[17:42:00] which mobilefrontend knows about and always inclueds at the top [17:42:30] New review: Siebrand; "Awesome" [mediawiki/extensions/MobileFrontend] (master) C: 2; - https://gerrit.wikimedia.org/r/68015 [17:42:41] mhm, can just be slapped on main page after all the fu[nc]ky transformations [17:43:01] ?? [17:44:21] MaxSem, should we keep that "zero" id around for the magic re-inclusion on the main page? [17:45:56] is it at the top or bottom? can it be made an integral part of the skin? [17:52:42] MaxSem, what is at the top or bottom? [17:52:56] that id is taking the whole

tag and includes it at the top [17:52:58] zero-landing-page [17:53:08] landing page is contained inside that div [17:53:50] added some bugs to https://github.com/orsa/TranslateWikiApp-iOS/issues?state=open [17:59:57] btw yurik - has a designer participated in Zero development? [18:00:47] MaxSem, not that i know [18:01:11] MaxSem and yurik, UX said dan and kul are talking with ux on thursday [18:01:41] MaxSem and yurik, i think they'll talk about end user plus projected carrier portal stuff [18:08:07] brion: feeling better today ? [18:08:35] bit better [18:09:22] got distracted with iOS stuff, loaded 7 beta on my iPod touch [18:09:35] mhurd: you're gonna love iOS 7 :) [18:09:38] brion: mhurd was showing me the same here [18:09:43] excellent [18:10:08] brion: it was a bit jarring at first but yeah, i really like it [18:10:13] btw, jdlrobson & yurik, what should be the order o CN and Zero banners, which of them on top? [18:10:34] MaxSem: it's a good question - that's never been specified :) [18:10:43] i thought CN is not shown for mobile? [18:10:45] also notice zero can disable central notice if it wanted [18:10:59] yurik: CN does run on mobile now - but currently only on beta i believe [18:11:44] mhm, we'd better decide upon this [18:12:53] we probably don't want to show any banners other than zero to zero users [18:13:03] b/c most CN banners are fundraiser related [18:13:16] brion: reading through the api diff last night it was interesting to see they're switching return type of methods which return instance of that objects same type to "instancetype" from "id". i think this is for better llvm static analysis goodness [18:13:36] Maryana, MaxSem, jdlrobson, yurik, i'll add a card to trello for zero to suppress other banners [18:13:37] Maryana, do we need a card for this? [18:13:44] lol [18:13:51] race condition! [18:13:56] cardception [18:14:01] brion: http://stackoverflow.com/a/14652187/135557 [18:14:57] brion: also looks like speech synthesis api is finally public [18:15:30] brion: cause i want apps that actually say "beep!" instead of just beeping, of course [18:16:11] this reminds me that i need to get back to mwalker about cn banners on mobile web [18:16:18] mwalker, you channel stalking? [18:21:26] mhurd: hehe [18:21:47] we could read out descriptions of newly uploaded images while you're walking to work [18:28:02] nice, instancetype makes since [18:28:06] *sense [18:28:08] cents? [18:28:15] cha-ching [18:29:10] yurik, so what are your thoughts about the main page? [18:29:31] MaxSem, sorry, meeting for half an hour [18:31:28] yurik: there wasn't a hangout attached to the invite. i just added one. [18:32:40] jdlrobson, yurik & dr0ptp4kt, anything we need to deploy today besides MF? [18:32:56] MaxSem, nothing "in sync" [18:33:20] i had a few patches, i should probably deploy them separately [18:34:33] MaxSem: I have just Iacfea0219ef21dc91d669f51140ec129c2f41f76 [18:34:38] although that's optional i believe [18:55:01] MaxSem, back [19:19:24] yurik, so what are your thoughts about the main page? [19:19:50] MaxSem, well, i am not very sure of what magic mobile does for the main page [19:19:54] i looked through the doc [19:20:09] you are grabing a number of

s by id [19:20:35] so i guess we could generalize that ability - with something like "mf-top-*' [19:28:02] yurik, I'd rather separate this functionality - that is, remove all mentions of Zero from MF and rely on hooks for adding stuff. I wonder if that landing page functionality can be plugged into SkinMobile [19:28:35] MaxSem, but mobile already rewrites main page in that function [19:28:45] i don't mind if it moves somewhere else [19:29:12] but it seems like it has hardcoded ID patterns for the CSS selector [19:29:24] yep, so 'zero-landing-page' should be added through some other means [19:30:11] MaxSem, exactly - so as an idea, we could rename that ID to "mf-top-...." [19:30:18] noooo [19:30:46] we need to stop doing this in BeforePageOutput [19:30:54] and by "..." -- do it just like mobile main page is already doing with the other tags [19:31:29] that is, zero landing page shouldn't be part of page content [19:31:36] it should be part of interface [19:31:44] MaxSem, parseMainPage() already takes all "mf-*" and re-adds them [19:31:54] yup [19:32:04] again - i don't know if that's the RIGHT (c) approach [19:32:10] and that's something we want to get rid of [19:32:16] i see [19:32:47] so the party's policy(c) is not to add further DOM manipulations [19:36:23] New review: MaxSem; "This code is very legacy and thus has a lot of dependencies, so can't be tested in a unit test way -..." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67361 [19:37:00] MaxSem, lol, well, i don't know about party policy, but Zero is based on DOM rewriting [19:37:25] and a lot of this rewriting can go away [19:40:50] MaxSem, how can it go away? Zero need to replace ALL links that are not zero-rated with warnings [19:41:24] yup, but you shouldn't need to add banners and landing page via DOM manipulations [19:41:51] MaxSem, that's true, and currently zero uses $out->add/prepend HTML [19:42:23] and I plan on adding a hook for you to provide Zero with our DOM tree so that you won't need to parse the HTML for the second time [19:42:43] MaxSem, that would be awesome! [19:42:53] what about footer links? [19:43:04] another hook! [19:43:25] modularity, MW style [19:44:32] bleh [19:44:47] New patchset: JGonera; "Nag overlay fixes" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68038 [19:45:29] grrr. wanted to refactor MobileFormatter but can't do that until awjr_away merges my previos commit [20:07:24] all right awjr_away, jdlrobson, jgonera & Maryana - we're live on testwiki, please test:) [20:07:40] yes, captain [20:09:58] New patchset: Jdlrobson; "Story 789: First time user upload tutorial revamp" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68013 [20:09:59] New patchset: Jdlrobson; "Beta: Make panels of uploads tutorial slide in and out" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68014 [20:10:00] New patchset: Jdlrobson; "Story 790: Standardise use of words" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68015 [20:10:30] New review: Jdlrobson; "Checked with Vibha. 2 minor fixes that will need to be done in a follow up but we shouldn't block on." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68013 [20:13:16] "Error, edit not saved." [20:14:29] mhm, preview doesn't work in opera either [20:14:49] preview isn't working for me at all [20:14:57] it just shows an empty screen [20:15:26] yurik, is 67947/2 ready for me to test? [20:15:37] {"warnings":{"main":{"*":"Unrecognized parameter: 'preview'"}},"parse":{"title":"San Francisco","text":{"*":"\n\n\n"}}} [20:15:48] mhmhmhmhmhm [20:16:02] dr0ptp4kt, take a look at it, but i suspect it has issues due to DOM rewriting [20:16:38] in other words, the hook will take the result of the special page and attempt to "fix" links [20:17:01] eghhhhhhhhhh [20:17:12] buuhhhhhhh [20:17:27] the fix for cite errors in preview caused it [20:17:40] jdlrobson, jgonera ^ [20:17:54] just making sure you hear our cries of agony [20:18:13] What did the api fix do? [20:18:19] can't edit [20:18:31] can't edit or can't preview? [20:18:35] both [20:18:38] grumble grumble [20:18:52] basically, it needs a section to enter section preview mode, however section=0 results in just section 0 being parsed. that is, an empty string [20:20:09] what article? [20:20:27] duh. the API needs to be reworked a bit [20:21:18] i was trying to edit [[san francisco]] [20:21:54] New patchset: MaxSem; "Temporary hack to fix previews" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68045 [20:21:59] jdlrobson, ^^^ [20:22:17] MaxSem: can you add documentation to that change? [20:22:23] what does preview: 1 do? [20:22:30] (that's different from preview true) [20:22:32] yurik, i see what you mean with regard to the page pipeline and clearHTML(). nonetheless, will check. [20:22:35] tfinc: headsup, I'm doing expense reports, should see you in about 30 mins [20:22:37] does it resurface the preview bug? [20:23:17] yes, the cite bug will resurface [20:23:50] ok we should add a comment there and reopen that bug then [20:24:20] what comment? [20:24:35] MaxSem: i'll do it 1s [20:24:54] doesn't seem that every API parameter is documented [20:25:03] MaxSem: why set preview as well? [20:25:10] can't we just remove the preview arg? [20:25:18] it doesn't hurt [20:25:43] e.g. https://gerrit.wikimedia.org/r/#/c/67310/ < revert it [20:26:04] reverting it would make more sense from a development perspective [20:26:09] guess with next API iteration it will be replaced (or not?) with sectionpreview or something.... [20:26:32] MaxSem: so could you just revert https://gerrit.wikimedia.org/r/#/c/67310/ with a little note to the why? [20:26:52] sorry if this seems pedantic I just think documenting decisions is important :) [20:27:35] New patchset: MaxSem; "Revert "Bug 49218: Hide errors in preview output"" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68046 [20:28:10] Change abandoned: MaxSem; "Just reverted the offending commit completely." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68045 [20:28:47] MaxSem, the API on test wiki returns 503 Service Unavailable for San Francisco article [20:28:51] I guess we had this problem [20:29:00] Maryana, did you try a smaller article? [20:29:03] the edit API? [20:29:21] perfect Max :) [20:29:29] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68046 [20:29:49] Maryana, yes, that's just testwiki crappiness... [20:29:54] I just added Hello to https://test.m.wikipedia.org/wiki/Page029 [20:29:59] and preview worked too [20:30:05] worked on a shorter article [20:30:12] both preview & edit [20:30:44] ok [20:32:24] but the bug is still there [20:32:29] the big red cite error [20:33:16] because core is not new enough [20:33:39] and because i've reverted the fix anyway [20:33:57] well, that was an interesting quit message [20:34:57] okay, updated once again [20:35:46] issues with viewport on edit mode for androids :( [20:36:01] viewport? [20:41:46] Error, edit not saved [20:42:59] j & j are working on some fixes [20:44:36] MaxSem, which article? [20:44:37] New patchset: Jdlrobson; "Regression: Stop locking viewport on overlays" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68048 [20:44:41] ^ jgonera MaxSem [20:44:43] SF [20:45:04] yeah, that's because testwiki API times out I guess [20:45:07] HTTP 503 [20:45:19] try something smaller [20:45:55] Change merged: JGonera; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68048 [20:47:01] does it needs to be deployed? [20:47:04] yes [20:47:05] MaxSem: yup [20:47:05] ^^^ [20:48:53] New review: Dr0ptp4kt; "Yuri, as you suspected, due to the way that parseLinksForZeroQueryString works due to it executing *..." [mediawiki/extensions/ZeroRatedMobileAccess] (master); V: -1 C: -1; - https://gerrit.wikimedia.org/r/67947 [20:49:07] ^ yurik, you were right. [20:49:36] dr0ptp4kt, what about? [20:50:19] deployed [20:50:23] yurik, about the hyperlink rewriting behavior. for once, the merge failure was good fortune. [20:50:55] yurik, not that the marginal case would really have been that bad, though. [20:52:18] yurik, funny enough, it would have let the user see the home page at Main_Page, but then get the warning after clicking on a link from within the main page, because none of the links would have been rewritten. [20:52:35] MaxSem: looks good [20:54:40] [Commons-iOS] montehurd opened pull request #86: Image of day transitions on old 3.5 inch device choppiness fix (master...choppyLogin) http://git.io/WaS2AQ [20:55:35] Maryana, does it look ok enough to scap? [20:55:53] sure, let's do it [20:56:20] DEPLOY [20:57:14] deploying always reminds me of this scene for some reason http://denerddotcom.files.wordpress.com/2013/04/flying-monkeys.jpg [20:57:25] haha [20:58:07] tfinc: did you see matthew's email about high res logo? [21:00:53] mhurd: i just read it. what's the highest resolution that we have ? [21:00:56] it is just the commons log [21:00:56] o [21:01:11] YuviPanda: i'm enjoying seeing you here each day [21:01:11] that's what i'm not sure about [21:01:19] mhurd: what's in the repot ? [21:01:21] tfinc: I bet I enjoy it more :D [21:01:31] tfinc: though currently I'm doing expense reports. [21:01:43] YuviPanda: come next (Wed) i'll actually be useful to you [21:01:53] soon enough :) [21:02:00] until then you'll have to do some fending for yourself :D but your good at that [21:02:21] yeah, indeed. [21:02:24] mhurd: is this our best https://github.com/wikimedia/Commons-iOS/blob/master/Commons-logo-72%402x.png ? [21:02:44] tfinc: today you can be useful by approving expense report! :) Will be in your inbox in about 5-10 mins [21:02:56] tfinc: i'll track down high res [21:03:36] tfinc: found sag: http://upload.wikimedia.org/wikipedia/en/4/4a/Commons-logo.svg [21:03:44] "svg" [21:04:26] tfinc: i'll just email them that [21:06:06] mhurd: thanks [21:06:14] mhurd: i can't wait to see it in print [21:06:22] tfinc: done :) [21:07:03] mhurd: but what's the highest quality icon we have in the repot ? [21:27:08] back stupid outage in the area [21:28:03] welcome back jcmish [21:28:12] thanks [21:28:15] i hate comcast [21:30:29] jcmish, we've finished deploying - can you check if everything's right? [21:30:39] yup [21:30:44] sorry guys [21:42:24] https://bugzilla.wikimedia.org/show_bug.cgi?id=43628 is still not fixed - just got sent to desktop commons :( [21:45:28] jcmish: I've been nomadic in recent times and the level of Comcast suckage varies from town to town but ime is never zero [21:47:41] omg they SUCK chrismcmahon [21:50:56] dr0ptp4kt, did some log digging, fun picture :( [21:53:29] yurik, you wanna do a google hangout? [21:54:58] dr0ptp4kt, in a sec, pls take a look at the email [21:55:16] yurik, from you to me? [21:55:20] dr0ptp4kt, there are many items that SHOULD have been identified as zero, but weren't by varnish [21:57:11] yurik, JOY! [21:57:20] dr0ptp4kt, see them? [21:57:32] yurik, yeah, have the email open [21:57:35] search for those IPs in the log [21:57:53] OPs claim that they have updated their log [21:57:57] i meant - varnish [21:58:05] yurik, i don't have access to the logs yet [21:58:18] yurik, are you saying i should look at the logfile section you sent me previously? [21:58:33] dr0ptp4kt, i sent you one just now, and the older are useless because varnish was broken [21:58:36] yurik, nm, i see the attachment! [21:59:00] yurik, hang tight, will take a quick look [22:00:45] New patchset: MaxSem; "Remove version from extension credits" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68112 [22:12:40] yurik, still looking [22:13:49] dr0ptp4kt, at least one of them has "https" [22:15:58] yurik, yeah, see that. [22:16:01] dr0ptp4kt, i just re-ran the script - almost all of them are https [22:17:41] dr0ptp4kt, and the only exception is a "zh" call [22:17:58] for 250-99 which does not whitelist anything except en & ru [22:19:45] dr0ptp4kt, so yes, it seems the issue here is HTTPS... need to figure it out at some point [22:19:47] New patchset: Jdlrobson; "Remove unnecessary typography rule" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67910 [22:20:09] New review: Jdlrobson; "Cherry picked (not dependent on previous commit)" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/67910 [22:20:25] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67910 [22:25:15] New patchset: Mattflaschen; "Link history page for displayed article, not title from URL." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68116 [22:25:47] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67938 [22:26:45] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67907 [22:27:55] New patchset: Jdlrobson; "Allow configuration of range in Special:Nearby" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67877 [22:28:26] MaxSem, do you know why https might not work? [22:28:30] in varnish [22:28:42] ?? [22:28:45] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/68112 [22:29:11] MaxSem, i'm looking at the logs, it seems there are many cases when X-CS is not set when user comes to HTTPS [22:29:11] yurik, yeah, that's strange. X-Subodmain is there, so i gotta think that there's termination of traffic to pick up the IP address. but then again maybe IP address extraction occurs in a different part of the stack than the request url itself. [22:30:56] MaxSem, found it in the zero log - the IP should have been identified as X-CS but wasn't. Looking for the cause. thought in case you might have known something. Actualy, brion might know, but he's not here [22:31:58] mhm, has that varnish ACL list been patched up on our cluster? [22:32:23] MaxSem, ?? are you talking about the ACL bug? [22:32:35] i think we had this issue before, let me run it against older olg [22:32:38] log [22:32:45] it looks similar, no? X_CS not being set [22:34:23] New review: JGonera; "Let's not create new files for a single CSS rule..." [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/67914 [22:37:08] dr0ptp4kt, MaxSem and what's interesting is that that traffic, even when it comes from opera mini, doesn't get identified as such [22:38:27] dr0ptp4kt, then again, https is NOT supported AFAK - carriers won't be able to do pocket inspection/zero rating on it [22:41:51] not sure about the Wikipedia Zero carriers, but some carriers deploy phones that allow transparent inspection of HTTPS. in other cases, if they're using proxy architecture, they would need to look at the CONNECT proxy request to examine the domain name, or use reverse DNS records. i don't know that carriers are expressly *not* supporting HTTPS. [22:44:07] New patchset: JGonera; "Add javascript specific stylesheet" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67916 [22:44:39] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67916 [22:57:53] New patchset: Jdlrobson; "Cleanup mf-common.less" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67913 [22:59:04] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67913 [23:00:07] New patchset: Jdlrobson; "Standardise on 'plumbing'" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67507 [23:00:45] New patchset: Jdlrobson; "Standardise on 'plumbing'" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67507 [23:00:51] did this get fixed in today's deployment, jdlrobson? https://bugzilla.wikimedia.org/show_bug.cgi?id=49030 [23:02:21] Maryana: yup [23:02:29] sweet [23:02:43] Maryana https://en.m.wikipedia.org/wiki/San_Francisco?useformat=mobile-wap [23:02:52] seems to work but no wap phone to test on :) [23:03:15] heh. we should get some [23:03:19] just for fun [23:05:32] New patchset: Jdlrobson; "Standardise on 'plumbing'" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67507 [23:06:07] Maryana: you crazy cat [23:06:19] what did i do? [23:06:25] New patchset: Jdlrobson; "Move talk add button to talk.less" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67911 [23:06:38] New review: Jdlrobson; "Rebased and +2" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/67911 [23:06:53] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67911 [23:07:27] New patchset: Jdlrobson; "Move overlay css to overlays.less" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67915 [23:07:53] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67507 [23:08:13] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/67915 [23:08:14] jgonera: http://developer.yahoo.com/blogs/ydn/many-users-javascript-disabled-14121.html [23:32:14] dr0ptp4kt, sent you PHPStorm license [23:32:35] yurik, thx [23:35:56] awjr_away, around? [23:36:11] anyone knows how authentication is done in mobile ATM? [23:37:15] the usual MW way [23:37:38] the only difference is that we force HTTPS [23:39:24] MaxSem, what about zero - apparently varnish does not support zero identification of HTTPS traffic [23:39:42] we need to rewrite a number of things, and get zero SSL cert