[00:04:07] awjr: never mind [00:07:02] preilly: i think this is the only other rev not in the deployment branch: https://www.mediawiki.org/wiki/Special:Code/MediaWiki/114318 [00:11:17] gnite everyone. [00:25:48] awjr: okay, beta logo is fixed [00:32:05] preilly: word [00:32:09] nice one [00:32:14] ok'd your revs [00:43:30] tfinc: do we have any numbers on how many people actually click 'Mobile view' from the desktop version of the site? [00:45:37] awjr: add a call like wfIncrStats( 'mobile view' ); [00:45:54] awjr: and we would add up with a pretty graph [00:46:14] oo fancy [00:46:45] like … if ( $useFromat=='mobile') wfIncrStats( 'mobile view' )? [00:46:51] but spelled right? [00:48:20] mobile.view [00:48:33] but, yeah — that's the idea [00:48:42] cool [00:48:49] preilly: where can i see graphs? [00:49:33] what's the wap format parameter ? http://en.wikipedia.org/wiki/Redox?useformat=mobile-wap doesn't work [00:49:45] awjr: https://graphite.wikimedia.org/dashboard/ [00:50:01] preilly: username/pw? [00:50:30] tfinc: hmm useformat=mobile-wap should work [00:50:33] awjr: use your labs account [00:51:56] yup, its broken [00:51:59] i can't get it to work [00:52:07] >_< [00:52:08] one sec [00:53:38] tfinc: i guess im not actually sure what mobile-wap is supposed to look like, but when i do useformat=mobile-wap i see something that looks like the mobile view [00:53:45] is that desired behavior or is that incorrect? [00:54:02] what should happen is that your output is broken up into "cards" [00:54:14] and you page through each article [00:55:54] pretty much its breaking up long articles as old phones browser only have so much ram for each page [00:56:01] so you have to break it down even more for them [00:56:15] tfinc i see where that happens - trying to figure out why it's busted [00:56:30] i think preilly is looking at that as well [00:56:34] so do coordinate [00:57:23] preilly: it doesnt look like 'useformat' gets honored around this: $this->contentFormat = self::parseContentFormat( self::$device['view_format'] ); [00:58:50] instead it just attempts to determine a format from the header if the device doesn't otherwise match. in other words, there doesn't appear to be a way to manually override [00:59:05] awjr: hmm [01:01:22] $this->contentFormat = self::parseContentFormat( ( $this->getUseFormat() == 'mobile-wap' ) ? 'mobile-wap' : self::$device['view_format'] ); [01:01:27] preilly that seems to work ^ [01:01:40] awjr: yeah [01:02:01] shall i make it less ugly and commit? [01:02:18] awjr: probably [01:02:23] awjr: and we should push it too [01:04:55] preilly: https://www.mediawiki.org/wiki/Special:Code/MediaWiki/114341 [01:05:40] awjr: ok'ed [01:06:12] pushing now [01:06:39] ok sync done [01:06:57] awjr: nice [01:07:10] tfinc: try now [01:07:45] awjr: should we get better about updating const VERSION = '0.6.1'; [01:07:54] awjr: or, should we just remove it? [01:08:09] heh we should probably just remove it [01:08:31] unless, preilly, there is a good version numbering scheme? [01:08:50] i feel like in this case it's kind of pointless, but i dunno [01:11:28] awjr: yeah, I'm on the fence [01:23:07] preilly: let's just ditch it. i imagine we'll just keep forgetting about it and if we can't think of a good reason to keep it, why bother [01:23:20] ok i gotta jet [01:23:22] talk to you all tomorrow [01:24:35] are any of you noticing pages more often then not being right aligned [01:24:36] ? [01:26:25] for instance. "Boston Red Sox" [01:32:53] * tfinc goes back to testing [14:17:45] yuvipanda: that bug is a nightmare [14:18:02] jdlrobson: yes, I agree [14:18:13] jdlrobson: it works fine on iOS [14:18:40] i noticed that -webkit-keyframe support in android is only limited to 2 steps [14:18:46] e.g. you can't do 0%, 50%, 100% [14:18:51] oooh [14:18:51] suggesting it is incomplete [14:20:05] jdlrobson: hmm, so we *have* to stick with the gif on android? [14:20:17] jdlrobson: and possibly override with the anim for desktop/iOS? [14:20:44] well the reason for not using the git was some androids dont support right? https://bugzilla.wikimedia.org/show_bug.cgi?id= [14:20:53] yes [14:20:57] bad link.. https://bugzilla.wikimedia.org/show_bug.cgi?id=33198 [14:21:02] so i guess older androids are fucked anyway? [14:21:11] :) [14:21:15] unless we figure out a way to do an animation with just two steps [14:21:29] well the animation works with 2 steps [14:21:41] as you say 0% rotate 0deg and 100% rotate 360deg [14:22:08] ahh [14:22:13] ill have a bit more of a think about it [14:22:17] jdlrobson: so that would actually do a non-choppy rotate? [14:26:33] yuvipanda, can you review jdlrobson's commit while Roan still sleeps? :P [14:26:58] MaxSem: on svn? I don't know enough MF to do that :P [14:27:07] MaxSem: you can, no? :) [14:27:10] MaxSem... yeh I read the code freeze email a bit late.. :S [14:27:18] i mailed him to say sorry.. was that a mistake :) [14:27:37] yuvipanda, it's mostly JS [14:28:05] and really minor MaxSem yuvipanda [14:28:26] oh heh [14:29:13] :) [14:29:16] reviewed 2 easy ones, not confident about 114360 [14:29:23] git from today, I guess? [14:30:23] from several hours from now [14:30:43] jdlrobson: you have an empty 'init' function [14:30:44] ? [14:30:51] i know.. for cleaner diffs [14:31:09] and i expect stuff will get moved their soon anyhow [14:31:23] cleaner diffs how? [14:31:49] ohh i remember now [14:31:56] its becasue the tests referenced the init function [14:32:05] ah [14:32:11] thus if I'd had to remove them there as well... [14:32:30] ill probably move initClearSearch into there [14:32:34] hmm [14:33:11] looks okay to me in terms of style, but i'm not confident enough of MF internals to 'ok' it. [14:33:20] note that i've never done MW CR before, so maybe i'm overthinking it [14:34:54] yuvipanda: im still stumped on the keyframes thing - its a bit of a bummer [14:35:10] jdlrobson: sigh. Android might end up being the IE6 of mobile [14:39:18] * jdlrobson needs some air [14:39:36] * yuvipanda gives jdlrobson his macbook air [16:09:55] greetings all [16:09:56] yuvipanda: ready to check in? [16:10:02] hi tfinc [16:10:09] tfinc: heya [16:10:15] yes [16:10:20] hmm, did you get that email that I sent a couple of days ago re my UDID changing? [16:10:36] Thehelpfulone: we'll make sure it's included in the RC that we'll put out today :) [16:11:05] great thanks [16:11:33] * yuvipanda continues with his 5.1 download [16:11:39] just 3 hours left! [16:11:39] ugh [16:11:46] 3 hours?! [16:11:54] use the delta software update from the device yuvipanda [16:11:59] oh? [16:12:01] that's what was so good about 5.0 [16:12:08] yes, settings - general - software update [16:14:54] yuvipanda: will the new app also support the "resolutionary" screen on the new iPad? [16:15:11] Thehelpfulone: we should :) [16:59:50] check in time .. [17:20:05] preilly, around? [17:21:37] sweet, so one of my usb ports is for internet and the other is updating my iPod (for 5.1 testing) so no android testing for me [17:21:38] grr [17:22:19] YuviPanda: I'm here btw :) [17:23:04] Jarry1250: heya [17:23:13] what ddo you mean by 'history doens't work'? [17:23:24] Jarry1250: thanks for the testing, btw :D [17:23:57] It's just blank for me. [17:24:14] Jarry1250: at all times? [17:24:18] even after navigating a few pages? [17:24:33] Seemed to be, I'll try again though. [17:25:26] Yup, blank. I'm thinking it could be relying on some sort of "page loading complete" hook that never fires? [17:26:08] Jarry1250: does the loading animation ever go away for you? [17:27:05] It's not an animation (just a still of what is clearly intended to be an animation), and it never goes away. [17:27:52] :( [17:28:00] Jarry1250: ahhh [17:28:04] Jarry1250: what device are you on? [17:28:57] MaxSem: yes [17:29:04] Samsung Galaxy Ace, running Android 2.3.3 [17:29:52] Jarry1250: hmm, maybe something broke today. let me give you a slightly older build.. [17:30:04] preilly, first 2 questions: does it require some specific vumi version, cause latest (0.3.1) fails even to install via easy_install [17:30:14] Jarry1250: (http://bit.ly/wpandroidbeta114) [17:30:15] 2) how to run tests for this? [17:30:17] can you try this? [17:30:19] kk, it'll take a couple of minutes to download though :P [17:30:41] MaxSem: talk to jerith [17:30:53] MaxSem: he can walk you through everything [17:30:58] okay [17:31:09] MaxSem: can you take a look at the VM? [17:31:18] soooo, jerith... [17:31:35] logging in [17:32:27] what to do in vumi-gw1? at least where it is installed? [17:33:02] Jarry1250: :) [17:33:49] YuviPanda: Downloading, 7%... [17:34:04] awjr: YuviPanda|afk : MaxSem : preilly : so jon sent his status update to mobile-l instead of mobile-tech .. how do you guys feel about that? i'd like to standardize on one so that we can keep it clear [17:34:24] MaxSem: it's in the default path [17:34:25] tfinc: it probably should've gone to mobile-tech [17:34:26] to me its a bit weird to send it to mobile-l as it sets a precedent of always sending it to mobile l [17:34:44] tfinc: mobile-tech [17:34:47] mobile-l seeems to be for devs+users, mobile-tech seems more appropriate [17:34:54] ok.. thats what i thought [17:34:55] thanks [17:38:12] awjr: I've got enwikimf_useformat=desktop; cookie [17:38:33] awjr: but, en.wikipedia.org is giving me the mobile version of the site [17:38:48] no good [17:38:54] one sec [17:39:37] preilly: is it possibly a browser cache issue? [17:39:50] preilly: also what device are you using? [17:40:10] awjr: chrome on osx [17:40:22] awjr: I'm in a meeting right now [17:41:00] preilly: ok can you send me a link to one of the pages behaving badly? i'll see what i can do to investigate [17:41:31] awjr: http://en.wikipedia.org/wiki/Katherine_Switzer [17:43:25] (YuviPanda: Incidentally, I can't seem to "upgrade" - I have to uninstall and reinstall.) [17:43:49] Jarry1250: yes. [17:43:54] signing issues, etc. [17:45:41] YuviPanda: Ah yes, history working now (and collapsible sections?) [17:46:00] Jarry1250: yes, they should work too! [17:46:06] ah, I think I figured out the problem with our nightly builds [17:46:19] Ah, could it be that collapsible sections were on the same onload trigger which is now firing but wasn't before? [17:46:22] I think so. [17:46:41] Jarry1250: no, our nightly builds weren't including 2 js files they should've [17:46:44] preilly: i am unable to replicate the problem you're seeing - let's chat when you're out of the meeting [17:46:45] YuviPanda: Glad to hear it :) [17:47:02] Jarry1250: also re - sharing to twitter. Do you have the twitter app insatlled? [17:52:50] heads up that the git migration is moving today [17:53:34] YuviPanda|afk: Oh, no, I just use the mobile site. That's probably it :) [17:53:50] Jarry1250: :) [17:53:51] blehhhh, git checkout of vumi also fails [17:54:06] Jarry1250: android's means we'll support all services installed in device [17:55:09] Yeah, I figured it was something like that - it was just the fact that you know, Android comes with a Twitter button, and you integrate the contacts, and all that shebang. [17:56:53] Jarry1250: yes, but you can turn off the contacts integration (it's a pain) [17:57:38] Oh, I had a play with that, yes, it just confused me a little because you'd have thought it was one Twitter authentication for all Android services. [18:00:28] Jarry1250: doesn't have that , IIRC [18:00:44] Jarry1250: could you update your feedback with what works/doesn't? :) [18:00:51] Oh, sure. [18:01:00] whee! awesome! [18:01:06] can someone take a quick look at this for sanity before I push it live: https://www.mediawiki.org/wiki/Special:Code/MediaWiki/114384 [18:01:20] MaxSem, preilly perhaps ^ [18:01:25] Jarry1250: can you also mail your iOS device's UDID to mobile-feedback-l@lists.wikimedia.org (a private list) so we can add your name to the list for today's RC? [18:01:33] awjr, too late [18:01:38] already seen [18:01:46] :d [18:01:47] er [18:01:49] "D [18:01:50] damn it [18:01:51] :D [18:01:52] there we go. [18:01:53] :D [18:02:03] i seemed to think I was 10 in 2011 [18:02:11] and tried to correct it to 2001 3 times and failed [18:02:28] okay, back to RTL bugs [18:04:30] Hey preilly, MaxSem. [18:04:58] Sorry for the delay. I was on a call with Google. [18:04:59] jerith: howdy [18:05:05] How're things? [18:05:13] jerith: can you help MaxSem a little bit? [18:05:57] jerith, hi [18:06:03] Sure. [18:06:20] MaxSem: vumi's a bit of a moving target at the moment. [18:07:02] how do I install it? setup script ends with "No eggs found in /tmp/easy_install-VHyyPE/python-smpp-0.1/egg-dist-tmp-vLUuAI (setup script problem?)" [18:07:13] We're running most of our production systems on the develop branch (which is the equivalent of "trunk" in svn-land). [18:07:29] I've tried installing from git, too [18:07:32] Setup's a bit painful right now. :-/ [18:07:33] same error [18:08:02] Is there a setup guide or something that you're following? [18:09:23] if there's one, I've not seen it. tried both sudo easy_install vumi and sudo python setup.py install on git checkout [18:09:59] This is probably the best place to start: http://vumi.readthedocs.org/en/latest/intro/tutorial01.html [18:10:29] (I think there are at least two others, though. Our docs are a mess.) [18:11:21] * jerith tries a fresh installation to see how much stuff actually works these days. [18:12:53] YuviPanda: Done, I thinl. [18:12:57] *think. [18:13:18] MaxSem: "pip install vumi" in a fresh virtualenv seems promising so far. [18:13:29] running it now [18:15:06] There's a long list of dependencies. :-/ [18:16:16] (Most of them aren't even necessary unless you're using the specific bits of vumi that want them.) [18:17:38] That seemed to work on my side. [18:17:56] * jerith is really embarrassed about how hard it is to get up and running with vumi. [18:18:29] jerith, also: how do I run tests for it (mostly vumi_wikipedia)? [18:19:04] vumi_wikipedia is in a separate package. [18:20:16] awjr: i'm loving the persistent cookie [18:20:16] awjr: FYI you can share a dashboard from graphite e.g., https://graphite.wikimedia.org/dashboard/temporary-8 [18:21:02] preilly: far out [18:21:19] ah yeah i see the 'share' icon [18:21:20] thanks [18:21:29] tfinc: glad to hear it - i hope we can get the lingering issues worked out soon [18:21:38] * tfinc gets distracted from budget meetings with pretty graphs [18:21:50] MaxSem: https://github.com/praekelt/vumi-wikipedia/ [18:22:03] awjr: how should i read these graphs [18:22:15] tfinc: i was just trying to figure out the same thing :p [18:22:16] awjr: is mobile cookie set desktop->mobile [18:22:20] jerith, I know, I've already forked it [18:22:41] awjr: and desktop cookie set mobile->desktop ? [18:22:43] tfinc: correct [18:23:09] awjr: thats a lot of people wanting to see the mobile view [18:23:26] tfinc: indeed, im surprised it's an order of magnitude more than desktop view [18:23:29] FTW! … kind of [18:23:43] what is likely telling us is that our device detection is not as good as we think it is [18:23:55] thats why i'm really happy to have this metric [18:24:09] indeed [18:24:15] MaxSem: Cool. There's no setup.py for it yet (see issue #4) so you'll have to add it to your PYTHONPATH or something. [18:24:39] awjr: keep an eye on this metric … overall it is bigger then i expected [18:24:56] yeah, same. i will [18:24:58] make sure that were not double counting [18:25:22] tfinc: the metric happens at the point the cookie gets set [18:25:57] yeah, i know. it'll make more sense when we see trends [18:26:05] jerith, your install docs are outdated: rabbitmq.setup.sh in in ./utils [18:26:10] binasher: thanks for graphite tool [18:26:14] MaxSem: Turns out vumi_wikipedia doesn't work with vumi 0.3. :-/ [18:27:30] * MaxSem is seriously doubtful that he should start hacking Python with this [18:27:50] :| [18:27:54] The vumi_wikipedia app is much simpler than the rest of vumi. [18:28:04] tfinc: actually, come to think of it, it *might* be double counting because of the separation of the domains [18:28:26] MaxSem: You also need to install redis, if you don't already have it. [18:28:31] tfinc: thinking [18:29:08] tfinc: actually no, it should be accurate [18:29:14] so far, I need to tweak vumi_wikipedia to use proper API to retrieve page plaintext [18:29:42] awjr: lets stop looking at it for an hour ;) [18:29:45] MaxSem: If you have vumi installed from github, you can just run "trial vumi_wikipedia" to run the tests. [18:29:54] Is there a proper plaintext API now? [18:29:56] tfinc: but… but.. PRETTY GRAPHS [18:30:07] awjr: put the graph DOWN [18:30:12] sigh [18:30:13] ok [18:30:22] * tfinc reloads it one more time in shame [18:30:23] :D [18:30:33] just one more! [18:30:34] s/shame/secret [18:30:35] jerith, yes - deployed yesterday :] [18:30:37] awjr: Want to make some pretty graphs and things for me? ;-) [18:30:45] \o/ [18:30:47] jerith: wfIncrStats() [18:30:49] * jerith hugs MaxSem. [18:30:51] will do it for you :) [18:31:12] awjr: My vumi code doesn't have that function. :-P [18:31:23] aww [18:31:30] I thought that trying to reproduce the same DOM parse madness that we already have in PHP is kind of pointless::) [18:31:35] https://wikimania2012.wikimedia.org/wiki/Submissions/Visual_Version_of_Wikipedia_for_Low_Cost_Electronic_Devices [18:31:39] kul: --^ [18:32:17] tfinc: oi [18:33:04] checking out the talk [18:33:41] twisted.python.reflect.ModuleNotFound: No module named 'vumi_wikipedia' [18:34:20] MaxSem: "export PYTHONPATH=/path/to/vumi-wikipedia" [18:34:41] * jerith really needs to get that setup.py sorted out so it can be installed properly. [18:36:14] awjr: what section are these mobile graphs in? i don't see a mobile seciton [18:36:23] awjr: i'm looking at the bottom left window [18:36:54] tfinc: i also was trying to figure it out. i wound up finding it by looking under '* [18:37:16] i was following the same syntax as is used for opt-in [18:37:24] but perhaps there is a better way to keep that stuff organized [18:37:28] preilly: binasher : can we code those better ? [18:38:11] tfinc: I thought that mobile. <- would make a section, but this is a binasher question [18:38:17] don't access it under * [18:38:47] wfIncrStats goes under the stats section [18:38:55] ohho [18:39:31] MaxSem: Are you winning yet? [18:39:37] there are so many sections now that going via * is slow as hell [18:39:47] exceptions.TypeError: Header entry 'User-Agent' should be list but found instance of instead [18:40:09] that seems to be package incompatibility [18:40:12] awjr: lets move it [18:40:14] MaxSem: Ah, right. That's because you're using a version of vumi that's too old. [18:40:33] tfinc: not sure how or if we can move it out of 'stats' [18:40:40] (That was a bug in the HTTP helper functions that got fixed after 0.3.) [18:40:44] i'll take a deeper look in a bit, tfinc [18:41:04] thanks, feel free to also just log a bug if its a tough change [18:41:07] * YuviPanda|afk lets a few profanities loose at the 4.0 emulator [18:41:11] and we can schedule it [18:41:27] MaxSem: "pip uninstall vumi; pip uninstall Twisted; pip install Twisted==10.1" will get you to a virtualenv without vumi in it. [18:42:06] (There's a bug in pip that makes it nuke half of Twisted when you uninstall vumi, so you need to uninstall/reinstall Twisted.) [18:42:38] Then install the develop branch of vumi from github. [18:42:59] Probably best to clone the repo and then "pip install /path/to/vumi" to do that. [18:44:02] Err, "pip install -e /path/to/vumi". That'll install it in an editable way, so you can just "git pull" to update if necessary. [18:44:58] * jerith takes the laptop downstairs to be closer to the delicious supper smells. [18:46:14] * preilly thinks jerith should take a break and actually eat his supper [18:46:47] supper isn't ready yet. [18:47:32] ah [18:57:24] MaxSem: Shout if you're stuck. Also, let me know if you've made it work. :-) [19:04:46] preilly: Also, today was a public holiday. This is the closest I've come to working all day. :-) [19:06:19] jerith: ha ha [19:08:29] preilly: If and when Google flies me over to Leftpondia for an on-site interview, could I pop by for a tour of WikiMedia Central? :-) [19:09:15] jerith: oh, course [19:09:23] jerith: s/oh/of [19:09:36] * jerith grins. [19:09:54] I'll let you know if and when. [19:10:00] jerith: okay, cool [19:22:04] MaxSem, awjr: can you take a look at https://www.mediawiki.org/wiki/Special:Code/MediaWiki/114393 [19:27:45] preilly: it seems that that WURFL detection wasn't even really being relied upon anyway, is that correct? [19:28:09] as $device gets reset a little later with $device = new DeviceDetection(); [19:28:23] awjr: yeah [19:28:51] preilly: ok'd [19:32:00] awjr: thanks! [19:32:06] np [19:36:54] is brion around? [19:37:01] yo [19:37:06] yo [19:37:21] brion: when was the last time you tried the app on your phone? [19:37:45] few days ago. i can pull it up, anything i should check for? [19:37:48] brion: looking to reproduce https://bugzilla.wikimedia.org/show_bug.cgi?id=35369 [19:38:00] eclipse refuses to let me do stuff with my ICS emu :( [19:38:36] YuviPanda : try turning the adb on and off [19:38:41] had the same thing happen to me [19:38:53] lifeeth: I restarted.. [19:38:57] * YuviPanda is from windowsland [19:38:57] OKai [19:39:03] but will try again [19:39:15] lifeeth: when you're not too busy frying circuits, check if you can repro that bug? [19:39:18] :) [19:39:29] YuviPanda : not until I go back to BLR :) [19:39:34] ah, forgot [19:40:47] lifeeth: sigh now the emulator-arm process just dies. [19:40:57] YuviPanda : got it [19:40:59] YuviPanda : same bug [19:41:03] doesn't scroll back up [19:41:08] sigh [19:41:08] okay [19:41:20] YuviPanda: can confirm i reproduce it on my galaxy nexus, yes [19:41:25] I am on 4.0.3 [19:41:26] brion: okay. [19:41:50] brion: think you'll have time to debug this if I can't get my emulators working? [19:42:12] probably needs to reset scroll position on the correct section when navigating [19:42:14] i'll take a peek [19:42:31] brion: \o/ [19:42:33] cool! [19:42:41] then I'll go check up on the RTL bug. [19:42:45] and then we can put out RCs! :) [19:47:04] Jarry1250: thanks a ton for the testing [19:47:04] :D [19:47:09] i've responded to the ones [19:47:19] major issue seems to be sharing URLs need to be proper. will poke at it [19:49:39] MaxSem: Ping? [19:49:48] pong [19:50:12] took a shower to fight brain overheating [19:50:48] looks like I've screwed that virtualenv and have to start from scratch [19:51:43] Ah. [19:52:00] So, when you start from scratch, do this. [19:52:39] Clone both the vumi and the vumi-wikipedia repos from github. [19:52:48] Create and activate new virtualenv. [19:53:15] "pip install -e vumi/" from the directory containing the cloned repos. [19:53:38] That should get you the right version of vumi and everything it depends on. [19:55:56] MaxSem: I'll be going to bed in a couple of hours, and I'm sadly unavailable tomorrow evening. [19:56:10] np [19:56:19] ImportError: No module named twisted.python [19:56:47] though it said that twisted is installed [19:56:47] That's odd. [19:57:01] Ah. Did you uninstall vumi at some point? [19:57:32] ah, reinstalling [19:57:55]