[00:10:15] New patchset: Jdlrobson; "add hamburger icon to special watchlist page" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/43469 [00:10:19] MaxSem: awjr entering exhaustion but hopefully this is cool ^ [00:13:35] why don't i see MobileFrontend under 'extensions' on either http://www.mediawiki.org/wiki/MediaWiki_1.21/wmf7 or http://www.mediawiki.org/wiki/MediaWiki_1.21/wmf8 ? [00:13:37] Reedy ^? [00:13:56] or ZeroRatedMobileAccess, for that matter [00:14:04] should i be worried? [00:16:49] Usually means that there has been no changes to the extension.. [00:17:20] For Zero [00:17:21] commit 2e9583ba015966264fa41e6a23784d30fd5a46fb [00:17:22] Author: Translation updater bot [00:17:22] Date: Fri Jan 4 21:31:49 2013 +0000 [00:17:24] on wmf8 [00:17:29] so last changed in.. wmf6? [00:17:45] New patchset: MaxSem; "add hamburger icon to special watchlist page" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/43469 [00:17:47] Reedy, awjr ahaha - that's because MF uses a dedicated deployment branch [00:17:58] commit 1ded0894cdbc089370a58ba41aeb4db53e5c7345 [00:17:58] Author: Translation updater bot [00:17:58] Date: Thu Jan 10 21:08:43 2013 +0000 [00:18:02] what about zero though? i pushed a change to wmf7 on zero last week [00:18:18] you pushed, you log it:P [00:18:23] Git log says otherwise [00:18:25] oh but i guess that was after the wmf7 branch was created [00:18:34] Did you push the same change to wmf6? [00:18:40] i think so [00:18:45] if so, the release notes will be there [00:18:45] i'd have to go back and look, but likely [00:19:00] ok, i'll stop worrying, sorry for the false alarm :) [00:19:02] http://www.mediawiki.org/wiki/MediaWiki_1.21/wmf6#ZeroRatedMobileAccess [00:19:03] heh [00:19:18] Basically, the revision should be listed where it first showed up [00:19:19] MaxSem are you looking at 43469? [00:19:38] awjr, just made a little tweak [00:19:41] if it's new in wmf8 and someone backports it wmf8, the change logs (when updated) will reflect that it was in wmf7 at some point [00:19:53] ok that makes sense [00:20:33] thanks Reedy :) [00:24:17] hmmm master is broken [00:24:20] for MobileFrontend [00:24:26] im getting js errors in non-beta [00:25:09] although beta seems OK [00:26:19] MaxSem: ihttps://gerrit.wikimedia.org/r/#/c/43469/ is that ready to go? [00:26:27] jdlrobson: absolutely not [00:26:31] the patchset is hella busted [00:26:35] how so? [00:26:39] im making screenshots [00:27:10] awjr: what are the symptoms [00:27:14] jdlrobson, it needs a frontend review by Juliusz anyway [00:27:23] MaxSem: i'm sitting with him now [00:28:01] jdlrobson: just check out the patchset and try it out - it breaks prod further and is obviously hella busted in beta [00:28:05] eek why are articles rendering as overlays now Max? [00:28:08] you'll see, it's hard to describe [00:28:08] they weren't when i submitted it [00:28:33] check out patch set 3 awjr [00:28:38] it's not busted [00:29:00] gee, did I break it? [00:29:15] New review: awjrichards; "There is something seriously wrong with this patchset. I haven't debugged yet, but check it out and ..." [mediawiki/extensions/MobileFrontend] (master); V: 0 C: -1; - https://gerrit.wikimedia.org/r/43469 [00:29:24] what did you change MaxSem ? [00:29:43] jdlrobson, simplified the isOverlay condition [00:29:49] * jdlrobson looks [00:30:01] missed the negation somewhere I guess [00:30:02] jdlrobson: there are js errors in production in master (and in that patchset) [00:30:03] gawd i wish there was an easier way to check diffs between patchset [00:30:16] jdlrobson: Uncaught bad key given: mobile-frontend-drawer-cancel [00:30:20] srsly [00:31:09] awjr: that seems to be a problem with master [00:31:11] not this patchset [00:31:23] jdlrobson: yeah, the problem is in master too - you're right [00:31:32] regression! [00:31:38] i'll look into it [00:31:41] looks like I need to sleep if I break stuff [00:31:46] :p [00:31:56] i can try to take a look at the diff between the patchsets and see what max was trying to do [00:32:07] because patchset 3 seems to at least function correctly [00:32:14] MaxSem: go to bed! [00:32:18] basically, I did $tpl->set( 'isOverlay', $context->isOverlay() && $specialPage ); [00:32:34] which looks damn plausibly [00:33:03] wrong logic Max [00:33:17] it's if !specialPage isOverlay = false [00:33:54] also go to bed MaxSem :-) [00:34:02] arggh [00:34:05] going:) [00:34:07] $context->isOverlay() === fase && !$specialPage ? [00:34:10] derp [00:34:10] good night [00:34:11] false [00:34:14] good night max [00:35:04] i don't like doing that kind of boolean check as a function argument - it's hard to spot/review/etc [00:35:04] awjr: i got a fix [00:35:07] cool [00:36:02] New patchset: JGonera; "update empty message (307)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/42251 [00:38:30] New patchset: Jdlrobson; "fix message support for stable (regression)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/44367 [00:38:31] awjr: ^ rather important fix [00:38:37] jdlrobson: on it [00:43:16] New review: awjrichards; "Still getting a JS error on the settings page - 'Uncaught TypeError: Cannot call method 'registerMod..." [mediawiki/extensions/MobileFrontend] (master); V: 0 C: -1; - https://gerrit.wikimedia.org/r/44367 [00:43:21] jdlrobson: ^ [00:43:28] huhzz [00:43:35] jdlrobson: were you fixing https://gerrit.wikimedia.org/r/#/c/43469/ or should i take a look? [00:43:54] nope wasn't looking at 43469 [00:43:58] not sure what Max was doing [00:44:34] mm that's a weird MobileOptions error.. [00:44:57] akkk i see what's happened [00:46:32] awjr: so there was a reason I was using wgResourceModules over $resourceLoader->register( $id, $module ); [00:49:14] our RL seems to be broken somehwat [00:49:22] oh? [00:50:11] you can't register modules that need to be loaded at the bottom for some reason [00:51:46] O_O [00:51:51] is that a RL bug? [00:51:59] or something whack with how we're doing things? [00:53:30] awjr: i'm not sure what's happened [00:53:38] but my patch worked and it's been amended [00:53:48] i have to go in 7 mins :/ [00:53:51] actually, i think i remember ori mentioning this when i was helping him debug some event logging stuff last week [00:53:56] i'll see if I can find a fix [00:53:59] yes ori did come across this [00:53:59] b [00:54:04] but no ones had a chance to look into it [00:54:16] jdlrobson: i can try to clean that up since i implemented those changes [00:54:27] if you wouldn't mind awjr [00:54:31] not at all [00:54:42] also it seems that the js is being loaded twice [00:55:01] 2 modules are added in the loop [00:55:06] one has just the javascript [00:55:11] the other has the css and the javascript of the one before [00:55:43] ok, i'll take a look [00:55:45] anyway i've got to go, i'll look into this first thing tomorrow if we haven't made any progress [00:56:12] sounds good, i'll do what i can right now [00:56:27] awjr: should merge https://gerrit.wikimedia.org/r/#/c/44367/ though [00:56:31] that fixes the other problem [00:56:47] anyway chat tomo [00:57:10] k have a good one [00:57:41] wait a minute awjr [00:57:45] why are we using registerModule at all [00:57:49] we don't have RL in stable [00:57:56] that shouldn't work.. [00:58:08] we *should* be using wgResourceModules as was originally intended [00:58:38] anyway must dash [01:19:46] New review: awjrichards; "Ok, aforementioned JS error is a result of regression introduced in Change-Id: I891f9b4fc1f4d06db0dd..." [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/44367 [01:19:46] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/44367 [01:20:48] i am so in love with git stash [01:53:12] awjr: did you just find about git stash ? [01:53:30] tfinc no i've been using it for a long time, just having a renewed love affair [02:36:51] New patchset: awjrichards; "add hamburger icon to special watchlist page" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/43469 [02:38:59] New review: awjrichards; "the patchset i just uploaded should fix the issue. however, i assume that Max introduced the previou..." [mediawiki/extensions/MobileFrontend] (master); V: 0 C: 0; - https://gerrit.wikimedia.org/r/43469 [02:42:36] gnight folks [06:07:13] New patchset: Jdlrobson; "don't use register (regression)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/44381 [06:10:09] New patchset: Jdlrobson; "add hamburger icon to special watchlist page" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/43469 [06:11:03] New review: Jdlrobson; "fixed whitespace problem in awjr patchset and merged" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/43469 [06:11:21] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/43469 [07:30:22] New review: Siebrand; "Tiny bit of trailing ws remaining." [mediawiki/extensions/MobileFrontend] (master); V: 0 C: 0; - https://gerrit.wikimedia.org/r/42251 [08:05:26] awjr: re git stash - the --include-untracked option can be wonderfully useful at times :) [09:03:41] * awjr waves [16:49:53] * awjr waves [16:52:54] yo [16:53:35] how's it giong MaxSem [16:53:38] *going [16:54:27] investigating GeoData weirdness [16:54:34] what's going on now? [16:54:53] also, my MF on labs commit was too naive:) [16:55:10] doh! [16:56:17] YuviPanda: i haven't looked closely at this yet, but discovered a module for ZNC that handles push notifications to mobile devices: http://wiki.znc.in/Push [17:05:39] MaxSem can you take a look at https://gerrit.wikimedia.org/r/#/c/43469/ ? I think we've got it all fixed now [17:06:18] lol, by reverting my change?:P [17:07:04] yeah, read the changes - comments explain why [17:08:41] actually that might not be ok, things still seem broken in non-beta [17:09:03] bgggg [17:09:05] oh MaxSem, apparently jon merged that already anyway [17:09:16] and prod is fine [17:09:20] * awjr needs hella more coffee [17:09:30] riiight, and I didn't notice it too:) [17:31:46] awjr: hey [17:31:47] just saw you remail [17:31:50] awjr: i investigated it [17:31:52] it is sortof useless [17:31:56] oh no [17:32:01] it can give you a notification of when someone pings you [17:32:02] but you've no way to actually reply [17:32:08] you'll need to boot up your IRC client [17:32:08] that's too bad, i had high hopes [17:32:13] and then go fidn that channel [17:32:14] and reply [17:32:51] interesting [17:33:01] but the paid android app lets you reply? [17:34:06] awjr yes [17:34:15] awjr: so, that is *not* running the IRC protocol between your phone and the server [17:34:42] Change abandoned: awjrichards; "I emailed a patch for this fix to Jon for review because I wasn't 100% sure it worked (it did, there..." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/44381 [17:34:56] eeeenteresting [17:35:19] Indeed [17:35:23] so what would be required [17:35:27] is for someone to take one of the opensource Android clients [17:35:31] and them modify it [17:35:37] to accept a form of push [17:35:41] and then modify znc push to support that [17:35:48] it should, in all, take a week, I suspect [17:36:05] also, I just switched back to a normal keyboard tham my macbook air's and wow does it make a differenc [17:36:05] e [17:36:23] I think my years of working from random places has done a lot of disservice to my body [17:36:34] (Going to an orthopedic this weekend. Hopefully I'll love for more than 5 more years :P) [17:37:04] YuviPanda: or make a lightweight irc client oriented specifically around responding to push notifications! [17:37:12] awjr: or that, yeah [17:37:18] awjr: but IRC as a protocol is badly suited for mobile [17:37:23] yah [17:37:40] needing to have a constant connection open kills battery [17:38:04] New patchset: awjrichards; "Returns mobile special page module loading to wgResourceModules" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/44442 [17:38:50] YuviPanda: but if it's only around responding to push notifications, and is iintended to work with a bouncer like znc… that might make it more mobile-friendly [17:38:55] although it's kinduva funky hack [17:39:10] awjr: sure, but ZNC talks IRC (the protocol) [17:39:13] on both sides [17:39:23] so unless we use some other protocol... that's not gonna work [17:39:29] i see what you mean [17:39:34] a really neat hack would involve using an IRC -> Jabber bridge [17:39:43] then you get a lot of it for free :P [17:40:00] that doesn't seem that far fetched to me [17:42:06] New patchset: awjrichards; "Returns mobile special page module loading to wgResourceModules" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/44442 [17:42:09] MaxSem can you take a look at ^ [17:42:15] that needs to go out today [17:42:18] already looking [17:42:35] thnx [17:43:29] mmm, I don't like this [17:43:43] extra cycles on every request [17:44:42] compared to doing it in the hook you mean? [17:44:47] yes [17:44:56] yeah i thought about that but couldn't think of a good way around it [17:45:05] short of just defining the special page modules the normal awy. [17:45:07] *way [17:45:46] for example, it gets executed on API calls which shouldn't care about this [17:46:18] New review: awjrichards; "Replaced by Change-Id: Ie1e088732a11b3fd101a96d2e1959bf7598a35d6" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/44381 [17:46:32] MaxSem can you think of an alternative? [17:46:51] debug why it doesn't work in a hook?:) [17:47:21] heh, yeah we need to figure out why RL is f*cked [17:47:36] actually, that's not the problem in this case [17:47:38] because it's a RL [17:47:53] the real problem is how we're dealing with RL in non-beta [17:48:04] the skin attaches the resources before that hook runs [17:48:33] i guess another possibility is to creaete a hook that runs in the skin [17:49:16] well, just make the RL load modules quickeer [17:49:30] you mean before the skin does it's thing? [17:49:43] fixing this in core in the next 11 minutes is not going to happen :p [17:50:00] in skin call ResourceLoader::getModules() or how is it called... [17:50:31] that is what i had it doing before [17:50:59] doesn't work? [17:51:19] but if you load the modules using $resrouceLoader->registerModule() or whatever in the resourceLoaderRegisterModules hook, the modules get loaded /after/ the skin does its thing [17:51:32] which is why that approach does not work for us in non-beta [17:52:29] so, SkinMobile::attachResources() runs before the ResourceLoaderRegisterModules hook does [17:53:05] i think we have two options if this is to get deployed today (which it kinda needs to): [17:53:21] 1) don't worry about it for now - merge it, and fix it later (but DEFINITELY fix it asap) [17:53:35] hahaha [17:53:44] 2) create a hook in the skin for fetching any last-minute modules [17:54:08] looks like you just need to call OutputPage::getResourceLoader() before adding these modules [17:54:08] actually a third solution would be to define all of the SpecialPage modules the typical way, rather than trying to dynamically load them [17:54:28] this will call this hook [17:54:32] 0_0 [17:54:37] lemme see... [17:55:00] kludge on a kludge on a kludge [17:55:08] welcome to MF! [17:57:07] MaxSem so just calling getResourceLoader() will cause the hook to run? or … ? [17:57:16] yes [17:57:21] interesting [17:57:26] funky hack. [17:57:41] not the first, not the last [18:01:06] oh, now i remember the OTHER problem [18:01:14] because that didn't totally work :p [18:01:30] using $resourceLoader->registerModule() for some reason won't load bottom scripts [18:01:41] im not sure of the details (i don't htink it's totally be debugged yet) [18:01:55] MaxSem ^ [18:02:21] whee [18:02:30] so... [18:02:33] doesn't matter cause we're past the deadline [18:02:45] we can't deploy what's in master without fixing this [18:03:03] so i think jcmish will make an exception, provided ew get this sorted immediately [18:03:36] so MaxSem, back to my original options - what do you think? [18:03:37] okay, let's merge whatever we have but create a blocking bug for it [18:03:49] and try to actually fix it [18:04:04] sounds good. let's try to get it fixed for next deployment [18:05:18] you've a typo in comment [18:05:25] ahahahahaha [18:05:32] shall i refer you to the epic thread on wikitech-l? [18:05:40] what is it? [18:05:42] i'll fix [18:05:46] invocatin [18:07:28] New patchset: awjrichards; "Returns mobile special page module loading to wgResourceModules" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/44442 [18:07:39] MaxSem ^ [18:07:54] Change merged: MaxSem; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/44442 [18:09:34] thanks MaxSem [18:10:06] MaxSem are you creating a bug or shall i? [18:10:29] you're more familiar with it, so I'd appreciate if you did it [18:10:34] sure thang [18:11:22] i am at least relieved to see master not throwing errors all over the place, and that i can now get out of the watchlist view :) [18:34:52] awjr: did you see 44068? [18:35:06] fatal crash in Chrome when clicking on the footer? [18:35:12] jcmish: bug? [18:35:18] yessir [18:35:22] no - looking [18:35:27] I'm gonna try to repro but just thought I'd ask [18:35:39] jcmish: also, we squeezed a changeset in 8 minutes after cutoff, but it's critical [18:35:45] https://gerrit.wikimedia.org/r/#/c/44442/ [18:36:19] hmm [18:36:31] that might be a problem with his local set up [18:36:50] because it is not happening for me in master [18:36:51] tfinc: I got the book :) [18:38:45] today seems to be spotty internet day [18:38:45] sigh [18:39:01] tfinc: Hello! just got off a call with howie. [18:39:18] MaxSem: https://office.wikimedia.org/wiki/Goals/2012-2013/Engineering/Max_Semenik [18:41:32] notnarayan: are you on the CIS wifi again? Seems spotty :P [18:41:46] no sir [18:41:54] YuviPanda: airtel 4G [18:42:01] ah :) [18:42:10] right. I'm still on Vodafone 3G [18:50:01] YuviPanda: i will be working with you brion and vibha soon :) [18:50:46] notnarayan: :D [18:51:26] hehe I know saw that one awjr [18:51:44] :) will you let it slide this once? [19:00:21] somebody asked me the other day for Firefox os app screenshots but I'm blanking on who [19:00:39] http://vimium.github.com/ [19:00:44] if you guys use chrome / chromium [19:00:47] and vim [19:00:48] *amazing* [19:01:01] sweet! [19:01:13] heh vim [19:01:26] is there an emacs version? [19:01:27] awjr: the 'o' and 'f' commands implementation sortof blew me away [19:01:29] yes [19:01:34] lol [19:02:14] http://www.fastcompany.com/3004896/software-developer-caught-outsourcing-his-own-job-china?partner=rss [19:02:17] awjr: actually I can't find one :P [19:02:19] did you guys see this? [19:02:23] jcmish: yes :P [19:02:28] people on Reddit are asking for an AMA from him :P [19:02:39] YuviPanda: I'm totally enjoying it [19:02:40] haha [19:02:47] :D [19:02:53] jcmish: yeah, and now im worried people are going to find me out :| [19:03:00] hahahahhaha [19:03:08] awjr: I was just thinking dangit [19:03:14] ehehhehe [19:03:14] I should have thought of guys [19:03:15] yeah, nobody knows I outsource my work to an Indian dude. [19:03:23] whoops. [19:03:26] HAHHAHAHA [19:03:34] dude YuviPanda, 'f' is AWESOME [19:03:40] awjr: try o [19:03:55] it's what firefox's 'aweseome bar' was suppsoed to be [19:04:04] o is less exciting (i always command-l for that) [19:04:12] but still sweet :) [19:04:14] hmm, I like this better than cmd-l [19:04:22] awjr: better still? You can customize keybindings :P [19:04:32] this is amazing, thanks for the tip [19:04:41] awjr: imagine navigating gerrit like you navigate vim [19:04:51] YuviPanda: you can, actually [19:04:56] oh :P [19:04:57] at least, a little bit [19:04:58] i didn't know [19:05:09] I think I've zero accepted patches in Gerrit :P [19:05:14] oh wait, you used to be able to… [19:05:35] hahaha [19:05:46] i wonder if the chrome plugin broke it [19:05:54] awjr: possible [19:06:00] there's a way to add exceptions [19:06:02] YuviPanda: well, in gerrit, press '?' [19:06:11] default exceptions for all of gmail for example [19:06:19] you can see the shortcuts, very vim-like [19:06:24] ah [19:06:24] nice [19:06:26] to me, it's the saving grace of gerrit [19:06:31] heh :P [19:06:49] awjr: one of the things the guys at the hackathon were working on was a gadget to add vim like keybindings to wiki [19:07:20] @_@ [19:07:25] remember when software came with keyboard overlays for your function keys? [19:07:29] YES! [19:07:36] dude [19:07:39] brion: no :P [19:07:39] good ol days [19:07:44] young'n [19:07:47] indeed [19:07:53] child! [19:08:17] awjr: I've always been using twitter from inside vim itself, and wordpress too. Now with this, things are a lot more complete :D [19:08:39] awjr: I wish it supported marks, registers and macros though [19:08:53] use Emacs [19:09:11] brion: nah, I already have a decent operating system :P [19:09:19] haha [19:09:27] lol [19:09:37] brion: I used to use emacs [19:09:39] switched [19:09:58] http://yuvi.in/blog/moving-to-vim-from-emacs.html [19:10:23] brion: you use emacs? [19:10:30] YuviPanda: i can't figure out what 'i' actually does in chrome... [19:10:36] awjr: me neither yet [19:10:38] I hate emacs [19:10:39] according to docs, 'enter insert mode' but… ?! [19:10:40] except annoy you greatly [19:10:43] YuviPanda: is this a rare thur that you join us? [19:10:52] tfinc: yes. I've no college today. [19:10:57] I use vi when I must, bbedit when I can ;) [19:11:00] YuviPanda: can that be true for ever ? [19:11:09] vi forever! [19:11:15] s/vi/vim [19:11:15] have more funeral stuff to do tomorrow, so will work a little bit today and a little bit tomorrow :P [19:11:17] tfinc: agreed :P [19:11:25] tfinc: saw the link I put up earlier that awjr is raving about? :P [19:11:32] the outsourcing guy? [19:11:36] tfinc: no not that [19:11:39] that was jcmish' link [19:11:40] resend [19:11:48]