[13:56:22] hi Yuvi [13:56:35] hey ragesoss [13:56:40] no response at all on commons vp [13:56:42] oh wlel [13:56:44] *well [13:57:03] app spam? [13:59:24] you need to annoy someone to start a dialogue [14:00:29] :) [14:02:41] i do not have an android device.. i should go and do it, really. i need a new phone [14:02:53] i get, i think, an HTC something [14:10:36] moogsi: where do you live? [14:11:57] if you don't have aren't on AT&T or Verizon in the US, probably a Nexus 4 is the best phone you can get for the price. [14:13:28] i am in the UK, i need to ask/look around really [14:14:34] £239/£279 https://play.google.com/store/devices/details?id=nexus_4_8gb [14:14:37] Both in stock [14:14:43] so you'd have it wednesday/thursday probably [14:22:36] moogsi: yes, get the nexus 4 :) [14:22:38] ragesoss: yeah, nothing at all [14:22:40] ragesoss: i'll ask for a watchlist notice later tody [14:22:42] *today [14:34:06] ooh nice i tohught it would be pricier than that [14:34:22] moogsi: nexus4 is the cheapestbestphone :) [14:34:40] i loves my nexus4 [14:34:42] except if you're in a 3rd world country, in which case you don't really exist [14:34:52] except it's too big and the back is too smooth and it keeps sliding around when i set it down [14:35:08] first world problems: my phone is too big [14:35:14] :P [14:35:15] third world problems: google won't ship a phone to me [14:38:39] my network is really pushing 4G [14:38:40] like _really_ [14:38:53] it's too expensive and i don't want it! [14:39:25] my last phone was LTE, it was quite nice but hspa+ on the nexus4 is good enough [15:01:34] ssoooo am i buying the phone, and buying a sim from somewhere else and insurance from somewhere else? [15:01:42] that's cheapest, right? i'm a cheapskate [15:01:53] unsure how things are in the UK [15:01:57] but that's the thing in here [15:02:33] or should i just get a contact with a provider and end up paying £600 for the phone :l [15:02:43] i was having an argument with my sister to this effect :P [15:03:09] she was like BUT YOU GET A FREE PHONE AT THE END SO IT'S NOT MORE [15:03:11] * moogsi sighs [15:05:13] i'm moving off contract plans, it makes me happy to buy the phone and sim separate \o/ [15:05:36] i *think* i don't have to replace my wife's phone too, it's a CDMA iPhone but should work on GSM too [15:09:50] I've resolved to not get any more subsidized phones. [15:09:57] Looking forward to a Nexus 5 or whatever. [15:11:49] i just know that as *soon* as i do it, i'll just instantly smash my phone into tiny pieces [15:11:58] :) [15:11:59] despite not having lost or broken a phone for like 5 years [15:12:14] i don't know how people break their phones so often [15:12:19] do they keep rocks in their phone pocket? [15:12:19] so i would need insurance [15:12:33] no, me either, it's always inside my jacket [15:12:46] i guess if someone tried to shoot me in the chest... [15:12:59] it is the metal cigarette case of our day [15:13:13] hehe [15:15:04] i guess if i do smash it during the warranty period i'll just get my friend to put a new chassis on it and send it back like I DON'T KNOW WHAT HAPPENED [15:15:07] god, i'm scum [15:15:13] i'm the reason the economy is such a mess [15:15:28] my HTC One XL had its backlight die after 7 months. Fortunately, I was able to get it replaced under manufacturer's warranty. [15:15:50] I've not had the greatest experience with HTC build quality. [15:16:20] only htc phone i've owned was my Nexus One [15:16:37] pretty solid machine, just got outdated and can't be upgraded anymore [15:16:37] my first HTC phone, the Incredible, got a crack on the LCD (but not the glass screen / touch digitizer), from basically just me having it my pocket. No serious trauma. [15:16:47] But they said I was out of luck. [15:16:59] :( [15:17:00] They will really take advantage of consumers every way they can. [15:18:10] indeed [15:18:14] perverse incentives i think [15:23:19] breakfast etc. catch y'all in a bit [15:24:19] they must be making a loss on the current nexus stuff [15:24:22] google [15:25:05] just another thing they will lovingly embrace until they are ready to tighten their deathgrip on everything [15:25:41] igooglinati, goog world order etc. [15:26:35] hey YuviPanda [15:26:41] hello MaxSem [15:26:45] moogsi: they're selling it at cost [15:27:20] have you found someone to flip a bit for you? [15:27:49] I did! [15:31:07] New patchset: Zfilipin; "Updated Ruby gems" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59150 [16:00:48] uuuuuuuuuuuuuuuuuuuuuuuh [16:00:56] I hate you ResourceLoader [16:03:28] New patchset: Mgrover; "changed element for searchbox to new name in order to fix automation tests" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59155 [16:23:27] New patchset: MaxSem; "Device-specific styles should be last in the list" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59157 [16:27:08] New review: MaxSem; "@todo: move this stuff to mw.o" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/59015 [16:27:40] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59015 [16:30:42] New patchset: Zfilipin; "changed element for searchbox to new name in order to fix automation tests" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59155 [16:33:53] New review: Zfilipin; "master on en.m.wikipedia.org: 7 scenarios (4 failed, 3 passed)" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/59155 [16:33:54] Change merged: Zfilipin; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59155 [16:36:01] New review: Cmcmahon; "maintenance" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/59150 [16:36:02] Change merged: Cmcmahon; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59150 [16:46:02] * awjr waves [16:46:34] yo [16:46:50] how was your weekend MaxSem? [16:47:30] partied on Saturday, worked on Sunday:) [16:47:36] heh [16:47:51] at least you got some party in :) [16:48:09] awjr, your help at https://gerrit.wikimedia.org/r/#/q/status:open+project:mediawiki/extensions/GeoData,n,z will be muchly appreciated:) [16:48:13] * SickPanda fixed UploadWizard bugs on Friday, then GitHub -> Gerrit bot on sat/sun [16:48:18] awjr: how was your weekend awjr? [16:48:19] cool MaxSem, i'll take a look shortly [16:48:32] actually, i'll take a look after standup, MaxSem [16:52:49] New review: Jdlrobson; "There should be no need at the current time for it to live across different domains. It will save wh..." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/58641 [17:01:14] jcmish: standup! [17:02:31] i hate google [17:02:33] brion: clearly, Picard :P [17:02:41] restarting my browser :D [17:11:44] jcmish: we're going to keep using the same hangout: https://plus.google.com/hangouts/_/8bc98160cdf0396821e2ba381099b56fde42cb0c [17:25:21] brion: ping [17:25:28] SickPanda: pong [17:25:31] don't get sick noooo [17:25:53] brion: I'm trying to stay up but might not fully make it to the app showcase [17:26:10] brion: my eyes are burning and I can hear my heart :| Not good, so hopefully I can sleep it off [17:26:14] brion: can you do the demos? :) [17:26:17] sure [17:26:25] thanks [17:26:29] categories, multiple uploads, anything else to show off? [17:26:54] there's a thin 'about' page and a way to turn off tracking [17:27:20] brion: there's also http://stat1001.wikimedia.org/limn-public-data/mobile/datafiles/alltime-numbers.csv which doesn't have any graphs yet, but shows you deletion numbers for app / web [17:27:21] ah nice [17:28:20] brion: that's pretty much it [17:28:28] brion: I'll try to stick around, but even if I am can you do the demos? [17:28:29] ok [17:28:51] brion: thanks :) [17:28:54] :D [17:29:44] brion: actually you should show single upload too. the last one was a *long* time ago [17:30:16] :D [17:40:53] Maryana: ping? [17:41:04] yo sickpanda [17:41:11] Maryana: can you get heather to update https://commons.wikimedia.org/wiki/User:Heatherawalls/sandbox [17:41:15] why are you working if you're sick? :-P [17:41:30] Maryana: i'm not so sick that I can't fully work :) [17:41:43] Maryana: but I'm sick enough that I'm taking 'lie down' breaks [17:42:06] brion is covering for me in the apps showcase, so it'll be fine, I think [17:42:32] i'm not sure what updates you guys want - the easiest thing to do would prolly just be to leave her a note on the talk page of that page with a bullet list of changes [17:43:02] then i can just ping her and tell her to check her sandbox talk :) [17:43:22] Maryana: ok [17:43:31] Maryana: I just want her to update the image to use the screenshot [17:43:34] than the mockup [17:43:36] Maryana: doing now [17:43:42] sweet, thx [17:44:48] did you also want to remove the last pic from the banner, too? i know both tfinc and brion were saying it might be kind of misleading to suggest that your images go straight from the app to wikipedia [17:45:14] yeah unless we actually add that feature :D [17:45:46] we could ask her to change it to a screencap of the front page of commons with your image as the PoD :D [17:46:09] MaxSem: I think we made a mistake supporting opera mini :) [17:46:10] I made a commons category with a bunch of screenies [17:46:24] http://commons.wikimedia.org/wiki/Category:Commons_app I think [17:46:31] opera mini works but is verrrrryyy weird [17:46:34] ah [17:46:35] actually https://commons.wikimedia.org/wiki/Category:Commons_Android_App_Screenshots [17:46:49] jdlrobson: in general or with the $wgMFVaryResources change? [17:46:56] They render link rel="alternative" tags o_O [17:47:03] awjr: this is with the RL change [17:47:13] previously we didn't support it - now we support v7+ [17:47:58] oh :( [17:48:07] jdlrobson, which version? [17:48:43] i was hoping we'd keep the device detection the same as what's currently in prod until we had this change out everywhere [17:52:50] guess my opera mini is not antiquated enough [17:53:24] jdlrobson: can shoot me off a list [17:53:29] of devices [17:55:48] brion: mhurd check the iOS section of http://commons.wikimedia.org/wiki/Commons:Mobile_app? [17:56:19] SickPanda: nice!!! [17:56:32] SickPanda: oh we can also accept files shared from other apps, if they support it [17:56:41] brion: add it! :D [17:56:46] will do [17:56:51] I wasn't sure how to do that without being inaccurate [17:56:55] needs examples of such apps [17:57:01] MaxSem: v7 something - very odd [17:57:15] jcmish: currently going through now [17:57:27] the most recent one is 7.5... [17:57:37] my nokia n95 won't turn on. Seem to have brought the Nokia E52 back to life though! \o/ [17:59:22] brion: is there an app meet in 1 min? [17:59:26] or is that switched to tomorrow? [17:59:35] SickPanda: in theory …. i think tfinc is on his way now :D [17:59:39] [Commons-iOS] brion pushed 1 new commit to master: http://git.io/z1_Dzg [17:59:39] Commons-iOS/master 4bc74ba Brion Vibber: Add new message files: fi, ksh, pa, pt-br, te, tr, uk, ur, zh-hant [17:59:40] heh, ok [18:00:59] tfinc: am I in the right hangout? [18:01:49] SickPanda: we see you i think [18:02:23] awjr jcmish annoyingly i can't find the squid logs which were showing our device traffic - do you know the link? [18:02:40] jdlrobson: are you looking for raw logs? [18:02:50] awjr: there was a page somewhere that had a breakdown [18:02:52] so not raw [18:02:57] jdlrobson: :( no sir I do not [18:02:59] i think it's on the reportcard [18:03:05] where's that awjr ? [18:03:10] * jdlrobson can never find these things [18:03:44] jdlrobson: there's http://reportcard.wmflabs.org/ and http://stats.wikimedia.org/ [18:03:51] iirc it's actually on stats, not the reportcard [18:04:01] jdlrobson: stat1:/a/squid/archive/mobile [18:04:10] ori-l: \o/ [18:04:22] \o/ is not a valid path [18:04:49] actually, it is [18:04:56] $ mkdir "\o/" [18:05:02] $ ls -la [18:05:10] drwxrwxr-x 2 olivneh wikidev 4096 Apr 15 18:04 \o [18:05:25] it goes from cheering to waving. kind of poignant, really. [18:05:40] ori-l: do you know where the breakdown of mobile browsers is? [18:05:52] i feel like i've seen it on stats.wikimedia.org but i'm not finding it [18:06:10] tfinc: I can't make the showcase [18:06:59] I need to do some stuff for my mom [18:07:06] jcmish: give me 10 minutes and i'll send the mail [18:07:06] did the analytics team produce one? i crunched this for a week's worth of sampled squid logs in dec '12. i still have that, but the numbers have likely shifted since [18:07:15] jdlrobson: sounds good [18:07:50] ori-l: actually if you have that handy, that would still be useful [18:08:49] ori-l: +1 [18:09:52] awjr / jdlrobson - sent [18:10:00] cool! thanks ori-l [18:10:25] iirc Maryana crunched this data? [18:11:44] jcmish: you got mail [18:12:43] was that the browser data or the edit data, ori-l? [18:14:16] ori-l: had a quick look it's the 'other devices' user agents i'm interested in :) [18:18:04] * jdlrobson is having a testing fail. None of the old devices are switching on [18:18:49] jcmish, are you around for the QA/design checkin today? [18:25:10] Maryana: i'll be covering for her [18:25:27] k, cool [18:25:51] brion: can you add a mini agenda to http://etherpad.wikimedia.org/MobileShowcase for apps ? [18:26:28] sure can [18:28:01] brion: https://gerrit.wikimedia.org/r/#/c/58332/ [18:29:55] New patchset: Jdlrobson; "Address nearby FIXME using colimit" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59171 [18:31:04] New patchset: Jdlrobson; "run qunit with debug=true parameter" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59172 [18:31:18] MaxSem awjr can you quickly +2 that small commit - i'd do myself but i guess that's bad ^ [18:31:47] done [18:31:50] jdlrobson, ping for QA/design checkin :) [18:31:52] cant right now jdlrobson, in a mtg [18:31:56] setting up in R34 [18:32:06] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59172 [18:32:15] thanks MaxSem :D [18:32:28] awjr, https://gist.github.com/MaxSem/5390242 [18:34:08] MaxSem: quick readthrough looks OK, i can take a closer look later [18:36:09] Maryana: http://en.m.wikipedia.beta.wmflabs.org/wiki/Main_Page [18:37:06] you can't not see "flabs" in that URL [18:38:01] MaxSem: im a little confused about the UA search feature in that script - how will whoever is searching already know what the ID is for a UA? [18:38:29] they visit that page, their UA gets saved and they see the ID [18:38:46] ah, and they tell us the ID? [18:38:55] yep [18:48:18] New review: Brion VIBBER; "(1 comment)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/58332 [18:48:29] yurik_: dr0ptp4kt : i'm adding you guys to the mobile dept showcase @ noon. just realized that you guys werent on it. drop in if you have the time. [18:48:57] tfinc, thx [18:56:13] jdlrobson, MaxSem: add your stuff to show off! http://etherpad.wikimedia.org/MobileShowcase [18:57:53] the biggest user-facing stuff we've done since we last had a showcase (q4 planning mtgs) includes: Special:Handshake, user friendly upload titles, instructing users on original work, page images on watchlist, CN support (though im not sure how we'd show that off), pages images on nearby, author/source infon on uploads, distance on nearby [18:58:58] New review: Brion VIBBER; "Trying this on my local instance in Chrome..." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/58332 [18:59:28] awjr, i'm grabbing some lunch w/vibha, so we'll be a few min late for the showcase - don't wait up :) [18:59:53] MaxSem, jdlrobson missing anything from that list? [19:00:29] jdlrobson: also can you walk everyone through that stuff? [19:02:15] heads up 15min delay for the showcase [19:03:04] ops team didnt book the room and they need to run over [19:04:18] jdlrobson: I'll e-mail you my scripts; you can run them against the most recent files on stat1 [19:04:40] tfinc: run over whom? :P [19:04:48] hehe [19:05:09] New patchset: Jdlrobson; "Story 394: Show CTA on nearby when no photo" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59175 [19:06:13] awjr: sure thing [19:06:19] \o/ [19:13:11] brion: tfinc can you have mobile web go first? I'm not feeling too good but want to see what they're up to... [19:26:20] ragesoss: moogsi see http://commons.wikimedia.org/wiki/Commons:Mobile_app [19:26:31] what needs to be added before it can be linked to from watchlist? [19:28:20] MaxSem: it was what to death? [19:28:30] bitrotten [19:28:33] ah [19:28:40] heh [19:28:40] ok [19:28:54] SickPanda: any questions for the webbies? [19:29:00] nah, it's cool [19:31:08] brion: we are counting the number of shares from people via eventlogging :) [19:31:13] SickPanda: since it's just one page for both platforms, you'll burn a lot of attention for iOS users if you do a watchlist notice before iOS app is approved. [19:31:32] SickPanda: but if you have a very explicit notice just about Android, go for it. [19:31:40] ah, right [19:31:41] true [19:32:16] brion: 3 people have opted out so far [19:32:50] brion: no [19:32:57] brion: you need to go to gallery app from outside [19:32:59] brion: and do it [19:33:44] number of tiems brion has said 'we will have to add that to iOS' - 5 so far [19:34:11] jdlrobson, awjr, Maryana_brb: /home/olivneh/user-agents-09-15-april-2013.log contains one user agent per line, representing requests from the 9th to the 15th, sampled at 1:100 [19:34:13] 6! [19:35:10] (i'm guessing that's the sample rate based on the the filename pattern, which is mobile-sampled-100.tsv.log-yyyymmdd.gz) [19:35:24] brion: I made it a lot more predictable :) [19:35:33] SickPanda: anything else to demo? [19:35:52] brion: not really, i think we covered it all [19:35:56] ok [19:45:05] brion: tfinc do you want to show the reportcard? [19:45:53] brion: also curl http://stat1001.wikimedia.org/limn-public-data/mobile/datafiles/alltime-numbers.csv [19:52:53] MaxSem: https://gerrit.wikimedia.org/r/#/c/59171/ [19:54:18] jdlrobson, my paranoia says keep the if [19:54:24] awjr: MaxSem yurik_ : how was the a/v for the showcase ? [19:54:31] today we used 2x ClearVoice [19:54:31] great for me, tfinc [19:54:38] quite good [19:54:39] one on each side of the table [19:54:42] excellent [19:54:44] i'll let chip know [19:54:47] they had the 2x clear voice in there last week, it seems to work relly well [19:54:48] MaxSem: haha it will be fine - it's in beta [19:54:58] we should avoid FIXME's [19:55:12] MaxSem: i could add event logging if you really want to be sure [19:55:43] s/\/\/.*$// [19:55:52] mission accomplished;) [19:58:15] awjr, if there are no objections from you, I'll publish that script on TS [19:58:30] MaxSem: lemme take one closer look [19:58:35] one sec.. [19:58:46] thanks [19:58:55] jdlrobson: btw, victor's girlfriend is in boston at the marathon. He just asked her to use mobile web uploader to upload pics [19:59:01] just passing on message from -commons [20:00:12] MaxSem: do you have the schema for the agents table? [20:00:45] create table agents(id integer primary key autoincrement not null, ua text); create unique index idx_ua on agents(ua); [20:01:32] will this still cause the id to increment? INSERT OR REPLACE INTO agents VALUES (NULL, ?) [20:01:54] it will, however I first for precense first;) [20:02:01] *I check [20:02:30] yeah, of the ua [20:03:08] i think this is fine MaxSem, i dont see anything that makes me nervous [20:03:23] cheers [20:03:34] thanks for slapping that together so quick :) [20:04:09] http://toolserver.org/~maxsem/ua.php [20:04:54] 404 [20:05:18] heh, NFS didn't update [20:05:31] now it's a 200 :) [20:05:32] thx [20:05:37] Toolserver is so Toolserver [20:05:43] lol [20:06:07] maybe, put this on staging? [20:06:18] that might be better MaxSem [20:06:20] this takes records from real world UAs, no? [20:06:24] this is timing out for me [20:07:00] dr0ptp4kt, it saves visitors' UAs and gives them reference IDs for reporting [20:07:04] which is less likely to cause problems… labs or toolserver? [20:07:16] s/cause/have [20:07:22] at this point, TS I'm afraid [20:07:55] Firefox says "Waiting for toolserver.org…" [20:08:08] 404 again :) [20:08:23] MaxSem, k, was just going to say be sure to use htmlentities() on the output. i've seen some XSS with UAs. you're probably already doing that, knowing you :) [20:08:36] keke [20:08:37] dr0ptp4kt: https://gist.github.com/MaxSem/5390242 [20:08:37] ^ using htmlentities, not XSS'ing people! :) [20:09:22] yo dawg i heard you like html [20:09:25] is the idea to give people an interface for reporting their UA when they encounter bugs? [20:09:28] so i put some html in your html and now you have an xss [20:09:35] hahaha [20:10:11] LOL [20:10:11] lulz [20:10:22] ori-l: to make it easy for testers to tell us their UA [20:10:27] ori-l, it's so that people don't have to save their UAs on mobile phones when reporting problems [20:10:32] brion: yo dawg, i heard you like security, so i escaped your escapes and now you have &amp; everywhere [20:10:38] hehe [20:10:41] hahaha [20:13:43] ha, i get really good laughs when (non-sql) entities are encoded on the inbound. [20:19:13] here we go: http://staging.wmflabs.org/ua.php [20:20:30] who uses FF/Mac?:P [20:21:13] not it [20:21:29] that's me :D [20:21:32] right, that was chrome [20:21:48] and now i see brion too;) [20:21:56] \o/ [20:34:19] only 4 agents harvested so far - how boring [20:34:37] MaxSem: i'll give you some more [20:36:57] MaxSem: if you want that to be semi-usable on mobile you might want to set a meta viewport :) [20:42:33] New patchset: Jdlrobson; "Rewrite search as an overlay" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/58332 [20:44:29] brion, done [20:44:42] New patchset: Jdlrobson; "Rewrite search as an overlay" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/58332 [20:44:47] \o/ [20:44:57] New review: Jdlrobson; "(1 comment)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/58332 [20:45:01] brion: ^ [20:45:18] \o/ [20:45:22] i'll test momentarily [20:48:04] New review: Jdlrobson; "eek.. is this the only way? I'm assuming there is a naming order thing." [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/59157 [20:50:25] jdlrobson, unfortunately there's no other way [20:50:57] all styles get sorted by name and grouped by group [20:51:14] and you can't even add your own group [20:51:25] so [20:51:28] ... [20:51:47] you can't achieve it without tinkering with core. again [20:57:10] well we got pics from Boylston Street within an hour [20:57:39] the uploader was going to use mobile web but of course the first thing that happens in that situation is everyone calls everyone else and you have no signal [20:58:53] New review: Brion VIBBER; "Looks fine to me, except a double-escaping problem. :)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/58332 [20:59:18] you can't really see anything in the pics, just some hazmat guys after the event... but really i don't think i'd have taken any pictures either... i'd be running [21:01:50] ohhhh... AP is saying the cell phone network has been shut down to prevent remote detonations [21:01:51] MaxSem: im having a little trouble grokking https://gerrit.wikimedia.org/r/#/c/51124/1/GeoData.body.php - is the idea that if one of the parts is already fractional, it's an invalid lat or long? [21:04:02] awjr, e.g. 23.4° is valid, 23.4° 5' is not [21:04:21] only the last number can be fractional [21:04:58] wait - do you mean, if there's only one part, it can be fractional? [21:05:45] actually this would work fine if tehre was only one part and it were fractional, or the last part was fractional [21:05:59] is that the idea? [21:06:49] yes [21:06:58] see https://gerrit.wikimedia.org/r/#/c/51124/1/tests/ParseCoordTest.php for more examples [21:07:10] yeah that's what im looking through now [21:12:51] i'm planning to reload the apache config for staging-zero.wmflabs.org. anyone have a thumbs up or thumbs down with doing so? trying to override some of the hard X-* headers it's setting for ZeroRatedMobileAccess [21:13:28] if you graceful it, nobody'll even notice:) [21:17:01] alright, "MaxSem told me to do it" :) [21:20:42] thanks, MaxSem, didn't cause a hiccup that i see [21:25:26] awjr: MaxSem what is our cache hit rate on mw.org before we make our x-device change? [21:25:43] dr0ptp4kt: yurik : did you guys find it useful attending the showcase today? [21:25:54] i'd like to involve the zero set of projects in the next one [21:26:05] I don't think that stats are available per-site [21:26:09] tfinc we don't have per project stats [21:26:14] only per-server [21:26:18] tfinc, yes, it was and is good. [21:26:21] tfinc, i didn't attend the showcase - i'm in s africa [21:26:29] awjr: how will you know that the cache hit rate is better/worse ? [21:26:47] but i dont think that's a problem tfinc, the main point of testing on mw.o is not to deteremine if cache hit rate is better/worse, but that our changes don't introduce regressions [21:26:49] dr0ptp4kt: good. then we'll pull you guys into the next one. [21:26:57] tfinc, thx [21:27:05] awjr: sure. but i want us to know what it is when we do deploy to the rest [21:27:05] once we have it deployed everywhere, we can look at whether or not our cache hit rate has improved [21:27:10] do we only have total aggregate now ? [21:27:15] yeah, that's easy tfinc, the answer is in ganglia [21:27:20] k [21:28:19] tfinc, would love some details on what was happening [21:29:05] tfinc to find the ratio, look at one of the mobile varnish boxes in ganglia (eg http://ganglia.wikimedia.org/latest/?c=Mobile%20caches%20eqiad&h=cp1041.eqiad.wmnet&m=cpu_report&r=hour&s=by%20name&hc=4&mc=2) and look at cache hits vs client requests [21:30:28] awjr: did you find out if there's a way to find larger articles? [21:31:19] jcmish there's Special:LongPages [21:31:26] Hohhot [21:31:31] but i think we want something more complex than just a long page [21:31:35] Opps [21:31:44] the longest pages on mw.o are not particularly useful for our purposes [21:31:52] jcmish: did you see the email i sent suggesting a few articles to use? [21:32:12] Yup saw that [21:32:16] groovy [21:32:45] the flow article i actually pulled out of the top 50 or top 100 from Special:LongPages [21:32:50] plus it has some complex elements [21:32:58] Just wondered if you found them some easy way [21:33:02] we could temp pull a few WP articles [21:33:15] :) wanted to add it to my arsenal! [21:33:18] the other two i asked around for :) they were suggested by sumana and robla [21:33:26] Ha cool [21:33:37] sumana also suggest importing specific articles from enwiki [21:33:41] yurik: we have yet to record these but dr0ptp4kt can fill you in on the big piece and then we can give you more info as needed [21:33:46] but i don't think it's really necessary for the simple testing [21:33:49] jdlrobson, have you seen my replies? [21:34:48] Yeah agreed awjr [21:34:52] tfinc, thx, i'm about to collapse, still a bit jetlagged. Gnight [21:38:45] peace, yurik [21:38:53] MaxSem: https://gerrit.wikimedia.org/r/#/q/geodata+is:open,n,z [21:41:34] awjr, awesome dude! [21:42:50] MaxSem: we should raise a bug for that addModuleStyles problem [21:42:58] i'm just grepping the code now [21:43:20] jdlrobson, but meanwhile... it's a showstopper for further deployment [21:43:34] https://commons.wikimedia.org/wiki/File:Commons-iOS_demo_2013-04-15.webm \o/ [21:45:41] Krinkle: is there any way to control the order of styles outputted that have been added via addModuleStyles ? [21:45:50] MaxSem: awjr did we get a link to point people at to capture the device info? [21:45:58] jdlrobson: No, first in first out. [21:46:08] addModuleStyles is discouraged. [21:46:10] MaxSem: slapped together a script and put it up somewhere, he should have a URL for ya [21:46:16] er jcmish ^ [21:46:16] Krinkle: we have styles that need to work without js [21:46:21] jcmish, http://staging.wmflabs.org/ua.php [21:46:28] so if you run addModuleStyles('foo') addModuleStyles('bar') the styles of bar get outputted after foo [21:46:29] ? [21:46:34] Need to work without js and need to not depend on js isn't the same. [21:46:42] jdlrobson, yes [21:46:44] Krinkle: both is what we need [21:46:47] Just add it with mw.loader? [21:47:03] mw.loader doesn't exist Krinkle on old blackberries [21:47:28] Thanks MaxSem [21:47:45] jdlrobson: so if you run addModuleStyles('foo') addModuleStyles('bar') the styles of bar get outputted after foo < Krinkle is this true? [21:47:48] Then none of the modules are there, so how can there be sophisticated dependencies? Basic output and skin stylesheet. What else is being loaded? [21:48:02] jdlrobson: Don [21:48:11] Krinkle: a device specific stylesheet that needs to run AFTER every other style [21:48:57] (to deal with device specific css problems such as bad box model support) [21:48:59] jdlrobson: I can't give a generic answer because I don't think this situation should exist. [21:49:21] What kind of device specific css is there that can't be in the same stylesheet? [21:49:58] Is there a way to style it that degrades gracefully? [21:51:00] Krinkle: not in some cases. For example on a Nokia N95 background: none gives a black background [21:52:51] jdlrobson: Does a transparent PNG work? [21:53:02] (is't 'none' the default?) [21:53:23] Krinkle: please don't start me on a conversation about css hacks :) [21:53:25] Is it black by default or black when you specify none [21:53:43] You can do it alone if you like. Just trying to help. [21:53:52] you forget i'm a purist - if i had my way we wouldn't support any browser that doesn't follow standards :) [21:54:39] Nokia N95 is from 2006/2007. They don't support javascript? [21:54:51] New review: Jdlrobson; "Can you put a comment in MobileFrontend explaining the importance of the name and why and maybe refe..." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59157 [21:55:18] Krinkle: it's not a case of whether they do support javascript it's a case of whether we _should support them. [21:56:11] jdlrobson: Wait, there is separate output by user agent? So you have control over the html? [21:56:28] Krinkle: no the css and js [21:56:31] it varies on X-Device header [21:57:27] What I'd do (in general, not for this specific thing) is come up with a very very basic stylesheet (google ie6 universal stylesheet, something like that) and server that to any browser that is not just unsupported in js but is also incapable of rending the basic CSS 2.1 skin styles [21:57:47] (and serve it instead of the skin stylesheet) [21:59:13] should probably only concern with text (make it not Times New Roman 12px with bright blue underlined links) [22:00:17] Krinkle: yes something like that could work - in fact we use less so in theory we could even go as far as serving modified the main skin with some additions [22:01:29] currently the mobile skin tends to work well cross device with a few slight quirks so i'm reluctant to use a universal stylesheet at the moment - i feel like with minor tweaks we should be able to support the same design [22:02:06] awjr: i'm about to give up with my device testing - my blackberry v5 won't switch on anymore either [22:02:22] v4 is fine though so i assume it will be okay :) [22:02:36] nokia n95 still won't get beyond the dark display with nokia on it :) [22:03:10] awjr, https://gerrit.wikimedia.org/r/#/c/59087/ [22:05:00] jdlrobson, come on up to 6 if you wanna join me & vibha [22:05:19] Maryana: sure i'll head on up [22:06:08] New patchset: MaxSem; "Device-specific styles should be last in the list" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59157 [22:06:12] jdlrobson, ^^^ [22:09:04] New patchset: Jdlrobson; "Device-specific styles should be last in the list" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59157 [22:09:24] Change merged: Jdlrobson; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59157 [22:30:41] New patchset: awjrichards; "Ensures phpunit run from the makefile respects MW_INSTALL_PATH" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59338 [22:31:09] New patchset: awjrichards; "Ensures phpunit run from the makefile respects MW_INSTALL_PATH" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59338 [22:46:01] MaxSem: https://gerrit.wikimedia.org/r/#/q/geodata+is:open,n,z :) i went ahead and updated your makefile, so if you like the change go ahead and merge [22:46:10] cool [22:47:24] MaxSem: similarly… can you take a look at https://gerrit.wikimedia.org/r/#/c/59338/2? not urgent, but would be nice to get in [22:47:50] why does it require a bootstrap while GeoData doesn't? [22:49:40] there's an XML file that sets some special magic for MobileFrontend when phpunit is run [22:50:01] and it needs to point at the phpunit bootstrap file [22:50:13] since GeoData doesn't ahve custom settings, doesn't need it [22:50:17] aha... so it's caused by the suite [22:50:21] yeah [22:57:52] argh [22:58:01] jdlrobson: the thumbs in nearby view are squashed again :( [22:59:04] Change merged: jenkins-bot; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59175 [23:02:08] mhurd: running about 5min late due to my existing meeting [23:03:24] tfinc: no worry! [23:06:46] awjr: in master? [23:06:54] yes jdlrobson [23:06:56] * jdlrobson has an hour to fix search bug :( [23:07:34] jdlrobson: i dont think it's a huge deal considering it's just in beta, so i would prioritize search bug over unsquashing nearby thumbs [23:07:54] i'm not sure why it's broken [23:08:14] awjr: can you checkout an older commit and work out what introduced it for me? [23:08:28] oops MaxSem i messed up resolving a merge conflict and broke GeoData tests [23:08:32] jdlrobson: sure in a sec [23:09:45] MaxSem: https://gerrit.wikimedia.org/r/#/c/59347/ [23:10:16] doned [23:10:19] thanks! [23:15:18] mhurd: i'm going to head out, got to start laundry at home while it's still light. :) will be online later to review & merge the about code when you're ready [23:29:39] New patchset: Jdlrobson; "Rewrite search as an overlay" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/58332 [23:29:47] * jdlrobson needs brion [23:30:19] ori-l: could you review https://gerrit.wikimedia.org/r/#/c/58332/ for me before our 5pm code cutoff? [23:30:24] brion took a look already [23:32:08] jdlrobson: the squished thumbs problem is weird; it looks like it was broken when the fix was merged (i wonder if we didn't notice it because the changeset hadn't been rebased against master in gerrit?) [23:32:52] last time the thumbs weren't squished was in 97faaad, which was before the merge for the fix [23:34:17] New patchset: Jdlrobson; "Story 438: Log uploads from nearby and watchlist" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/58641 [23:34:26] New review: Jdlrobson; "rebased" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/58641 [23:34:26] Change merged: Jdlrobson; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/58641 [23:34:52] awjr: o_O [23:35:10] awjr: file:///Applications/TweetDeck.app/Contents/Resources/htdocs/default.html# [23:35:15] https://twitter.com/leebehrens/status/323942521786994688 [23:35:17] ^ awjr [23:35:27] WE NEED TO MERGE https://gerrit.wikimedia.org/r/#/c/58332/ for my Sanity :D [23:39:53] sanity? what do you need that for? [23:43:22] looking jdlrobson [23:43:24] too many people moaning/upset about it [23:48:52] MaxSem: so you really want me to leave the if statement on https://gerrit.wikimedia.org/r/#/c/59171/ ? [23:49:13] it wouldn't hurt [23:49:31] and will actually prevent problems in unusual situations [23:49:58] can be merged without it of course, but more checking is always good [23:50:18] MaxSem: personally i'd rather see people complaining about it being broken in beta and fixing it [23:50:23] (the api side) [23:50:29] hah [23:50:40] do we see many of such complaints? [23:50:43] MaxSem: if it was going to stable i'd be inclined to agree with you [23:51:00] MaxSem: we could also add event logging in future if you really see this as being a problme [23:51:02] stuff tends to move to stable eventually [23:51:18] we have no evidence that it should not return coordinates correct? [23:51:32] have you ever seen an api response to that type of url that doesn't have them? [23:51:43] or is there any situation in which you believe it could happen? [23:52:10] i think the important thing is to get this change merged as currently the list of nearby results is not correct [23:52:18] * jdlrobson notes we now have 8 minutes till cutoff [23:53:06] okay [23:53:59] MaxSem: I can work to setup an event logging schema that catches problems with the api if you are worried about it [23:54:04] jdlrobson, merge conflict [23:54:25] we should log all exceptions anyway [23:54:36] MaxSem: will fix now [23:54:39] MaxSem: agreed [23:54:44] New review: awjrichards; "(1 comment)" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 1; - https://gerrit.wikimedia.org/r/58332 [23:55:21] New patchset: Jdlrobson; "Address nearby FIXME using colimit" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59171 [23:56:21] so awjr i've also worked out the problem with squished images.. [23:56:30] oh? [23:56:46] simple fix.. [23:57:27] cool - i still dont really understand what happened... [23:58:11] awjr: actually no i haven't [23:58:26] heh [23:59:21] Change merged: MaxSem; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/59171 [23:59:35] I think they are supposed to be somewhat squished.. i'd be interested to know whether the same images are squished on watchlist..