[00:19:13] New patchset: awjrichards; "Fix titles with formatting" [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5294 [00:19:14] New patchset: awjrichards; "Change name of file to jquery.min.js remove the version reference" [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5295 [00:19:15] New patchset: awjrichards; "Pass RequestContext" [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5296 [00:19:15] New patchset: awjrichards; "Restore message lost in merges and refactorings" [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5297 [00:19:16] New patchset: awjrichards; "Localisation updates from http://translatewiki.net." [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5298 [00:19:57] preilly: when merging from master -> remote branch, is it better to squash the merged commits into one or just push all the merged commits as individual commits? [00:52:57] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (contact-us-redesign); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5298 [00:53:02] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (contact-us-redesign); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5297 [00:53:06] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (contact-us-redesign); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5296 [00:53:13] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (contact-us-redesign); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5295 [00:53:17] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (contact-us-redesign); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5294 [00:53:20] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5298 [00:53:21] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5297 [00:53:22] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5296 [00:53:23] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5295 [00:53:25] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5294 [10:07:23] * YuviPanda looks around [10:27:45] YuviPanda: are we ready to merge yet? [10:28:14] good spot on the ios typo [10:31:13] jdlrobson: heya [10:31:18] jdlrobson: I think so. [10:31:32] jdlrobson: want to look over the commits once more? [10:31:35] we're close to a hundred :) [10:31:59] https://github.com/yuvipanda/WikipediaMobile/tree/api-move yeh? [10:32:15] jdlrobson: or https://github.com/wikimedia/WikipediaMobile/pull/208 [10:32:16] sae thing [10:32:17] yeah [10:32:21] did you say ics had choppy fixed positioning still? [10:32:33] jdlrobson: it wasn't really 'choppy' [10:32:48] jdlrobson: essentially, *while* scrolling, the search bar moves down a pixel or so [10:32:52] and when you stop it goes back [10:33:04] mm i wonder why that is [10:33:18] it feels like the 'kinetic scrolling' algorithm is being applied to it as well [10:33:28] if I scroll faster, i see more movement [10:33:34] (more than a few pixels) [10:34:53] only suggestion so far is you might want to git rebase interactive and fixup https://github.com/yuvipanda/WikipediaMobile/commit/ceb8177d193811b65e9e408581fbc2527c772b0a [10:35:42] I'm guessing that's okay, not sure we should be 'hiding' mistakes behind squashes [10:37:00] same with https://github.com/yuvipanda/WikipediaMobile/commit/c7658c65824eea7d630dbfa511b63e8c2328d743 [10:37:23] well these mistakes are just sloppy and i don't think add anything to the commit log [10:37:32] apart from make it harder to follow [10:37:39] imo [10:37:57] but that aside i think we're good to merge [10:38:07] this is the 'history is sacred' vs 'that is pointless' bikeshed (of sorts) [10:38:31] jdlrobson: \o/ [10:38:37] jdlrobson: accept the pull req and merge it in? [10:38:58] wait let me do a rebase [10:39:07] k [10:39:13] ping me when it is time [10:39:58] jdlrobson: brion has made a few changes to master that're preventing a clean rebase :( [10:40:10] investigating [10:40:11] \o/ [10:40:18] gonna have lots of fun with those existing pull requests :D [10:43:12] heh [10:43:15] jdlrobson: rebased and pushed [10:46:34] sorry 1s just fixing a MFE bug [10:46:40] ok :) [10:48:32] k looking into it now [10:48:39] the bug effected the app as well [10:48:39] https://twitter.com/#!/Shaddez/status/192455772901347328 [10:49:21] my h1b got approved yesterday btw [10:49:23] \o/ [10:49:25] \o/ [10:49:50] jdlrobson: congratulations :) [10:50:07] still no physical visa... have to go to embassy for that. i'm not sure where that puts me in terms of eta but guess things are going to go really quickly now [10:50:25] right ICS testing first [10:50:52] that's an MFE bug? [10:51:02] mobile frontend [10:51:03] sorry [10:51:17] yeah, but was wondering how that'd be a MobileFrontend bug [10:51:24] unless you meant the 'two line looks weird' bit [10:52:27] well i think it might be due to common.css [10:52:35] it's present in the mobile site too [10:52:38] so when i click settings i lose my scroll position in an article - is that normal [10:52:57] right. no it isn't normal [10:53:05] that's because we're doing window.scrollTo [10:53:10] instead of $(content).scrollTop [10:53:13] we should start writing unit tests for these sorts of things [10:53:48] yeah [10:53:57] should be easier to do so now, than before (which depended too much on DOM) [10:54:34] also save page seems to be broken.. [10:54:38] yes it is [10:54:42] i've not started work on it [10:54:48] k as long as thats known [10:54:56] actually weirdly it saved [10:55:00] but didn't tell me it saved [10:55:37] so, half the code is there. but it doesn't take into account dynamic DOM loading, and so images in sections not expanded won't be saved [10:55:40] also did we decide what to do about MediaWiki:Common.css ? [10:55:44] no [10:56:07] jdlrobson: it's different for every language wiki, so I don't think we can get away with including a static version of it [10:56:18] doh [10:56:27] ok i'm gonna merge [10:56:28] so, yay to app breakages because of CSS changes from other people? [10:56:54] i'm wondering how to deal with MediaWiki:Common.css. They might easily conflict with app styles... [10:57:11] we might have to namespace classes/ids used in the app then... [10:57:12] just a thought... when i merge [10:57:16] irc is going to go crazy [10:57:21] hehe [10:57:43] just doing a bit more ios testing [10:57:45] (4.2) [10:57:52] ok [11:00:42] mm ios 4.2 isn't working at all [11:00:53] let me reboot xcode [11:01:23] seems fine on my sim? [11:01:24] ok [11:06:43] ahh i was running from wrong respository :D [11:06:50] so yeh ios doesn't have the scroll problem [11:06:51] :D [11:06:59] (when i load settings it keeps my place in the article) [11:07:09] jdlrobson: so, as I said - that is because we're using window.scroll [11:07:19] k as long as we know how to fix it :) [11:07:26] its looking awesome [11:07:30] ready for irc to go crazy? [11:07:35] indeed sir! [11:08:09] ! [11:08:30] mm [11:08:32] [WikipediaMobile] jdlrobson pushed 93 new commits to master: http://git.io/eORvZQ [11:08:32] [WikipediaMobile/master] Initial cut of pages being rendered via the API - YuviPanda [11:08:32] [WikipediaMobile/master] Get rid of the tree-conversion code - YuviPanda [11:08:32] [WikipediaMobile/master] Added MFE as a Submodule - YuviPanda [11:08:33] where's the madness? [11:08:35] ah there [11:08:41] heh [11:08:42] oh wait [11:08:45] not yet [11:08:49] it's jenkins that'll go mad [11:08:49] Project WikipediaMobile - Nightly builds build #348: SUCCESS in 9.2 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/348/ [11:08:50] * yuvipanda: Initial cut of pages being rendered via the API [11:08:50] * yuvipanda: Get rid of the tree-conversion code [11:08:51] * yuvipanda: Added MFE as a Submodule [11:08:51] * yuvipanda: Top level toggling works [11:08:52] * yuvipanda: Temperoarily hide extra header [11:08:52] * yuvipanda: Added Patch file needed to make MF work with app [11:08:53] * yuvipanda: Move back to local version of MF's application.js [11:08:53] * yuvipanda: Make only h2 collapsible [11:08:54] * yuvipanda: Drop the .m prefix [11:08:54] * yuvipanda: Don't depend on history for current URL [11:08:55] * yuvipanda: Spacing fixes [11:08:55] * yuvipanda: Refactor to make page's source more generic [11:08:56] * yuvipanda: Refactor page to be language independent [11:08:56] * yuvipanda: Redid Saved Pages to work with Page objects. Broken in iOS [11:08:57] * yuvipanda: Fixes to make RTL languages display as RTL [11:08:57] * yuvipanda: Make Language Links work via the API [11:08:58] * yuvipanda: Fixes to make loading read-in not painful [11:08:59] muahhaha [11:09:09] * yuvipanda: Added warning about code in comments [11:09:09] * yuvipanda: Fixed bash misunderstanding [11:09:10] * yuvipanda: Open attribution links under maps in external browser [11:09:10] * yuvipanda: Make sure that 'Show more results' is visible [11:09:11] * yuvipanda: ReadItLater issues fixed by changing modal presentation targets and eliminating previous ex nihilo method. Associated issues also fixed. Completing registration needs testing with a working RIL API key. [11:09:11] * yuvipanda: Added Rolken to AUTHORS [11:09:12] * yuvipanda: Support all orientations for RIL on iPad [11:09:12] * yuvipanda: Fix scrolling after orient change on iOS 4.x [11:09:13] * yuvipanda: Increasing keypress timeout to 300. Works better on iPhone3G [11:09:13] * yuvipanda: Fix for saved pages never persisting [11:09:14] * yuvipanda: Does not crash on 2.x -> 3.x app upgrade [11:09:14] * yuvipanda: Use noheadings in API call [11:09:15] * yuvipanda: Add elements to the DOM only on section expansion [11:09:15] * yuvipanda: Updated to latest commit on MF [11:09:16] * yuvipanda: More 'Proper' Section expansion implementation [11:09:16] * yuvipanda: Fixes to work with iScroll [11:09:17] * yuvipanda: Add hide buttons too to sections [11:09:18] * yuvipanda: No more zepto. [11:09:21] best log ever [11:09:24] more tings [11:09:29] * yuvipanda: Override scrollTo on android [11:09:29] * yuvipanda: Working badly hacked version of scrollTo for android [11:09:30] * yuvipanda: Removed all traces of doScrollHack [11:09:30] * yuvipanda: fixes for Ice Cream Sandwich android 4.* [11:09:31] * yuvipanda: remove unnecessary show/hide [11:09:31] * yuvipanda: Add detector classes to html element [11:09:32] * yuvipanda: Remove stray alert [11:09:32] * yuvipanda: Use jQuery to set classes instead of plain js [11:09:33] * yuvipanda: device specific fixes for android 2 [11:09:33] * yuvipanda: append missing semi colons [11:09:34] * yuvipanda: fixed ICS bug with content flashes [11:09:34] * yuvipanda: prevent horizontal scrolling in android ics [11:09:35] * yuvipanda: separate ios specific css interfering with android ics [11:09:35] * yuvipanda: only apply overflow-y to ios devices [11:09:36] * yuvipanda: Fix iOS 5.x scrolling [11:09:41] huzzah [11:09:44] all done [11:09:49] \o/ [11:10:03] now to go fix another scroll bug :) [11:10:07] so i'm keen to get https://github.com/wikimedia/WikipediaMobile/pull/210 merged - could you have a quick poke at it and my comments before doing that? [11:10:18] i expect lots of rebasing will now be required [11:10:29] jdlrobson: yes, and i'm not sure if I'd want to do it for 1.2 [11:10:40] why not? [11:10:49] since we have too much already? :) [11:10:56] yes :) [11:11:10] very sensible [11:11:16] i'm a sucker for new shiny things [11:11:18] i'd want this in 1.3, along with moving off geonames into the API [11:11:24] don't you mean shiny new bugs? [11:11:25] :D [11:11:43] nothing shiny about those! :PP [11:11:43] so I'm wondering how to deal with pull requests that we want to do 'later' [11:21:14] merge into a branch e.g. 1.3 ? [11:22:51] jdlrobson: but then that branch will have to be kept in sync with master [11:23:16] well lets use labels then.. [11:23:56] makes sense [11:23:58] done [11:24:11] annoyingly those don't show up in pull requests [11:24:13] just issues [11:24:28] we should respond to https://github.com/wikimedia/WikipediaMobile/issues/202 [11:25:25] was just looking at that [11:26:10] jdlrobson: also https://github.com/wikimedia/WikipediaMobile/pull/198 [11:26:10] ? [11:27:21] yeh what's your thoughts on that [11:27:36] well, we 'fixed' that on iOS [11:27:41] we did? [11:27:43] but not a fan of the approach taken here [11:27:51] jdlrobson: yeah, iOS no longer shows the navbar when on overlays [11:27:53] it just seems weird [11:28:00] yeh hiding the nav bar would make more sense [11:28:03] but you can't do that in android [11:28:10] yes [11:30:08] jdlrobson: the new code has setCurrentPage and setErrorPage, and I think it can be handled there. [11:30:19] haven't thought about that yet, currently figuring out how to do saved pages properly [11:30:27] k [11:30:30] jdlrobson: closing that pull request seems to be the way to go for me. [11:30:40] in terms of saved pages.. [11:30:50] why not just do one big additional ajax request for all of it [11:30:57] go for it then YuviPanda [11:31:07] jdlrobson: ahem, that's what we're doing now :) [11:31:14] jdlrobson: problem is images. [11:31:17] and where to save them [11:31:21] ah i see [11:31:25] well can base64 them... [11:31:28] so [11:31:31] that's what we're doing on iOS :) [11:31:34] :) [11:31:44] android doesn't support that on 2.x [11:31:45] oh wait [11:31:50] maybe we can get the API to do this for us [11:31:55] hmmm [11:31:56] :D [11:32:16] i'm not sure though [11:32:24] jdlrobson: thoughts on that idea? [11:32:45] jdlrobson: i'll note that the usual way of converting to base64 from canvas is what we're using on iOS. It seems to cause some perf problems [11:32:46] 2.x doesn't support base64? [11:32:51] jdlrobson: it does. [11:32:55] mm [11:32:58] it doesn't support converting images to base64 via canvas [11:33:13] i guess the question is are images essential offline [11:34:05] why not as a first pass forget about them and get offline working that aside [11:34:09] I'd say they are - it'd look broken otherwise? [11:34:21] plus we've been supporting that already, and you know how much people love having features taken away [11:34:26] ok [11:34:27] mm [11:34:56] though *I* am a fan of killing saved pages though :) [11:35:36] jdlrobson: we're including clicktracking to see how many people use that [11:35:58] well there seem to be 2 options - api base64s them on your behalf or we do it client side using canvas [11:36:23] alternatively we could look into the file api but i'm not sure if that works with phonegap [11:36:28] jdlrobson: client side using canvas is not doable on android 2.x [11:36:31] hmm [11:36:48] jdlrobson: so, on iOS we use the file API + canvas conversion [11:36:53] jdlrobson: so we save a single file. [11:37:05] so wait a minute android 2.x never had images saved offline? [11:37:15] jdlrobson: android has a java plugin that does the saving [11:37:26] while iOS is all js, uses canvas + file API [11:37:45] it's messy super-platform specific code [11:37:48] yeh [11:37:50] i hate that stuff [11:38:01] sure we can't take it away from 2.x ? [11:38:26] i'd want to take it away from all of them :) But not for this release - tfinc loves saved pages. [11:38:27] if they have an internet connection they can still get the images [11:38:51] i'm also worried about localStorage limits with images [11:39:14] jdlrobson: we're not using localStorage [11:39:28] we're using localStorage only for boolean/simple-string settings [11:39:35] jdlrobson: all our saved pages are done via files [11:39:40] ok.. sorry i've not really explored the offline code much as you can see [11:39:49] will have to explore [11:40:05] jdlrobson: i think having the API itself give us bas64 would be awesome [11:40:07] * YuviPanda pokes MaxSem  [11:47:48] jdlrobson: https://code.google.com/p/android/issues/detail?id=7901 [11:49:58] eek "Still no support for jpeg (tested on icecream sandwich)" [11:50:02] so yeh canvas is out [11:50:12] jdlrobson: and no support for jpeg on iOS either. [11:50:19] jdlrobson: we're converting to PNG before saving on iOS right now [11:50:25] so eek [11:50:30] eek indeed [11:52:27] jdlrobson: I think getting base64 from the API would be awesome [11:52:47] sounds like it would take a lot of processing power [11:54:01] i'm guessing MaxSem would know [11:54:16] jdlrobson: also, the APi does not handle things in the File: domain currently [11:54:28] jdlrobson: this means that we need to now build our own Image handling code [11:54:33] i.e. what to do when someone taps an image [11:56:10] mm [11:56:18] filed as https://bugzilla.wikimedia.org/show_bug.cgi?id=36036 [11:56:31] I could just punt them to the browser for this release [11:56:45] but seems icky [11:59:46] New patchset: Jdlrobson; "simplify border rule into single line" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5312 [11:59:46] New patchset: Jdlrobson; "prevent long search inputs forcing a new line" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5313 [12:05:59] yuvi - can you not use an images intent? [12:08:55] jdlrobson: on android, yes [12:09:24] oooh, actually - you can't do that without saving it on file :) [12:13:37] Project WikipediaMobile - Nightly builds build #349: SUCCESS in 7.5 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/349/ [12:28:13] New patchset: Jdlrobson; "style errors better" [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5316 [12:29:30] New review: Jdlrobson; "A few outstanding problems" [mediawiki/extensions/MobileFrontend] (contact-us-redesign) C: 0; - https://gerrit.wikimedia.org/r/5316 [12:40:18] who poked meeeee? [12:40:34] It was I, master! [12:41:16] MaxSem: thoughts on API that gives us base64 representations of files? Or better yet, an option to mobileview where it converts all images into data URIs so that it doesn't need additional requests [12:41:16] ? [12:41:46] mmm [12:42:03] are single requests so slow? [12:45:26] Project WiktionaryMobile - Nightly builds build #91: SUCCESS in 7.2 sec: https://integration.mediawiki.org/ci/job/WiktionaryMobile%20-%20Nightly%20builds/91/ [12:45:37] MaxSem: no [12:45:40] this is for our 'saved pages' feature [12:45:54] so we're just wondering how feasible this will be on the server side [12:46:12] at least, not easily [12:47:19] it would require either hooking up to parser or magic remapper of URLs back to FS paths or fetching of images via HTTP [12:47:36] that sounds ugly. [12:48:44] parser doesn't provide us with sizes of all thumbnails involved [12:48:53] just which images are used on page [12:49:30] MaxSem: or adding a data property to action=query & prop=imageinfo [12:49:37] that gives us the data as well? [12:49:37] so the answer is yes it's possible but not as a quick hack like action=mobileview which was created in one evening [12:49:41] hmm [12:50:01] MaxSem: i guess this would be just as messy (fetch over HTTP, base64, return) [12:50:41] there's such thing as remote repos e.g. Commons [12:51:00] for them everything's e ven more fun [12:51:21] yes, that 'idea' went downhill pretty fast :) [12:51:54] well, the idea actually makes sense but requires design and planning [12:52:26] New patchset: Jdlrobson; "use one button for toggling" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5319 [12:52:27] yeah, and I'm not sure how much of a benefit it'll give us [12:52:29] ^ YuviPanda [12:52:36] \o/ [12:55:01] YuviPanda: any thoughts on references? [12:55:21] jdlrobson: so, if we load the references section into the DOM by default [12:55:32] I think the reference reveal code from MF should work unmodified (mostly) [12:55:54] i'm going to update the references code to take a parent element in the init [12:56:15] it currently does $( 'sup a' ).unbind('click').click( clickReference ) [12:56:23] should probably do $( 'sup a', parent ).unbind('click').click( clickReference ) [12:56:43] otherwise lots of needless rebindings [12:56:47] hmm, that should help with loading it on every section reveal [12:57:49] ill make the change and send to you for review [12:58:14] jdlrobson: okay! [13:14:32] New patchset: Jdlrobson; "allow initialisation function to target a specific element" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5320 [13:17:43] brb [14:06:00] hi guys [14:06:33] is the mobile app setting an explicit user agent as of 1.0.3 for both the android and the ios version? [14:07:58] diederik: yes [14:08:26] thanks! [14:08:41] when was 1.0.3 released? [14:15:25] diederik: quite a while ago - almost two months. I could check the exact date if you want me to... [14:15:31] diederik: iOS was much more recent though [14:15:52] i am most interested in the release date of iOS [14:16:18] * YuviPanda hunts for exact date [14:16:26] thanks [14:16:32] roughly a week and a half ago, looking for exact date [14:16:52] diederik: ~Apr 5 [14:17:00] about two weeks ago [14:18:05] thanks! [14:18:22] diederik: :) [14:18:33] can you give me an example of such a user agent string? [14:19:15] diederik: essentially, we just prepend WikipediaMobile/ to the actual device's User agent [14:19:30] okay [14:19:35] so the user agent differs for each device/OS combo, but will be prefixed with WikipediaMobile/ [14:19:43] cool [14:19:50] again thanks :D [14:24:03] MaxSem: is your samsung jet handy? https://bugzilla.wikimedia.org/show_bug.cgi?id=30829 [14:24:56] MaxSem has a Jet? [14:25:32] jdlrobson, I lookd at it immediately after it was pushed [14:26:00] but what was the result [14:26:01] ? [14:26:32] with images disabled can you see the search button? [14:33:42] hmm, mobile-geo looks dead [14:34:18] memcached went offline? [14:39:31] hahaha $wgMemCachedTimeout = 500000; [14:41:45] dammit, bot funny [14:41:49] *not [14:42:00] it doesn't work anyway [14:47:12] jdlrobson, I can't check it until the Labs instance is fixed [14:47:17] k no worries [14:47:48] MaxSem i can put it on http://mobile-geo.wmflabs.org/w/index.php?title=Main_Page&mobileaction=toggle_view_mobile if that helps? [14:47:56] oh, after a reboot and 30 seconds' load it finally "works" [14:48:26] trying [14:49:33] jdlrobson, nope - still empty space clicking on which works [14:50:25] really? [14:50:30] why putting a simple button here is heretically unacceptable [14:50:55] well i noticed some browsers completely destroy any element that involves an element [14:51:03] there is a button there [14:51:14] element that involves an element? [14:51:26] lol sorry any element that involves an image [14:51:46] do any images show alt text? [14:51:50] no [14:51:57] waitasec... [14:51:57] yeh so some browsers if they encounter it just throws it away rather than showing the alt text [14:52:02] which is bad and the browsers fault imo [14:52:26] my android phone does that for example [14:59:10] meh, quite blurry: http://tinypic.com/view.php?pic=ddj8xy&s=5 [15:15:07] jdlrobson, is https://bugzilla.wikimedia.org/show_bug.cgi?id=34878 fixed? [15:15:16] * jdlrobson looks [15:15:25] well i wondered that myself [15:15:35] it's certainly not as important any more [15:15:43] the references code is only in the beta [15:16:01] and it doesn't actually remove the references section (although easily could...) [15:16:21] actually saying that... [15:16:26] let me check some thing [15:24:09] updated MaxSem [15:24:21] cool [15:27:32] YuviPanda: is this still relevant? https://bugzilla.wikimedia.org/show_bug.cgi?id=29862 [15:27:51] nope [15:31:14] i'll close then [15:31:32] should i close as WONTFIX or INVALID ... [15:31:40] jdlrobson: anything referencing the old one are all closed as 'INVALID' i guess [15:31:45] agreed [15:32:05] i'm moving MFE bugs from https://bugzilla.wikimedia.org/buglist.cgi?list_id=108980&field0-0-0=bug_severity&resolution=---&resolution=LATER&resolution=DUPLICATE&query_format=advanced&bug_status=UNCONFIRMED&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&type0-0-0=notequals&value0-0-0=enhancement&product=Wikimedia%20Mobile [15:37:19] YuviPanda: is network.js still required? [15:37:29] jdlrobson: i remember rming it [15:37:33] mm [15:37:47] that's what i thought [15:37:51] but it's still in my repo [15:38:06] hmm [15:38:20] jdlrobson: not in mine? [15:38:27] are you doing it from the wrong repo again? :P [15:38:54] yes [15:38:57] :) [15:39:38] how do i force german language? [15:39:52] i want to test something [15:40:23] uselang=de [15:40:39] oh, app [15:41:01] jdlrobson: the app? Change your app settings in the os settings? [15:41:10] its ok i did it via settings [15:41:15] i wanted to just change the js [15:41:19] but not important any more [15:42:49] YuviPanda: Cannot read property 'sections' of undefined at file:///android_asset/www/js/page.js:17 - interesting.. [15:43:06] jdlrobson: i assume you opened a 'File' page? [15:43:23] or a page that didn't exist. [15:43:46] should handle the error case in the latter, and the former - we need to figure out what to do (action=mobileview doesn't do file:) [15:44:08] i rebuilt and it's gone away [15:44:13] ... [15:44:34] i was looking into https://bugzilla.wikimedia.org/show_bug.cgi?id=35938 and trying to load a page beginning with a number [15:44:38] but now i do it and it works fine :/ [15:48:11] jdlrobson: close as 'works for me'? [15:48:32] i always leave a bit of time in case they want to comment [15:49:12] On Meta, the document needs to be in English when you want to translate it [15:49:30] jdlrobson: there is always 'reopen' :) [15:50:14] true [15:50:24] done [15:50:50] mm i've just given myself loads more bugs [15:50:52] \o/ [15:51:01] annoyingly none are easily fixed [15:53:08] jdlrobson: :D [15:54:09] jdlrobson: thoughts on https://bugzilla.wikimedia.org/show_bug.cgi?id=33648 [15:55:21] yeh i have no idea on this one - basically we need ResourceLoader [15:55:36] as Webfonts has a load of css and javascript [15:56:19] preilly would be the best one to advise on this i guess [16:02:54] New patchset: Jdlrobson; "force desktop site when oldid and diff parameters present" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5334 [16:06:39] jdlrobson, but why oldid? [16:07:16] good point [16:08:08] pushed [16:08:13] New patchset: Jdlrobson; "force desktop site when diff parameter present" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5334 [16:16:09] gotta go [16:19:27] New review: MaxSem; "(no comment)" [mediawiki/extensions/MobileFrontend] (master) C: 1; - https://gerrit.wikimedia.org/r/5334 [16:29:02] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5334 [16:29:05] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5334 [16:37:06] greetings all [16:37:16] YuviPanda: furious activity on bugzilla over the last 24 hours [16:37:23] YuviPanda: you guys are keeping my inbox busy [16:37:27] :) [16:39:31] tfinc: we also merged in the api-move branch. quite a lot of 'tings' from the mw-jenkinsbot [16:39:39] YuviPanda: nice [16:39:53] if i find time today i'll build a new apk/ipa to test [16:40:12] tfinc: so we also managed to fix our android nightlies :) [16:40:41] these inline style bugs are killing us [16:41:49] they are. [16:42:03] tfinc: the solution is to deploy semantic mediawiki to enwiki [16:42:04] all the wikis [16:42:08] i'm sure Reedy would approve [16:42:19] YuviPanda: doesn't that both fix and create all problems? [16:42:26] :D [16:42:33] jons h1b got approved … woot! [16:42:48] tfinc: also, if we're including MediaWiki:Common.css, we're now making it super easy for anyone with edit access to that to break our app entirely [16:43:32] tfinc: yes, he mentioned it earlier :) [16:43:37] pretty awesome :) [16:43:51] there he is [16:43:53] jdlrobson: congrats [16:43:58] YuviPanda: not *anyone* literally, only admins, and most are usually conscious while doing css changes [16:44:10] srikanthlogic: 'anyone who has access' [16:44:26] srikanthlogic: and no, I don't want to burden them as well with having to check with what classes and ids the app uses before making a change :) [16:44:47] srikanthlogic: solution is namespacing everything on our end, no change in community's process :) [16:47:19] jdlrobson: i just sent you a mail but i'll mention it here as well. inline styles are *killing* us [16:47:38] yep i've noticed - you saw my mail to wikitech ? [16:47:53] ohh [16:48:03] I've only gone through bugs in todays mail [16:50:02] i see it now [16:50:04] cool [16:50:15] feel free to add anything if I've missed some key points out [16:50:16] also .. jdlrobson congrats on the H1B [16:50:23] yey! :) [16:50:33] i'm not too clear what the next steps are... is it quite fast from here? [16:50:36] that was far less panful then either of us could have anticipated [16:50:52] i need to think about flights and terminating my flat in the UK and a whole host of other things [16:50:58] seems like you need to get your passport visa next [16:51:06] let chat over pm about that [16:58:03] tfinc: i moved a host of bugs over to the right projects - guess you've noticed that though [16:58:09] there are still quite a few left [16:58:45] jdlrobson: yes. i've been meaning to do an MF triage like that for a while [16:58:48] thanks for doing it [16:58:53] there's still a few left [16:58:59] jdlrobson: what are your thoughts on bug layout ? [16:59:02] we need to clean that up [16:59:36] tfinc: it's been working quite well for the app [16:59:37] well there should be Mobile App, Mobile Site, Mobile Both and Mobile Unknown imo [17:00:05] jdlrobson: sure but we have to module it in the bugzilla hierarchy [17:00:13] products->components->versions-> etc [17:00:14] I'd worry your average person wouldn't know that extensions/MobileFrontend is where to find mobile bugs [17:00:33] jdlrobson: correct … thats why the generic Wikimedia Mobile product exists [17:00:39] all should be under Wikimedia Mobile then [17:01:09] e.g. Wikipedia App should be a component and m.wikipedia.org should be a component [17:02:53] jdlrobson: if we do that then versions are meaningless [17:03:09] and the ease of triage by components as we have them in the wikipedia app goes away [17:03:17] YuviPanda: sounds like the components have been working well for you [17:03:25] YuviPanda: what does it make easier .. what does it make harder [17:03:48] awjr, how's your HTMLForm rewrite doing? [17:04:04] tfinc: components as such have mostly been useless tbh [17:04:06] I've found a great application for it: https://bugzilla.wikimedia.org/show_bug.cgi?id=34751 [17:04:12] tfinc: having the product as such helps. [17:04:18] MaxSem: pretty good - i spent yesterday afternoon playing around with figuring out how to edit articles on remote wikis [17:04:25] tfinc: because there's a very easy way to say 'get me all the app bugs' [17:04:30] today im going to abstract the proof of concept i worked out. [17:04:43] i think tomorrow for my 20% day i might try moving some of my HTMLForm hacks back into core [17:05:07] still, I like tables with their simplicity [17:05:21] MaxSem tables will still be supported [17:05:34] btw tfinc how are you feeling today? [17:05:35] of course [17:05:37] any better? [17:05:46] but tables are good [17:06:13] MaxSem i think tables are good for tabular data but for managing other html elements tables are a nightmare [17:06:15] I'm no designer so don't care about rightness or wrongness [17:06:20] hahaha [17:06:31] MaxSem if i had it my way everything would be green text on a black background [17:06:46] jdlrobson: somewhat better .. still feeling off [17:07:10] I would have waged war on you, then - my eyes are hurt by inverse colors [17:07:10] YuviPanda: if the app was a component .. could you do the same ? [17:07:15] tfinc: yes [17:07:37] ok. so it sounds like collapsing the hierarchy isn't necessarily a bad thing then [17:07:38] tfinc: also, i'm not sure using milestones has been great for us. [17:07:44] i prefer tracking bugs [17:08:05] * MaxSem goes through BZ patches [17:08:09] YuviPanda: i like how easily i can see other bugs blocking a tracking bug [17:08:14] milestones don't do that well [17:08:16] yeah [17:08:23] plus milestones don't let you comment on the milestone itself [17:08:25] plus they're hard to share [17:09:31] +1 to everything YuviPanda has said [17:10:13] hmm ok [17:10:49] so consensus so far is to collapse under Wikimedia Mobile, make everything that is a product a component, and drop the use of milestones [17:11:15] in favor of tracking bugs [17:12:10] and stop using MobileFrontend extension - use component Wikipedia Mobile site instead? [17:12:31] (or if there's a way to make them the same even better...) [17:13:04] jdlrobson i do not think that's such a good idea [17:13:25] i think it's important that MobileFrontend extension stay under the MediaWiki extensions in bz [17:13:25] +1 to Arthur [17:13:52] but the rest i agree with :) [17:14:06] this is a case where bugzilla completely fails [17:14:18] we should be able to have multiple entry points for the same end point [17:14:28] yep [17:14:41] as in … bugs entered into the extension or top level product should end up in the same place [17:14:48] bugzilla for the fail [17:14:51] tool purposebuilt for mozilla fails for other people? You don't say... [17:14:59] :D [17:15:30] haha [17:18:22] so the only consensus that we have is to move the app back under the Wikimedia mobile product [17:24:36] awjr: why did you need to figure out how to edit articles on remote wikis we already had the code for that? [17:24:54] preilly: we already have the code for editing articles on local wikis [17:25:25] preilly if we have code for editing articles on remote wikis, i did not realize - where is it? [17:25:28] awjr: you mean remote [17:26:20] preilly: where is there code for editing remote wikis? [17:28:18] awjr: it's not in the current version anymore [17:28:25] awjr: let me find it again [17:28:32] preilly: cool thanks [17:28:45] tfinc: philinje thoughts on getting some UX work done on https://bugzilla.wikimedia.org/show_bug.cgi?id=36017 [17:29:00] our current 'initial open' experience is pretty bad [17:29:34] YuviPanda: i've been thinking more and more about some of the confused reviews in the app stores and how a help screen could fix that [17:29:41] +1 [17:29:47] for instance … we still have people asking for easy language switching [17:29:55] … which we've had for all of time [17:30:01] yes … some people can't same to find it [17:30:06] s/yes/yet [17:30:33] +1 [17:31:12] YuviPanda: get philinje to draw one up for us [17:31:26] * YuviPanda pokes philinje  [17:31:33] * tfinc goes to do a tech screen for fundraising [17:31:39] ok, in a call [17:34:49] philinje: no rush … lets just get a conversation going [17:35:19] awjr: is this fixed? https://bugzilla.wikimedia.org/show_bug.cgi?id=35250 [17:35:28] tfinc: this blocks 1.2 final release, since we don't really have the current Mobile Home page available via API [17:35:30] philinje: ^ [17:36:11] YuviPanda philinje let me get involved in copy text if you need some [17:36:33] jdlrobson: im not sure i havent tried cookie stuff on a windows mobile phone in a while [17:36:36] jdlrobson: philinje is apparently in a call, am assuming he'll get back to us once that is done [17:37:02] YuviPanda: or he'll schedule a time to talk to you guys about it ;) [17:37:13] yes. [17:37:17] get back about getting back :) [17:40:05] YuviPanda, you should use action=parse for main page, as it doesn't need per-section loading [17:40:22] MaxSem: it doesn't do the changes that MF does, no [17:40:23] ? [17:41:24] &mobileformat=html&mainpagetransform [17:42:15] MaxSem: right, so there is an API method [17:42:33] so now it boils down to 'do we show default home page by default or do sometihng else that is faster' [17:43:52] New review: awjrichards; "i'll take a stab at making the changes you mentioned in your comment." [mediawiki/extensions/MobileFrontend] (contact-us-redesign); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5316 [17:43:55] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5316 [17:43:59] YuviPanda: when are you going to update phone gap? [17:45:10] preilly: 1.6.0 had a few issues, 1.6.1 is supposed to be fix them. I'm unsure if I want to do it this cycle or wait for the next one. It's got quite a few breaking changes, so a lot of our plugins will need to be touched.' [17:45:18] *supposed to fix [17:45:35] https://github.com/apache/incubator-cordova-android/tags [17:45:47] 1.6.1 is available now [17:46:16] YuviPanda: hmm [17:46:57] also, there are quite a few code changes for this release already (API move), and I'm not sure if we should do a breaking platform upgrade with it as well [17:47:11] (1.4.x -> later is breaking) [17:47:20] YuviPanda: got it [17:48:30] yowser.. what happened to http://en.m.wikipedia.org/w/index.php?title=Special:UserLogin [17:48:45] jdlrobson: damn [17:55:03] https://appworld.blackberry.com/webstore/content/reviews/105171/?lang=en [17:55:06] \o/ [17:55:38] YuviPanda: has the api branch been succesfully merged into WikipediaMobile yet? [17:55:44] pfhayes: happened yesterday [17:55:49] YuviPanda: cool [17:55:53] err [17:55:56] by yesterday I mean today [17:56:00] YuviPanda: :) [17:57:12] BB are jerks [17:58:22] goddamit, we're in 2012 - it's not fashionable to use "Best viewed with Netscape" banners or "you're running an unsupported broser" warnings anymore [17:58:48] MaxSem: are those notices wrapped in tags? [17:58:53] because i hope so. [17:59:03] * MaxSem bites awjr [17:59:08] hahaha [17:59:30] goddamn bees. hope these don't sting [18:01:41] heh [18:01:51] BEST VIEWED IN NETSCAPE NAVIGATOR 2.0 GOLD [18:02:28] brion: btw, I don't see the save page code getting more platform agnostic. [18:02:37] it is possible that for perf reasons we might have to go less platform agnostic [18:02:43] really? it should just be stashing the data we got from the api [18:02:56] feels less snurchy [18:03:14] brion: so, problem is images [18:03:42] hmm [18:04:11] can we fetch the images via XHR and get the original data? or is that even more painful :( [18:04:18] brion: for iOS, we're using data URIs. I talked to MaxSem about having the API itself give us a 'single request' that gives us images as data uris, but we figured it is quite a fair bit of work [18:04:35] brion: hmm, that is where I'd probably move to. [18:04:53] brion: problem last time was I kept running into segfaults when I tried doing that with android last time [18:04:57] blugh [18:05:00] it ate a week [18:05:18] brion: so went back to using urlcache in android and dataURIs in iOS [18:05:31] brion: i'll probably try again, atleast for iOS. And it is possible this time it will work [18:05:41] the zepto segfault, for example - vanished when I removed it this time :( [18:05:58] \o/ wtf mystery segfaults :) [18:06:52] yeah [18:07:00] philinje: finishing up a tech phone screen [18:09:41] hmm [18:10:00] YuviPanda, does clicking on reference footnote numbers work as expected in the latest app for you? [18:10:06] brion: master? [18:10:07] brion: nope :D [18:10:13] brion: we haven't gotten to that yet [18:10:20] ok [18:10:25] just checking if that's known :D [18:10:45] brion: we'll be reusing the code from the new mobile reference reveal bits [18:11:06] +1 [18:14:07] YuviPanda, and loading saved pages is broken too for now? [18:14:17] yes [18:14:24] :( ok [18:15:18] hmm [18:15:30] map view on my galaxy nexus covers the markers on the map, so you can't see the markers [18:15:54] ? [18:16:07] tiles have a larger zindex than the markers? [18:16:09] screenshot? [18:18:00] moment [18:18:38] i'm also guessing that it'll be fine if we handle hashlinks only to section headers? [18:19:54] most hashlinks will be to section headers (or references directly) [18:20:03] occasionally there'll be specific ones to a specific id, but rare [18:20:39] http://bug-attachment.wikimedia.org/attachment.cgi?id=10441 [18:21:09] right [18:21:14] we'll handle references with reference reveal [18:21:19] and have code for handling sectionh eaders [18:21:36] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5312 [18:21:39] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5312 [18:22:06] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5313 [18:22:09] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5313 [18:22:36] http://bug-attachment.wikimedia.org/attachment.cgi?id=10442 <- here's one showing some of the markers actually there whiel scrolling :) [18:22:51] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5319 [18:22:54] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5319 [18:23:51] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5320 [18:23:53] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5320 [18:24:26] brion: it seems to me that just a few tiles have failed loading? [18:24:41] YuviPanda, that's during scrolling, so those tiles haven't been loaded yet [18:24:50] as soon as you let it go, they load up and cover the markers and the zoom controls [18:25:32] oh damn i opened the same attachment twice and was trying to find differences [18:25:41] brion: i see it [18:25:44] were you seeing this before? [18:25:56] first time i've seen it [18:30:02] right [18:30:05] MOAR REGRESSIONS [18:30:07] * YuviPanda stabs self [18:30:21] heh [18:33:43] New patchset: awjrichards; "Error handling tweaks to display field-specific error messages above field containing the error and disable display of the 'error stack' above the form fields." [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5343 [18:35:19] New review: awjrichards; "Take a look at https://gerrit.wikimedia.org/r/#change,5343 for your requested error handling tweaks." [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5316 [18:57:55] I hate how one single hack to fix one tiny hack can take a few hours to a day and still never really work [18:59:42] welcome to software [19:00:20] brion: indeed. [19:00:24] heh [19:01:01] worst part is [19:01:06] this worked a few days back [19:01:30] exact same code [19:01:31] and now it doesn't [19:01:32] :( [19:16:00] hmm, our skins aren't supposed to show login/logout links although we have code for it? [19:40:55] am off [19:40:56] gnite everyone [19:51:41] ewww https://en.m.wikipedia.org/w/index.php?title=White_Sea_%E2%80%93_Baltic_Canal [20:03:27] MaxSem: thats pretty b ad [20:03:28] bad* [20:03:38] is it another inline style issue ? [20:04:13] I'm asking for my rewrite that should fix it [20:04:37] should they be in normal skin in addition to beta? [20:05:01] is it much work to make both work? [20:05:58] hehe, you were talking about images, I were talking about loin links [20:06:03] yes! [20:06:53] well, I'm not touching image flow, but would like to know if we're supposed to display login links [20:10:56] MaxSem: those aren't going to come out of beta into the new nav shows up [20:11:10] so i wouldn't worry about anything user facing for now [20:11:52] okay [20:12:12] reminder to myself to rewrite our login form [20:18:31] MaxSem: https://gerrit.wikimedia.org/r/5355 [20:20:52] New patchset: preilly; "Add options to hook in order to render warning on desktop view click from zero rated sites" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5356 [20:20:57] MaxSem: ^^ [20:21:06] awjr: ^^ [20:21:48] awwww, I'm rewriting all this [20:22:08] MaxSem: what do you mean? [20:22:28] that there will be no ApplicationTremplate, for example [20:22:46] MaxSem: okay, well you'll need to port this too :) [20:22:59] rawr [20:23:52] MaxSem: ha ha [20:25:41] New patchset: preilly; "Add options to hook in order to render warning on desktop view click from zero rated sites" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5356 [20:29:40] preilly, do we still need the SOPA notice? [20:29:47] MaxSem: nope [20:29:55] *lilling* [20:30:02] killing [20:30:31] should I keep the notice functionality? [20:30:45] that is probably a good idea [20:33:08] preilly, you may want to rename onBeforePageDisplay() now that it handles a different hook [20:34:30] MaxSem: yeah [20:36:40] preilly im looking at the application.js changes. if you want to abort the 'stopMobileRedirect' cookie from getting written, do you also want to stop the mf_mobileFormat cookie from expiring? [20:37:29] awjr: I don't think so [20:37:35] k [20:38:44] preilly should hookOptions be passed to js as an array so it can support more than just 'toggle_view_desktop' or… cross that bridge when we get to it? [21:15:55] New patchset: MaxSem; "Almost finished skin rewrite" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5361 [21:16:09] preilly, awjr ^^^ [21:16:26] it's not complete, but I'd like to know your opinion [21:17:13] +593, -796 <-- win [21:30:12] MaxSem why change from public/protected properties to private? [21:30:35] because I created a few accessors [21:31:26] i think we should be really careful about private and static properties/methods as they make extending our code a lot more difficult [21:31:54] well, I'm waging war on static [21:32:10] i think things should be private only if there's a really good reason to keep that thing unique to the class - in general my preference is to default to protected if public is no appropriate [21:32:18] MaxSem that is a good war to wage [21:32:50] there are 2 good reasons: [21:33:03] 1) to hide implementation details [21:33:32] 2) to put properties behind accessors [21:34:12] what implementation details need hiding? [21:34:21] #2 is applicable to protected properties as well [21:37:32] for example, is this thing just a variable? or it needs to call another function? or it caches something? hiding stuff behind accessors actually helps descendant classes [21:38:42] by the way, what classes to you envision to be extended by 3rd party that doesn't have access to our repo to tweak access levels if needed? [21:39:21] s/to you/do you/ [21:40:23] MaxSem: im more thinking about someone who might want to extend MobileFrontend without messing directly with the MobileFrontend code [21:40:32] for instance, like I've been doing with HTMLForm [21:40:46] extending HTMLForm is kind of a nightmare because of sloppily used static methods and properties [21:40:57] but, extending HTMLForm is enormously useful for us [21:42:19] i guess i don't mind private properties as much as long as we're very consistent about accessors. i'm way more concerned about private methods (eg extMobileFrontend::getOptInOutCookie() or extMobileFrontend::removeQueryStringParameter(), etc) [21:43:01] im not sure any of the methods extMobileFrontend really ought to be private [21:43:19] s/methods/methods in [21:43:41] preilly what are your thoughts? [21:44:03] lol, PHP vs. enterprisey war [21:45:05] New patchset: L10n-bot; "Localisation updates from http://translatewiki.net." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5375 [21:45:09] New review: gerrit2; "Auto-approving/merging l10n updates" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5375 [21:45:09] Change merged: gerrit2; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5375 [21:50:53] awjr: yeah, I agree [21:51:19] awjr: cross the bridge when we get to it regarding hookOptions [21:51:27] aye [21:51:47] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5356 [21:51:50] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5356 [21:52:36] awjr: thanks! [21:54:21] any comments on anything not related to visibility? [21:55:24] MaxSem sorry got pulled away - haven't finished looking through the code yet [21:55:40] no hurry about that [21:56:08] I'm not even sure if it should be deployed next Monday [22:12:59] awjr: can you vet this on ICS https://twitter.com/#!/thirstforwine/status/193097552357363712 ? [22:16:38] tfinc: what was the actual reported issue? [22:17:08] tfinc nm found it [22:17:50] thanks. i'm curious if its just his device [22:17:55] if so i'll file a bug [22:18:02] its its all of them then its higher priority [22:18:43] tfinc i have no problem scrolling in native browser, chrome, or dolphin on ICS [22:18:58] tfinc: is there an ICS tablet in the office? [22:24:01] awjr: good question [22:24:04] i know the xoom runs 3.0 [22:24:12] id have to check the galaxy tab [22:24:55] tfinc: i am however having problems with the beta opt-in [22:25:23] i need to debug a bit to see what's going on [22:25:56] * awjr goes in search of microusb cable [22:32:08] preilly: i'm camping out next to r35 if you have anything to chat about [22:32:18] tfinc: not really [22:32:21] k [22:33:23] hmm Special:MobileOptions/BeatOptIn is not setting the opt-in cookie for me [22:34:09] oh i wonder if it's because it's posting to en.wikipedia.org instead of en.m.wikipedia.org [22:34:43] MaxSem ^ [22:36:09] hmm, yes - fail [22:36:35] MaxSem both buttons point to non-mobile site [22:38:39] we should set $wgServer appropriately on mobile site [22:39:11] MaxSem there are some helper function in ExtMobileFrontend to help you get the correct URL [22:39:17] why we don't by the way? [22:39:21] yeah, getMobileUrl() [22:39:44] MaxSem because the .m domains are served by the same app servers as the non-.m domains [22:39:49] but hell knows how many more places are subtly broken due to that [22:39:57] the domain magic happens at the webserver/proxy level [22:40:48] that's not a problem, after all, the same webservers serve WP and mw.o, for example [22:40:51] plus there are lot of places where we display both mobile links and desktop links from within MF so modifying the $wgServer global for that purpose seems messy and error-prone [22:41:26] hmm [22:41:30] that's why we have getMobileUrl() and getDesktopUrl() to ensure that you're getting the right kind of URL [22:42:52] awjr, whee - HTMLForm is also broken by this [22:42:59] :( [22:43:00] it uses $wgServer [22:43:17] MaxSem but you can set your own action with HTMLForm [22:43:24] thou shalt not lie [22:43:34] i shant [22:43:44] in your _________ [22:43:56] if user sees a domain, we shouldn't pretend we're on a different one [22:43:59] lol [22:44:08] MaxSem ? [22:44:36] lemme plug this hole [22:44:39] k [22:45:04] but we *seriously* should consider using proper $wgServer [22:46:04] in the magical future there will be no separate mobile domains and we will all live free amongst the unicorns [22:46:07] then it wont matter. [22:46:29] because we'll be too busy hunting unicorns for their tasty tasty unicorn meat [22:56:06] sigh to my router [22:56:21] New patchset: MaxSem; "Fixed forms submitting from mobile to desktop site" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5391 [22:56:27] awjr, ^^^ [23:00:08] New patchset: MaxSem; "Make sure that getMobileUrl(0 works for protocol-relative URLs" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5392 [23:02:22] MaxSem there's something wrong in the first changeset [23:02:46] im not sure what it is yet but it's causing undefined index errors for me when using a mobile-specific domain [23:05:17] * MaxSem sets up multidomain [23:05:59] i dont understand why that changeset would cause problems though since you only made changes in the special pages [23:06:17] but im seeing the problem when i checkout your changeset and go to any mobile page [23:07:16] oh interesting but the call stack shows SpecialMobileOptions getting called [23:08:19] MaxSem found it [23:09:05] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 0 C: -1; - https://gerrit.wikimedia.org/r/5391 [23:10:43] perhaps we should make the getMobileUrl method a little smarter to do URL expansion if necessary [23:10:47] i've been bitten by that before [23:13:38] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5392 [23:13:40] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5392 [23:14:07] awjr, what's your setup? the change ^^^ checks for protocol-relative URLs [23:14:36] MaxSem yeah https://gerrit.wikimedia.org/r/5392 is fine, the other is not [23:14:54] so what's the difference? [23:15:05] MaxSem i set up my /etc/hosts file to look something like this: [23:15:07] 192.168.169.128 mobiletest.localhost.org [23:15:07] 192.168.169.128 mobiletest.m.localhost.org [23:15:17] (192.168.169.128 is the ip of my testing virt) [23:15:24] then, in LocalSettings: [23:15:39] $wgMobileUrlTemplate = "%h0.m.%h1.%h2"; [23:16:13] MaxSem so 'desktop' url for my local testing instance is mobiletest.localhost.org, mobile URL is mobiletest.m.localhost.org [23:16:35] awjr, same thing [23:16:48] also i have E_NOTICE on [23:17:41] regardless, if you have a two-domain setup, the patchset you submitted breaks things like the disable images link [23:17:56] you need to wrap $t->getLocalURL() in wfExpandUrl() [23:18:18] how do I force mobile view on a domain? [23:18:58] that requires some url rewrite magic. i don't have that set up locally [23:19:29] cause now mobile view links lead me to mobile domain but it displays desktop content [23:19:53] maybe i do have that set up. [23:21:30] oh i do [23:21:41] MaxSem i set up to different virtualhosts in apache [23:21:49] let me show you [23:22:10] hmm, reprod [23:22:15] awjr: why not just two host entries [23:22:52] http://pastie.org/3819697 [23:22:55] MaxSem ^ [23:24:06] preilly didnt occur to me [23:24:18] slapping together two vhosts seemed expedient [23:24:39] actually preilly i do have two host entries [23:25:01] + 2 different apache vhosts, one of which sets X_DEVICE header for the mobile domain [23:25:30] my solution: if ( preg_match( '/\bm\./', $_SERVER['HTTP_HOST'] ) ) $wgMFConfigProperties['useFormat'] = 'mobile'; [23:25:39] awjr: you could do that with one ghost [23:25:44] s/ghost/vhost [23:25:53] awjr: but, I guess it really doesn't matter [23:26:20] preilly iirc i was trying to do it with one vhost but was having trouble getting it working right so figured i'd just make two separate vhosts in the interest of time [23:26:31] i forget what i was trying to do though [23:26:47] no worries [23:39:38] New patchset: MaxSem; "Fixed forms submitting from mobile to desktop site" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5391 [23:40:11] simply switched it to use local URLs more [23:52:20] New patchset: awjrichards; "Modest cleanup of html handling; getFeedbackHtml() now returns html rather than piping it directly to getOutput. This now handled in execute() method." [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5394 [23:53:38] New patchset: awjrichards; "Modest cleanup of html handling; getFeedbackHtml() now returns html rather than piping it directly to getOutput. This now handled in execute() method." [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5394 [23:54:32] New review: awjrichards; "Still some work to do with the API client, namely around error handling and documentation." [mediawiki/extensions/MobileFrontend] (contact-us-redesign); V: 0 C: 0; - https://gerrit.wikimedia.org/r/5394 [23:56:34] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5391 [23:56:36] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5391