[00:00:13] Ks0stm: open up the app and go to settings -> about, check if it says 3.1 or 3.1.1 [00:00:38] 3.1 [00:00:46] dang. same thing i get :( [00:00:50] i'm not seeing the right version in the store [00:00:51] 1) wmlContext currently gets set by what's in useFormat. should it draw from the cookie instead (if present), falling back on what's in use format, and then falling back on the format specified for a particular device? [00:00:52] YuviPanda: and yours actually installed 3.1.1? [00:01:12] i'm seeing 3.1 which is *not* what i pushed [00:01:22] brion: yes [00:01:37] 2) when 'useformat=' appears in the query string, does that only happen for that particular request? (imo it should, as opposed to lasting for the session) [00:01:40] tfinc: i downloaded from the store, i have 3.1.1 [00:01:44] the app store is confused [00:01:46] yes [00:01:58] i wonder if its something with the old link that were passing around [00:02:22] whats the contact info for developer support? [00:02:26] 3) should i move cookie handling stuff to a SpecialPage, or should we just adopt another query string parameter for handling cookie stuff (particularly if we can't rely on javascript) [00:02:32] tfinc: i just updated the app. [00:02:40] I also saw a 'cloud' button, however [00:02:43] and actually … 4) should we change the cookie name to avoid confusion with ?useformat [00:02:48] preilly ^ [00:02:53] Ks0stm: going to sound cliché, but try restarting? [00:03:06] awjr: 1) makes sense [00:03:11] i'm seeing the same .. item you want is no longer available [00:03:12] awjr: 2) yes [00:03:30] awjr: 3) new query string use mobile action maybe [00:03:33] YuviPanda: OTRS ticket 2012040410015312 is having the same issues I am, by the way, and 2012040410012726 seems to be related [00:04:21] I don't have OTRS setup. tfinc should I ask for OTRS access or can someone forward me those? [00:04:37] Ks0stm: can you forward those to yuvi [00:04:44] those accounts can't easily be created [00:04:49] s/easily/quickly [00:04:58] I don't think I can without the permission of the people they're from [00:05:02] ah [00:06:14] Ks0stm: can you cc YuviPanda ? [00:06:30] tfinc: i haven't found a way in my portal to 'contact apple' or anything. have you? [00:06:38] checking [00:06:39] YuviPanda: Restarting didn't work either [00:06:53] preilly i have some changes i need to get in for jdlrobson before he starts his day, so i'll probably tackle the cookie stuff first thing tomorrow. [00:06:56] there is a contact us form [00:07:11] ok, can't update from within itunes either -- it gives me the same 'not available' [00:07:28] tfinc: i see a 'technical support' bit [00:07:33] https://developer.apple.com/membercenter [00:08:06] wow … the contact us from don't actually let you contact them [00:08:19] tfinc: https://itunesconnect.apple.com/WebObjects/iTunesConnect.woa/wa/jumpTo?page=contactUs [00:08:29] preilly: thats where i am [00:08:52] preilly: do you see a combination that actually allows you to contact them? [00:08:53] tfinc: I don't have access to iTunesConnect [00:09:38] failed app store update attempt doesn't log anything to the ipad's console as seen in xcode... [00:10:30] maplebed & Ryan_Lane : if we can't figure this out then i won't be able to climb [00:10:38] "successful" "update" that reinstalls 3.1 just logs that it kills the app for reinstallation, nothing unusual [00:11:32] Apple Developer Relations phone numbers: [00:11:33] (800) 633 2152 [00:11:33] (408) 974 4897 [00:13:26] preilly or MaxSem can one of you please review: https://gerrit.wikimedia.org/r/#change,4249,patchset=1 [00:13:37] preilly: their office is closed [00:14:01] tfinc: damn [00:14:29] of course it is. *sighhhhh* [00:14:34] awjr, trailing ws again [00:14:41] blargh [00:14:58] ok ..found the general error form [00:15:38] it's not our exact problem but its worth a shot [00:15:44] googling around finds plenty of people with the issue [00:15:47] but no resolutions [00:15:48] :| [00:17:21] beg apple to update your app again and see if it works is all we got left i guess :P :) [00:18:03] ok .. contact has been sent to apple [00:18:10] this is such a pain [00:18:39] tfinc: totally [00:18:46] could they have screwed up staging the correct IPA ? [00:19:10] entirely possible [00:19:13] awjr: 238 » » if ( !isset( $extMobileFrontend )) { [00:19:15] 'there are two hard problems in computer science: naming things, cache invalidation and off by one errors' [00:19:20] they put 3.1 into the 3.1.1 slot allowing someone to download it but not udpdate [00:19:24] hehe +1 [00:20:00] awjr: do you not like stylize.php [00:20:20] good luck guys, i'm gonna try out star wars kinect ;) [00:20:20] :( [00:20:53] maybe, the problem is in commit tool? [00:21:46] MaxSem: explain? [00:21:55] YuviPanda: that's 3 problems ;) [00:22:08] oops, off by one! [00:22:15] it doesn't show w/s clearly? [00:22:19] YuviPanda: i just got 3.1.1 through the store [00:22:31] tfinc: I got 3.1.1 through the store when I first tried it [00:22:42] but i had to load it from here : http://itunes.apple.com/us/app/wikipedia-mobile/id324715238?ls=1&mt=8 [00:23:20] download it through itunes [00:23:23] then sync [00:23:28] works great now [00:23:49] Ks0stm: ^ [00:24:22] YuviPanda: When I tried updating through iTunes it gave me an error [00:24:35] tfinc: can you tell Ks0stm the exact steps you followed? [00:24:36] and this i on 5.1 [00:24:39] is* [00:25:13] sure [00:26:15] 1) load this link - http://itunes.apple.com/us/app/wikipedia-mobile/id324715238?ls=1&mt=8 [00:26:24] 2) click update in iTunes [00:26:26] 3) sync [00:27:36] YuviPanda: i'm going to delete it and try it one more time [00:27:47] ok [00:28:57] reinstall worked perfectly [00:29:00] no issues [00:29:05] tfinc: Do you have OTRS access? [00:29:08] yes [00:29:39] Tickets 2012040410015312 & 2012040410012726 seem related [00:30:28] Ks0stm: search is not finding those [00:30:33] pass me a direct link to them [00:30:45] Ks0stm: were you able to verify that my workflow worked for you? [00:30:47] that's https://ticket.wikimedia.org/otrs/index.pl?Action=AgentTicketZoom&TicketNumber=2012040410015312 and https://ticket.wikimedia.org/otrs/index.pl?Action=AgentTicketZoom&TicketNumber=2012040410012726 tfinc [00:31:02] they're in the mobile subqueue of info-en tfinc [00:31:03] YuviPanda: this confirms that i *didn't* screw up uploading a new ipa :D [00:31:10] :P [00:31:25] Thehelpfulone: thanks. that works [00:31:30] no problem :) [00:31:32] do all mobile developers have info-en mobile access tfinc? it might be a good idea to give it to those who can answer the mobile tickets [00:31:45] Thehelpfulone: i don't think thats a good idea [00:31:54] 90% of the emails i get through otrs are spam [00:32:02] and i've already told phillipe about that [00:32:15] until there is less spam i can't take up the devs time on this [00:32:23] do you subscribe to the mobile queue tfinc? [00:32:27] yes [00:32:29] most of the time we sort through the spam [00:32:34] I'm on it weekly [00:32:46] but to be honest .. i hate reading it as its mostly a waste of my time [00:32:53] our mobile feedback list is tons better [00:33:13] tfinc: It seems to have downloaded to iTunes this time around...it gave me an error earlier...and now it worked on my iPhone too, so it seems to be working [00:33:42] tfinc: I might recommend that you just check it weekly - the actual queue as opposed to looking at your emails [00:33:58] most of the time the spam will be cleaned by volunteers first, so then the devs have real stuff to look at [00:34:04] Thehelpfulone: thats what i do [00:34:25] but its so much email noise that i dred logging in [00:34:28] and there's still spam? hmm, I always clean up as much spam as I can on all the subqeues first [00:34:43] which queues do you have access to? [00:34:50] probably too many that scare you off! [00:34:54] Ks0stm: hmm, so seems to be a caching issue? [00:35:22] I don't know, but it's working now [00:35:47] Ks0stm: Thehelpfulone both otrs contacts have been responded to [00:35:48] :D [00:36:06] YuviPanda: remember … apple is perfect … they can do no wrong [00:37:26] sure, let me sit twiddle my thumbs and feel bad while they continue to do no wrong [00:37:26] sigh [00:38:01] so lets assess [00:38:07] we know that the right IPA is in the store [00:38:15] we know that using my workflow users can get the update [00:38:43] what other workflows are causing our users issues ? [00:39:15] tfinc: the same workflow you used is what Ks0stm tried a while ago, I guess [00:39:16] Thehelpfulone: just mobile [00:39:20] and it didn't work [00:39:34] but it did work now [00:39:49] yes [00:40:25] YuviPanda: pick some of the twitter users and get them to try again [00:40:29] we have plenty to choose from [00:40:31] yes just tweeted [00:41:25] good [00:41:27] i like twitter for this [00:41:40] jeezus .. did i just say i like twitter for something [00:41:43] * tfinc puts the keyboard down [00:41:54] tfinc: there's nothing wrong with twitter [00:41:55] :D [00:42:09] * YuviPanda looks at his 23,000 tweets [00:42:38] why are we still not in the search index? [00:42:40] this is annoying [00:42:50] i can't do an install directly on the iPad [00:43:38] because i can't find it [00:44:13] siiighhhhh [00:44:53] its times like this that make me not want us to have an iOS app [00:45:07] +1 [00:46:22] I can't install it [00:46:27] YuviPanda: so I'm scheduled to leave with Ryan_Lane in about 15min to go climbing with maplebed . if you need me to stay then i'm happy to but i'm not sure what else i can do right now [00:46:30] devgeeks: what happens ? [00:46:34] be specific [00:46:35] I can see it available as an update [00:46:43] who here has a gun? [00:46:51] but then it says the item I am trying to buy is no longer there [00:46:54] tfinc: well, I see nothing we can do other than email apple. [00:47:04] devgeeks: your seeing the same issue that has been reported [00:47:05] YuviPanda: we live in america. even if we don't have one, we just just drive to the store [00:47:06] tfinc: which you did [00:47:22] Ryan_Lane: do you have to be over 21 to buy one of those too? :D [00:47:26] stabbing feels better. more personal. [00:47:32] YuviPanda: nope. only 18 [00:47:36] Ryan_Lane: o_O [00:47:38] O_o [00:47:41] "this is for 5.1!!" *stabbity* [00:47:42] o_o [00:47:43] O_O [00:47:47] Ryan_Lane: ok [00:47:50] but don't even think about drinking unless your 21 [00:48:03] that takes all the fun away from owning a gun [00:48:19] tfinc: hold my beer. watch this. [00:48:24] * tfinc hums america fuck yeah [00:48:32] hollywood is so full of lies [00:48:35] why am I surprised? [00:49:38] New patchset: awjrichards; "Added validation for edittoken handling Added method to write form data to specified wiki page Integrated method with onSuccess()" [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/4318 [00:49:58] Change abandoned: awjrichards; "Replaced by https://gerrit.wikimedia.org/r/4318" [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/4249 [00:50:03] tfinc: you emailed apple, right? [00:50:11] yes [00:50:18] tfinc: we just need to figure out what we're going to communicate to people with this issue [00:50:43] MaxSem, preilly: https://gerrit.wikimedia.org/r/4318 [00:50:52] 'we are seeing intermittent installation failures. This seems to be an issue with the app store. We're in touch with apple and trying to fix this. Stay tuned' [00:50:52] ? [00:50:55] tfinc: ^ [00:51:01] hmm [00:51:09] yeah [00:51:16] lets spread the blame since we have no control here [00:52:17] does deleting the old one help/ [00:52:30] or is it just a too bad so sad issue [00:52:37] devgeeks: try it? [00:52:55] YuviPanda: https://twitter.com/#!/jeremyschofield :D [00:52:57] k [00:53:08] five? you aim too low sir [00:53:18] it had 'curses' *and* a grammatical mistake :P [00:53:24] this is true [00:53:28] i should have asked for more :D [00:53:31] :P [00:53:40] there are many things i need to learn from you YuviPanda [00:53:52] and there are many that i should not [00:53:53] ;) [00:54:06] i really should do https://xkcd.com/249/ [00:54:09] with my computer [00:54:29] deleting and installing worked [00:55:00] devgeeks: did it show up on search? [00:55:00] oooh... "save to read it later", hehe [00:55:05] devgeeks: :) [00:55:06] FTW [00:55:10] I just went back in via updates [00:55:29] shoulda tried searching [00:56:07] were showing up in searches now 'Wikipedia mobile' [00:56:19] whew [00:56:44] Maybe it's actually fixed... maybe it only worked because it's working not cause I deleted [00:57:09] blame spread. [00:57:19] hah [00:57:42] I *really*, *really* want to see apple become irrelevant before I die [00:57:49] pfft [00:57:49] say, go the way of RIM [00:57:52] or Palm [00:58:02] trying hard to be cool and failing miserably [00:58:04] And leave us with... google???? [00:58:05] I think RIM, not Palm [00:58:14] *shudder* [00:59:07] 'look guys, you're cool! Buy this! THIS!' [00:59:15] and silence [00:59:23] I would like to see some ACTUAL open source phone-ness [00:59:34] Not google's "version" of it [00:59:51] devgeeks: http://en.wikipedia.org/wiki/Boot_to_Gecko [00:59:59] devgeeks: CyanogenMod is reasonably open sourcer [01:00:40] tfinc: i'll timebox a 3 hour tomorrow and see if i can get our app to run 'natively' on the browser. [01:00:47] cyanogen is just a hack on google ickiness [01:00:55] devgeeks: take a look at boot to gecko [01:01:05] tfinc: I see your Mozilla and raise you a UBUNTU! [01:01:06] http://en.wikipedia.org/wiki/Ubuntu_Mobile [01:01:11] hehe [01:01:47] * devgeeks <3 Canonical [01:01:58] YuviPanda: thats lower priority then the decoupling [01:02:16] tfinc: okay. Will move it around accordingly. [01:02:37] YuviPanda: what else would you like me to help with before i head out? [01:02:54] tfinc: looks good. Enjoy your climb! [01:03:11] hey tfinc - text me when you leave, would you? [01:03:19] maplebed: ok [01:03:25] or rather, text me when you're at west oakland. [01:03:31] Ryan_Lane: lets leave in the next 5-10min [01:03:35] that's a more reliable method of getting me to meet you on time. [01:03:39] maplebed: but you want us to get out at ashby right? [01:03:41] tfinc: beta on friday looks okay to you? [01:03:45] ok [01:03:48] tfinc: correct - ashby station. [01:03:48] and by beta I mean beta to the android market [01:04:10] we shall now call them 'beta web' and 'beta app' [01:04:25] to reduce confusion [01:04:31] i don't like releases on fridays [01:04:49] lets push early monday [01:04:56] ah [01:04:57] right [01:04:58] unless you can make it thur [01:05:00] which i doubt ;) [01:05:03] nope, that's too close [01:05:08] it's alread thur here [01:05:11] and i've a bus ride bakc [01:05:12] *back [01:05:29] tfinc: monday morning PDT then. [01:05:47] and I think I should be able to make this one all by myself? Thanks to the fact that I now hve access [01:05:48] nice [01:06:10] heads up to everyone .. make sure you read awjr 's mail about a deployment freeze [01:06:26] YuviPanda: does the beta need any MF changes ? [01:06:49] tfinc: well, it needs a few changes to master. Doesn't need any deployments [01:06:57] tfinc: also, we need to figure out what to do with Home page. [01:07:06] i'll have to make it mirror MF's for now. [01:07:20] that's 'coupling' [01:08:46] YuviPanda: make me feel dirty to have to do that [01:09:04] so we need to figure out a strategy for that [01:09:09] but it can just be blank for the first beta [01:09:15] ah yes [01:09:17] i think [01:09:19] best would be [01:09:20] to have a page [01:09:22] that tells people [01:09:24] how to give us feedback [01:09:36] :D [01:09:42] and load that instead. [01:10:46] that might work. lets chat about it tomorrow [01:10:50] Ryan_Lane: i'm spinning down here [01:11:13] tfinc: alright [01:11:16] will file a bug [01:11:28] * YuviPanda keeps an eye on the twitter stream [01:11:33] tfinc: enjoy climbing. [01:11:40] try not to die :) [01:11:50] will do :D [01:11:55] maplebed will treat us well [01:12:06] YuviPanda: i'll have my phone on me in case you need to get a hold of me [01:12:07] he's a bed made of maples [01:12:08] how can he not [01:12:13] tfinc: yes, will do. [01:16:43] YuviPanda: yes, a bed of maples [01:17:05] i'm not sure if maples are a real thing [01:17:07] or imaginary. [01:17:09] like canada [01:19:16] ha ha ah [01:19:17] ha [01:37:14] Travel tiem [01:37:15] brb [01:37:19] in a day [02:22:47] there have been at least a few new iphone app OTRS tickets recently [02:22:47] http://dpaste.com/726993/plain/ [07:36:21] Launchpad hasn't noticed the vumi-wikipedia source package that I uploaded last night, but won't let me upload it again. [08:31:40] Change abandoned: Jdlrobson; "Brion is right. I hadn't even thought of using the -webkit prefix. Thanks!" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4253 [08:41:40] New patchset: Jdlrobson; "add -webkit-background size for browsers that do not support" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4321 [10:06:17] New patchset: Jdlrobson; "make box shadow at top of footer full width" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4322 [10:06:17] New patchset: Jdlrobson; "prevent search font size increasing on focus" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4323 [10:06:18] New patchset: Jdlrobson; "increase font size of links at bottom of footer and color of more/less" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4172 [10:06:19] New patchset: Jdlrobson; "increase size of wikipedia logo in footer" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4171 [11:01:16] New review: Jdlrobson; "This seems to have broken the outputted html leaving an unclosed div element. Most noticeably on the..." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4198 [11:06:06] New patchset: Jdlrobson; "allow enabling of beta via query string" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4327 [11:34:35] New patchset: Jdlrobson; "redesign footer" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4328 [11:35:02] New review: Jdlrobson; "This is not finalised. Please do not merge." [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/4328 [12:26:07] [WikipediaMobile] jdlrobson pushed 1 new commit to master: http://git.io/E66BuQ [12:26:07] [WikipediaMobile/master] Merge pull request #197 from wikimedia/v3.1.1-iOS - Jon Robson [12:26:20] Project WikipediaMobile - Nightly builds build #275: SUCCESS in 6.7 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/275/ [12:26:20] * yuvipanda: Move WebKit settings to proper location. [12:26:21] * yuvipanda: Bumped version number for iOS [14:44:47] New patchset: Jdlrobson; "redesign footer" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4328 [14:45:31] New review: Jdlrobson; "Still not finalised. Just sharing." [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/4328 [14:54:10] jdlrobson, what do you think about switching MF to HTML5? [14:55:57] hello MaxSem [14:56:06] you mean using section elements etc [14:56:16] and header tags [14:56:28] no, new tags wil definitely cause problems [14:56:43] so far, I'm interested in doctype [14:56:48] we've already done that no? [14:57:00] yep we have [14:57:30] and it's deployed [14:57:30] and one line below, [14:57:35] ROFL [14:57:45] lol [14:58:06] although.. http://validator.w3.org/check?uri=http%3A%2F%2Fen.m.wikipedia.org%2F&charset=%28detect+automatically%29&doctype=Inline&group=0 [14:59:37] * MaxSem waves to YuviPanda [14:59:46] * YuviPanda waves to MaxSem  [14:59:47] we've been discussing HTML5 [14:59:50] aha [14:59:58] * YuviPanda looks at logs [15:00:38] speaking of what's deployed, I don't see the desktop view link anymore [15:02:39] in the beta MaxSem ? [15:02:45] no [15:02:53] well, I was interested in
etc in the API [15:03:07] the desktop view link is in the footer of the deployed regular mobile site [15:03:15] on the beta it is in under the W logo [15:03:39] hmm, and more/less switch at the bottom does nothing [15:03:50] which page are you looking at? [15:03:56] and what device? [15:04:01] this sounds strange [15:04:13] Desktop Opera [15:04:56] on the live site or dev version? [15:05:10] live site, all pages [15:05:15] * jdlrobson boots up [15:06:04] jdlrobson: when do you think the CSS refactoring will make it to master? [15:06:22] i thought you wanted to do it YuviPanda to get up to speed with gerrit? [15:06:37] yes, and i have gerrit set up [15:06:45] jdlrobson: but I thought you wanted to get RL in [15:07:00] i've done that - it's just waiting for review and merging [15:07:04] ah [15:07:09] i'll wait for that to merge [15:07:12] * jdlrobson hates opera [15:07:16] jdlrobson: link? [15:07:46] https://gerrit.wikimedia.org/r/#change,4226 [15:08:54] i'm sure this has been done to death but looking at diffs sucks [15:09:16] yeh quite a few unit tests seem to be breaking in opera a few related to toggling [15:09:18] will look into it [15:10:50] actually opera desktop is working fine for me [15:11:08] v11.61 [15:12:04] now confused [15:28:54] jdlrobson, http://i39.tinypic.com/2iuexhi.png [15:30:08] MaxSem what am I looking at? [15:30:36] footer and its DOM tree in latest Opera [15:30:43] the ::after ? [15:30:52] note that strange tag I've highlighted [15:30:58] the ::after I guess is something opera adds in for debugging purposes that relates to css3 ::after pseudo [15:31:45] http://coding.smashingmagazine.com/2011/07/13/learning-to-use-the-before-and-after-pseudo-elements-in-css/ [15:32:13] so clicking on more does nothing but revealing several-pixel-wide blank space [15:32:15] there is a rule #footer .links:after which creates a new div after the element that matches #footer .links - interesting that opera dev tools shows that [15:34:52] i'm confused sorry [15:35:10] the footer doesn't toggle for you? it is just always open? [15:35:26] yup [15:35:35] Can you try the latest footer as well and see if that has the same problem? http://mobile-geo.wmflabs.org/w/index.php?title=Main_Page&useformat=mobile#section_footer [15:35:39] what version of opera are you running? [15:35:43] the latest? [15:35:52] this one WFM [15:37:08] mm very odd [15:37:17] maybe there was some bad html somewhere [15:37:29] 11.62 [15:37:57] k i've updated to 11.62 [15:38:43] footer works for me there too :/ [15:38:54] is it caching some old css potentially? [15:40:44] brb [15:52:10] New patchset: MaxSem; "Unbreak sticky cookies after change 4198" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4209 [15:53:31] New patchset: MaxSem; "WML view had lost its search field some time ago :0" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4333 [17:09:12] MaxSem: is MF broken in master right now? [17:09:42] define broken:) [17:10:32] preilly, are you referring to https://gerrit.wikimedia.org/r/4333 ? [17:10:52] MaxSem: nope [17:11:20] MaxSem: PHP Fatal error: Call to a member function isSpecial() on a non-object in /home/preilly/core/extensions/MobileFrontend/MobileFrontend.body.php on line 1009 [17:11:35] mmmm [17:13:54] wait a sec, I've got wikidata office hours [17:15:20] MaxSem: is it broken for you too? [17:15:31] I'm not using beta [17:16:37] MaxSem: not beta [17:16:46] MaxSem: I get a 500 on anything [17:16:48] how to repro? [17:16:54] works for me [17:18:58] MaxSem: try setting a mf_useformat desktop cookie [17:19:10] and the page load will be a 500 [17:19:22] preilly, my fix to it is still not merged [17:20:01] I have this cookie and still don't see mobile site [17:20:15] if I have a desktop cookie [17:20:18] I get a 500 [17:21:32] greetings all [17:21:38] [WikipediaMobile] brion pushed 2 new commits to master: http://git.io/uTgHTw [17:21:38] [WikipediaMobile/master] bug 35710 - about page links now open in external browser - Brion Vibber [17:21:38] [WikipediaMobile/master] Merge pull request #199 from brion/about-link-fix - Brion Vibber [17:21:51] Project WikipediaMobile - Nightly builds build #276: SUCCESS in 7.1 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/276/ [17:21:52] brion: bug 35710 - about page links now open in external browser [17:22:01] MaxSem: shouldDisplayMobileView seems messed up [17:22:10] I've set that cookie and still don't see the problem [17:22:49] messed up merge? [17:22:59] heatherw: worst case we can just present inside a desktop browser [17:23:18] the stretched out version? [17:23:32] MaxSem: I've got the version from master with no local changes [17:23:48] MaxSem: and it throws a 500 if I've got the desktop cookie set [17:23:51] preilly: did you see my comment about sections being broken? [17:23:57] * MaxSem sesses with cookies once again [17:24:18] jdlrobson: yes [17:24:28] cool just checking it was on your radar [17:24:32] jdlrobson: I don't know why that change would effect that at all [17:24:39] jdlrobson: could it be something else [17:24:40] me neither it's very strange [17:24:45] i guess it's exposing something that wasn't exposed before [17:24:48] jdlrobson: like maybe the copyright change [17:25:00] it was that commit - if you check out the commit before it works fine [17:25:03] but on that commit it breaks [17:25:46] preilly, still can't repro. setup differences? [17:25:51] jdlrobson: I've got no idea [17:26:19] MaxSem: do you have your unmerged changes locally? [17:26:41] nope, I'm on master [17:27:46] MaxSem: and you are sure you've got the desktop cookie set? [17:28:25] from Dragonfly: Cookie: optin=1; wiki_session=00d837382a78e9d2302b10dff41aa3ab; wikiUserID=2; wikiUserName=MaxSem; wikiToken=edc48fdcdf6b8bd23fdfad7cab57845a; mf_useformat=mobile; mfsecure=1 [17:28:37] MaxSem: so you don't [17:28:49] MaxSem: mf_useformat=desktop [17:28:50] frak, I clicked on mobile view again [17:29:37] ( :: )))))))) [17:29:57] optin=1; wiki_session=00d837382a78e9d2302b10dff41aa3ab; wikiUserID=2; wikiUserName=MaxSem; wikiToken=edc48fdcdf6b8bd23fdfad7cab57845a; mf_useformat=desktop; mfsecure=1 [17:30:01] MaxSem: so, on line 130 if ( !$this->shouldDisplayMobileView() ) {…. returns differently then on 388 [17:30:11] MaxSem: and that request works? [17:30:47] the reviews on the app store so far are fascinating [17:31:04] as we expected … some people love it … some people hate change [17:31:05] tfinc: in what way? [17:32:15] "New Update RUINED a good app" vs "Wow! What an improvement!" [17:32:30] preilly, try git fetch ssh://maxsem@gerrit.wikimedia.org:29418/mediawiki/extensions/MobileFrontend refs/changes/09/4209/2 && git checkout FETCH_HEAD [17:32:45] eh, git fetch ssh://preilly@gerrit.wikimedia.org:29418/mediawiki/extensions/MobileFrontend refs/changes/09/4209/2 && git checkout FETCH_HEAD [17:33:29] MaxSem: do you have https://gerrit.wikimedia.org/r/#change,4209 locally? [17:33:37] did we get that update issue sorted yet? [17:34:09] preilly, no [17:34:18] brion: the market looks to have finally resolved its own caching issues [17:34:28] MaxSem: but, your site works with that cookie set? [17:34:29] tfinc: excellent. i can confirm i got a 3.1.1 update on my ipad \o/ [17:34:35] woot. [17:34:36] MaxSem: that makes no sense at all [17:34:36] preilly, yes [17:34:54] * MaxSem blames git, as always [17:35:00] brion: so comments that are noise because they are simply change averse. others mention real limitation or non iOS practices which are justified [17:35:03] some* [17:35:22] *nod* [17:35:28] MaxSem: what is the last commit that you have [17:35:31] the monthly metrics call ... is it now? Should i dial in? [17:35:33] MaxSem: if you git log [17:35:38] jdlrobson: it's now [17:35:50] i'm skipping it, i've been at enough metrics meetings :) [17:36:10] brion: are you at home? [17:36:13] yeah [17:37:18] heatherw: we won't be able to use a real device [17:37:28] so you can either [17:37:37] 1) navigate within eriks browser [17:37:41] 2) show the blog post [17:37:44] 3) … other? [17:37:52] show my slides? [17:38:11] heatherw: does erik have them ? [17:38:19] tfinc: yes [17:38:27] YuviPanda: very polarized reviews on the iTunes store [17:38:30] heatherw: sure [17:38:40] hmm, I saw zero when I checked a fwe hours back [17:38:51] 25+ reviews for 3.1.1 [17:38:58] i still see 0 for current version :| [17:39:01] and 14 'overall' [17:39:22] YuviPanda: look at the US store [17:39:24] MaxSem: what does git log say for you? [17:39:34] sorry, was afk [17:39:35] wow, these are separated by geographic region? [17:39:38] yes [17:39:44] just like the android store is [17:39:47] this is no different [17:39:57] Author: Jon Robson [17:39:57] Date: Tue Apr 3 00:27:06 2012 +0100 [17:39:57] removed minified javascript [17:40:17] tfinc: link? [17:40:29] Android market is segmented by language, IIRC. Not language [17:40:44] MaxSem: that isn't right [17:40:50] MaxSem: it should be: [17:40:53] commit afed3c6757090a28cdf12a5f04a06855a0ca02fe [17:40:54] Author: awjrichards [17:40:55] Date: Wed Apr 4 11:52:53 2012 -0700 [17:40:56] Fixes bug 35686 - when looking for the 'mobile token' in the hostname, the str_replace now wraps the mob [17:40:57] [17:40:58] Change-Id: I5bfa27fd46de233cf8f08eee60575b9381151ac6 [17:40:59] YuviPanda: go to iTunes [17:41:06] YuviPanda: click one the store link [17:41:16] YuviPanda: scroll to the bottom right where you'll see a flag [17:41:24] YuviPanda: click on the flag [17:41:26] yup, I can't read git log [17:41:47] better [17:41:49] MaxSem: so, that is what you have? [17:41:51] preilly, I'm indeed at that Arthur's change [17:42:07] MaxSem: this makes no sense [17:42:12] which is tip of master [17:42:25] okay that's some polarization [17:42:33] YuviPanda: now you see what i mean [17:42:46] preilly, what are your MF settings? [17:43:53] MaxSem: nothing special [17:44:23] MaxSem: so, the issue with mine is that shouldDisplayMobileView returns differently on 130 and 388 [17:45:44] quite possible, since it doesn't always load cookies now, so getUseFormat() behaves chaotically [17:46:06] MaxSem: okay, well that is what breaks it [17:46:26] YuviPanda tfinc would be interesting to know whether these are ipad/iphone users and what version [17:46:28] preilly, have you tried checking out my fix? [17:46:38] MaxSem: not yet [17:46:48] you should [17:46:53] MaxSem, preilly: im beginning work on changing how the cookies work [17:46:53] jdlrobson: we've never been on the iPad before [17:47:01] jdlrobson: so its not that [17:47:05] which includes totally decoupling useformat from cookie stuff [17:47:11] mm [17:47:15] awjr: so, master is broken currently [17:47:18] :( [17:47:26] if MaxSem's fix fixes it, let's get it merged for now [17:47:35] awjr: if you've got a cookie set for desktop [17:49:17] preilly what's the problem? [17:49:44] tfinc: hitting back on settings takes me back [17:51:58] preilly, MaxSem are you getting 404s on js resources on the mobile view? [17:52:08] and I *do* see a scrolling indicator [17:52:29] tfinc: is there any way to find out which iOS version these guys are running? [17:52:41] let me check [17:52:41] awjr, yes - on *.min.js [17:52:57] MaxSem you need to make a config change [17:53:00] ?version= also looks weird [17:53:02] hang on one sec [17:54:17] MaxSem i hacked in a way to set properties in MobileFrontend.body.php with a config var - so you can use that to disable minification, which should take care of the 404s [17:54:38] YuviPanda: i don't see a way to view it yet [17:54:43] another FAIL of the iTunes store [17:54:56] MaxSem try: $wgMFConfigProperties = array( 'minifyJS' => false ); [17:54:58] it just continues to cause us problems [17:55:36] ?version= is not a big deal, but if you dont like seeing it, set $wgMobileResourceVersion to some arbitrary string [17:55:49] awjr, yup, it helped [17:55:55] tfinc: :( [17:56:10] shouldn't the version have some sensible defaults? [17:56:14] YuviPanda: ok i found out what they have for us [17:56:18] now we've people reporting problems with no way of contacting them or of figuring out what version they are. [17:56:20] hmm? [17:56:26] MaxSem personally i think null is pretty sensible :p [17:56:38] * MaxSem bites awjr [17:56:43] lol [17:56:45] it's really only useful for forcing cache invalidation [17:57:22] MaxSem: feel free to set it to something more sensible if you like [17:57:25] YuviPanda: its still mostly useless [17:57:33] :( [17:57:37] YuviPanda: the only advantage is that i can select reviews by country [17:57:41] no iOS version filter [17:57:49] is there a way to contact these people? [17:57:59] YuviPanda: nope [17:58:01] the guy with the huge review at the end - I'm not able to repro most of his stuff [17:58:01] :( [17:58:33] YuviPanda: i wish they would just tweet their reviews to us [17:59:23] tfinc: thoughts on what we could do? [17:59:36] give it another 24hrs so we can let the dust settle [17:59:46] we don't have enough reviews to make sense of this yet [18:00:10] ok [18:03:04] the uk reviews are especially brutal [18:03:53] you know those brits [18:03:59] they know how to write a stern letter to the times [18:04:28] 'I read the reviews and didn't update' [18:04:35] i think Apple pushed out v3.1 to a lot of folks [18:04:58] 'sorry, v3.1.1 is no different from v3.1' [18:05:27] UK reviews suggest that all of them are on v3.1 [18:05:45] SadPanda: correct [18:06:02] or people who downloaded 3.1 .. and didn't bother updated [18:06:07] and just filed a review against 3.1.1 [18:07:01] I feel utterly powerless [18:08:18] SadPanda: put the reviews down. lets get to 1.2 .. we'll get back to this once we have more reviews [18:08:35] some poeple probably got 3.1 thinking it was 3.1.1 [18:08:40] bluh [18:08:48] brion: that happened to you too yesterday, I think [18:08:58] yeah [18:09:20] both my ipads updated correctly today though, so hopefully things will even out soon [18:09:26] hope so. [18:09:43] brion: they will. this is why we will shift our energy to moving forward on what we know needs to get done [18:09:57] once the upgrade mess settles we'll actually know what problems are around [18:10:02] \o/ [18:13:09] tfinc: the blog post is out, btw [18:13:46] YuviPanda: woot! [18:17:04] New patchset: MaxSem; "Reanmed Mobile.php --> SkinMobile.php per conventions" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4345 [18:17:55] YuviPanda, is switching to GeoData planned for 1.3? [18:18:38] good question. [18:18:42] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/4345 [18:18:44] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4345 [18:18:57] we'll worry about 1.3 when we get 1.2 out the door ;) [18:20:06] * MaxSem wants it deployed in his lifetime :} [18:20:39] tfinc: I have a potential solution [18:20:49] It is possible that we're not restricting things to 4.3+ [18:20:50] for? [18:20:51] correctly [18:20:59] and these people are trying it out on older OSes [18:21:01] on which it fails [18:21:07] YuviPanda: intresting [18:21:35] our current req is 4.2+ [18:21:43] yes [18:21:51] i checked and we have it set to 4.2+ [18:22:02] is there someone you know with an earlier OS who can check if they have it available? [18:22:29] YuviPanda: i have an older iPhone at hoem [18:22:31] home* [18:22:37] possibly another in the office [18:22:44] would be nice if we could rule that out [18:22:47] New patchset: preilly; "No need to do this check twice" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4346 [18:23:04] i'll check post metrics meeting [18:23:10] sure [18:23:20] MaxSem, awjr: please approve https://gerrit.wikimedia.org/r/#change,4346 [18:27:56] New patchset: Jdlrobson; "redesign footer" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4328 [18:28:24] New review: Jdlrobson; "Don't merge" [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/4328 [18:28:26] New review: awjrichards; "The change looks fine but can you please make a more specific and clear commit message?" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: -1; - https://gerrit.wikimedia.org/r/4346 [18:29:49] preilly: the change looks fine - can you just amend with a more specific commit message/ [18:38:24] New patchset: preilly; "No need to call the shouldDisplayMobileView method twice since it is already called in the requestContextCreateSkin hook handler Change-Id: I767e7d3cbe120e5c7b7489b20af1cc41e528cdf7" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4346 [18:38:26] Is there a place where I can learn about how mobile stats are calculated ? [18:38:58] awjr: is that specific enough for you? [18:39:29] preilly: that'll do :) thanks [18:39:38] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/4346 [18:39:40] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4346 [18:40:04] http://stats.wikimedia.org/EN_India/TablesPageViewsMonthlyMobile.htm --> seeing 100-200% growth MoM makes me both feel awe / there is a bug [18:40:05] surfing through git log is much more useful with specific commit messages [18:46:06] I *finally* have a vumi-wikipedia deb being built on Launchpad. [18:46:45] I think it silently threw away the previous uploads because it didn't like the version number format or something. [18:48:08] jerith: ah, I see [18:48:22] jerith: how long before they're ready? [18:48:50] About 25 minutes until the build starts, it claims. [18:50:36] * MaxSem is still unsure if his vumi changes are ok [18:51:14] MaxSem: My big commercial deadline was today, and I'm busy tying up the loose ends from that. [18:51:27] np [18:52:04] I have a lovely four-day weekend ahead of me, so I'll do a proper review tomorrow or Saturday. [18:52:16] YuviPanda: http://thenextweb.com/apps/2012/04/05/wikipedia-updates-ios-and-android-apps-ditches-google-maps-for-openstreetmaps/?utm_source=twitterfeed&utm_medium=twitter [18:53:01] hehe [18:53:01] 'the company behind the project has released updates' [18:53:27] yeah .. thats factually incorrect [18:54:16] do you want to poke them about it? I think we should... [18:55:26] sure [19:00:48] YuviPanda: i talked to jay. he's going to get it fixed [19:01:17] cool [19:03:04] preilly: This is the PPA to use for the version of redis we depend on: https://launchpad.net/~rwky/+archive/redis [19:03:23] jerith: okay, great [19:06:33] New patchset: Jdlrobson; "hide close button in references box for firefox browser" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4354 [19:06:34] New patchset: Jdlrobson; "give capable browsers jQuery" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4355 [19:16:31] preilly: i don't know what's happen but think the weird section stuff is happening in headingTransformCallbackXHTML in MobileFormatter.php I think closeElement / openElement are not being used properly. Gonna grab some food now but maybe look again later [19:17:16] jdlrobson: okay [19:18:46] jdlrobson: thoughts on what to do for the main page? There's no API access to our current main page [19:18:55] preilly: i am on a call, done in a half hour [19:19:22] philinje: can we just get coffee later today instead? [19:19:29] sure [19:20:50] jdlrobson, what's wrong with headingTransformCallbackXHTML? [19:20:57] http://teaandkittens.co.uk/ [19:21:21] MaxSem: can't really speak now but on current dev version it's all messed up - footer goes into last
[19:21:26] one of the divs doesn't get closed [19:21:58] bizarrely problem was exposed with preilly 's fancy 841497fc9bb1c24b6139f0783e9e43cab3389a85 [19:22:12] tfinc: poke me when you've a moment, need to talk about berlin [19:22:13] jerith: wow [19:22:44] MaxSem: could it be div content related [19:22:57] awjr: what is your wgMobileUrlTemplate set to? [19:23:52] awjr: also, what does your wgMFConfigProperties look like as well [19:25:40] awjr: are you around? [19:26:27] ahaha http://validator.w3.org/check?uri=https%3A%2F%2Fen.m.wikipedia.org%2Fw%2Findex.php%3Ftitle%3DBarack_Obama%26useformat%3Dmobile&charset=%28detect+automatically%29&doctype=Inline&group=0 [19:27:25] 05 02:22:46 < jeremyb> there have been at least a few new iphone app OTRS tickets recently [19:27:28] 05 02:22:47 < jeremyb> http://dpaste.com/726993/plain/ [19:27:54] MaxSem: what is the ahaha about? [19:28:06] preilly, the number of errors [19:28:26] only 7 on desktop version [19:29:04] MaxSem: do you have wgMFConfigProperties set in your LocalSettings? [19:29:21] preilly, $wgMFConfigProperties = array( 'minifyJS' => false ); [19:30:34] MaxSem: do you have wgMobileUrlTemplate configured? [19:30:38] no [19:30:50] awjr: ping [19:33:44] preilly: python-vumi-wikipedia deb built and published in the PPA. The newest python-vumi is still waiting to build, but the only thing that changes is the version of redis-server it depends on. [19:34:39] jerith: okay, great [19:34:57] jerith: so, when are you going to move past the PPA phase? [19:36:13] I need to clean stuff up and put it all in a new PPA with better package metadata. [19:36:25] jerith: okay, makes sense [19:36:57] MaxSem: are you looking into the div issue that jdlrobson reported? [19:37:33] preilly, yes - looks like closing of
previously masked the problem with unclosed divs in section headings [19:37:56] I'll probably do something on that over the weekend, but the current PPA should be fine for getting stuff set up. [19:41:08] jerith: okay, great! [19:41:36] MaxSem: so, it should be an easy fix right? [19:41:48] yup [19:42:23] YuviPanda: http://techcrunch.com/2012/04/05/wikipedias-mobile-apps-drop-google-maps-for-openstreetmap/ [19:42:46] tfinc: we should've gone with the map showing altacraz :P [19:43:02] haha [19:43:05] it's also interesting how these people spin it without understanding why we do things [19:43:23] are you surprised ? [19:43:26] because you shouldn't be [19:43:41] i'm not. [19:44:08] what was left of my 'respect' for these 'journalists' went out when they basically printed whatever we wanted, whenever we wanted when I was at the startup [19:44:14] more press-release than 'journalism' [19:45:56] were catching a lot of twitter traffic post tech crunch [19:46:24] preilly hey im around-ish, was making lunch [19:46:34] preilly: Realistically, the only things that are likely to change are metadata and version numbers. The latter might require an uninstall/reinstall of the packages if it looks like the new packages are lower versions. [19:46:55] awjr: what is your wgMobileUrlTemplate set to? Also, what does your wgMFConfigProperties look like as well? [19:47:11] jerith: okay, that makes sense [19:47:35] YuviPanda: i'm getting a tweet every couple of seconds [19:47:37] jerith: is the new PPA with the redis version change built? [19:48:19] preilly: Built, but not yet published. [19:49:07] But all that does is depend on "redis-server (>= 2:2.4.10-whatever)" instead of "redis-server". [19:49:34] tfinc: what search are you using, exactly? [19:49:46] So it'll just fail to install if you aren't using redis from the PPA I linked earlier. [19:50:04] i always have 'Wikipedia mobile' & 'Wikimedia mobile' running [19:50:10] preilly: my wgMobileUrlTemplate = '%h0.m.%h1.%h2' but you should only do that if you actually have a mobile domain set up locally - otherwise you should leave it at default [19:50:14] preilly: $wgMFConfigProperties = array( 'minifyJS' => false ); [19:50:22] hmm, i've both as well [19:50:48] New patchset: MaxSem; "Fix unclosed div" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4361 [19:50:57] here we go^^^ [19:51:38] ah refresh now sees it [19:51:50] tfinc: for the RL meeting - should i just call r35? [19:52:30] i'll spin up a conf line. keep in mind that your optional on that meeting [19:52:36] so no harm if you have other things to do [19:53:04] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/4361 [19:53:07] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4361 [19:53:15] tfinc: i'm happy to not go to the meeting as long as someone can mail out notes [19:53:17] actually .. video may be even better as i want roan to give us an overview of how resource loader works [19:54:41] awjr: so, in beforePageRedirect the isSpecial( 'Search' ) section messes up zero rated sites [19:54:57] awjr: getMobileUrl [19:55:24] awjr: makes http://preilly.zero.mediawiki.org/index.php/Quark into http://preilly.m.zero.mediawiki/index.php/Quark [19:56:21] preilly what is desired behavior? [19:56:46] awjr: that it remains zero [19:57:12] preilly i dont really know how zero interacts with mobilefrontend [19:57:18] awjr: so, http://en.zero.wikipedia.org/wiki/Quark do a search on that page [19:57:32] awjr: it redirects to http://en.m.wikipedia.org/wiki/Quark [19:57:51] preilly actually it did not redirect me to .m. [19:57:54] it kept me on zero [19:58:13] ah [19:58:14] preilly [19:58:16] awjr: type Quark press enter [19:58:21] if you click 'enter' it will send you .m. [19:58:22] yeah [19:58:26] awjr: the results don't [19:58:32] awjr: those are relative [19:59:03] ok so how does zero relate to mobilefrontend though? i thought it was its own ext? [19:59:05] awjr: so, enter or pressing the search [19:59:17] awjr: it is it's own extension [19:59:49] awjr: but, the beforePageRedirect methods isSpecial( 'Search' )… section redirects to .m. [19:59:56] awjr: in MF [20:00:15] preilly: so the problem is that shouldDisplayMobileView is returning true when it should be returning false? [20:01:35] office wifi again? [20:01:36] preilly: so the problem is that shouldDisplayMobileView is returning true when it should be returning false? [20:01:53] awjr: yes [20:02:22] preilly my guess is that isMobileDevice is returning true. [20:02:46] oh dumb that's not right [20:02:50] YuviPanda: https://twitter.com/#!/evula/status/187966223806504960 :D [20:02:58] or perhaps you have a cookie set for mobile view? [20:03:51] tfinc: saw that :) You had started the conversation so left it on :) [20:04:13] hmm nope [20:05:29] tfinc: who does comment approval duty? [20:05:37] preilly: it's because isMobileDevice() is returning true [20:05:49] so MF thinks you should be seeing the mobile version of the page [20:05:56] awjr: i'll take notes [20:06:15] isMobileDevice() is returning true because, at least on en.wikipeida, X-Device is getting set [20:06:18] tfinc awesome thanks [20:06:20] preilly: http://etherpad.wikimedia.org/ResourceLoader-Mobile [20:07:15] awjr: well, it should be the mobile view [20:07:27] awjr: just not on the .m. domain [20:08:37] preilly then zero sites should have $wgMobileUrlTemplate = ''; configured rather than the default [20:09:05] awjr: that wouldn't work [20:09:13] preilly, MaxSem: Should we be returning a Zero URL with the SMS content? [20:09:15] preilly why not? [20:09:43] awjr: because it's the same backend [20:10:11] preilly well then that's complicated :( [20:11:21] 'OSM is also backed by Google, as it also participates in GSoC' [20:11:27] 'Isn't OSM backed by google? So it circles around a new name but lands on the googles doorstop no?' [20:11:27] wow [20:12:54] preilly not sure what to tell you - when you're out of the meeting perhaps we should brainstorm a solution [20:13:33] awjr: okay [20:14:06] What is bootstrap() ? [20:14:37] the function that roan just explained at the end of the initialization pass [20:14:52] keep in mind that these are just rough notes for roan is talking about here [20:15:03] i'm not trying to make this exhaustive [20:15:11] He is talking right now? [20:15:16] Krinkle: yes [20:15:27] Oh, I didn't realize these were talk notes [20:15:36] I thought you were planning something over the net [20:15:42] I won't add things he's not saying then [20:15:48] ie.. nothing since I'm not there :P [20:16:00] nope. thats why i just sent it directly to preilly in this channel [20:16:04] otherwise i would have said something [20:16:05] no worries [20:16:19] * Krinkle clicks all links [20:25:00] YuviPanda: hi [20:25:04] philinje: hi [20:26:04] YuviPanda: some of the user reactions to the new app are to be expected, but just wondering, did you test on all versions of iOS? [20:26:48] we didn't test extensively on iOS 5.0 and had beta testers trying it out as well [20:27:04] bizarre stuff, eh? [20:27:06] err [20:27:08] we *did* [20:27:09] not didn't [20:27:27] we didn't test extensively on 5.1 and hence the v3.1 issue [20:27:42] amazing what some people are seeing [20:27:57] philinje: it is partly consistent with my theory that they are seeing a v3.1 push [20:27:59] instead of a 3.1.1 [20:28:18] even more reason to get 3.2 out the door [20:28:23] ok - so the App store just automatically attaches the latest version number to the reviews [20:28:29] awjr: what about something like: http://pastebin.mozilla.org/1555204 [20:28:39] philinje: apparently. [20:28:49] philinje: i'm downloading a 4.3 sim now, will test on it as well [20:29:16] preilly i was thinking this might actually be a good place to insert a hook [20:29:21] does Perfecto have all versions? [20:30:06] preilly: plus the redirect stuff in that code snippet is WMF-specific [20:30:34] preilly: however, the general idea would certainly work and could be a good quick fix [20:30:35] awjr: as, is isZeroRatedRequest [20:30:51] preilly yeah, exactly, which is why i think a hook would be better [20:31:02] philinje: i think it does. haven't tried it out. [20:31:17] awjr: how do you envision that working? [20:31:41] will do if the sim fails [20:31:44] * awjr thinks for a minute [20:31:46] ah, if we only had QA support :( [20:32:12] philinje: good thing we have a candidate going to second stage interview with YuviPanda :) [20:32:22] cool! [20:32:24] preilly we could provide a hook in getMobileUrl to give other ext's the opportunity to modify the URL [20:32:35] tfinc: poke me when you've a moment or two, will need to talk about that + berlin [20:32:39] k [20:33:03] \o/ hooray on the QA front [20:33:40] preilly or we could provide a hook getMobileUrl to see if we should even bother modifying the URL [20:34:09] the first is probably more flexible, the latter probably more typical use-case, i would guess [20:34:47] [WikipediaMobile] brion pushed 2 new commits to master: http://git.io/sCCFwg [20:34:47] [WikipediaMobile/master] Bug 35734 - Update iOS splash screen images - Brion Vibber [20:34:47] [WikipediaMobile/master] Merge pull request #200 from brion/splash-update - Brion Vibber [20:35:00] Project WikipediaMobile - Nightly builds build #277: SUCCESS in 6.4 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/277/ [20:35:00] brion: Bug 35734 - Update iOS splash screen images [20:36:17] YuviPanda: we should ask Lindsey to make some tweaks to the UI, but anyway that will come out of the new nab work [20:36:57] philinje: are we going to unify Web UI and App UI going forward? [20:37:16] I personally think we should... [20:37:20] that's the theory, but in practice there will be some differences [20:38:14] also, it's hard to make the iPhone browser fit, just because it always has the toolbar at the bottom [20:39:04] :) [20:39:32] philinje: we're going to be reusing lots of it though. [20:39:41] 1.2 will have new styles and new reference reveal [20:40:00] YuviPanda: Did you notice the CC license / copyright stuff is added to In the News as bullets [20:40:17] yes, that seems to be a recent issue on MobileFrontend [20:40:22] that I think is being fixed [20:40:36] ok [20:47:25] YuviPanda, when I follow an interwiki to a different language in the app, it still searches in my language, this is inconsistent [20:47:38] MaxSem: yes. very. [20:47:47] that's what the language selector is supposed to fix [20:48:01] good to hear [20:48:06] nobody's been able to agree on what exactly we need to be doing there [20:48:22] tthere is 'current page's language' 'Current Wikipedia's language' and 'Current UI language' [20:48:26] needs to be a bit unified [21:06:55] YuviPanda: language selector was originally part of the nab design, but will probably follow right after [21:07:00] nav [21:07:15] philinje: when's nav coming up? [21:07:34] we are working on a mock-up that should be posted tomorrow [21:07:57] coool! [21:08:07] there has been a bunch of discussion and research so far, which you can see on the project page under Initial menu [21:08:53] YuviPanda: just wondering - it seems the iOS app has become slower [21:09:04] scrolling, for example, is more jerky [21:09:48] philinje: it is possible that it is slower on 4.3. My simulator is about to finish downloading, will check that [21:09:57] but on my device and on the 5.x simulator it was pretty smooth [21:10:44] interesting - a couple weeks ago it was very smooth on the same device [21:11:51] which is 4.3.3 [21:12:28] philinje: are you running the market release? [21:12:30] err [21:12:32] app store release? [21:12:43] i am now, but before not (of course) [21:12:46] philinje: does the other functionality work? [21:12:47] awjr: ping [21:12:48] awjr: http://pastebin.mozilla.org/1555259 [21:13:04] YuviPanda: yes [21:13:14] hmm [21:13:16] that's interesting [21:13:44] awjr: and http://pastebin.mozilla.org/1555262 [21:14:00] YuviPanda: but Nearby is a little unreliable [21:14:50] philinje: jumpy? [21:16:04] awjr: ping [21:16:17] preilly pong - looking at the code snippets now [21:16:19] main problem seems to be zoom level - opens very zoomed out sometimes - like the whole world [21:16:33] I'm using WiFi only [21:16:59] philinje: it does that if it has no idea where you are. [21:17:03] is this an iPhone or iPod Touch? [21:17:13] awjr: I'm heading off to grab some food [21:17:19] preilly ok [21:17:22] iPhone 4 [21:17:26] awjr: but, let me know what you think in a PM [21:19:49] YuviPanda: when I zoom in to a tpical level, no articles appear, though they were appearing slowly in the zoomed out view [21:20:21] philinje: can you tell me where this bug happens and what language setting your phone is on? [21:20:41] YuviPanda: here in SF and English [21:21:14] philinje: is it intermittent? [21:21:45] yes [21:21:52] [WikipediaMobile] brion pushed 2 new commits to master: http://git.io/D2u9WQ [21:21:52] [WikipediaMobile/master] Bug 35741 - iOS save/share menu should offer to open page in Safari - Brion Vibber [21:21:52] [WikipediaMobile/master] Merge pull request #201 from brion/open-in-safari - Brion Vibber [21:22:04] Project WikipediaMobile - Nightly builds build #278: SUCCESS in 6.4 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/278/ [21:22:05] brion: Bug 35741 - iOS save/share menu should offer to open page in Safari [21:22:12] YuviPanda: but i haven't tested it a lot, just a few times over the past few weeks [21:24:45] since I got the app from the app store, it seems like it doesn't work consistently [21:27:02] tfinc, what will be my next big mission? [21:27:22] philinje: :( [21:27:27] philinje: can you check which version you have? [21:27:50] 3.1.1 [22:01:22] tfinc: lost you [22:12:41] YuviPanda: Just saw a tweet. Was gonna ask you guys last night when I was struggling with google what you use for mapping [22:12:44] heh [22:12:47] devgeeks: :D [22:12:50] leaflet.js is amazing [22:13:21] I was looking at that, but since I actively don't want interaction I found a static map service for open Maps [22:14:36] Still tempted to set up my own [22:14:57] Still not ENTIRELY sure I understand the MapQuest licensing [22:14:59] hehe [22:16:07] devgeeks: mapquest licensing != mapquest open tiles license [22:16:17] plus it is super simple to set up your own. Plus they'll look much sexier [22:16:30] They'll look better if I do my own? [22:19:06] http://open.mapquestapi.com/staticmap [22:19:16] That's what I am using at the moment [22:20:21] devgeeks: you haven't seen http://maps.stamen.com/ [22:20:22] ? [22:20:24] they're all OSM [22:20:42] omg [22:20:47] I had seen those [22:20:56] Didn't know they could actually be USED [22:21:18] tfinc: so options for home page - 1. Call into MF 2. Make a new API call that gives you custom front page content 3. Use feature feeds, etc to build our own. [22:21:20] (3) is overkill [22:21:32] devgeeks: well, you couldn't use those tiles - I'm pretty sure stamen's servers won't like it [22:21:43] haha [22:21:49] devgeeks: http://mapbox.com/tilemill/ [22:22:06] I *knew* I shoulda asked you last night [22:22:11] devgeeks: IT IS AN AMAZING PIECE OF SOFTWARE [22:22:12] :P [22:22:25] devgeeks: plus setting up serving is easy too. Or you can pay mapbox to do that for you [22:24:29] devgeeks: also, you can haz offline maps :P [22:24:40] heh [22:24:52] How big would offline maps be though [22:25:01] devgeeks: depends on how big an area you want them for [22:25:05] and at what zoom levels [22:25:06] right [22:25:08] and with what detail [22:25:12] New review: Jdlrobson; "This doesn't seem to have solved the problem. Now the element with id content_wrapper doesn't contai..." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4361 [22:25:19] a friend made one for my city (Chennai). Was 3MB [22:57:05] awjr: http://pastebin.mozilla.org/1555384 [22:57:47] awjr: actually http://pastebin.mozilla.org/1555387 [22:58:36] awjr: minus the, "+ var_dump( $assembleUrl ); + echo '
';" [22:58:40] yeah [22:59:40] preilly will that work regardless of the location of the mobile token? [23:00:51] preilly: also, i thought if we're looking at zero, shouldDisplayMobileView() == true? [23:02:47] awjr: it is [23:02:54] awjr: paste is wrong [23:03:10] preilly roger [23:04:00] awjr: does the rest look fine to you? [23:04:26] preilly yeah i think so, assuming that will work regardless of location of mobile token [23:05:04] also perhaps the $subdomain var should be renamed to something like $subdomain_token or something since it looks like it maps to the token rather than the full subdomain [23:05:07] preilly ^ [23:06:27] awjr: $subdomainTokenReplacement [23:06:34] preilly awesome [23:07:26] New patchset: awjrichards; "Removed checkUseFormatCookie() method and simplified with cookie getter/setters that don't care about extMobileFrontend::useFormat" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4376 [23:07:53] awjr: okay, so — http://pastebin.mozilla.org/1555398 looks good? [23:07:57] preilly, MaxSem please take a look at ^ - it makes the sticky cookies a little less sticky. [23:08:26] preilly aye just update the var name [23:08:54] awjr: it is in that link [23:09:18] preilly no 1. if ( wfRunHooks( 'GetMobileUrl', array( &$subdomain ) ) ) { [23:09:29] mm [23:09:34] awjr: in this one [23:09:35] http://pastebin.mozilla.org/1555400 [23:10:55] preilly that looks better. but you know i just realized… won't ( '/%h[0-9]\.{0,1}/ only match single letter mobile tokens? [23:11:36] i know i wrote that regex but it looks .. incorrect to me now [23:12:31] awjr: do you expect a host name with more than 9 parts? [23:12:49] preilly jeez i hope not [23:13:10] New patchset: preilly; "Add a GetMobileUrl hook in order to override the mobile host portion of the $wgMobileUrlTemplate template variable (e.g., replace .m. with .zero., etc.)." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4377 [23:13:17] awjr, MaxSem: please review ^^ [23:14:07] preilly perhaps we should functionalize $mobileToken = preg_replace( "/%h[0-9]\.{0,1}/", "", $mobileUrlHostTemplate ); [23:14:47] awjr: make a getMobileToken method? [23:14:53] yeah [23:14:59] awjr: that's fine [23:15:03] awjr: want me to do it? [23:15:13] preilly sure that'd be swell [23:17:57] preilly, !empty( $subdomain ) <-- variable not defined [23:18:12] i hate that about PHP. [23:18:41] awjr: http://pastebin.mozilla.org/1555419 [23:19:05] preilly: looks good [23:19:07] preilly, wfProfileOut() after return [23:19:17] oh except for that... [23:19:29] preilly should i be wrapping my methods in those profile statements/ [23:19:43] New patchset: preilly; "Add a GetMobileUrl hook in order to override the mobile host portion of the $wgMobileUrlTemplate template variable (e.g., replace .m. with .zero., etc.)." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4377 [23:20:08] New patchset: preilly; "Add a GetMobileUrl hook in order to override the mobile host portion of the $wgMobileUrlTemplate template variable (e.g., replace .m. with .zero., etc.)." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4377 [23:20:44] preilly: trailing whitespace [23:20:48] New patchset: preilly; "Add a GetMobileUrl hook in order to override the mobile host portion of the $wgMobileUrlTemplate template variable (e.g., replace .m. with .zero., etc.)." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4377 [23:21:30] New patchset: preilly; "Add a GetMobileUrl hook in order to override the mobile host portion of the $wgMobileUrlTemplate template variable (e.g., replace .m. with .zero., etc.)." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4377 [23:22:08] awjr, MaxSem: how does ^^ look? [23:23:03] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/4377 [23:23:06] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4377 [23:23:11] preilly ^ [23:25:03] awjr: thanks! [23:25:20] preilly can you take a look at https://gerrit.wikimedia.org/r/#change,4376 ? [23:25:25] MaxSem too ^ [23:28:56] New patchset: preilly; "Add documentation to getMobileToken method" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4378 [23:29:40] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/4378 [23:29:42] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4378 [23:30:15] awjr: *looking* [23:30:35] awjr: toggle_view-desktop, toggle_view-mobile seems weird to me [23:30:50] preilly just the syntax? [23:31:08] awjr: yeah syntax [23:31:14] awjr: mixing of _ and - [23:31:57] preilly yeah it's a convention i use for myself where "_" is a separator use inside of a particular thing and "-" separates particular things [23:32:05] but im definitely not married to it for this [23:32:37] i guess the other mobileactions just always use "_" [23:32:47] preilly would you prefer "toggle_view_desktop"? [23:34:00] New patchset: awjrichards; "Removed checkUseFormatCookie() method and simplified with cookie getter/setters that don't care about extMobileFrontend::useFormat" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4376 [23:34:13] awjr: I think so [23:34:32] preilly see new patchset [23:35:01] mmm, I was writing inline comments to the previous patchset [23:35:04] awjr: yes sir [23:35:18] doh [23:35:26] MaxSem go ahead, i'll check 'em out in the first patchset [23:35:28] when you're done [23:36:35] awjr: Primarcy? [23:36:37] New review: MaxSem; "(no comment)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4376 [23:36:41] MaxSem: any idea when aciton=mobileview's noheadings option hit production? [23:36:57] awjr: did you mean, "primacy"? [23:37:05] preilly: yes [23:37:06] YuviPanda, it didn't? [23:37:21] awjr: line 1489 [23:37:35] thanks [23:37:38] MaxSem: not when i see the api sandbox, no [23:38:17] YuviPanda, in two weeks, apparently - there will be no deployments next monday [23:38:42] hmm, I thought there was one going to be on next thursday? [23:39:03] ask teh deployers;) [23:39:10] MaxSem: per your comment on line 538 - is that new from what it used to be? [23:39:28] YuviPanda: yeah we found out yesterday that there's a deployment freeze next week [23:39:52] awjr: so when's the next deployment going to be? [23:39:54] YuviPanda: checkout email with subject "Notes and impact analysis on mobile team from 'deploying from git' meeting" [23:40:04] YuviPanda: the following Tuesday, I think [23:40:20] ah, I read that and for some reason registerd as 'following Thursday' [23:40:24] now this makes sense [23:40:57] awjr, previously it was added only when current URL contained an explicit &useformat [23:41:07] MaxSem hmm [23:41:50] is there a way to diff between patchsets? [23:42:00] MaxSem: ah yeah i see now it will get overridden by the cookie, which is probably dumb [23:43:48] MaxSem per the htmlspecialchars comment, isnt' that not necessary with the hardcoded value? [23:43:56] awjr: shouldSetUseFormatCookie() looks like it might still be too aggressive.. is it a problem if desktop users get a mobile view cookie when they follow a mobile url? [23:44:58] binasher the only way you'd possibly get a cookie set is if mobileaction=toggle_view_{desktop|mobile} [23:45:06] in the query string [23:46:00] i suppose shouldSetUseFormatCookie() is a bit of a misnomer because there's a check that happens earlier to determine whether or not the view should even be toggled [23:46:28] binasher ^ [23:47:05] what happens if i click on a link that includes mobileaction=toggle_view_mobile and then try browsing en.wikipedia.org later? [23:47:07] awjr, even if $wgRequest contains only percent-encoded URLs that can't result in XSS, you still have to replace &s with &s for validity [23:47:29] yeah, the function name was a bit confusing [23:48:16] this seems a lot more complicated than the old pre-sticky way of choosing views [23:48:48] awjr, MaxSem: please review https://gerrit.wikimedia.org/r/#change,4379 [23:48:49] binasher you're right, then you would still see mobile view because we changed the cookies to use * subdomain [23:48:56] binasher indeed but before the view toggling wasn't sticky [23:50:33] preilly, 1) shouldn't it be public static function getMobileUrl( >>&<<$subdomainTokenReplacement )? [23:50:59] 2) you should probably make the handler return false to abort further processing [23:52:18] MaxSem: 1) no [23:52:45] New patchset: awjrichards; "Removed checkUseFormatCookie() method and simplified with cookie getter/setters that don't care about extMobileFrontend::useFormat" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/4376 [23:52:47] MaxSem: 2) maybe [23:53:00] i think the first point might confuse some users, and people using the mobile view thru the non-mobile vip isn't ideal from a caching viewpoint. i'm not sure if stickiness is needed in that direction for us, since anyone can go to m.wiki in any browser and use the mobile site [23:53:45] MaxSem: so, you want: http://pastebin.mozilla.org/1555454 [23:54:14] preilly, vice versa [23:54:16] binasher: preilly and i came up with the * subdomain cookies for a good reason last week but i'm blanking on what it was [23:54:46] to disallow other extensions to modify what you've already changed to your needs [23:55:02] binasher what makes this surprisingly challenging is finding the right path that will let this work sanely on WMF infrastructure but be usable out-of-the-box [23:55:06] MaxSem: so, http://pastebin.mozilla.org/1555456 [23:55:33] since presumably most people who would otherwise use this extension won't have as complex of a setup as we do. [23:56:28] preilly, yes, but else is not needed [23:56:39] MaxSem: that is wrong [23:57:40] MaxSem: if I return false it doesn't work because in MF it's if ( wfRunHooks( 'GetMobileUrl', array( &$subdomainTokenReplacement ) ) ) {... [23:57:42] i wonder if it would be worthwhile triggering the m redirect on the squids if the mobile cookie is set [23:58:39] binasher how difficult would that be to implement? [23:58:48] not difficult [23:59:07] mmm, right preilly [23:59:23] so if mf_useformat=mobile, but the user is trying to access non- .m, they would be redirected to .m [23:59:30] and we already have the inverse rule in place [23:59:56] MaxSem: so, can you just approve it? [23:59:57] binasher i think that would resolve the issue