[16:47:53] @trustadd .*@mediawiki/Yuvipandan admin [16:47:53] Successfuly added .*@mediawiki/Yuvipandan [16:48:11] yuvipanda: done [16:48:21] you should read some guide on meta :) [16:48:30] @commands [16:48:30] Commands: channellist, trusted, trustadd, trustdel, infobot-off, refresh, infobot-on, drop, whoami, add, reload, logon, logoff [16:48:38] this is basic stuff you can use [16:48:42] petan: coool! [16:49:09] databse of keys is exported to http://bots.wmflabs.org/~petrb/db/ [17:53:06] howdy [17:53:32] good morning, philinje [17:54:21] philinje: good morning [17:54:37] philinje: the channel is currently being logged by wm-bot, any objections? [17:55:51] hi, sorry [17:56:30] tomasz and patrick are not here, so I'll try to have a quick call with you guys [18:00:20] apologies [18:00:21] got dc [18:00:25] yuvipanda, start Skype [18:00:30] MaxSem: am here [18:00:34] on skyp [18:05:19] tfinc, hi [18:05:22] greetings all [18:05:24] we've just finished [18:05:25] philinje: https://bugzilla.wikimedia.org/show_bug.cgi?id=33326 and https://bugzilla.wikimedia.org/show_bug.cgi?id=32905 to start with then. [18:05:27] sweet [18:05:29] philinje: and you've figured out https://bugzilla.wikimedia.org/show_bug.cgi?id=33289 [18:05:30] ? [18:05:33] thanks for running it philinje [18:07:02] * tfinc wonders if we should trust philinje or pchang [18:07:04] :) [18:07:19] tfinc: or tomaszf [18:07:20] ? [18:07:21] :D [18:07:24] seriously [18:07:26] can't i just pick one [18:09:07] what have we been hearing about RC3 ? [18:09:16] tfinc: philinje the channel is now being logged by wm-bot. Any objections? [18:09:16] the one that doesn't suck [18:09:23] uhoh [18:09:30] thats it .. i'm outta here [18:09:40] :D [18:09:56] we should put that in the topic [18:10:01] and let the other people know [18:10:06] what made you log it [18:10:15] true [18:10:24] tfinc: well, preilly gave me the API link a few days ago [18:10:32] i didn't save it, and then went to look for logs [18:10:34] didn't find [18:10:51] then was in #wikimedia-labs when petan was talking about the bot, so asked him to turn it on here as well [18:12:09] meh [18:13:05] http://meta.wikimedia.org/wiki/WM-Bot if you guys want to do anything to it [18:13:28] thanks [18:13:43] my transcripts are very unreliable [18:15:05] preilly, are you here already? [18:15:25] MaxSem: yes [18:15:37] good morning [18:15:51] MaxSem: morning [18:15:54] your review, sir: https://www.mediawiki.org/wiki/User:MaxSem/CR/ZeroRated [18:23:01] brion: yuvi will handle the nearby me bugs [18:24:24] ok [18:24:27] brion: i'd need to sign / get a maps key to get this to work on my dev setup, right? [18:24:33] MaxSem: how did you end up with Undefined index: RUSSIAN FEDERATION in D:\Projects\MediaWiki\extensions\ZeroRatedMobileAccess\ZeroRatedMobileAccess.php on line 84 [18:24:40] yuvipanda: just to see the map tiles :) [18:24:47] lemme recall... [18:24:59] the geo lookup is actally separate so you can test everything else, including tapping on the items [18:25:14] but you might want to set up your debug key for maps if you can [18:25:16] Main_Page?renderZeroRatedLandingPage=1&country=ru [18:25:28] brion: ah, okay. then I'd just have to fake a location via adb and everything'll be cool [18:26:41] :) [18:27:13] MaxSem: try RU [18:27:49] same thing, getFullCountryNameFromCode() normalises case [18:28:28] MaxSem: hmm that is weird it works for me [18:28:41] MaxSem: oh, wait you don't have the wikipage [18:31:31] tfinc: any other highish-prio bugs y'all want me to take a peek at? only things currently assigned to me on android look low-prio [18:34:34] MaxSem: you was your 20% time ? [18:34:41] brion: i'll take a look at the queue [18:34:50] tx [18:34:51] \ [18:34:52] yes, a lot of CR [18:35:15] still need nore experience with it though [18:35:42] also, iterated through a few instances on labs [18:36:52] one before last was funny: I installed all the needed packages, went to /var/www, attempted to create a directory for MW and received a message that I'm outta space:P [18:37:21] MaxSem: been there :) [18:38:22] yuvipanda: did you tackle the near by me touch surface bug ? [18:38:33] brion: can you get our blinking cursor back? [18:38:38] i'll open a bug [18:38:42] heh [18:38:47] we lost it long ago and miss it [18:38:57] https://bugzilla.wikimedia.org/show_bug.cgi?id=32917 [18:38:58] i don't know if i can do that without dropping the position: fixed [18:39:03] tfinc: am on it right now., [18:39:10] experiment and let us know [18:39:12] tfinc: setting up debug keys, since I guess i'll need them anyway [18:39:49] yuvipanda: if your doing that then resolve this https://bugzilla.wikimedia.org/show_bug.cgi?id=33133 [18:40:09] create a password less official cert [18:40:46] tfinc: aha, sure. [18:41:29] tfinc: assigned [18:41:41] then pass on that cert to hashar [18:41:42] s [18:41:48] so that he can sign the builds with it [18:42:17] just don't put a pass on it [18:43:43] ROFL http://www.extremetech.com/computing/110061-eff-reverse-engineers-carrier-iq [18:43:53] * tfinc hugs the eff [18:44:52] tfinc: sure [18:52:18] tfinc: ok try this branch and tell me what you think https://github.com/brion/Wikipedia/tree/absolute [18:52:31] brion: will do [18:52:32] https://github.com/brion/Wikipedia/commits/absolute [18:55:08] so that basically just swaps from position: fixed to position: absolute and drops the evet handlers that broke the cursor, since they're no longer needed [18:55:26] downside (or upside): gives more screen space for reading text since the header doesn't float at the top of the screen all the time [19:00:35] MaxSem: see https://www.mediawiki.org/wiki/User:MaxSem/CR/ZeroRated#i18n [19:02:46] preilly, btw - why you don't credit yourself by name? [19:04:17] MaxSem: like 'author' => array( 'Patrick Reilly' ),… ? [19:04:24] yup [19:05:03] MaxSem: no sure [19:05:12] tfinc: this should also fix the problem some people saw with an extra 'view' of the search input box floating over/near it [19:05:13] s/no/not [19:11:10] tfinc: any thoughts? [19:29:46] sent a message to mobile-l asking for more feedback [20:20:27] ok .. lets see if this internet is more stable [20:20:55] brion: checking out the changes now [20:21:07] whee [20:38:50] brion: should it work in froyo? [20:39:29] it should [21:09:14] [WikipediaMobile] brion pushed 1 new commit to master: http://git.io/AV7j0A [21:09:14] [WikipediaMobile/master] Update messages per bug 33352: - Brion Vibber [21:09:27] Project WikipediaMobile - Nightly builds build #30: SUCCESS in 6.8 sec: http://integration.mediawiki.org/ci/job/WikipediaMobile%20-%20Nightly%20builds/30/ [21:09:27] brion: Update messages per bug 33352: [21:32:33] brion: it certainly fixes the input box issue [21:32:42] but it sucks to lose all the functionality [21:32:51] * tfinc is clearly getting distracted at home [21:35:53] well i'm not sure if it's a plus or a minus so far :) [21:36:03] feedback has been mostly positive on the theory at least [21:36:13] but that's 2/3 data points eh :) [21:36:14] brion: can we bring that back on 'tap'? [21:36:40] that would match up with what brandon has been wanting to do [21:36:48] to think of it like the kindle app [21:37:29] that way its there if you want it [21:37:32] otherwise its just gone [21:38:05] plus we should bind 'search' button to it too [21:39:49] could try it [21:40:45] brion: can you mock that up? [21:41:00] sure i'll ping ya when it's ready [21:41:05] sweet [21:46:59] tfinc: why an option? [21:47:38] * tfinc ponders [21:47:56] tfinc: any scenario when you'd *want* to use HTTP? [21:48:03] my only worry is that we haven't tried https at scale [21:48:08] not really [21:48:20] tfinc: well, let's try it then :) [21:48:46] sure [21:49:04] tfinc: i'm sure you can talk to the ops folks and check on that [21:50:29] we can always bug Ryan_Lane :) [21:51:15] yuvipanda: hows https://bugzilla.wikimedia.org/show_bug.cgi?id=33280 going ? [21:51:45] tfinc: bad news: that gives us the same 'sends clicks beneath' bug :( [21:52:03] tfinc: am in javaland now. no passwordless keys :( [21:52:11] or keystore, rather [21:52:22] javaland for near me touch surface ? [21:52:55] keysigning and that, yes? [21:53:03] yeah [21:53:15] tfinc: will do that after this [21:54:13] tfinc: did you play with different search strategies y'day? [21:54:26] i did, inconsistent results [21:54:34] i go the suggestions for gerrit co [21:54:37] which was nice [21:54:47] but you'll never get them for 'go' [21:54:49] or anything that's short [21:54:59] but i still couldn't find the article that mentions 'charlie sava' [21:55:57] tfinc: the other apk? [21:56:11] let me see which one i tried [21:57:02] brion: what was "that" (bad news)? [21:57:15] i did first exact, then full text [21:57:25] i think given some tweaking we could make it better [21:57:32] but its not good enough yet [21:58:07] philinje: any time we have the header float over the iframe, stuff breaks [21:58:09] * tfinc wonders how ct is doing hiring someone for search [21:58:20] doesn't matter if it's position:fixed or if we manually position it after a tap [21:58:28] other things are more urgent [21:58:46] i meant in terms of the search options [21:58:52] yup [21:58:53] tfinc: did you try the 'pure' full text one? [21:58:54] brion: i saw the email [21:58:56] not a priority right now [21:59:48] so when the header re-appears, clicks go through it? [22:00:36] philinje: right. just the same as position:fixed [22:00:51] brion: sounds like we could live with the header staying in place at the top of the page [22:01:21] staying in place == scrolls off the pgae? or stays fixed on screen? [22:01:40] scrolls off, and use the hardware search key [22:02:07] philinje: i don't know yet. from our research people have said that we should always have search available at the top [22:02:11] we already use the hardware search key (assuming anyone actually uses it) [22:02:23] note however that it doesn't set the focus properly, so you see neither a focus ring nor blinking cursor [22:02:32] i would just scroll back up to top myself [22:02:43] note also that search key is being obsoleted, it doesn't exist in android 4 [22:02:55] but keyboard appears (with hardware search key) [22:03:09] yes, that's all. keybord yes (because we force it), focus no [22:03:20] on my phone typing sets the focus, but some may not based on feedback i've seen so far [22:03:35] yes, it's annoying [22:03:48] where are we on the focus problem? [22:05:01] that is the focus problem. you saw the mail? [22:06:24] ok i'm gonna try another variant using http://cubiq.org/iscroll-4 for the scrolling :P [22:07:56] brion: i thought the focus problem was a byproduct of another fix [22:08:36] well, that part of the focus problem sorta is. but it's all related [22:11:33] basically the workaround for the clicks-go-through cancels the normal handling of the event, then forces focus [22:11:37] and the search button forces focus [22:11:45] both have to open the software keyboard manually iirc [22:11:58] and the focus set doesn't quite work on its own [22:12:49] damn Android! [22:12:54] :) [22:14:26] is there a conflict with opening the keyboard? [22:17:54] ? not that i know of [22:18:00] not opening the keyboard doesn't help, anyway [22:18:35] not having much luck with iscroll, doesn't seem to be compat with iframe contents [22:19:12] hmm [22:19:17] that's too bad [22:20:38] hmm, maybe i can inject it inside the iframe [22:21:35] brion: do you deploy / test from eclipse? [22:21:40] yes [22:21:53] brion: hmm, what do you do about those 'override' errors that pop up? [22:22:01] i dont get them [22:22:05] :| [22:22:11] iirc: make sure it's compiling with java 1.6 [22:22:28] well, am compiling from the commandline [22:22:39] brion: if I understand what happened in bug 32319, the workaround caused the focus problem, but the workaround was needed because the search bar was floating [22:22:41] * yuvipanda debugs [22:23:03] philinje: right. in that case we wouldn't have a focus problem without the workaround for the floating bar clickthrough [22:23:35] brion: i am on 1.6.x [22:23:41] *shrug* i dunno then [22:23:55] my vote is to go with fixed position i.e. scroll off, and reverse the workaround [22:23:58] brion: you've 'never' had that issue? [22:24:12] i had it once, changed the version, and it went away [22:24:24] okay, let me mess around [22:25:13] then hardware key for search should set focus, right? [22:25:44] that seems straightforward and simple [22:26:05] philinje: no, that doesn't set focus [22:26:12] not straightforward at all [22:26:36] so the workaround is not the whole problem [22:27:03] i've tried a few things including this suggestion for a java-side focus tweak: http://stackoverflow.com/questions/5662828/android-webview-setting-html-field-focus-using-javascript [22:27:14] no, we have no way to programmatically set focus that works [22:30:34] wow, sounds severe [22:32:52] brion: by programmatically, are you including the hardware key for search? [22:33:05] philinje: since it must programmatically set focus, yes [22:34:04] *looks at feet* [22:34:53] :) [22:41:19] gotta run some errands around the house -- ping me if anything else to poke at and i'll take a look in a bit! [22:42:05] brion: i just noticed in another Android app that when setting focus to a search field, some text inside is automatically highlighted, which then is replaced when you start typing [22:42:37] so there is no cursor until you start typing, and it is very obvious the search field is in focus [22:42:47] hmm, might be worth trying [22:42:56] drop in a bz and assign it to me [22:43:01] ok [23:04:43] brion: bug 33358 [23:08:51] yuvipanda: are we currently asking the user's permission to use the current location? [23:09:10] philinje: only on install, not on start of activity [23:09:44] install or first boot-up of app [23:09:51] install [23:09:54] install [23:10:14]

[23:10:31] In order to receive location updates from NETWORK_PROVIDER or GPS_PROVIDER, you must request user permission by declaring either the ACCESS_COARSE_LOCATION or ACCESS_FINE_LOCATION permission, respectively, in your Android manifest file. [23:10:39] from http://developer.android.com/guide/topics/location/obtaining-user-location.html [23:10:49] most apps ask when they load up, right? [23:11:11] philinje: no? [23:11:21] philinje: if you have your GPS turned off, I think they'd ask you to turn it on [23:11:29] philinje: not typically in an Android workflow [23:11:45] s/they'd/they could/ [23:11:49] i see, i am thinking of iPhone apps [23:12:11] yuvipanda: yeah, I guess you could but it wouldn't be a secure action [23:12:28] preilly: secure action as in? [23:12:38] philinje: i think web apps will ask, on android too. [23:13:15] yuvipanda: I mean you would not be prevented from doing it in app [23:13:27] yuvipanda: e.g, you could ask yes or no and still call the API [23:14:00] i think they usually fall back to just network provider's estimate of location if you don't turn on GPS [23:18:00] okay, learnt enough java to fix the nearby me bug :) [23:19:04] :) [23:19:11] * yuvipanda sends pull request, goes to fix other bugs [23:22:14] did you plug into the Matrix? [23:22:34] sorry [23:22:45] philinje: no, I just had to inherit from a class, and create a whole new file and a full new class [23:22:52] just so I could have 'onClick' on my Dialog objects [23:23:13] just realized that we don't show current location on the map [23:23:21] philinje: ? [23:23:24] current location as in? [23:23:36] user's current location [23:23:59] or am I mistaken? [23:24:04] philinje: well, wouldn't we be showing them a map…? [23:24:11] which begins with where they are.. [23:24:28] but yeah, perhaps a 'my location' button that takes them back to their current location is good [23:24:41] that's what I thought, but the blue spot is not appearing [23:24:44] for me anyway [23:25:08] also for some reason it opened in a different location [23:25:21] philinje: rather than your current location? [23:25:32] yes [23:25:40] hmm, my gps doesn't work, so am testing on emu. And not noticing anything different. [23:25:54] is the blue spot there? [23:26:12] philinje: didn't setup my debug key, so i got no map. [23:26:24] I'm assuming gmaps is what gives you the blue dot, so no, it isn't there. [23:26:36] philinje: let me setup the key and try. [23:26:41] not sure [23:27:36] in some other apps, there is a menu button for My location [23:27:55] yes [23:27:59] i have a feeling we never implemented the function [23:28:02] right now i think 'near me' doesn't have any menu [23:28:06] i don't see code for it [23:28:09] philinje: add bug? [23:28:16] ok [23:29:26] hmm, moving to https slows down the app considerably? [23:29:28] or is it just me/ [23:29:31] * yuvipanda reinstalls [23:29:37] no, that is likely [23:29:42] yeah [23:29:48] but *this* slowly? [23:29:50] doubt that [23:30:17] i think there is a risk in turning it on, as Tomasz sayd earlier [23:30:42] he was talking about load on our servers, I guess? [23:30:46] rather than load on client devices [23:31:19] yes, load on our servers [23:31:49] probably we should deal with HTTPS when we implement login on the app [23:32:54] well, i've already committed and pushed that (just now). [23:33:00] login is a generic function for accessing a user profile and getting certain results [23:33:08] philinje: we could talk to the ops folks before merging it [23:34:16] ok, i will try, but let's hold off on that regardless [23:34:56] philinje: we could just do https for articles, without having to worry about login and stuff for now, no? [23:35:01] philinje: will also help test the servers... [23:35:37] well, yes, but it may expose problems which we will be blamed for [23:36:12] we have plenty to worry about as it is [23:37:48] philinje: ah, right. [23:39:41] bug 33359 [23:41:52] philinje: 33359 or https://bugzilla.wikimedia.org/show_bug.cgi?id=32905 first? [23:44:08] 32905 but depends on if you are in Java land [23:45:28] philinje: well, i'll do 32905 first then. [23:46:35] adding a new bug that is an enhancement, but could be good as part of Nearby work [23:47:58] philinje: ok [23:48:15] philinje: we *are* shipping in one of the search variants with 1.0, right? [23:50:05] what do you mean by search variant [23:50:36] philinje: suggestions vs fulltext [23:51:32] suggestions for now, but let;s evaluate once we get through all the urgent stuff [23:51:43] philinje: okay [23:51:45] at the rate you're moving, we could get a lot done! [23:51:53] philinje: what else is the urgent stuff? [23:51:55] focus bug? [23:53:31] yes, and Nearby stuff [23:53:40] okay [23:54:14] hmm, languages now appears in order but there's a ~500ms lag before it shows up [23:54:35] 33360 is an enhancement, so not so urgent, but could be quick if done with 33359 [23:54:50] languages in Settings? [23:55:30] philinje: https://bugzilla.wikimedia.org/show_bug.cgi?id=32905 [23:56:22] philinje: in this case, we literally *are* plugging into the matrix :D [23:56:26] the sitematrix api, that is :) [23:57:27] not seeing the delay [23:57:54] philinje: err, the fix causes the delay [23:58:01] you do see it jumping around.. [23:58:19] anyway, i'm going to cache that list... [23:58:41] not seeing it jump around, myself [23:59:30] philinje: well, it depends on your connection. [23:59:38] if it's fast enough, it won't jump [23:59:46] else there's a race condition happening, and it will.