson? [22:25:46] needs more disucssion [22:25:50] (design) [22:26:01] did you get vibha's email, jdlrobson? [22:26:23] she sent us both a mock of the menu item (don't think the copy is right, but other than that looks ok) [22:29:22] Maryana: i haven't checked my email in a while - i only check it at specific breaks - I just looked at it [22:29:24] looks BIG [22:30:10] we should put any assets in card or at least in mobile-tech though [22:44:10] awjr: i replied on your bug [22:44:16] looks like the API is being inconsistent like normal [22:50:06] jdlrobson: if it's consistently 'inconsistent'... [22:51:20] awjr: please don't encourage more gaffer tape - this should only be a last resort. We should strive to make our API's as easy to follow as possible as an API doing this sort of thing can put off a potential developer from using it. jgonera will agree [22:51:37] jgonera: context: https://bugzilla.wikimedia.org/show_bug.cgi?id=48512 [22:51:39] meh, so fixit :) [22:52:09] awjr: at very least it should pass an error [22:52:23] well,no results is different than an error state [22:52:31] but i dont really know how that api works [22:52:55] awjr: true that's why I say it should return {"query":{"pages":{}}, "limits":{"coordinates":500} } - the structure of the empty state is completely inconsistent [22:53:13] yeah, i agree that that would make a lot more sense [22:53:21] jdlrobson: https://mingle.corp.wikimedia.org/projects/mobile/cards/495 [22:53:29] unblocked now? [22:53:38] but it's frustrating that we keep finding bugs because of api 'inconsistencies', you know? [22:53:54] awjr: yes.. but it's not my job to fix the API. I'm a frontend developer [22:54:02] I just report how I use it and problems I come up against [22:54:12] we should make the API friendlier to the developers that use it [22:54:17] we should expect the inconsistencies and be defensive about these things (and test for the different possible states) [22:54:57] awjr: i disagree. https://en.m.wikipedia.org/wiki/Broken_windows_theory [22:55:17] im not sure i follow how the broken windows theory applies here? [22:55:19] if we are defensive we are accepting it and these things will never change [22:55:29] they will encourage more hacks and messy code [22:56:21] http://www.codinghorror.com/blog/2005/06/the-broken-window-theory.html [22:56:55] I agree that the API is inconsistent, but I'm not sure if we should only complain without changing it ourselves. on the other hand, I understand it's difficult to find time to deal with it [22:57:06] jdlrobson: either way we need to *something* about it - either fixing the API or working with the API in its existing state. [22:57:41] jdlrobson: the benefit to being defensive is that if something in the future causes a bug or breakage in the API, our code can fail gracefully [22:57:45] awjr: i would fix the API myself but I imagine my investment in doing that would be much more significant then a backend engineer [22:57:47] imo BOTH should happen [22:58:31] jdlrobson: im not saying you specifically should fix the API [22:59:39] but i think the person owning a feature should be responsible for seeing that it gets done (if they believe it needs to be done) - so filing bugs, finding someone to help out, etc [22:59:51] awjr: all i'm saying is this behavior needs to stop ( grep -r FIXME javascripts/ is the proof) [23:00:19] jdlrobson: you are preaching to the choir :) [23:00:38] ok good i don't see where we are disagreeing - i'm simply saying i'm not getting the support to fix these API problem [23:00:56] what kind of support do you need to fix these api problems? [23:01:36] awjr: at the very least I need someone to entertain the idea rather than tell me "this is the mediawiki way" or "it just works that way" [23:02:00] jdlrobson: ok, i am entertaining the idea - what is next? [23:02:05] in this particular example can i really guarantee that if the page key is not present it means empty list - or maybe it means something else [23:02:46] * awjr shrugs [23:02:54] awjr: I guess the API code would be the starting point [23:03:12] who is the maintainer, jdlrobson [23:03:14] ? [23:03:26] good question awjr that's never been clear to me [23:03:50] i assume MaxSem knows that code well though? [23:04:55] uhoh, is MaxSem the maintainer? [23:07:56] ok so yeah it looks like that particular api is in GeoData, which is a MaxSem thing [23:08:14] have you talked to max about the inconsistency, jdlrobson? [23:08:18] but not API itself [23:09:22] * MaxSem wonders how many clients will be broken by such a change [23:09:26] oh, core API is where the problem is? [23:09:42] because it's an action=query [23:09:42] ? [23:09:56] yeah - I simply return it a page set [23:10:19] huh who does core api stuff these days? Reedy? [23:10:32] yurik? [23:10:47] lots of clients might check for query or pages existence [23:11:11] presumably they would, yeah [23:11:13] so it looks like a yet another API 2.0 wish [23:13:07] so jdlrobson improving the API is not out of the question, but it can't happen as quickly as we develop our features [23:13:31] MaxSem: who is leading the api 2.0 charge? [23:13:47] is there a charge? [23:14:05] I thought there's some guerilla movement [23:14:14] but not a cavalry charge [23:14:18] oh i see [23:14:37] i remember hearing grumblings about it maybe a year ago [23:14:41] but i can't remember from where... [23:18:03] yurik: was that something you were/are involved in? [23:19:07] awjr, api 2.0 is on my plate [23:19:15] jdlrobson: talk to yurik :) [23:19:18] that's my primary target at hackathon [23:19:22] awesome [23:20:10] awjr, i have commited a large number of structural changes to the api a few months ago, but had to put it on hold while sorting out the zero mess [23:20:37] now that it is gaining some semblance of normalcy, i will jump back to the api [23:21:12] the api 2.0 spec is on RFC page, and it still needs a big brainstorming in terms of content api -- what jdlrobson wants [23:21:55] in principle I agree that this particular suggestion makes sense [23:22:36] can someone review https://gerrit.wikimedia.org/r/63972 please? [23:22:36] yurik: i struggle approaching it from an RFC page. Is there any better way we could more iteratively describe a API 2 ? [23:22:46] the best way I can contribute is by doing a grep of FIXME's in MobileFrontend [23:23:14] I tried participating in the RFC but just got too lost [23:23:53] jdlrobson, most of the api changes in the rfc is about the core api structure, errors, warnings, i18n, etc etc etc [23:24:18] what you want is content stuff - ability to parse stuff on the server (if i remember/understood correctly) [23:24:25] New patchset: Jdlrobson; "Move ResourceLoader modules to their own file" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63997 [23:24:40] yurik: there's a variety of problems [23:24:41] so your stuff is at the bottom of that rfc, and needs to be moved somewhere else where it can be speced out [23:25:06] yurik: my main issues are inconsistencies - what constitutes an error, what an error looks like [23:25:21] yurik: this is a classic example - https://bugzilla.wikimedia.org/show_bug.cgi?id=48261 [23:25:43] i switched off from this discussion as i felt like i was talking to a wall [23:26:07] (no offense to brad here - i just felt like we were just in disagreement and couldn't find a middle ground) [23:26:35] i feel like every time i complain about something in the API I'm told to sit in the corner and face the wall [23:26:41] it's not a nice feeling [23:27:17] awjr: MaxSem can you look at https://gerrit.wikimedia.org/r/63997 and merge rather quickly? quite a significant change that will result in merge conflicts if left too long [23:27:22] i ahven't paid attention to that bug, will review. I aggree that consistency is lacking, but we haven't had a way to put together a cleanup list -- i kind of started it with a cleanup subpage from the rfc [23:28:01] but i absolutelly agree we need to figure out the best way to move forward [23:28:20] anomie (Josh) is coming from the bot writer's perspective [23:28:23] New review: MaxSem; "(1 comment)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63997 [23:28:48] yurik, you and me are also bot writers [23:28:50] and so was I when I was writing it. Now we have lots of content-oriented api requests [23:28:54] both retired;) [23:29:00] yep [23:29:24] so api need to be revised to suit content / AJAX style usage better [23:29:48] are we really goingg to keep only JSON? [23:30:13] moreover, in reality we really ought to switch to a very AJAXy wiki, with very little static content :) [23:30:16] MaxSem, yep [23:31:15] and api is perfectly positioned (and is already being used) for that [23:33:44] New review: Jdlrobson; "(1 comment)" [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/63997 [23:34:29] thanks MaxSem [23:35:30] MaxSem, I still (hope) to get back into bots [23:35:36] interwiki is done for [23:35:41] but there are other minor things :) [23:36:15] ugh, the iw sorting saga in AWB - remeber, Reedy?:P [23:36:27] it's still there! :D [23:36:33] Shame wikidata had to go and ruin it all [23:38:28] ok MaxSem, https://gerrit.wikimedia.org/r/#/c/63972/ +2'd [23:38:37] thanks! [23:41:56] yurik, just noticed memcached locking in Zero - there's PoolCounter for that [23:45:59] New patchset: Jdlrobson; "Move ResourceLoader modules to their own file" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63997 [23:46:00] MaxSem: awjr fixed ^ [23:50:46] man oh man, i wish i hadn't created that global var $localBasePath [23:50:57] or rather, i should've named it differently [23:51:03] $mfLocalBasePath or something [23:51:11] that seems scary [23:52:18] MaxSem, i copied it from event system by ori -- will revise it [23:53:37] awjr: you referring to my patchset? [23:54:10] New review: awjrichards; "(1 comment)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/63997 [23:54:24] jdlrobson: yes, but nothing wrong that you did [23:54:33] awjr: we can use this opportunity to fix it.. :) [23:54:38] it just made me realize something wrong that i did [23:54:44] yeah, once your patchset is merged i'll fix it [23:56:39] sweet [23:56:59] so we good to merge that awjr ? [23:57:07] im looking closely at it now, jdlrobson [23:57:11] thanks :) [23:59:00] so jdlrobson using that RL hook in that way is discouraged (I don't really know why, but the docs say so) https://www.mediawiki.org/wiki/Manual:Hooks/ResourceLoaderRegisterModules [23:59:19] why not just use $wgResourceModules?