[00:03:39] any Howies around the third floor? [00:04:05] MaxSem: i think he's hanging out with jdlrobson and/or brion [00:04:33] we're past our window already [00:05:24] their all standing our jdlrobson [00:05:31] around* [00:05:39] if we had video you'd see that quick [00:09:14] MaxSem: if you need to roll back and get out of here, let jdlrobson and howief know [00:11:04] how, if they're not responding... [00:11:06] hey all [00:11:07] sorry [00:11:11] i'm here working with jon [00:11:39] MaxSem: you can ask someone like tomasz to harass them IRL if they are not responding [00:12:03] howief, we're past our deployment window [00:12:04] awjr: i'm happy to throw various projectiles at them. i have many. [00:12:08] max, go ahead and roll back [00:12:18] we're done with test [00:12:18] thanks [00:12:22] tfinc duly noted! [00:12:35] there's always mobile-testing [00:13:26] and i believe jcmish is working on something to make it easier for product to see the changes and signoff on them rather than having to deal with testwiki and/or performance pain on mobile-testing [00:15:14] although, that doesn't really help anyone at the present moment.. [00:17:31] okay, we're back to ol good MF [00:19:26] howief how did things look, what was the outcome of the testing? [00:23:20] jdlrobson made some tweaks to the font size and line height [00:23:24] things are looking good now [00:23:47] we're looking into the language selector since that didn't appear when I was testing [00:24:01] New patchset: Jdlrobson; "typography changes" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29518 [00:24:05] i suspect a bad merge... [00:25:33] cool thanks howief - so other than language selector, things are ok to go to production tomorrow? [00:25:41] New patchset: Jdlrobson; "promote language selection from beta" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29519 [00:27:23] if the features look like they do on jdlrobson's local machine :) [00:29:38] howief ok will you or maryana be available during the deployment window tomorrow for signoff? (also 3-5) [00:30:08] i'm not going to be available, but maryana should be [00:30:29] ok cool - as long as one of you is available to signoff [00:30:48] yup [00:30:58] i may ask munaf to verify the typography changes [00:33:39] brion or awjr or preilly or jdlrobson, could you take a look at the discussion in https://gerrit.wikimedia.org/r/#/c/28863/ ? I would appreciate a second pair of eyes on it [00:33:57] looking [00:36:03] MaxSem: about the localization & authors bit? [00:36:10] yup [00:37:09] hmm [00:37:16] MaxSem: i tend to agree with siebrand here, stick the library credits in with the extension credits in $wgExtensionCredits and keep the description clean of names [00:37:21] easy fix [00:39:49] brion, thanks [03:52:19] New patchset: Jdlrobson; "cleanup footer" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29532 [03:52:19] New patchset: Jdlrobson; "reduce left and right margins of articles by 5px" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29533 [04:19:17] [bugs] NEW [Bug 41293] Rewrite iOS Save Page functionality in native code (priority: Unprioritized severity: normal) https://bugzilla.wikimedia.org/show_bug.cgi?id=41293 [07:37:58] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29532 [07:37:58] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29519 [07:37:59] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29518 [07:38:25] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29476 [07:38:53] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29533 [16:51:23] [bugs] NEW [Bug 34293] Android app should allow pinch and zoom (priority: Highest severity: enhancement) https://bugzilla.wikimedia.org/show_bug.cgi?id=34293 [16:52:49] MaxSem: ping [16:52:52] pong [16:53:13] MaxSem: can you make the change the Jon requested on https://gerrit.wikimedia.org/r/#/c/29441/ [16:55:22] New patchset: MaxSem; "Back out the preference for core classes" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29441 [16:55:28] preilly, ^^ [16:55:49] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29441 [16:56:05] MaxSem: thanks! [17:01:03] teehee i'm streaming the apple event on my Apple TV. i feel like i'm part of a cult <3 [17:01:46] lol [17:16:01] yuvipanda: hey lets talk here. I was working on this bug https://bugzilla.wikimedia.org/show_bug.cgi?id=34293 and needed some help. [17:17:44] prageck_: hey [17:17:44] * yuvipanda looks [17:18:09] prageck_: excellent! [17:18:13] how can I help? [17:19:09] yay 13" with retina displayyy [17:19:34] soon i won't be the only one [17:19:51] yuvipanda: did you see the comment I have made on bugzilla ? [17:24:53] prageck_: about the search header zooming? [17:25:28] [bugs] RESOLVED [Bug 32292] MobileFrontend extension needs access to a notifications system for 'status' changes (priority: Unprioritized severity: enhancement) https://bugzilla.wikimedia.org/show_bug.cgi?id=32292 [17:25:44] there is a way to actually zoom just the content [17:25:58] you detect the pinch gesture on the content div and just increase the font size [17:26:13] we sortof did this for WLM [17:26:23] trouble is, it works only on Android 3.x+ [17:26:44] brion: https://bugzilla.wikimedia.org/show_bug.cgi?id=31675 is that fixed now? [17:26:59] looking [17:27:04] yuvipanda: could you please direct me towards the WLM app [17:27:05] jdlrobson: partially done. we're not natively using all the loader bits [17:27:44] prageck_: check your email [17:27:44] [bugs] NEW [Bug 31675] Port MobileFrontend to use ResourceLoader for JS and CSS (priority: Unprioritized severity: enhancement) https://bugzilla.wikimedia.org/show_bug.cgi?id=31675 [17:27:53] prageck_: look for messages from "bugzilla-daemon" [17:28:04] prageck_: you should look at http://stackoverflow.com/questions/11183174/simplest-way-to-detect-a-pinch [17:28:17] prageck_: https://github.com/wikimedia/WLMMobile is the WLM app [17:28:22] i'm looking for the specific commits, one second [17:28:32] brion: can you update the bug with the latest status please? [17:28:37] jdlrobson: just did :) [17:28:52] prageck_: there have been useful comments on the bug you commented on :) [17:28:53] ahh your always one step ahead of me ;-) [17:31:53] jdlrobson: just to clarify on the language selector, we're not going with overlay for this deployment? [17:31:53] sumanah: yes I just went through them [17:32:09] prageck_: see 02bbeb490e217631d71e125b29827fd58822708f and 8df2893162f14774976fad57f3c2b897edfdec1a [17:32:15] and dda82939ffc52d00bc860e69344d1f1bd25a697f [17:32:23] howie said you guys decided on a flat list of languages instead of a separate page/overlay thingy [17:32:39] that plugin was backed out, due to issues with the map, but I think it can be easily used here [17:32:44] prageck_: how comfortable are you with git? [17:33:07] yuvipanda: I am pretty used to it, so that would not be a problem. [17:33:22] sweet! [17:33:22] new iMac is thinner, but not retina display. bah! [17:33:30] prageck_: so it should be a good idea to port code from there to here! [17:33:51] Maryana: no I can boot up my machine if you want to try it out [17:34:43] yuvipanda should I submit the pull request with my current fix, or should I implement the functionality we just discussed and then submit the request ? [17:35:20] prageck_: I think porting the PinchZoom stuff would be the best course of action [17:35:36] prageck_: having the searchbar zoom would be awkward (as you probably noticed) [17:37:52] yuvipanda: alright I would look into the links you pointed me to, and try fixing it. Thanks a lot for the help :D [17:39:59] prageck_: glad to help :) I'll be on IRC, and if I'm not you can email me (yuvipanda@gmail.com) or the mobile-l list :) [17:40:48] [bugs] REOPENED [Bug 39467] Mobile feedback to OTRS should include article in the summary (priority: High severity: enhancement) https://bugzilla.wikimedia.org/show_bug.cgi?id=39467 [17:52:27] MaxSem: https://gerrit.wikimedia.org/r/#/c/28658/ < how would you do this out of interest? [17:53:59] function cleanupPresentationalFlags( $enable = true ) { $this->whateverFlag = $enable; } [17:54:29] MaxSem: I don't follow [17:54:53] New patchset: Jdlrobson; "help mobile validate (experimental feature for beta)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28658 [17:55:57] also, there has to be a performance consideration: this would require separate caching for beta and nonbeta HTML, decreasing lifetime in cache and increasing appserver load (which is already considerable) [17:55:58] MaxSem also had worries about styling using align/center getting broken in mobile - but I don't see this as a big deal and I think encouraging editors to use classes to get around this can only be a good thing [17:56:40] iPad mini screen resolution is as expected; we shouldn't need to do anything special to support it. yay! [17:58:05] MaxSem: Why? [17:59:53] MaxSem: why would separate caching for beta and non-beta HTML suddenly be an issue? [17:59:55] preilly, because as it stands, this change would mix beta and non-beta output for action=mobileview [18:00:25] MaxSem: how would stripping non valid attributes mix content? [18:01:12] because mobileview assumes that beta and nonbeta HTML are the same [18:01:18] MaxSem: what I mean I guess is why couldn't we just always do this on all content before render [18:03:11] that's a different question. Jon did this just for beta so I'm explaining what needs to be done to make it work right and what consequences it will have [18:04:48] some of this could be easily added to Sanitizer in core instead [18:04:49] MaxSem: so just drop this line 225 » » if ( MobileContext::singleton()->isBetaGroupMember() ) { [18:05:40] preilly, there was actually a merit in doing this just for beta - nobody knows how widespread would be the changes [18:08:37] and frankly, the only reason to do it I've heard so far is to ease development by reducing the number of validation errors [18:09:38] jdlrobson, I could help you straighten this stuff after I'm done with my current tasks [18:09:49] MaxSem: I also think it helps drive improving wikitext markup to be more mobile friendly [18:09:58] so probably beta is the better place to do this for the time being [18:10:12] but thanks MaxSem that's great -it's not urgent I just think it would be a step in the right direction [18:10:13] no it won't. it just creates weird differences between mobile and desktop [18:10:20] New patchset: Jdlrobson; "left menu adjustments for screen widths > 700px" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29620 [18:10:20] New patchset: Jdlrobson; "help mobile validate (experimental feature for beta)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28658 [18:10:34] argg [18:10:34] dirty branch [18:11:11] New patchset: Jdlrobson; "left menu adjustments for screen widths > 700px" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29620 [18:12:04] MaxSem: We're mobile - part of our job I believe is to expose weaknesses in core and encourage people to be more future thinking! [18:12:19] New patchset: Brion VIBBER; "Mobile replacement for Special:Watchlist, initial version." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29621 [18:12:52] jdlrobson, I've left a few comments [18:12:59] thanks MaxSem [18:13:21] MaxSem: you wrote on the wrong patchset [18:13:49] you're submitting them too often:P [18:13:49] (I just posted patch set 1 again accidentally :-S [18:14:18] i removed style - i was just putting it in there to check it was working [18:14:23] frankly, after I noticed that you want to remove inline styles, I was tempted to abandon it on the spot:angry: [18:14:33] New patchset: Jdlrobson; "help mobile validate (experimental feature for beta)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/28658 [18:15:31] MaxSem: we need to drive that conversation soon though - at some point in future we need to turn off inline styles and get styles written in wiki page stylesheets [18:15:46] we might need to write some sort of bot that repopulates said stylesheets [18:15:51] *populates [18:16:06] exactly - start discussion, not introduce guerilla changes;) [18:16:34] off topic but i don't see how beta is guerrilla [18:16:46] pushing straight to production would be guerrilla :) [18:17:46] shit doesn't get done unless you completely break things anyhow - there's no incentive [18:17:47] h [18:17:47] I'll look at integrating as much of this into core as possible [18:18:00] hence why I ended up fixing all of the portal pages single handed! [18:18:14] jdlrobson, see Sanitizer::fixDeprecatedAttributes() [18:18:33] thanks MaxSem [18:19:02] we might as well add fixPresentationalTags() [18:20:39] jdlrobson, how about using some of our research time for this? [18:21:09] agreed MaxSem - :) [18:28:17] [bugs] NEW [Bug 40013] Connectivity interferes with saved articles (saved pages do not work offline) (priority: Highest severity: normal) https://bugzilla.wikimedia.org/show_bug.cgi?id=40013 [18:35:27] hey jdlrobson? [18:36:30] hey jcmish [18:36:40] howdy :) [18:36:53] can you spare me a few minutes for qunit? [18:43:10] * tfinc has been loving qunit and grunt for his personal projects  [18:43:15] jcmish: are you a grunt user ? [18:44:34] hahahha [18:44:34] yes yes I am [18:44:34] i think we use explore using it for our js code [18:44:34] especially for automation [18:44:35] yup [18:44:57] yup so I'm working on my automated ruby tests [18:45:03] but I wanted to stop and do this first [18:45:26] jcmish: sure i'll Skype you in 5 [18:45:48] jdlrobson [18:46:21] yuvipanda: preilly is having issues with firefox os - the goodscroll class is getting added rather than the badscroll class how can he fix this? [18:47:45] jcmish: now good? [18:48:02] let me run down to take my tea off [18:48:04] 1 min [18:48:18] yuvipanda: are we clear to push the WP iOS app udpate? [18:49:12] tfinc: +1 for qunit and grunt [18:49:26] Krinkle: glad to see you use it as well [18:49:42] Krinkle: i use quint, grunt, and phantomjs for my toolkit [18:49:50] back jdlrobson [18:49:57] tfinc: I'm actually planning on converting our ant/xml build scripts into grunt and put them in mediawiki/core.git instead of [integration/jenkins.git]/shared/_something [18:50:14] Krinkle: nice. i'd love to use grunt for mw [18:50:28] or rather in tools/grunt-tasks.git, and submodule in mediawiki/core (since we don't want to duplicate it all over the place) [18:50:58] tfinc: 'grunt lint', 'grunt qunit', 'grunt testswarm' :) [18:51:14] where lint would be our grunt implementation of the "Universal Linter" we have in Jenkins now (that will be able to lint puppet files, php, css, javascript etc.) [18:51:57] tfinc: Last week I met "the" cowboy that created grunt (at the jQuery summit in DC) [18:51:59] jcmish: just finding ar oom [18:52:17] no prob [19:12:13] Krinkle: nice. what motivated him? [19:12:52] tfinc: not sure, other than node is cool and xml sucks [19:13:03] (for build logic) [19:15:31] especially with package.json/npm awesomeness [19:15:35] preilly, will you be around during today's deployment? [19:15:37] tfinc: been having intermittent power issues. [19:15:49] tfinc: can you point me to the place where older IPAs were archived? [19:16:16] yuvipand1: you'll have to build them as i don't see them on download http://dumps.wikimedia.org/iOS/ [19:16:18] only 3.1.2 is there [19:16:42] http://dumps.wikimedia.org/iOS/archive/ [19:16:47] alright [19:17:04] now to find out which branch brion was working off [19:19:56] * tfinc reads up on todays apple news [19:20:19] yuvipand1: v3.2.1-iOS [19:21:42] was actually brion/v3.2-iOS-6-fixes [19:26:02] bug report… maybe? http://meta.wikimedia.org/wiki/Talk:Mobile_projects#Always_show_2_copies_of_lists_in_.22References.22 [19:26:23] Maryana: was fixed, should go out as an update soon [19:26:40] I can dig up a bug number if you want [19:26:40] sweet! thanks, yuvi :) [19:28:14] tfinc: see my email about the qa meeting? [19:28:29] awjr: about you not being able to attend ? [19:28:29] yes [19:31:19] tfinc rather than email my thoughts re qa, i just added notes to the etherpad. [19:33:52] MaxSem: yes [19:34:09] cool - then we can update Zero [19:37:03] power's out, brb [19:40:10] mhm, if a file belonging to foo:foo with mask 644 resides in a root-owned directory with mask 755, user foo will be unable to write it, right? [19:42:17] yuvipand1: i spy a brion in the office [19:42:27] brion: http://208.80.153.133/ [19:42:41] i really want to use the new iMac in person [19:42:42] hi tfinc & co, look what a nice Ohloh stats page I put together for the Analytics team: https://www.ohloh.net/p/wmf-analytics [19:42:44] it looks very nice [19:42:49] brion: it's the vumi instance in labs [19:42:50] would you like on for the mobile team? :) [19:42:55] hey qgil [19:42:57] qgil: please do [19:43:16] ok, after I sort out the happy worls of MW extensions :P [19:43:27] world [19:44:41] preilly: http://www.apple.com/macbook-pro/specs-retina/ 15'' has 512 ssd [19:44:50] tfinc: nice! [19:52:25] * yuvipanda pours himself a strong drink and heads to xcode [19:53:45] Question: is all the MediaWiki code (no extensions) found at https://gerrit.wikimedia.org/r/p/mediawiki/core.git alone? [19:54:00] oops wrong channel [20:00:42] tfinc: so, i'm testing it out - and even on 3.2 saved pages don't seem to persist?! [20:07:26] * yuvipanda pokes brion  [20:07:59] yuvipanda: meeting! but poke away [20:08:26] brion: did saved pages persist in 3.2? [20:08:40] on iOS 5.x [20:08:40] yuvipanda: in theory i think they should have, but not on iOS 6 [20:08:47] can you confirm? [20:09:11] I remmeber that we did testing when we released to make sure they did work, but right now it doesn't :| [20:09:12] so i'm confused [20:09:16] might be because I force quit the app [20:09:35] uggggggh [20:10:48] brion: does persist. [20:10:53] :) [20:10:53] \o/ yay [20:10:56] doesn't persist only if you force close the app [20:11:08] wait, this is only 5.x. iOS 6 is still downloading :) [20:18:05] iOS 6 is 580MB [20:21:53] perfect time for the power to go out. Again [20:22:05] * yuvipanda sighs [20:24:48] :( [20:25:02] 18 minutes more it says. my UPS will give out any second now. [20:46:52] i think we are good to go for deployment [20:47:23] \o/ [20:47:33] \o/ [20:49:14] * jdlrobson tries to think of an obscure browser to test on [20:49:42] jdlrobson: if you want to, try UCBrowser [20:49:48] anyone got a kindle? [20:49:54] has pretty big marketshare in Indian mobile phones [20:50:07] goddamnit there goes the power again. [20:56:46] yuvipand1: nice work on cordova update [21:01:09] woohoo [21:04:56] Strict Standards: Only variables should be passed by reference in /Users/jrobson/Sites/w/extensions/MobileFrontend/includes/skins/SkinMobileBase.php on line 100 [21:05:04] where's that come from? [21:08:00] ahhh it was watch list related I think [21:10:39] jdlrobson: yeah that should be set() not setRef() [21:11:39] New patchset: Brion VIBBER; "Mobile replacement for Special:Watchlist, initial version." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29621 [21:12:36] New review: Brion VIBBER; "Patchset 2 changes a 'setRef' to 'set' to avoid a strict mode warning." [mediawiki/extensions/MobileFrontend] (master); V: 0 C: 0; - https://gerrit.wikimedia.org/r/29621 [21:31:12] New patchset: Jdlrobson; "define height of search button by padding and font-size" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29703 [21:34:35] ohh dear.. looks like navigation broke somehow [21:35:13] New review: Brion VIBBER; "Much prettier :D" [mediawiki/extensions/MobileFrontend] (master); V: 0 C: 2; - https://gerrit.wikimedia.org/r/29703 [21:35:13] Change merged: Brion VIBBER; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29703 [21:35:39] jdlrobson: seems ok on master to me, what's wrong? [21:35:53] When I click the left menu it is not opening at all [21:36:02] let me cleanup my branch [21:36:07] you didn't leave JS off did you? :) [21:36:52] seems to be loaded twice.. [21:36:59] are you running last mediawiki and mobile frontend? [21:37:14] lemme update my mediawiki [21:37:42] yeah seems fine here [21:37:44] mm [21:38:06] so for some reason mf-navigation.js is being loaded twice for me [21:38:11] dingdong, deployment in 20 minutes! [21:38:20] aiee [21:38:26] :) [21:38:26] and application.js [21:38:27] do we have everything ready? [21:38:45] do we have anything we don't want to deploy?:P [21:38:58] MaxSem: ^ I've just encountered an issue on my local machine which Brion cannot replicate that is loaded resources twice [21:41:32] deploy all the things! [21:52:22] Warning: Cannot modify header information - headers already sent by (output started at /Users/jrobson/Sites/w/includes/OutputPage.php:1998) in /Users/jrobson/Sites/w/includes/WebResponse.php on line 38 [21:52:22] < brion [21:54:29] jdlrobson, what's the call stack? [21:55:42] something's calling WebResponse::header() somewhere very late during execution [21:55:51] we need to know _what_ [21:56:11] or how to repro? [21:56:44] ok it's okay.. i just forgot to run the maintenance scripts [21:56:52] things look good again [21:58:47] MaxSem let's go go go! https://www.youtube.com/watch?v=SESI19h4wDo [21:59:12] http://opbeat.files.wordpress.com/2012/08/fisi_640.jpg [22:00:07] [bugs] NEW [Bug 41333] MF loads JS twice, has weird errors when database error occurs in jobqueue (priority: Unprioritized severity: normal) https://bugzilla.wikimedia.org/show_bug.cgi?id=41333 [22:00:27] if it fits, it ships [22:00:54] I accidentally my MobileFrontend [22:01:23] Maryana: hey did we get a design mockup for watch list drill down screen yet? istr someone working on this, i may as well get started on implementing it soon :) [22:01:37] check the mingle card :) [22:01:53] https://mingle.corp.wikimedia.org/projects/mobile/cards/231 [22:02:07] hey guys, we have one last sync to run as part of the e3/aft deploy -- have you guys started deploying yet? if not, mind waiting another couple of minutes? [22:02:26] brion: if you scroll aaaaalll the way down, there's a tiny attachments section. top seekrit. [22:02:27] who's mlitn? [22:02:43] matthias muelle(sp?) [22:02:57] i see it! [22:03:04] also looks like http://www.mediawiki.org/wiki/Mobile_design/Watchlist_view has been updated yay! [22:03:10] yep [22:03:17] tx [22:03:32] just doin m'job, good sir [22:03:33] mm found a weird bug on settings [22:04:00] Clicking save on settings page when return to is also a special page doesn't redirect [22:04:00] http://4wnu.localtunnel.com/w/index.php/Special:MobileOptions?title=Special:MobileOptions&returnto=Special:MobileMenu [22:04:34] i can confirm :( [22:05:17] brion: problem in core? [22:05:25] looking... [22:05:35] $context->getOutput()->redirect( $request->getText( 'returnto' ) ); [22:05:47] L180 SpecialMobileOptions.php [22:06:08] hmm [22:06:31] that seems to take a URL, not a page title [22:06:47] note: we can't currently deploy because mlitn is running scap [22:06:50] i bet the : in the special: is confusing it [22:07:33] jdlrobson: i think i've got a fix in the works, lemme test it [22:09:42] New patchset: Brion VIBBER; "Use URL instead of article name as returnto; fixes redirection" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29714 [22:09:43] jdlrobson: try that [22:09:45] MaxSem: 15:09 (mlitn) it should be nearing completion (srv285: Done) [22:10:15] MaxSem: i didn't realize the sync was already invading your slot, so am retracting the request for an additional sync. i'll find another time. [22:10:26] the scap, rather. [22:10:26] thanks ori-l :) [22:13:37] MaxSem: https://bugzilla.wikimedia.org/show_bug.cgi?id=38273 < fixed now? [22:15:13] [02:14:47] (FIXED) MobileFrontend corrupts parser cache for regular page views - https://bugzilla.wikimedia.org/38273 +comment (Max Semenik) [22:15:38] found another bug - no close section link in last section [22:18:40] okay, we can deploy now [22:19:03] awjr: our 1:1 is right within the deployment … delay or keep? [22:19:16] tfinc i just sent you an email about it - can we reschedule for tmrw? [22:19:29] awjr: wow you beat me by 3min. i'm falling behind [22:19:35] awjr: pick a time on my calendar and we'll do it [22:19:42] thn [22:19:42] x [22:19:57] tfinc: preilly: http://etherpad.wikimedia.org/MobilePageLoadStats [22:20:11] awjr: MaxSem jdlrobson : i asked brion to run some front end perf tests --^ [22:20:21] these are the kind of stats i'm going to want us always measuring on a dedicated dashboard [22:20:26] but for now the one offs are good enough [22:20:34] tfinc very cool! [22:20:44] brion: thanks [22:20:49] whee [22:20:57] performance is a feature that we can't forget about [22:20:59] MaxSem: did you update Zero to master too? [22:21:11] preilly, waitasec [22:21:11] tfinc: hence why we need the ori-l logger... [22:21:32] does somebody need to +2 on https://gerrit.wikimedia.org/r/#/c/29714/ ? [22:21:55] jdlrobson should be able to +2 now :) [22:22:09] \o/ [22:22:25] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29714 [22:22:37] thx [22:23:07] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29620 [22:23:22] brion: np [22:26:28] preilly, awjr, jcmish, jdlrobson, brion, Maryana, MF and Zero are live o n testwiki [22:26:37] squee [22:26:46] … zero? [22:26:51] New patchset: Jdlrobson; "generate back to top link on the last section" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29718 [22:27:06] ^ bug I just discovered. Probably not too severe. [22:27:38] messages are not up to date - but that's fine right? [22:27:42] jdlrobson, tests? [22:29:27] sorry, internets fail [22:29:40] jdlrobson: update the test [22:32:29] so far no surprises on test, things seem to be working for me [22:32:46] yeah, looks good! [22:33:14] munaf is testing out the design changes. he says the search bar could be a little wider to be symmetrical on both sides, but other than that it's fine [22:33:36] * brion sidebar: android 4.1.2 update for nexus 7! whee [22:33:56] brion: same here all looks well [22:34:23] are the 2 changes merged by preilly needed for today's deployment? [22:34:38] i think we want that redirect fix [22:34:53] not sure about the other one [22:35:00] hmm, looks like the ribbon is still there in beta, and the login button isn't styled [22:35:50] New patchset: Jdlrobson; "generate back to top link on the last section" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29718 [22:35:55] MaxSem: I would like them plus ^ [22:36:29] Maryana: correct. the ribbon has not been removed yet as the story is for the next iteration which we haven't started. The login code never got finished [22:37:18] wait, i thought we agreed that these were the changes going out today? http://etherpad.wikimedia.org/Mobile23Oct12 [22:38:01] brion: can you glance at [Bug 41333] MobileFrontend extension loads JavaScript twice, has weird errors when database error occurs in job queue [22:38:06] found a bug - http://test.m.wikipedia.org/w/index.php?title=Main_Page&mobileaction=toggle_view_mobile#section_language [22:38:15] brion: https://bugzilla.wikimedia.org/show_bug.cgi?id=41333 [22:38:20] it seems to be reporting itself as a language [22:38:34] preilly: i filed it, what you want me to look at on it? :) [22:39:05] MaxSem: seems to be some assets not showing up [22:39:14] should be icons on http://test.m.wikipedia.org/w/index.php?title=Special:UserLogin&returnto=Main+Page&returntoquery= [22:39:44] http://test.wikipedia.org/w/static-1.21wmf2/extensions/MobileFrontend/stylesheets/common/images/W.png?2012-10-23T22:25:00Z [22:39:50] those may not come through until we scap... [22:40:54] jdlrobson, we've got a JS error [22:40:54] document.getElementById( 'results' ).ontouchstart = hideKeyboard; [22:41:04] what browser MaxSem ? [22:41:09] document.getElementById( 'results' ).ontouchstart = hideKeyboard; [22:41:11] WFM [22:41:12] MaxSem to get the assets on test, run sync-common on srv193 [22:41:23] brr document.getElementById( 'results' ) is null [22:41:26] jdlrobson, FF [22:42:13] just sync-common the files [22:42:22] MaxSem: not getting that on my FF [22:42:24] oh never mind awjr already said that [22:42:36] MaxSem: I don't see that either [22:42:43] sunched [22:42:50] *synched [22:43:35] I like the word sunched better [22:43:38] does anyone else see the footer rendering funny on http://test.m.wikipedia.org/w/index.php?title=Special:MobileOptions&returnto=Main%20Page when they open the left menu? [22:43:39] beta/nonbeta differences? [22:44:18] jdlrobson: yeah it looks pretty weird [22:44:39] argghghgh html validation [22:44:42] this is why I need it [22:44:50] jdlrobson :D [22:45:34] found it [22:47:42] still not fixed it though [22:47:42] something funny going on [22:49:51] New patchset: Jdlrobson; "fix another bad merge" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29723 [22:49:51] sigh [22:49:55] ^ MaxSem that needs merging [22:50:35] was it caused by manual conflict resolution? [22:51:40] i'm not sure MaxSem [22:51:48] tfinc: https://twitter.com/auroraeosrose [22:52:33] but the real cause was the neglected backlog of code review.. [22:52:54] Change merged: MaxSem; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29723 [22:52:54] i'm gonna be less moany when this deployment is out the way.. promise. [22:54:21] tfinc: http://emsmith.net/again-on-the-job-market [22:54:41] speaking of which, are we ready to step it up to beta? [22:55:03] i'm happy w/the way it looks on test [22:56:23] wheeeeee [22:57:44] I've updated it with latest merges [22:58:28] preilly, how does Zero look on test? [22:58:47] MaxSem: seems good to go [22:58:56] New patchset: Jdlrobson; "improve html validation" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29726 [22:59:00] ^ not important for this deployment [22:59:04] just did it whilst i was there [23:00:35] maxsem: what is this zero thing you keep mentioning? is that some kind of new cool kid terminology for production? [23:00:40] :) [23:00:56] WP zero [23:01:11] ahhh [23:01:14] gotcha [23:01:24] it's calorie-free [23:01:38] but with all the good wikipedia flavor you're used to [23:01:59] mmm, refreshing! [23:02:14] ha ha ha [23:02:21] nerds! [23:02:30] are we good to deploy everywhere? [23:02:51] i think we are. any objections? [23:03:05] crickets [23:05:06] i wish we'd removed the ribbon from beta & not released whacky unstyled login just yet, but oh well. let's not let that stop us from going to production [23:05:35] Maryana: it ads character ;-) [23:05:38] we'll get the ribbon next time [23:05:43] that's what I tell myself [23:05:58] tfinc: this is the other guy http://thefrozenfire.com/about/ [23:05:58] haha. well, i guess people *are* signing up for experimental features [23:06:04] so experimental they don't actually do anything! [23:06:19] preilly: let me switch to pm [23:06:50] tfinc: sorry wrong winodw [23:06:58] s/winodw/window [23:07:40] Maryana: we'll just say it's a very small version of http://en.wikipedia.org/wiki/Clippy but it bluescreened [23:08:08] "it looks like you're trying to start a resume!" [23:08:23] brion: OMG you just didn't [23:08:27] oh yeah you did [23:08:40] * Maryana shudders [23:09:31] I'm scapping in a couple of minutes, last call for objections:P [23:09:53] make it so [23:13:01] http://www.youtube.com/watch?v=2kuna7sC5ek&t=23 [23:13:04] that's all i have to say [23:13:45] jdlrobson: where is the goat path? [23:14:28] Maryana: i'm still confused with what you mean by login - it looks fine to me [23:15:14] jdlrobson: success message (which you should see on the cluster due to CentralAuth) and logout message are not totally styled yet [23:15:16] scapping (hides) [23:15:21] \o/ [23:15:34] um. well, for starters, the button looks like this: [23:15:45] that will go away once messages are sync'd to the cluster [23:15:46] Maryana: that's a problem with test [23:15:58] ah, ok [23:16:27] will the funky alignment of the login page boxes & text get fixed, too? [23:16:29] also Maryana the stuff to do with removing the ribbon from beta was never scheduled for this iteration so i'm a bit confused why you expected it to be.. [23:17:03] heh. login is also in iteration 3 [23:17:15] Maryana: login was in this iteration but never got finished [23:17:24] it rolled over [23:17:28] login was originally in iteratoin 2 but was moved to iteration 3 since it wasn't finished in iteration 2 [23:17:53] also bear in mind the beta is always 'unfinished' hence the caption on the settings page "By joining the beta, you will get access to experimental features, at the risk of encountering bugs and issues." ;-) [23:18:28] i'm not too concerned by sticking to an ironclad iteration release schedule. i guess i'm just still super confused by what's going out in a deployment. which seems… less than ideal [23:18:43] it is :( [23:18:49] we'll get that fixed [23:18:56] no biggie [23:19:10] it's just hard to know what to test when everything seems kinda up in the air [23:19:21] and i'm not sure what's a bug and what's a feature :) [23:19:39] we'll all get the hang of it after a couple more of these :D [23:19:52] brion: true dat [23:20:02] yupyup [23:20:19] consensus achieved! [23:20:45] MaxSem: did https://gerrit.wikimedia.org/r/#/c/29723/ get merged on test? i'm still seeing issues [23:21:11] cache issues? [23:22:15] doesn't seem to be.. [23:23:23] MaxSem: did you update to master on deployment branch? [23:25:03] jdlrobson, http://dpaste.org/R3bOG/ [23:27:36] Out of interest... Do you guys know of anything in your code that'd be calling action=parse on numerous special pages? [23:27:56] MaxSem: ^^ [23:28:27] action=parse, not just Parser? [23:29:05] 2012-10-23 23:26:22 srv297 eswiki: [73e068db] /w/api.php?action=parse&format=xml&page=Especial:Contribuciones/Ekonomista Exception from line 104 of /usr/local/apache/common-local/php-1.21wmf1/includes/WikiPage.php: Invalid or virtual namespace -1 given. [23:29:05] Stuff like that [23:30:06] no, I don't think we do anything like that... [23:30:12] MaxSem: there's still a serious issue with this footer [23:30:21] Reedy, MF doesn't call parse directly or use xml [23:30:50] a tag is not getting closed for some reason [23:30:50] Similar for WikipediaMobile? [23:30:54] I'm just making sure ;) [23:31:07] Reedy: mobile should be using action=mobileparse [23:31:19] mobileview [23:31:19] or action=mobileview rather [23:31:24] yep [23:31:25] jdlrobson: are we just missing a div from a bad merge [23:32:06] Right-o [23:32:18] The 10s of requests per second are just concerning ;) [23:32:32] Reedy: if it's not some sort of server-side bot i'd check for gadgets or something maybe [23:32:42] is it only on eswiki or on others as well? [23:32:47] Indeed [23:32:49] I've seen them all over [23:32:55] enwiki was giving 10+ per second this morning [23:33:02] damn [23:33:02] [23:33:16] ^ so I made the api die out instead, rather than propogating exceptions [23:33:23] running to get the boy back in a bit [23:34:56] A gadget does seem very likely, but also strange [23:35:11] Reedy: they're not coming from a common IP range or user-agent or anything? [23:35:25] got it [23:35:28] Without hacking it in, I've no idea how to get that info? [23:35:46] Reedy: look at the IP making the request [23:35:55] hm, we need better error logging still I guess :P [23:36:25] Reedy doesn't the error log log the client ip? [23:36:26] yeaaah [23:36:44] Nope [23:36:52] exceptions really ought to be cross-referenced with url, ip, user-agent etc [23:36:52] And if it's not fatal, we don't get a stack trace [23:37:08] New patchset: Jdlrobson; "mimic if statement above" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29730 [23:37:12] sigh ^ preilly MaxSem [23:37:15] https://bugzilla.wikimedia.org/show_bug.cgi?id=40166#c2 < on proper fatals we do get some useful info [23:37:56] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29730 [23:38:19] I'll deploy it after the scap concludes [23:38:19] Reedy: can you catch any of them in the sampled access logs, or are they ungreppable due to being POSTs or something? [23:38:28] MaxSem: okay great thanks! [23:38:47] looks like there's URL parameters, so should be findable [23:39:33] It's if/where/how we log it... [23:42:04] brion: T=36ms format=json action=mobileview page=Special%20Air%20Service sections=all prop=sections%7Ctext sectionprop=level%7Cline noheadings=yes [23:42:31] hmm [23:42:48] no IPs recorded? [23:43:09] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29718 [23:43:30] brion: yeah, I wasn't sure about posting an IP in a public channel [23:44:21] That one resolves back to an AT&T request [23:45:03] Reedy: well that's an action=mobileview so that's one of ours, looks legit enough [23:45:32] It's a weird special page [23:45:47] 'Special Air Service', a regular content page [23:46:16] oh [23:46:16] Duh [23:46:21] It's been a long day ;) [23:46:22] :) [23:47:09] is scap still running? daaaamn this used to be faster [23:47:33] last time i ran scap it took nearly an hour [23:47:51] jeez [23:48:25] ok some of my hits on en.m.wikipedia.org are returning new-style HTML, but old-style CSS [23:48:30] is this the scap article on wikipedia? http://en.wikipedia.org/wiki/Separation_of_content_and_presentation ? if so, it could use some help [23:49:22] Maryana: it's wmf devops jargon. :) short for sync-common-all-php [23:49:37] brion that is likely a symptom of scap not finishing and/or mobile cache needing to be cleared [23:49:43] o i c [23:49:46] :) [23:49:49] awjr: yeah we'll see when it's done :D [23:50:47] as i recall back in the day, sync-common-all synced a bunch of files and was slow, so I added a faster job that only synced the files for mediawiki…. hence -php [23:50:57] and now that's super slow. :P :) [23:50:58] whee Fatal error: Call to a member function getLocalURL() on a non-object in /usr/local/apache/common-local/php-1.21wmf1/extensions/MobileFrontend/i [23:50:58] ncludes/skins/SkinMobile.php on line 149 [23:51:16] red alert! [23:51:27] poop. [23:51:27] easy fix, i'm on it [23:51:28] i'm still seeing old wikipedia but debug=true shows correct wikipedia ( http://en.m.wikipedia.org/wiki/The_Publishers_Association?debug=true ) [23:51:41] i'm *really* glad i didn't set the blog post to be published today. "check out our awesome new mobile changes! we worked so hard on them! now you have a floating back button overlaid over all your article titles. enjoy. sincerely, wmf mobile team." [23:51:56] jdlrobson: looks like scap is still running [23:53:31] Maryana: this is typical for a deployment and that's not a helpful comment.. these big deployments are usually pretty stressful [23:53:52] just joking :) i know it's temporary [23:54:37] New patchset: Brion VIBBER; "Fix for fatal error in returnto handling" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29731 [23:55:08] MaxSem: that should handle it. sounds like most cases shouldn't encounter the bug though, it should be rare. i …. hope :) [23:55:49] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/29731 [23:56:38] brion: thanks for the quick fix [23:56:46] whee [23:57:33] https://wikitech.wikimedia.org/view/Wikimedia_binaries :D that brings back memories [23:58:12] i still cant believe wikitech still has certificate issues >_< [23:59:27] urgh [23:59:43] format=json action=mobileview page=Special:Search%3Fsearch%3Dorange%2Bsocks%2Bkilling%26go%3DGo sections=all prop=sections%7Ctext sectionprop=level%7Cline noheadings=yes