[00:14:12] awjr: preilly: which one of you guys would like to do round 3 of our qa candidate ? [00:14:26] he's already spoken to yuvi and me [00:14:32] notes are in job vite [00:14:59] tfinc: in person? [00:15:02] phone [00:15:29] MaxSem: https://gerrit.wikimedia.org/r/5145 [00:16:32] awjr: is/did this go out today https://gerrit.wikimedia.org/r/#change,4899 ? [00:16:46] New patchset: preilly; "Add new 'BeforePageDisplayMobile' hook to beforePageDisplayHTML method" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5146 [00:16:55] MaxSem: and https://gerrit.wikimedia.org/r/5146 [00:16:56] its a bit hard to tell as http://www.mediawiki.org/wiki/Extension:MobileFrontend/Deployments is empty for today [00:17:31] tfinc: it likely did, no one added their changes to the page :( [00:17:36] lemme double check [00:17:43] awjr: you in for the qa interview ? [00:17:51] it'll be a final before we bring the candidate on [00:17:59] i just want to get one more opinion [00:18:00] tfinc when is it? [00:18:09] the earlier the better [00:18:14] i want it to happen this week [00:18:23] tfinc im happy to do it [00:18:27] k [00:18:32] i'll get angela to schedule [00:18:48] tfinc that change did get deployed. [00:19:17] tfinc: Change Ice9c7672: stylesheets/common.css is in [00:19:42] tfinc: e.g., https://gerrit.wikimedia.org/r/#change,4899 [00:19:51] preilly: what's the easiest way for me to see all the changes that went out without having to bug you guys to update the wiki page ? [00:20:05] on a weekly basis [00:22:20] tfinc: git log on origin/wmf/1.20wmf1 [00:23:58] tfinc: or, maybe https://gerrit.wikimedia.org/r/#q,status:merged+project:mediawiki/core+branch:wmf/1.20wmf1+topic:wmf/1.20wmf1,n,z [00:26:51] tfinc: so, you could do this: git log b246830720baa9030bcfe20d961a8ce585927691..6329cd0ea5d03ffb83de91352886f5e85389b4a6 --graph [00:27:06] tfinc: based off of the values in https://gerrit.wikimedia.org/r/#patch,unified,5136,1,extensions/MobileFrontend [00:27:22] tfinc: e.g., -Subproject commit b246830720baa9030bcfe20d961a8ce585927691 +Subproject commit 6329cd0ea5d03ffb83de91352886f5e85389b4a6 [00:28:36] tfinc: for output like: http://pastebin.mozilla.org/1573865 [00:33:50] is it just me or are page load on mobile super slow? [00:34:00] its bits [00:50:06] New patchset: awjrichards; "Resolves bug 36024" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5147 [00:50:56] preilly ^ [00:51:36] grumblegrumble whitespace [00:54:21] New patchset: awjrichards; "Resolves bug 36024" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5147 [01:03:30] awjr: I'm not too sure about the use of $_SERVER['HTTP_HOST'] [01:03:50] preilly i wasnt either but it has been used elsewhere in MF [01:03:52] awjr: also why is it $useMobileFormat == 'true' [01:04:20] awjr: shouldn't getUseFormatCookie cast? [01:04:25] preilly i changed the cookie name to mf_mobileFormat but didnt change all the format names [01:04:38] s/format/var [01:05:30] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5147 [01:05:32] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5147 [01:06:17] preilly it could cast, but i think doing string evaluation is better/less error prone since cookie values will be returned as strings [01:07:13] preilly since it's getting late, are you ok with me waiting til tomorrow am to push that change to produciton? [01:09:07] awjr: that's totally fine by me [01:09:29] groovy thanks [01:16:50] awjr: np [01:19:02] awjr: can you look at this: https://gerrit.wikimedia.org/r/#change,5146 [01:19:31] awjr: as, well as — https://gerrit.wikimedia.org/r/#change,5145 [01:21:42] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5146 [01:21:44] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5146 [01:21:48] tfinc: wow, http://ganglia.wikimedia.org/latest/?r=hour&cs=&ce=&m=&s=by+name&c=Mobile%2520caches%2520eqiad&tab=m&vn= [01:22:39] Current Load Avg (15, 5, 1m): [01:22:39] 23%, 38%, 95% [01:23:08] yes [01:23:15] thats why i was saying that it felt really slow [01:25:37] tfinc: argh [01:26:49] I'm stepping out. preilly is going to continue debugging this [01:26:53] preilly: we should set an alarm for page times [01:26:59] so that we can be alerted about this [01:27:50] preilly i can hang out for a bit longer - is there anything i can help with? [01:43:52] awjr: nope I'm good [01:44:48] k [01:45:30] awjr: Tell someone you love them today, because life is short. [01:45:46] awjr: But SHOUT it at them in German, because life is also terrifying. [01:45:49] preilly: i love you [01:45:51] awjr: "Ich liebe dich!" [01:45:56] hahaha [01:46:18] awjr: that was from someones Facebook status [06:14:16] hmm, a git bisect that ended up useless [06:14:19] * YuviPanda goes to difing [06:30:42] YuviPanda: any word from RiL ? [06:30:53] deafening silence, mostly [06:31:03] but Rolken helped fix the rotation bugs [06:31:20] oh it's fixed? [06:33:11] would love to see the fix so I can fix the plugin... [06:39:40] ah interesting [06:40:48] nice job [06:41:37] Still a shame the orientations are hardcoded... but better than nuthin [07:16:05] binary search debugging [07:16:06] ugh [07:35:42] YuviPanda, how's your elbow? [07:58:04] MaxSem: not broken. Took painkillers yesterday. Alright now [07:58:09] MaxSem: apparently a 'pinched nerve' [07:58:14] eww [07:58:46] MaxSem: it's still in general a bit weak, but I can type without discomfort and move it around without discomfort. [07:58:48] so not bad [08:00:24] 'blind' debugging hurts more though. [08:00:30] 'fail silently' is just such a wonderful idea... [08:00:36] hey YuviPanda [08:00:38] jdlrobson: thanks for moving that [08:00:41] heya jdlrobson [08:00:47] what did i move heatherw ? [08:00:55] the feedback [08:01:09] ahh yes i did [08:01:24] i pinged phil but didn't hear back [08:01:41] sorry i'm currently a bit confused as something I expected to get deployed didn't get fixed :( [08:01:52] :( [08:02:20] YuviPanda: sorry about your elbow [08:02:47] heatherw: it's fine now :) [08:02:50] (atleast for now) [08:09:42] YuviPanda: slept today? [08:09:47] jdlrobson: yes. [08:09:53] \o/ [08:09:57] jdlrobson: I think i've 'reset' to a saner sleep cycle [08:10:13] i was going to help you get the api stuff merged today [08:10:23] any of brion's feedback you want me to specifically look into [08:10:29] jdlrobson: yeah, am trying to split 'doScrollHack' into three functions [08:10:48] setupScrolling, scrollTo and resetToTop [08:10:56] i suspect - there's extra space at the bottom of the screen, presumably reserved for the iOS toolbar?- is due to the change in app.css [08:10:58] though i'm guessing it should just be two [08:11:11] yeah, but that isn't there on android 2.x [08:11:16] so I need to figure out what's happening there [08:11:22] also entire scrolling on 4.x is super weird [08:11:24] i'm also wondering if we might need a different stylesheet for ios/android [08:11:27] touching things makes them scroll [08:11:35] +1 [08:12:05] jdlrobson: do you have a 4.x emulator? [08:12:24] can you look into why touches are being translated to scrolls? plus a huge amount of whitespace at the bottom... [08:13:02] yeh ill look into the whitespace problems to start off with [08:13:06] ok [08:13:07] only have emulators mind you! [08:13:29] ok! [08:18:44] New patchset: Jdlrobson; "show page titles again" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5158 [08:19:57] scrolling issues are incredibly irritating. sigh [08:20:02] :/ [08:20:28] if mf-to-app.patch needed any more btw [08:20:38] jdlrobson: nope, should get rid of it [08:21:40] scrollTop has absolutely no effect whatsoever [08:21:49] we had this bug in 3.x, fixed with an ugly hack [08:21:54] but my emu is on 2.x [08:21:55] :( [08:22:00] which didn't have this bug [08:22:00] sigh [08:25:06] jdlrobson, cough [08:25:18] MaxSem: strepsil [08:25:23] howdy [08:25:27] what did i break? [08:29:00] jdlrobson, you should have just moved FirstHeading from the other condition [08:29:03] jdlrobson: removed the patch. [08:29:27] yes.. silly me MaxSem - ill commit ammend [08:29:56] actually MaxSem they are slightly different [08:30:08] jdlrobson, just poke me next time, it was me who broke it after all [08:30:41] what's different? [08:30:41] what is the different between getPageTitle and getTitle ? [08:30:57] title is Title object [08:31:04] we seem to have a title and pageTitle passed to the application template [08:31:07] do we need both? [08:31:24] pageTitle is what should be displayed in

[08:31:31] ok i see [08:32:11] I'm tempted to start killing the еуьздфеуы кшпре тщцб иец [08:32:20] *the templates right now [08:32:40] k resent MaxSem [08:32:42] New patchset: Jdlrobson; "show page titles again" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5158 [08:32:48] re-sent (not resent :)) [08:33:08] but for the record MaxSem you might have broke it but i code reviewed it :) [08:33:29] YuviPanda: - i've found the problem with toggling [08:33:31] i'll fix that now [08:33:48] mmm, undefined variable [08:33:48] jdlrobson: 4.x? [08:34:13] YuviPanda: - no the issue with the icons not changing [08:34:17] ah, right [08:34:18] ok [08:34:57] or not - but it displays pointless "Main Page" on main pages [08:35:11] meh, I'm gonna rewrite it anyway [08:35:16] yeh MaxSem - i couldn't decide if that was pointless or not [08:35:19] we didn't use to have it [08:35:36] YuviPanda: was the reason for using a different toggle purely to handle dynamic loading? [08:35:43] jdlrobson: yes [08:36:05] what is actually different? [08:36:14] jdlrobson: everything? :) [08:36:23] ok, off with their heads [08:36:25] it isn't really a 'toggle' anymore. It actually loads the content into DOM [08:36:26] sorry looking at the wrong file lol [08:36:34] jdlrobson: it's in chrome.js now [08:36:36] not toggle.js [08:36:47] i'm sure we can use the same toggle.js ... [08:38:12] 1. what exactly does it give us over the current 16 line implemnetation? [08:38:17] 2. How would it handle dynamic expansion? [08:38:49] aren't they doing two different things now? [08:39:01] toggle.js is doing toggling, while this is doing full on loading + toggling [08:39:37] plus there's the matter of having to reset scroll after every expansion on iOS 4.x that we handled in the branch by patching toggle.js [08:42:30] ok so the main issue is the current 16 line implementation is wrong [08:42:38] :D [08:42:47] it adds the openSection class to content block [08:42:51] but not to the heading [08:43:01] also the css in beta_common.css is linked to toggle.js [08:43:12] so I worry if there are changes to toggle.js in future it will break the app [08:43:26] git submodules are pinned to a commit [08:43:28] also expansion would be handled by just adding two click handlers to a section [08:43:45] one for loading the section via the api (your code) the other for handling the expansion [08:44:14] That does seem code reuse for the sake of code reuse to me jdlrobson [08:44:20] also a lot of the code in toggle.js was for the app anyway - so i expect it to go away [08:44:45] this will no longer be dependent on MF html structure (yay). beta_common.css is pinned to a commit (git submodule) [08:46:59] i dunno.. it just feels very wrong. I guess because I see the future as being the app is exactly the same as the mobile site [08:48:37] jdlrobson: so it'll be the same when MF has dynamic content loading [08:49:12] sure but they should feed each other - if the current code doesn't allow for dynamic content loading we make that possible [08:50:08] right, but until then why block the app on an MF feature? [08:56:00] New review: MaxSem; "(no comment)" [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/5158 [08:56:16] I'm rewriting all this stuff anyway [09:02:08] YuviPanda: sent you a pull [09:02:28] jdlrobson, we currently output both lang and xml:lang attributes in , the latter isn't needed for HTML5, right? [09:02:39] also xmlns [09:03:21] jdlrobson: i'll cherrypick those two commits [09:03:39] jdlrobson: that was the most minor issue though. scrolling is, as usual, the major pain point. [09:03:46] just the lang attribute yes MaxSem [09:04:03] thx [09:06:48] jdlrobson: the README modification. I added that instruction to the top of the file :) [09:06:56] thx [09:07:15] *had added. Been there for a while [09:07:34] ahh ... sorry I didn't see it. Guess I was scanning for mobile browser development [09:08:39] looking in ICS now [09:08:53] jdlrobson: ok [09:09:23] including the weird scrolling :) [09:10:08] jdlrobson: i'm currently trying to split doScrollHack into 'setupScrolling' and 'scrollTo'. Working on a branch. [09:10:25] [09:10:33] do we really need this? [09:12:18] YuviPanda: Apple has approved 3.1.2 [09:12:44] flyingclimber: nice. [09:12:48] i shall tweet [09:13:01] I haven't released it yet [09:13:05] ok [09:13:09] poke me when you do [09:13:18] i can either release it now and have you monitor ( as its 2AM here ) [09:13:23] or wait till the morning [09:13:41] flyingclimber: your call :) [09:14:35] how much of my morning will you be up for? [09:15:00] i'll be up until about the same time as yesterday. [09:15:09] that would be till around noon PDT, I guess? [09:15:35] then we can release tomorrow [09:15:42] i'll be up by 8PDT [09:15:48] alright [09:15:48] and in the office by 9PDT [09:16:13] * YuviPanda stashes tweet [09:16:17] MaxSem: - I think that was for validation purposes see http://oli.jp/2009/html5-charset/ -could use instead [09:16:51] * flyingclimber tries to compile ganglia for arm one more time before crashing  [09:40:45] * flyingclimber crashes [10:03:40] jdlrobson: any luck with scrolling? [10:05:25] not making much progress no :( [10:05:52] ok [10:06:05] i'm doing the split and modifying things one commit at a time. will poke when it's done [10:14:50] this is incredibly irritating [10:17:24] agreed... [10:17:31] YuviPanda: where does #menu get removed for android? [10:17:57] jdlrobson: it isn't. [10:18:08] items are not rendered, so afaik it shouldn't affect anything [10:18:17] #menu never gets added [10:21:41] got it [10:23:52] YuviPanda: https://github.com/jdlrobson/WikipediaMobile/commit/dbe7a2578fbe6ede07a20b3c847be6617f626ff7 < sorry it's not the scroll fix :( [10:24:19] heh, okay [10:24:25] will cherry-pick [10:27:18] jdlrobson: https://github.com/yuvipanda/WikipediaMobile/commit/1260be8013dbd7cda78c61804b7b16218bb9cd5c [10:27:48] jdlrobson: i need to have a ~200ms delay before calling scroller.refresh. I am guessing that is because of the time taken to construct DOM + show it [10:28:05] is there any way I can do it without just hardcoding 200? [10:28:11] .show, etc don't return a deferred... [10:29:50] * YuviPanda continues with that splitting branch for android fixes [10:31:17] well looking at iscroll.prototype.refresh [10:31:22] it checks a property zoomed [10:31:34] and if not set it does ._resetPos(200); [10:31:47] where 200 is a time for the scroll duration [10:32:10] i don't know if that is coincidence of related [10:32:13] jdlrobson: are you looking at iscroll-lite? [10:32:20] or iscroll? [10:32:24] iscroll4 [10:32:50] check iscroll-lite.js instead. [10:32:57] i do see the resetPos but no conditionals around it though [10:33:15] yeh same line is there [10:33:16] hat._resetPos(200); [10:33:20] grep for that._resetPos(200); [10:34:24] hmm, i'm not sure how this should be related though [10:34:40] resetPos is called from refresh, but my timeout call is making refresh get called 200ms later [10:35:48] 1. scrollTop does not work on android 2.x now (Not sure about 4.x) [10:35:55] 2. It *was* working on 2.x before, but not on 4.x [10:35:57] 3. WTF [10:36:19] :( [10:36:35] it has absolutely no effect and is always zeor [10:36:36] *zero [10:36:56] bug 35369 specifically states this was a problem but only on 4.x device [10:36:56] s [10:38:47] we did the hiding/showing fix previously... [10:39:09] no you are right... we used scrollTop [10:39:11] mm [10:39:35] jdlrobson: hiding showing with a timeout was for 4.x and 3.x [10:39:38] it was fine on 2.x before [10:40:19] nope, doesn't help [10:44:57] btw do you notice the scrolling problem over infoboxes? [10:45:09] jdlrobson: on? [10:45:12] which platform? [10:45:20] well in ICS when i scroll down it is very jerky [10:45:26] it seems to be near the top of the article [10:45:35] ICS for me is completely unusable on the phone [10:45:39] :( [10:45:42] every touch seems to be interpreted differently [10:45:44] sometimes as a touch [10:45:47] sometimes as a scroll [10:45:59] yeh i'm noticing that [10:46:05] sometimes it opens up another article [10:46:07] can't even toggle :( [10:46:22] yeh [10:47:55] btw don't cherry pick the commit i posted earlier.. ill send you a pull when ready [10:48:39] jdlrobson: too late, but not a problem. [10:48:49] jdlrobson: also i'm wondering why that happened as well. it wasn't around a few days back [10:48:56] jdlrobson: i'm doing git bisect to pin point the ICS isssues now [10:49:32] i made a few adjustments -> https://github.com/jdlrobson/WikipediaMobile/commit/0a18d6194f5f1690ff76edd5b8187961644167ee [10:50:10] bisect has been mostly made useless because of the noScroll parameter astagi introduces, which I did not remove completely during the rebase [10:50:33] jdlrobson: why is this a problem now, when it was not at 1.1 release? which change caused this? [11:02:20] jdlrobson: 13469482cea2f810ef0d12765599b0e1abc7c721 is what introduced that [11:02:28] (the whitespace) [11:02:40] still unable to pinpoint the scroll issues though [11:03:23] 13469482cea2f810ef0d12765599b0e1abc7c721 caused what? [11:03:58] that css rule is needed for ios but not android ^^ [11:04:15] my commit does the same just also bears in mind android has a native menu [11:04:18] yeah, looks like it. [11:04:19] okay [11:04:23] the scrolling thing on IOS is doing my head in  [11:04:28] ICS sorry [11:04:29] the 200ms? [11:04:30] oh [11:04:31] ok [11:04:36] just the fact it doesn't work [11:04:47] it's schizophrenic [11:04:59] it works for its own definition of 'work' [11:05:08] jdlrobson: also, why was there no blank space bar on android 2.x? [11:05:25] no blank space bar? [11:05:37] whitespace at the bottom where the navbar is on iOS? [11:08:50] oh i see [11:08:56] i'm not sure [11:09:12] with that commit is there still blank space? [11:10:35] jdlrobson: no there isn't. [11:10:41] but i'm still confused about why it was not present on 2.x [11:11:04] * YuviPanda checks on 2.x [11:11:12] yeh that is strange [11:11:27] we were probably exploiting some weird browser bug [11:12:32] one which I'm sure will come to bite us again someday later [11:12:43] we've essentially been fixing the same bugs over and over again [11:16:27] okay, so the new commit does not cut off content on 2.x [11:16:31] so that seems okay for now [11:16:41] jdlrobson: i've cherry picked your new commit and dropped your old one [11:17:37] cool - im trying to work out on which commit ICS got broken [11:18:12] then i'll break for some lunch [11:18:46] jdlrobson: okay. git bisect? [11:32:02] so https://github.com/jdlrobson/WikipediaMobile/commit/2230bc59854dc0d4af3dc70cae99916b82909ffa seems to be where it goes all wrong in iScroll in ICS [11:32:15] jdlrobson: iScroll? [11:32:24] well scrolling in general [11:32:29] jdlrobson: iScroll shouldn't be on in iCS at all. [11:32:29] ok [11:32:31] before that commit it seems to work fine [11:32:36] then in this commit it all goes wrong [11:32:45] maybe related to the doScrollHack call [11:36:10] jdlrobson: getting rid of that has no effect [11:36:23] yeh i've noticed [11:38:43] going afk for a bit [11:39:28] me too [11:39:50] for some reason i've started testing on a jonas brothers article [11:39:56] and i'm getting a bit tired of it [11:40:01] and all its bizarre problems [12:27:53] jdlrobson: true, the bizareness of the problems is very irritating. Sigh [12:28:05] yeh [12:28:08] * YuviPanda puts a bullet through the screen [12:28:13] :( [12:28:23] it is sortof like that 127 hours movie, except I can't really cut my arms off [12:32:43] haha [12:32:58] we'll work it out [12:33:57] jdlrobson: so, window.scrollTo works [12:34:07] jdlrobson: but .scrollTop of everything and anything is always zero [12:34:11] and has no effect whatsoever [12:34:45] which was distinctly *not* the behavior we were seeing on 1.1 release [12:35:10] I don't even know what it'll do on 4.x, considering 4.x's current schizophrenic state. [12:46:00] jdlrobson: https://github.com/yuvipanda/WikipediaMobile/tree/api-move-scroll [12:46:07] gets rid of doScrollHack [12:46:20] and there's a hack to fix the '2.x android doesn't scroll to top' [12:46:23] but it's a hack [12:47:38] well that seems like a nicer hack.. [12:48:28] jdlrobson: well, i don't know how that'll behave on 4.x [12:48:50] plus i'm still unable to reconcile *how* this does not work now [12:48:58] while I very clearly remember it working before [12:49:04] plus us putting in a hack for 4.x [12:49:04] ugh [12:50:52] * Reedy hacks YuviPanda [12:51:28] well 4.x is a nightmare anyway [12:51:31] not even toggling seems to work [12:51:55] touching doesn't work [12:51:59] it just jumps about randomly [12:54:41] jdlrobson: also, https://groups.google.com/forum/#!topic/phonegap/BFyBYjQp3a8 [12:54:55] jdlrobson: plus I managed to fix that 'whitespace while scrolling lists' bug :) [12:54:57] a while back [13:01:22] jdlrobson: i shall merge in the commits from api-move-scroll to api-move. Any objections? [13:02:49] nope [13:02:53] jdlrobson: ok [13:03:23] jdlrobson: pushed. You might want to pull if you're doing work on top of api-move [13:18:25] jdlrobson: i'm leaving ICS to you for now and working on fastclick [13:18:41] ICS makes me cry [13:18:55] new bug in ICS :( - when you click a search item/load first page it is only when you touch the screen that the article loads [13:19:26] ... [13:19:45] jdlrobson: it is incredibly irritating because we're fighting with bugs that have nothing to do with us/our app [13:19:48] it also seems to be massively cut off [13:19:54] we're fighting whatever we're building on top of [13:19:57] as if it has a height imposed on it [13:19:58] jdlrobson: cut off? [13:20:02] hmm [13:20:08] so I can't scroll all the way down the article [13:20:11] i was/am not seeing that [13:20:14] (but i can scroll now so that's an improvement) [13:20:23] jdlrobson: is this api-move? [13:20:26] or with your changes on top? [13:20:58] this is with your latest [13:21:09] hmm, I did not test on ICS [13:21:12] * YuviPanda does [13:22:22] jdlrobson: yeah, it sortof scrolls. [13:22:26] still unable to expand anything [13:22:34] and I see what you mean by 'height set' [13:23:59] jdlrobson: not much of a difference though, I was seeing it even before the changes [13:24:02] it is still schizophrenic [13:25:09] ok migt have a fix for the height problem [13:25:12] and the scrolling [13:25:14] bear with me [13:25:24] still need to work out toggling.. [13:25:27] ok [13:26:42] i might break 2.x in the process ;-) [13:27:11] and iOS 5.x [13:27:13] and iOS 4.x [13:27:21] no they shouldn't be effected.. [13:27:27] hehe 'shouldn't' [13:28:39] so far my fix is very android specific [13:28:39] ok [13:31:29] MaxSem: poke back when around [13:45:47] think i've fixed ICS YuviPanda [13:45:52] wooo [13:46:02] very apt that radiohead creep has come on to spotify [13:46:22] i'm just cutting down my commit to minimum possible [13:46:28] ok [13:48:04] but essentially - in android i make #content position relative [13:48:13] and have to override several css definitions [13:50:57] jdlrobson: do you know the cause? [13:51:33] i have several candidates [13:55:31] still haven't fixed the problem in ICS where it requires a touch of the mouse to show the page [13:55:54] jdlrobson: is it possibly due to the transform3D applied to make scrolling work properly in iOS? [13:56:00] it is in deed [13:56:16] .android .scroller > * { webkit-transform: none; } [13:56:21] is what's messing up scrolling [13:56:42] pull request almost ready [13:56:45] ok [13:57:13] * YuviPanda continues rewriting doFocusHack [14:04:30] sent [14:04:32] YuviPanda: ^ [14:05:43] jdlrobson: looking [14:06:20] jdlrobson: why is there a + setCurrentPage(page); [14:06:22] in app.js? [14:06:34] * YuviPanda looks for where it is [14:06:42] argh ive committed wrong [14:06:50] not sure how i managed that [14:06:55] 1s [14:07:00] but setCurrentPage is correct [14:07:08] just don't merge [14:07:17] i've not touched setLocalPage yet, so not sure how that would even work in current state :D [14:07:19] yeah not [14:11:30] ok done now https://github.com/jdlrobson/WikipediaMobile/tree/api [14:11:35] key commit is - https://github.com/jdlrobson/WikipediaMobile/commit/e9c32063604ed6df9fa3ab057c1ee2db41adbcd7 [14:12:36] looking [14:12:36] can you confirm that fixes scrolling for you in ICS ? [14:13:22] jdlrobson: checking [14:14:50] jdlrobson: scrolling is fine, expansion doesn't work [14:14:57] no? [14:15:01] i see the classes change [14:15:04] (arrow keys change) [14:15:06] even without other commits? [14:15:08] but no content turns up [14:15:22] *even with other commits ? [14:15:27] what are you testing on? [14:15:49] jdlrobson: i pulled your 'api' branch on to api-move [14:15:53] applied cleanly (no merge commits) [14:16:40] what device you testing on? ill check it out [14:16:57] jdlrobson: nexus s [14:16:58] ics [14:17:11] physical or emulator? [14:17:11] jdlrobson: does it work for you on the emulator? [14:17:34] i'm testing on a physical device [14:17:45] i couldn't get a working 4.x emulator [14:17:48] my device is 4.x [14:17:57] yeh works for me on the emulator but im using 4.0.3 [14:18:32] this is 4.0.3 as well, I think [14:18:33] * YuviPanda checks [14:18:47] yup [14:18:49] 4.0.3 [14:19:05] a bug that is not reproducible on the emulator but is there on the real device? [14:21:48] mm [14:21:55] we've seen it all now [14:21:55] * YuviPanda debugs [14:22:30] what article are you looking at? [14:22:32] jdlrobson: are we encountering this many bugs because of the platform we're building on or the app being crappily done? [14:22:39] jdlrobson: main page and 'Mumbai' [14:23:53] so page.js is doing its job well, the html is being produced [14:24:19] mm... no problem with mumbai in emulator for me :/ [14:24:43] jdlrobson: did this break things in 2.x? :D [14:24:55] I didnt check [14:24:57] i hope not.. [14:25:05] otherwise i might have to run into a wall [14:25:20] jdlrobson: so it *is* being added to the DOM [14:25:24] .html tells me it is there [14:25:27] yet it isn't [14:26:04] * YuviPanda updates MF [14:26:42] no luck [14:26:49] mm [14:26:52] but it scrolls right? [14:27:06] it does [14:27:11] not schizophrenic [14:27:47] calling show/hide or slideup/slidedown has no effect either [14:28:07] well that's progress [14:28:33] jdlrobson: why was the setCurrentPage required? [14:28:49] the main reason was that toggling didn't seem to work without it [14:28:54] but... [14:29:00] it's unrelated and shouldn't even be called :| [14:29:14] (on the Main Page) [14:29:40] why is loadLocalPage called on the main page? [14:29:47] loadLocalPage is called only for the error pages [14:30:22] jdlrobson: also, setCurrentPage expects a 'page' obejct [14:30:25] jdlrobson: you're passing it a string [14:31:43] that would definitely blow up if that line was encountered :) [14:31:51] whoops [14:32:13] and commenting that out seems to have no effect on scrolling [14:32:15] or the missing sections [14:33:11] going afk for a bit so I don't punch my keyboard and injure my hand again [14:33:26] jdlrobson: try ant clean and then install to see if section expansion stops working? [14:34:11] i do ant clean by default [14:34:14] and ant uninstall [14:34:23] i don't trust ant otherwise [14:36:49] jdlrobson: so... [14:36:50] it works [14:36:53] I turned my phone off [14:36:55] then back on [14:37:09] sections expand [14:37:11] and scrolling work [14:37:12] s [14:37:17] ... [14:38:28] argsfgsgsgsgfsgsfgfsfs [14:38:35] * jdlrobson 's head explodes [14:40:01] jdlrobson: so... [14:40:09] do we merge things in now? [14:40:12] i do see the left scroll [14:40:21] yes as long as it works fine in android 2.* [14:40:24] and the flash of no content [14:40:25] oh yeah [14:40:25] that [14:40:28] * YuviPanda starts emulator [14:40:53] the flash of content is annoying [14:41:24] ill try debug that [14:41:37] also homepage - we should do something there right? [14:41:42] even if it just the logo [14:42:02] jdlrobson: i'm okay with leaving it as it is for now until we figure out what to do - which should be before we ship anything publicly [14:42:15] yep sounds good [14:42:29] jdlrobson: want to look at the 'can scroll entire app a little bit to the left' bug brion raised? [14:42:37] possibly a meta tag messup [14:42:42] if i have to :D [14:42:45] ;-) [14:42:52] im gonna grab a coffee first though [14:43:03] nah, i'll take care of the [14:43:05] *that [14:43:14] the flash of content is tougher, so :D [14:43:16] *no content [14:43:52] jdlrobson: i'm not picking up your last commit (loadCachedPage) [14:44:53] you fully fetched? [14:44:54] https://github.com/jdlrobson/WikipediaMobile/tree/api [14:45:38] jdlrobson: yeah, just cherrypicked the first two. [14:45:50] https://github.com/jdlrobson/WikipediaMobile/commit/05749b1616fec82ee0aeb5697fabaa0b99916064 as i mentioned will blow up if hit [14:45:52] plus has no effect [14:45:54] so not picking it up [14:45:59] np just make sure you get this one -> https://github.com/jdlrobson/WikipediaMobile/commit/e9c32063604ed6df9fa3ab057c1ee2db41adbcd7 [14:46:06] rather than the one i submitted before [14:46:19] (the one i submitted before just changed index.html) [14:48:27] jdlrobson: sorry, got disconnected [14:48:30] jdlrobson: so, 2.x [14:48:32] content is cut off at the bottom [14:48:36] looks like 52px [14:48:49] grr [14:49:56] can you try menu { display: none; } and see if that works (note menu not .menu) [14:49:57] jdlrobson: also, iOS seems to have an extra 52px, but i guess that's okay [14:50:02] yeah doing it [14:50:11] ioS does!?! [14:50:12] it shouldn't a [14:50:14] aarragagagaf [14:50:14] a [14:50:31] extra 52px padding [14:50:34] at end of content [14:50:36] the margin? [14:50:45] yes, I think so. [14:50:55] which version of ios? [14:50:59] 5.1 is the device [14:51:02] let me check 4.x on sim [14:51:50] jdlrobson: hahahahahahaha [14:51:50] so [14:51:52] it is fine on 4. [14:51:53] x [14:51:57] on the simulator [14:52:00] * YuviPanda checks 5.x on sim [14:52:01] yeh that's what i was testing on [14:52:20] so i think we need to add classes to the body tag so we can be more specific with our styling rules [14:52:23] we currently have android [14:52:34] jdlrobson: so 5.x on simulator also has extra margins [14:52:39] but ios-4 ios-5, android-2 classes are probably going to be useful [14:52:52] yup [14:53:23] jdlrobson: want me to add them now? [14:53:33] if you can easily that would be sweet [14:54:45] jdlrobson: i'm going to put them on body [14:55:12] android, iOS, iOS-4, iOS-5, android-2, android-3, android-4 [14:55:46] excellent [14:55:51] actually html would be better [14:56:05] (in case we want to target the html tag for some reason) [14:57:31] jdlrobson: ok [15:06:51] platform-stub doesn't seem to be executed for iOS? [15:06:54] shoot me [15:07:45] so it executes except that console.log won't work [15:07:46] fine [15:13:40] jdlrobson: done and pushed to api-move [15:14:33] note that teh DOM is not guaranteed by that time, hence am doing things without jquery. [15:14:40] on second thought that sounds stupid [15:16:03] but i guess it'll do, unless you've objections [15:22:16] jdlrobson: stepping afk for dinner. [15:22:35] I think we can merge once this css issue has been fixed :) [15:22:41] k i might not be around when you get back - ill send you a pull if i fix anything [15:23:09] jdlrobson: cool [15:23:29] jdlrobson: i'd like you to at least look at that commit though, just double checking if i'm not doing anything stupid [15:24:18] add detector classes? [15:24:24] yeah? [15:24:44] well there is an alert in there... [15:24:48] guessing that's not wanted? [15:24:55] ugh [15:24:57] fixed [15:25:11] also why not just use $("html").addClass("foo bar") [15:26:31] jdlrobson: this was outside of any DOM Ready functions so wasn't sure how to safely handle jquery objects. but saying it out feels stupid, so I'll go fix that [15:36:51] jdlrobson: fixed those. [15:36:53] * YuviPanda heads out [15:39:38] * MaxSem weeps [15:39:47] if ( !empty( $_SERVER['HTTP_APPLICATION_VERSION'] ) && [15:39:48] strpos( $_SERVER['HTTP_APPLICATION_VERSION'], 'Wikipedia Mobile' ) !== false ) { [15:39:48] self::$hideSearchBox = true; [15:39:48] if ( strpos( $_SERVER['HTTP_APPLICATION_VERSION'], 'Android' ) !== false ) { [15:39:48] self::$hideLogo = true; [16:05:00] jdlrobson: ready to caht? [16:05:01] chat* [16:05:05] yup [16:18:19] YuviPanda: hey there [16:18:39] i didn't make it into the office by 9 so i'll have to release the app and then commute in [16:23:34] * tfinc wonders if YuviPanda is still around [16:28:47] tfinc: he went afk an hour back [16:29:06] k, i'll be back on in 30min … hopefully he'll be back then [16:29:37] any new updates to MF recently? I see css bug for left justification of images resurfacing http://ta.m.wikipedia.org/ [16:30:10] or any changes to css off late ? might be a different bug as well [16:34:47] tfinc: just got back. will be back in about 40 mins. Feel free to release, I'll be available for a couple of hours after that [16:34:49] brb [17:11:09] MaxSem: what is the issue with: https://gerrit.wikimedia.org/r/#change,5158 [17:13:04] preilly, it still doesn't do what it should and I've started moving stuff into the skin anyway, I explained it to Jon on IRC [17:15:29] MaxSem: did you explain it in a PM or in this channel? [17:15:36] here [17:16:22] MaxSem: so, is he fixing it or are you? [17:16:35] MaxSem: also, what are you moving into the skin? [17:16:46] HTML generation [17:17:19] to be precise, to a BaseTemplate-derived template [17:17:52] preilly: what did we find about the varnish cache cpu issue ? [17:19:10] tfinc: it appeared to have been a SYN flood [17:19:23] tfinc: it wasn't anything on the actual application side [17:19:33] tfinc: it should be fixed now [17:19:35] someone DoSes us? [17:19:50] MaxSem: not sure [17:20:11] preilly: i see the mails about restarts [17:20:17] did we do anything else [17:20:18] ? [17:20:46] tfinc: the box had been up for 46 days [17:21:06] tfinc: did a diet-upgrade and pulled in security patches for varnish [17:21:12] tfinc: and rebooted [17:21:17] k [17:21:22] s/diet/dist [17:21:31] damn autocorrect [17:21:58] MaxSem: where are you with your BaseTemplate changes? [17:22:10] MaxSem: is there anything that I could look at now? [17:22:39] hmm, I'm far from a sensible commit [17:23:02] MaxSem: define far? [17:23:10] MaxSem: in days…. or hours [17:23:17] MaxSem, preilly: my changes to fix the cookie-related issue we saw yesterday are live on test.wikipedia.org [17:23:26] if you guys have a second, would you mind taking a look? [17:23:28] awjr: okay, I'll take a look [17:24:13] awjr: seems legit [17:24:25] booyakasha [17:25:20] MaxSem: should core have a method to get the robots policy [17:25:26] MaxSem: can you confirm that the changes fix the issue you were seeing yetserday? [17:25:54] awjr, to confirm it, I need to see it on enwiki;) [17:26:02] MaxSem: something like getRobotpolicy [17:26:25] MaxSem: ok - but you can see now that setting the mobile view cookie on test will not cause you to see the mobile view on enwiki [17:27:13] awjr, confirmed [17:27:52] ok gonna deploy [17:27:55] thanks guys [17:29:10] MaxSem: we could add a getRobotpolicy to OutputPage.php to return the values of, "private $mIndexPolicy = 'index'; and private $mFollowPolicy = 'follow';" then we wouldn't need to DOM parse those out the page content [17:29:32] MaxSem: What do you think? [17:30:08] preilly, doable, but not quite urgent [17:30:59] at least, we need a steady core rollout schedule for this [17:33:53] preilly, MaxSem: are these safe for deployment? https://gerrit.wikimedia.org/r/#change,5133, https://gerrit.wikimedia.org/r/#change,5146 [17:35:20] awjr: yeah, I think so [17:35:32] MaxSem: how do you feel about 5133? [17:36:03] MaxSem: are the titles going to be messed up? [17:36:16] preilly, they already are [17:36:24] or not? [17:36:37] MaxSem: e.g., do we need https://gerrit.wikimedia.org/r/#change,5158 fixed and deployed as well [17:36:39] * MaxSem tries to remember how they looked before [17:37:38] awjr, from my, committer's POV 5133 is a _complete_ change, not needing further commits to be deployable. I'll defer the opinion about its quality to others [17:37:59] preilly, I can fix it [17:38:29] but what's our current procedure for amending other users' patchsets? [17:39:24] MaxSem: let's have him abandon it and you make a new commit [17:40:06] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: -1; - https://gerrit.wikimedia.org/r/5158 [17:40:25] Change abandoned: preilly; "(no reason)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5158 [17:40:38] * YuviPanda looks around [17:40:41] tfinc: heya [17:40:43] released? [17:40:47] YuviPanda: there you are [17:40:56] i was waiting for you to get back [17:41:06] as i'll get sucked into meetings shortly [17:41:07] tfinc: I came back and said i'll take another 40 mins? [17:41:12] not sure if that got through :( [17:41:22] but am here now [17:41:24] tfinc: released? [17:41:31] yeah, i saw that right before leaving [17:41:35] * tfinc goes to release  [17:41:46] ok [17:42:13] awjr: so, can you wait a little bit for MaxSem to make that change for titles before deploying 5133 and 5146? [17:43:15] preilly: yes [17:43:20] awjr: okay, great! [17:43:44] released [17:44:00] YuviPanda: --^ [17:44:04] preilly: since we have to deploy from master for 1.20+, i think we should set up a staging or bleeding-edge branch as an intermediary between people's local dev and master [17:44:12] ooo [17:44:18] so that way we can keep master totally deployable [17:44:49] YuviPanda: we have our 1:1 set to 12-1 tomorrow. do you want to move it up given your new hours? [17:45:06] tfinc: that would be great [17:45:32] awjr: hmm [17:45:39] tfinc: thanks [17:46:04] preilly: the alternative i guess is being more careful about what gets merged into master, but that seems error-prone. [17:46:11] awjr: so, would we always git review to that branch then cherry pick from that branch to master ? [17:46:18] preilly yeah [17:46:32] awjr: let's hold off on that just yet [17:46:41] awjr: I want to think about it a bit more [17:46:49] preilly: for sure [17:46:55] awjr: in principal it seems like a good idea [17:47:00] preilly but im glad i dbl checked about the state of master because i almost deployed those changes [17:47:04] tfinc: tweeting. [17:47:15] awjr: well those changes are okay [17:47:22] YuviPanda: its still processing [17:47:29] YuviPanda: pickup the update .. THEN tweet [17:47:29] awjr: it's just that we also want to fix one more thing [17:47:36] right. [17:47:40] preilly: roger [17:47:49] composing, keeping in draft, pickup iPod touch [17:47:57] preilly, can you reopen https://gerrit.wikimedia.org/r/#change,5158 - I think I can recommit into it [17:48:18] awjr: Roger Wilco [17:48:37] Change restored: preilly; "(no reason)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5158 [17:48:50] MaxSem: Patch Set 2: Restored [17:49:06] New patchset: MaxSem; "Show titles for all pages but main page" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5158 [17:49:33] the ritual to do this was quite funny [17:50:52] including altering that long command line copied from Gerrit [17:51:11] MaxSem: how so? [17:51:22] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5158 [17:51:25] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5158 [17:51:53] I had to create the branch explicitly on my end [17:52:15] MaxSem: well, you should have been able to fetch and amend [17:52:30] MaxSem: why wouldn't that have worked? [17:52:46] awjr: Okay, I think we are in good shape now. [17:52:53] preilly: cool [17:54:42] brion, need your advice - I'm currently porting MobileFrontend to MW's skinning infrastructure, including QuickTemplate. however, I've doubts about QuickTremplate's usefullness: templates are there for skins to do similar things, however MF's skin will need to do everything differently [17:54:58] preilly: https://gerrit.wikimedia.org/r/5180 [17:55:20] preilly: are there any production issues right now? I'm seeing seemingly random pages hang on load on a iPhone4S (iOS 5.0) [17:55:31] i also have one tweet saying the same [17:57:05] preilly: i'm seeing the same slowdown in the simulator when loading http://en.m.wikipedia.org/wiki/Recto_and_verso [17:57:16] MaxSem, feel free to do a totally different implementation as necessary :) [17:57:32] i suspect the basic interface of 'set parameters on it, ask for html back' will still be useful [17:57:41] even with the insides changed from a simple template to something else [17:58:55] i'm getting more twitter reports of safari users seeing issues [17:59:20] brion, thanks [17:59:30] awjr: so, this http://pastebin.mozilla.org/1575109 is what you pushed right? [17:59:43] preilly it hasn't pushed yet [17:59:47] i just got all the changes up on test [17:59:49] actually [17:59:59] does php-1.20wmf1 point to test2? [18:00:07] awjr: yes [18:00:18] ok so changes from master are on both test and test2 [18:00:26] please take a look and confirm their ok-ness [18:00:38] preilly: https://twitter.com/#!/oakapplegardens/status/192311111339622400 [18:00:54] shit test2 also doesn't have a .m properly configured [18:04:04] preilly do you know which host test2 runs on? [18:04:09] i need to sync-common [18:04:48] hmm, still not seeing it [18:05:25] okay, app has made it [18:06:14] awjr: you can look at the output of the page [18:06:54] preilly eh? [18:06:57] awjr: [18:07:17] preilly where do you see that? [18:07:18] awjr: [18:07:31] awjr: actually it doesn't seem to just be one box [18:07:39] preilly oh [18:07:43] awjr: that is at the bottom of all pages [18:07:53] preilly i had no idea [18:07:54] how handy [18:07:58] awjr: ha ha [18:09:30] preilly: yeah test2 is a regular cluster wiki - it's not actually useful for staging like test. [18:09:45] so preilly MaxSem, those changes are live on test.wikipedia.org [18:09:51] please confirm before i sync to the cluster [18:10:05] [18:10:05] Dear Wikimedia Foundation, [18:10:06] The following app has been approved. The status has changed to Ready for Sale. [18:10:14] App Name: Wikipedia Mobile [18:10:14] App Version Number: 3.1.2 [18:10:16] \o/ [18:11:02] "Sale":) [18:11:29] we're going to get rich off of that. [18:12:04] hmm, try typing "foo" in search bar on test [18:12:18] sale :) [18:12:33] MaxSem: yeah, hitting 'enter' doesn't work [18:12:44] i'm pretty sure it had been working before [18:12:47] awjr, and search finds page called ";*Foobar/Special:Search/*;Foobar/;Special:Search/;foobar" [18:12:58] jdlrobson, ^^ [18:13:02] tfinc: it's bits [18:13:11] MaxSem: to be fair, that page actually exists [18:13:21] * MaxSem chokes [18:13:28] http://test.wikipedia.org/wiki/;*Foobar/Special:Search/*;Foobar/;Special:Search/;foobar [18:13:49] whats wrong with bits today that wasn't wrong yesterday ? [18:14:06] awjr, fixed:P [18:14:07] tfinc: hold on [18:14:25] MaxSem: hahaha nice work. [18:14:33] MaxSem: hitting 'enter' from the search bar is still broken though [18:15:01] also there seems to be an image missing from the search bar [18:15:43] tfinc: it's stuck on Pragma: no-cache [18:15:43] User-Agent: Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_3_3 like Mac OS X; en-us) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8J2 Safari/6533.18.5 [18:15:45] Accept: */* [18:15:46] Cache-Control: no-cache [18:15:47] Referer: http://en.m.wikipedia.org/wiki/Recto_and_verso [18:15:48] http://bits.wikimedia.org/w/extensions-1.19/MobileFrontend/javascripts/jquery-1.7.1.min.js [18:16:27] preilly: that resoure no longer exists [18:16:31] *resource [18:17:37] awjr: http://bits.wikimedia.org/w/extensions-1.19/MobileFrontend/javascripts/jquery-1.7.1.min.js [18:17:57] interesting [18:18:43] preilly why would that still exist after we synchronized a change to the cluster yesterday that should have gotten rid of it and replaced it with jquery-1.7.2.min.js? [18:19:00] 1.7.2 appears to be there as well. [18:19:46] awjr: [18:19:59] awjr: that is what the page has [18:20:07] MaxSem awjr looks fine to me [18:20:15] hitting enter gives you the top search result on search [18:20:17] i can see all images [18:20:34] jdlrobson: can you check in opera [18:20:48] Opera looks fine [18:20:49] mini mobile or desktop? [18:20:56] jdlrobon: hitting enter is not working for me in FF [18:21:13] and i am getting a js error [18:21:24] opera desktop looks fine [18:21:29] results is not defined [18:21:30] j()beta_o...4612982 (line 1) [18:21:30] D = Object { originalEvent=Event mousedown, type="mousedown", timeStamp=3246998311, more...} [18:21:30] H()jquery...4612982 (line 3) [18:21:30] c = Object { originalEvent=Event mousedown, type="mousedown", timeStamp=3246998311, more...} [18:21:30] H()jquery...4612982 (line 3) [18:21:30] a = mousedown clientX=1895, clientY=21 [18:21:31] ...ction=opensearch&limit="+x+"&namespace=0&format=xml&search="+u;l.ajax({url:url,s... [18:21:32] * jdlrobson boots firefox [18:21:51] beta or normal? [18:22:00] jdlrobson: i think i got that error when i clicked 'x' on the search box [18:22:01] beta [18:22:09] tfinc: okay, it is fixed now [18:22:24] jdlrobson: yeah, search is wonky in beta on FF on os x [18:22:30] tfinc: the resource bits.wikimedia.org/w/extensions-1.19/MobileFrontend/javascripts/jquery-1.7.1.min.js was deleted [18:23:01] tfinc: but, the mobile varnish cache had the old one bits.wikimedia.org/w/extensions-1.19/MobileFrontend/javascripts/jquery-1.7.1.min.js cached in it's response [18:23:06] yeh so on firefox i can see pressing enter doesn't work [18:23:13] but everything else looks ok and no errors [18:23:22] tfinc: so, we cleared the varnish cache for mobile and now they are in sync [18:23:28] (firefox 11) [18:24:17] jdlrobson: it appears to be working fine for me in chrome [18:24:27] jdlrobson: did you try clicking the 'x' in the search bar after entering some text? [18:24:38] im using FF 11 as well [18:25:37] jdlrobson: we have had reports from carriers that the enter key isn't working for search [18:25:46] jdlrobson: they are on opera mobile [18:26:20] Project WiktionaryMobile - Nightly builds build #88: SUCCESS in 7.6 sec: https://integration.mediawiki.org/ci/job/WiktionaryMobile%20-%20Nightly%20builds/88/ [18:26:21] ah i get results not defined there [18:26:21] * pfhayes: Trying to make buttons work on Galaxy Nexus [18:26:21] * pfhayes: More button styling for Galaxy Nexus [18:26:22] * pfhayes: Removing unnecesary style for Hide button [18:26:22] * pfhayes: Adding comment [18:27:20] Project WiktionaryMobile - Nightly builds build #89: SUCCESS in 7.3 sec: https://integration.mediawiki.org/ci/job/WiktionaryMobile%20-%20Nightly%20builds/89/ [18:27:21] * hy.cheng: disables select text on android 3.0+ [18:27:21] * hy.cheng: select text now removes, without the plugin [18:27:22] * hy.cheng: handles undefined androidversion [18:28:31] k this is weird [18:29:05] [WikipediaMobile] brion pushed 1 new commit to master: http://git.io/SsV0mA [18:29:05] [WikipediaMobile/master] (bug 36026) Allow platforms to override the page & list actions menus - Brion Vibber [18:29:15] ah no its not [18:29:19] Project WikipediaMobile - Nightly builds build #345: SUCCESS in 8.1 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/345/ [18:29:20] Brion VIBBER: (bug 36026) Allow platforms to override the page & list actions menus [18:33:49] New patchset: Jdlrobson; "make sure results defined" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5184 [18:34:06] preilly: varnish cache purge again ? [18:34:07] ugh [18:34:09] Project WiktionaryMobile - Nightly builds build #90: SUCCESS in 7.3 sec: https://integration.mediawiki.org/ci/job/WiktionaryMobile%20-%20Nightly%20builds/90/ [18:34:10] * pfhayes: disables select text on android 3.0+ [18:34:10] * pfhayes: select text now removes, without the plugin [18:34:11] * pfhayes: handles undefined androidversion [18:34:51] YuviPanda: hows the iOS rollout looking ? [18:35:01] preilly: can we hold off on the varnish cache purge [18:35:10] tfinc: looks good on my device. am monitoring twitter [18:35:11] until i get these changes deployed? (a few minutes) [18:35:15] YuviPanda: thanks [18:35:17] and refreshing reviews [18:35:30] YuviPanda: keep me updated [18:37:07] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5184 [18:37:13] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5184 [18:40:24] awjr: just purge the cache again [18:40:51] awjr: we needed it purged as safari just failed miserably [18:42:08] preilly: keep on eye on @WikimedaMoible. more weird issues [18:42:18] kk [18:42:26] tfinc: are you seeing more complaints? [18:42:28] s/on/an [18:43:18] preilly: i just saw one more but it looks browser scrolling related. it came up on my "Wikipedia mobile" search [18:43:23] let me message them and find out more details [18:44:25] tfinc: what is the search you are doing? [18:45:08] tfinc: so, like: https://twitter.com/#!/search/%40WikimediaMobile [18:49:27] preilly: this is the one that i was triaging https://twitter.com/#!/JayDeeDaz/status/192321643320852480 [18:49:31] preilly: https://gerrit.wikimedia.org/r/5185 [18:49:44] of course now that twitter actually loads i can see that he was using chrome beta on android [18:49:58] i wish tweet deck showed me that [18:50:10] tfinc: ha ha [18:50:58] awjr: can you do a git log dd00aa5a601e35765e7b53bbf193e6dca28425e6..9a46db57223e9b4a737a3e997a1dc600ebf9068b --graph [18:51:05] awjr: and post that to a paste bin [18:51:37] preilly: fail - http://pastebin.com/Pzmw2rAC [18:52:18] awjr: you need to be in the wmf/1.20wmf1 branch for that to work [18:52:33] preilly: i was [18:53:30] awjr: you sure about that? [18:53:57] preilly yeah [wmf/1.20wmf1] is the branch name [18:55:46] preilly: for the 1.20wmf1 branch, have you just been minifying the JS on fenari? [18:56:22] awjr: it should be minified in the repo [18:56:26] mmmm [18:56:32] mmagnificent: http://dpaste.com/733227/ [18:56:57] at last, write-only PHP code that is not regexps! [18:58:18] preilly: awjrichards@fenari:/home/wikipedia/common/php-1.20wmf1/extensions/MobileFrontend$ git status [18:58:18] # On branch master [18:58:18] # Untracked files: [18:58:18] # (use "git add ..." to include in what will be committed) [18:58:18] # [18:58:19] # javascripts/application.min.js [18:58:19] # javascripts/banner.min.js [18:58:20] # javascripts/beta_application.min.js [18:58:20] # javascripts/beta_opensearch.min.js [18:58:21] # javascripts/opensearch.min.js [18:58:21] # javascripts/references.min.js [18:58:22] # javascripts/toggle.min.js [18:58:30] the minified files do not exist in the repo. [18:58:52] awjr: hmm [18:59:00] awjr: they should [18:59:21] preilly iirc we had made a decision a while ago to *not* have the minified files in the repo [18:59:30] but that was before the new deployment methods [18:59:42] awjr: I wasn't aware of that [18:59:53] awjr: we have them in the svn repo [18:59:56] awjr: don't we? [19:00:32] preilly we do in the deployment branch of svn [19:00:59] New patchset: Jdlrobson; "just update address bar" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5186 [19:01:00] New patchset: Jdlrobson; "remove fennec specific hack into firefox.css" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5187 [19:01:45] tfinc: so far nothing on twitter or on the stores [19:01:52] switching between us and uk stores [19:02:26] jdlrobson: your fix is live on testwiki [19:02:27] YuviPanda: reveiws? [19:02:42] thx tfinc [19:02:47] nothing yet. [19:02:51] *awjr [19:03:14] jdlrobson: one more task assignment coming your way [19:03:45] so yes awjr now clearing works without breaking anything [19:03:51] \o/ [19:03:57] jdlrobson: does enter work now? [19:04:11] no.. did you merge that awjr - its quite recent so guessing not? [19:04:18] [WikipediaMobile] brion pushed 1 new commit to master: http://git.io/6bmwkg [19:04:18] [WikipediaMobile/master] (bug 36027) Make save/share and sources popup menus more tablet-y on Playbook - Brion Vibber [19:04:22] https://gerrit.wikimedia.org/r/5186 should fix the enter problem... i hope [19:04:27] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5186 [19:04:29] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5186 [19:04:31] Project WikipediaMobile - Nightly builds build #346: SUCCESS in 6.7 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/346/ [19:04:31] Brion VIBBER: (bug 36027) Make save/share and sources popup menus more tablet-y on Playbook [19:04:32] preilly, just how insane QuickTemplate is? http://dpaste.com/733227/ [19:04:36] so preilly in opera what exactly was reported as happening when enter was pressed? [19:04:47] jdlrobson: nothing happened [19:04:54] jdlrobson i did not [19:04:58] but i see that patrick just merged it [19:05:05] i'll get it out in a few mins [19:05:06] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5187 [19:05:08] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5187 [19:05:36] preilly is that supposed to be deployed ^? [19:05:40] jdlrobson ^ [19:05:40] nom nom nom [19:05:45] awjr: yes please [19:05:47] k [19:07:08] preilly: is this beta or normal site? [19:07:39] so preilly im not sure what to do about the lack of minified files on master. java doesn't appear to be installed on fenari, and we haven't had them in git for a long time [19:07:44] i can minify locally and scp to fenari [19:07:44] jdlrobson: normal [19:07:51] k think i have the problem [19:07:54] should be an easy fix [19:08:06] awjr: just minify and check in to master [19:09:03] / TODO: not working in opera mini 6.5 :D [19:09:12] (that's a line in opensearch.js) [19:09:30] awjr: I know that it's a pain but until we get resource loader working for JS we'll need it [19:09:49] preilly no worries [19:10:36] New patchset: awjrichards; "Adding minified JS files" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5188 [19:10:44] preilly ^ [19:11:32] preilly delay that [19:11:42] accidentally checked in jquery-1.7.2.min.min.js [19:11:43] ... [19:12:15] heh heh [19:12:50] New patchset: awjrichards; "Adding minified JS files" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5188 [19:13:03] ok preilly ^ [19:14:08] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5188 [19:14:10] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5188 [19:14:15] jdlrobson: are you working on the enter key issue? [19:14:21] yep just need to test [19:14:23] which is always a pain [19:14:25] jdlrobson: e.g., will that come quickly [19:14:31] yes 5 mins [19:15:00] jdlrobson: okay, coolio [19:15:07] awjr: you'll probably want that one too [19:15:19] awjr: and it will require a new round of minifying [19:16:48] k [19:16:56] im going to whip up some lunch real quick then. [19:16:56] brb [19:17:31] ok actually preilly it doesn't seem as easy as i thought :/ [19:21:18] best solution would be not to serve opensearch.js to opera mini users [19:21:29] a better solution is going to take a bit more time [19:22:00] back [19:24:18] so preilly awjr what should i do about the search problem? [19:24:31] jdlrobson: what is the issue? [19:24:45] tfinc: still no tweets (am tracking @ replies, plus search for anything with 'Wikipedia') or reviews (US / UK). Will be around for ~15 mins or so before crashing [19:24:46] well the main issue is that opera mini doesn't support keypress events [19:25:18] and something is stopping the form from submitting altogether [19:25:24] jdlrobson: since it is not a regression (afaik - it appears to exist in production right now anyway) let's not worry about it for this push. [19:25:45] i was just trying to push out a bug fix for cookie stuff, but turned out it was dependent on some other changes in master hence this slew of other functionality going out right now as well. [19:26:05] awjr: okay [19:26:10] the beta search actually works [19:26:32] i can push a change which detects the browser and doesn't run the code [19:26:40] but that would only be a short term fix [19:26:49] jdlrobson: would that make it work? [19:27:00] it would make the non-javascript search work yes [19:27:22] jdlrobson: is this like… a 2 minute fix? [19:27:29] yep i can push now [19:28:00] sent [19:28:21] New patchset: Jdlrobson; "don't run opensearch.js on opera mini" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5191 [19:28:33] preilly: can you merge approve/merge the changes i've pushed to the 1.20wmf1 branch? [19:29:42] jdlrobson: did you test that change on opera mini? i do not have it set up [19:30:10] http://www.opera.com/developer/tools/mini/ [19:30:17] can double check there when on test but yes i ran on mobile-geo [19:30:22] which is why it took so long [19:32:16] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5191 [19:32:18] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5191 [19:32:52] jdlrobson: can you also minify that js please? [19:33:00] we had to re-add the min js to repo [19:33:02] jdlrobson: http://i3.kym-cdn.com/photos/images/original/000/257/633/af8.jpg [19:33:30] make minified [19:33:59] awjr: roan will need to do it [19:34:07] preilly: eh? [19:34:35] awjr: in regards to awjr: preilly: can you merge approve/merge the changes i've pushed to the 1.20wmf1 branch? [19:34:50] preilly: oh sorry [19:34:53] i'm confused.. i thought we remove minified js from the repository and we built it for deployments? [19:35:08] jdlrobson we did and that worked fine for svn deployments but not for deployments from git [19:35:16] ahh ok [19:35:20] jdlrobson: yeah, what he said [19:35:32] ill push them now [19:35:56] jdlrobson: can i suggest a change to the Makefile? [19:36:25] of coure [19:36:27] *course [19:36:32] New patchset: Jdlrobson; "update minified js" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5192 [19:36:34] make clean ? [19:36:42] jdlrobson instead of listing out each individual JS file you could do: -o '.js$:.min.js' *.js [19:36:59] which will turn all js files into .min.js files [19:37:15] oh but the only problem with that is that it seems to re-minify jquery-1.7.2.min.js [19:37:22] so maybe not the best idea [19:37:27] and all the other min.js [19:37:32] i think it's better explicit as well [19:37:40] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5192 [19:37:41] jdlrobson: it actually does not do the other min.js's [19:37:43] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5192 [19:39:15] YuviPanda: did you tedx talk get posted? [19:39:34] good question. let me poke [19:40:13] poked [19:43:29] tfinc: still no reviews. [19:43:32] US/UK [19:43:35] tfinc: I think i'll crash [19:43:41] k [19:43:42] and check up on it tomorrow morning [19:44:05] did respond to a few oldish tweets, one of whom got back. No 'reviews' as such though [19:44:06] tfinc: gnite [19:44:13] jdlrobson, preilly: changes live on testwiki [19:44:36] awjr: okay, looking [19:45:35] awjr: seems okay to me [19:46:18] preilly me too [19:46:23] looks fine. enter seems to blur on opera mini but clicking search button causes a normal search [19:46:23] jdlrobson: satisfied? [19:46:27] MaxSem: what do you think? [19:46:32] so i'd say satisified [19:46:55] lgtm [19:46:55] (although search is redirecting to incubator i can assume its going to the right place) [19:47:32] jdlrobson search should not be redirecting to incubator - how can i repro? [19:48:16] http://www.opera.com/developer/tools/mini/ [19:48:27] go to http://test.wikipedia.org?useformat=mobile [19:48:36] type in foo [19:48:38] click search button [19:49:33] takes me to incubator.wikimedia.org/wiki/W/test/Foo [19:49:49] oh jdlrobson yeah, ok that is because of device detection and test not being set up right [19:49:53] ok that should be ok [19:49:58] k cool that's what i assumed [19:50:04] i'm happy then [19:50:10] jdlrobson from a desktop browser in mobile view, search does not redirect to incubator [19:50:31] correct [19:52:43] binasher: can you please flush the mobile varnish cache/ [19:52:59] it was flushed about 5 minutes ago fwiw [19:53:29] binasher yeah i know, it was flushed to fix an issue with safari (i think?) but we just did a signifigncat software push [19:53:36] er significant [19:55:00] flushed again [19:55:15] thanks binasher [19:55:22] awjr: this workflow is pretty broken though [19:55:29] binasher agreed. [19:55:30] i have solid evidence that carriers are doing their own caching [19:55:30] binasher: YES [19:55:43] flushing the varnish cache is useless for a large number of users [19:55:49] :( [19:55:56] so isn't solving anything [19:56:07] need to fix how mobilefrontend references resources [19:56:27] binasher well at least for now it allows us to ensure whether or not our changes work when we deploy [19:56:48] but it is still a crappo workflow [19:56:54] for you.. not for the public at large [19:57:12] preilly, jdlrobson, MaxSem: changes are pushed [19:57:33] binasher: what would be a sound solution? [19:58:51] lots of people are still trying to load http://bits.wikimedia.org/w/extensions-1.19/MobileFrontend/javascripts/jquery-1.7.1.min.js [19:59:11] which no longer exists [20:00:00] binasher are you suggesting deprecating changes for a release cycle or two before totally removing them? [20:01:42] we could symlink 1.7.1 to 1.7.2 [20:01:51] i don't think that solves the problem of requests to flush varnish [20:02:10] I think this is a bigger topic than we can handle right now [20:02:21] I suggest that we set-up a meeting about this [20:02:30] preilly binasher, i will be in town next week [20:02:31] and come up with a few ideas [20:02:32] you could stop putting versioned links in things in content that gets cached [20:02:46] talk to the resourceloader people [20:03:25] resource loader adds versions as well [20:03:37] e.g., http://bits.wikimedia.org/en.wikipedia.org/load.php?debug=false&lang=en&modules=jquery%2Cmediawiki&only=scripts&skin=vector&version=20120315T172939Z [20:03:47] thats a query param [20:04:09] binasher: oh, you mean version number in the jquery file name [20:04:27] binasher: that's a jquery nuance [20:04:39] binasher: nothing that we do specifically [20:05:05] I had to add it yesterday, as a weird 404 got stuck in varnishes somehow [20:05:23] MaxSem: what was the 404? [20:06:04] preilly: awjr: talk to the resourceloader devs about how to handle versioning in a way that doesn't require cache flushes [20:06:37] binasher, http://dpaste.com/732777/ [20:07:13] MaxSem: that is because that file didn't exist on bits [20:07:23] test.wikipedia.org doesn't go thru varnish [20:07:27] MaxSem: it was a new file and hadn't been sync'ed yet [20:07:34] preilly, it did and worked fine from the US [20:07:40] but not from Europe [20:08:39] MaxSem: I think you are wrong [20:08:51] MaxSem: /w/extensions-1.19 is mapped differently in Apache on test [20:09:21] MaxSem: even if the files are on Fenari they aren't in the correct path for that directory until you sync common [20:09:25] the only thing I know is that this thingie worked fine for awjr and failed for me and TS [20:09:49] MaxSem: well, that would be really really weird [20:10:05] I specifically checked on Toolserver to exclude browser issues [20:10:10] i think this is the trigger of what brought down bits / the site last night [20:10:25] binasher: all the 404's? [20:10:46] binasher: on that resource... [20:11:49] the behavior i saw when trying to request 1.7.1 from bits earlier was that it would hang indefinitely even though the apaches would return an immediate 404 [20:12:11] binasher: it would finally time out and return a 404 for me [20:12:18] binasher: but, after a great deal of time [20:12:33] we should probably name it jquery.min.js now [20:12:46] and remove the 1.7.2 [20:12:49] but a request to apache directly would result in an immediate 404 [20:12:58] yeah [20:13:40] weird [20:13:51] varnish serializes requests to a busy uncachable url by default [20:14:13] binasher: can you duplicate the behavior to a non existent resource [20:14:40] binasher: or does it have to be requested N times to cause the serialization to occur [20:17:43] New patchset: preilly; "Change name of file to jquery.min.js remove the version reference" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5195 [20:24:45] preilly: it should be duplicatable [20:38:51] New patchset: awjrichards; "Merging from master at b58aad17506cb153aecfc372c679b71e785c9585" [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5199 [20:39:25] preilly can you take a look at ^ it's merging master -> contact-us-redesign [20:39:51] i dont know why it set the topic as bug/36041 though... [21:03:14] MaxSem: [Bug 36046] New: Diablo II article title shows as Diablo II (DISPLAYTITLE issue?) [21:04:52] on it [21:11:24] New patchset: MaxSem; "Fix titles with formatting" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5203 [21:11:31] preilly, ^^^ [21:14:00] MaxSem: Topic 2012/fix-title should be the bug number right? [21:14:04] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5203 [21:14:06] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5203 [21:14:11] ehm [21:14:26] awjr: can we get https://gerrit.wikimedia.org/r/5203 out the door [21:14:36] awjr: it seems like a sort of big deal for me [21:22:47] preilly: looking [21:23:10] preilly oh you already merged it [21:23:14] preilly did you deploy it already? [21:24:30] awjr: no I didn't deploy it [21:24:37] awjr: I wanted you to do that [21:24:40] :) [21:24:44] preilly ok on it [21:25:10] awjr: cool [21:29:32] preilly: done [21:31:34] awjr: is it deployed [21:31:43] preilly: yes [21:32:26] [WikipediaMobile] brion pushed 1 new commit to master: http://git.io/LfwOdw [21:32:26] [WikipediaMobile/master] First stab at modernizing WinPhone7 project - Brion Vibber [21:32:41] Project WikipediaMobile - Nightly builds build #347: SUCCESS in 7.7 sec: https://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/347/ [21:32:41] Brion VIBBER: First stab at modernizing WinPhone7 project [21:35:29] awjr: what do you think of https://gerrit.wikimedia.org/r/#change,5195 [21:36:40] preilly: what if we left jquery-1.7.2.min.js but symlinked it to jquery.min.js? it is nice having the version # in there so you know what it is at a glance [21:37:04] although i suppose you could always just look in jquery.min.js to find out [21:38:07] meh [21:38:09] preilly im ok with it [21:39:09] awjr: okay [21:39:18] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5195 [21:39:21] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5195 [21:57:39] jdlrobson you there/ [21:57:44] yup [21:57:51] hello [21:58:13] i synchronized contact-us-redesign branch with master, just waiting for someone to approve and merge (ahem preilly) [21:58:35] sweet [21:58:43] it would be great if you could take a look and make sure i didn't destroy any of the changes you had made (it was a burly merge) [21:58:57] i'll try to sync with master more frequently to prevent such scary merges [21:59:22] anyway, now im going through MaxSem's changes and am going to start merging the work we've done [21:59:44] jdlrobson the merge commit is here: https://gerrit.wikimedia.org/r/#change,5199 [21:59:48] no worries - ill look at it with a fresh face tomorrow morning and build on it where/if needed [22:00:11] jdlrobson ok cool [22:00:22] i'll send you an update before im done today to let you know where i've left off [22:01:10] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (contact-us-redesign); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5199 [22:01:12] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5199 [22:01:23] \o/ [22:02:51] awesome [22:31:13] * MaxSem prepares a first rewritten skin pageview.... [22:33:29] MaxSem: ? [22:34:01] preilly, nevermind - it's still hours from completion [22:34:16] at least it had just one fatal :P [22:34:18] MaxSem: what are you actually doing? [22:34:33] class SkinMobileTemplate extends BaseTemplate { [22:34:46] MaxSem: Are you just moving all the HTML generation to the skin [22:34:51] yes [22:35:04] MaxSem: so nothing really performance related as such [22:35:14] yeah [22:36:07] performance should remain approximately the same [22:36:09] MaxSem is there a way to get HTMLForm to not display the form in a table/ [22:36:10] ? [22:37:29] awjr, doesn't look so - even separate field classes have getTableRow() [22:37:38] MaxSem blech [22:37:49] i am remembering why i was avoiding using HTMLForm before [22:38:25] you can use it in one column [22:38:44] MaxSem what is the advantage using HTMLForm gives us aside from some built in magic like validation? [22:39:04] MaxSem: interesting, how do i make it render in one column? [22:39:05] code becomes shorter [22:39:53] textarea field type in HTMLForm doesn't appear to support placeholer [22:40:42] awjr, see my code: by default it places label on the same row as field, but it's easy to overcome [22:41:07] MaxSem sure but it still creates a separate column (presumably for labels) [22:41:25] it's narrow enough [22:41:25] awjr: what happens if you add placeholder [22:41:34] preilly nothing [22:41:42] HTMLForm just ignores it [22:42:00] awjr: did you spell it placeholer or placeholder [22:42:10] placeholder :p [22:42:49] awjr: http://svn.wikimedia.org/viewvc/mediawiki/trunk/phase3/includes/HTMLForm.php?view=markup line 1238 [22:43:08] awjr, have you tried 'default'? [22:43:18] MaxSem default works [22:43:34] but does not give desired effect [22:44:21] preilly: line 1238 is unique to text fields not text areas [22:44:34] awjr: yeah, I know [22:45:13] awjr: probably because the placeholder attribute is new in HTML5 [22:45:27] yeah [22:45:42] i guess i could add it to HTMLForm [22:45:59] i'll have to see how jdlrobson feels about dealing with the form in a table - personally, don't like it. [22:46:04] awjr: yeah, you probably should [22:46:16] awjr: forms in tables is fine [22:47:29] preilly in principle sure, but with HTMLForm you dont really have control over the table that gets built. i think it will always create a two-column table with the leftmost column intended for labels and the other for fields [22:47:41] awjr: I was joking [22:47:53] preilly oic [22:48:05] awjr: what about extending it and replacing displaySection [22:48:51] preilly: intriguing. [22:49:58] awjr: then you could use
and and stuff [22:50:23] awjr: e.g., http://www.alistapart.com/articles/prettyaccessibleforms/ [22:50:49] awjr: some pretty stuff like: http://www.alistapart.com/d/prettyaccessibleforms/example_3/ [22:51:03] very pretty [22:51:44] awjr: yeah, I'd just try extending it [22:53:10] preilly: i'll give it a shot [22:53:52] did something happen to the opt-in or am i just doing it wrong? [22:55:12] it works! http://i39.tinypic.com/14dpyqu.png XD [22:55:31] hahahaha [22:55:43] heatherw, Special:MobileOptions/BetaOptIn [22:55:44] poopapedia [22:55:46] i like it. [22:56:45] S:MO/BOI [23:04:30] awjr: so, class MobileHTMLForm extends HTMLForm… seems to work fine [23:04:48] awjr: and just overriding the displaySection method [23:05:29] preilly aye [23:05:39] just add support for placeholder attrib in core [23:06:48] awjr: also, you can change that in your derived class as well [23:06:57] awjr: that way we don't have to wait for the core change to go out [23:07:08] thanks MaxSem it's not working on the simulator [23:07:18] preilly: for sure. but that's a silly thing not to have in core [23:07:25] awjr: yeah, totally [23:07:31] heatherw, ?? [23:07:34] preilly but i'll do it locally for MF as well [23:07:35] awjr: I think doing both is good for now [23:07:39] yah [23:08:37] MaxSem: i go there, i choose optin, the page reloads with no change [23:09:01] what if you hard-refresh? [23:11:42] First iOS app review is in 5 star :) [23:11:55] now we just need more of those [23:16:04] tfinc: nice! [23:23:00] curses. HTMLForm maps fields types -> handler classes in a static property [23:23:17] awjr: but, to self right? [23:26:04] awjr, you can specify class directly [23:26:28] preilly yes but my overloading of $typeMappings does not appear to be working [23:26:32] getting the last 30% of android users to upgrade to the new app is a snail race [23:26:44] i wonder if other apps have this problem [23:26:46] preilly wouldn't i need to use to late static bindings to make it work right? [23:26:48] this is why i like the web [23:26:50] no stupid versions [23:26:56] * tfinc says no to versions [23:33:30] preilly oops nm it is working i was just doing something dumb. [23:49:34] preilly: the table-ness of the HTMLForm is built in to every single form field type. to override it, i will have to overload every field type we will use [23:52:01] awjr: how so? [23:52:27] preilly well, each field type is responsible for building it's own table row [23:52:48] awjr: hence overriding getTableRow [23:53:27] preilly right, but that will have to be overridden for each field type we use [23:53:56] awjr: I don't follow [23:54:14] preilly getTableRow is implemented for each field type [23:54:45] preilly if we want to make HTMLForm return us a from in divs rather than in a table, we will have to override getTableRow for each form element type we want to use [23:55:44] awjr: I don't think so [23:56:17] awjr: I only see three references to 'td' and four to 'tr' [23:56:25] awjr: I must be missing something [23:58:10] preilly each field type invokes getTableRow(), a few of them override it themselves [23:58:47] awjr: yeah, but that doesn't seem like a big deal to me [23:59:01] awjr: there are only six references to it [23:59:32] preilly getTableRow() gets called for every field type. so if we want to override it, we have to create our own field types.. [23:59:47] brb