[00:01:21] upload works locally but i'm not seeing the image in the article until i reload [00:01:24] (in chrome) [00:02:11] that's no different to master though :D [00:02:59] New review: Brion VIBBER; "Patch Set 1: Verified+2 Code-Review+2" [mediawiki/extensions/MobileFrontend] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/49195 [00:03:01] Change merged: Brion VIBBER; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/49195 [00:06:51] brion, what browser? [00:06:59] jgonera: tested chrome [00:07:09] on mac [00:11:16] brion, and you don't see the image until you reload? anything in the JS console? [00:11:37] ugh, why can i not stay logged in to any WM project today? [00:11:54] lemme check [00:12:54] jgonera: nope nothing in console [00:13:04] it pops up a toast "Success! Your image is now live on this page." though [00:13:22] and the image doesn't show up, weird, it does for me [00:14:16] lemme make sure everything's up to date [00:15:07] yep, no difference after update to latest core & MF code [00:15:28] hm, I'll have a look at it when you're in the office [00:16:02] safari doesn't show the upload button at all [00:17:09] hm, neither does chrome anymore [00:17:15] does it stop offering to upload after a certain number of images? [00:17:48] no, it shouldn't [00:18:16] there we go, if i remove the images now safari and chrome prompt me to add a first image [00:18:26] still doesn't show the image tho [00:18:29] until reload [00:19:11] oh, ok, you mean on the article [00:19:18] yes, it shows up only if no image is present [00:19:32] I'll double check that later [00:20:35] hmmm, then for some reason i am seeing it sometimes when there is an image ;) [00:25:39] yes, that's a bug ;) [00:26:14] hm, the image shows up for me when I upload it on chrome... [00:26:24] without reloading [00:36:31] awjr: lost you [00:36:52] again? [00:39:23] awjr: i'm going to grab a wire [00:39:25] tfinc are you on wired? [00:39:25] :) [00:44:28] awjr: http://www.dragonsatwork.com/ [01:00:40] [Commons-iOS] brion pushed 1 new commit to master: http://git.io/Ff_1Qw [01:00:40] Commons-iOS/master ae65200 Brion Vibber: pretty fade-in on images loading in the cells [01:06:48] is tomasz around? [01:07:03] tfinc_: ? [01:13:41] Ryan_Lane: i certainly am not [01:13:49] Ryan_Lane: sup ? [01:36:56] [Commons-iOS] brion pushed 1 new commit to master: http://git.io/-vhDUg [01:36:56] Commons-iOS/master e4f4784 Brion Vibber: (Breaking data format!) Save file source for logging... [13:53:00] [android-commons] yuvipanda pushed 1 new commit to master: http://git.io/IoIKsQ [13:53:00] android-commons/master a89e2c8 YuviPanda: Add Device / Platform / App info to EventLogging [13:53:32] Project Android-Commons (mobile) - Nightly builds build #73: SUCCESS in 39 sec: https://integration.mediawiki.org/ci/job/Android-Commons%20(mobile)%20-%20Nightly%20builds/73/ [13:53:33] yuvipanda: Add Device / Platform / App info to EventLogging [15:26:41] MaxSem: around? [15:26:50] no [15:26:56] damn. [15:26:56] precisely there;) [15:27:01] haha :P [15:27:40] MaxSem: I remember you asking me to not use action=parse for getting page HTML. still true? [15:27:52] yes [15:28:04] if it's for the app [15:28:25] MaxSem: well, I'll be parsing info out of the HTML [15:28:31] {{Information}} precisely [15:28:45] what for? [15:28:58] MaxSem: showing description, etc for images [15:29:03] from inside {{Information}} [15:29:17] * MaxSem facepalms [15:29:45] is there a better way than parsing the HTML looking for https://commons.wikimedia.org/wiki/Commons:Machine-readable_data [15:29:56] I'm not going to parse WikiText [15:30:04] MaxSem: ^ [15:30:42] how many pages do you want to parse? [15:31:10] MaxSem: ? [15:31:16] MaxSem: all of a user's contributions, to start with :P [15:31:24] and any other image they happen to see [15:32:11] okay, user goes somewhere this information is needed - how many pages do you want to parse? [15:32:43] MaxSem: 'how many pages' as in? [15:32:50] however many pages the user visits? :D [15:32:59] at once [15:33:29] MaxSem: I'm currently running a query with limit set to 500 for actually getting contributions :P [15:33:39] phewww [15:33:53] and you want to request HTML for them all? [15:33:53] MaxSem: I'm just starting to write the code [15:34:33] MaxSem: If possible, yes. [15:34:37] so I could either pre-cache them [15:34:42] or fire off a request when the user views it [15:34:53] [15:35:08] now multiply this by 500 [15:35:16] yeah, I realize. [15:35:25] that is why I am asking you what I should be doing :) [15:35:34] this is nowhere near acceptable, despite even all the caching [15:36:11] well, I can tell you what you shouldn't be doing;) [15:37:05] maybe I could just use action=mobileview and request just that one section? [15:37:18] MaxSem: but I can't guarantee that the section number would be consistent, so can't... [15:37:44] MaxSem: and I suppose it isn't possible to API-ize this, and have a commonsinformation prop maybe? [15:37:51] action=mobileview has exactly the same worst-case performance [15:38:13] well, it's possible, of course [15:38:23] but not in 5 minutes [15:38:34] is that a 'one hour' thing? a 'one week' thing? a 'one month' thing? [15:40:53] MaxSem: ^ [15:41:17] depends on requirements... [15:41:30] more like 1-2 weeks [15:41:39] lazy [15:41:56] Reedy, what about requirements? [15:42:04] Wanted: Everything [15:42:06] DONE. [15:42:10] could be less, of cource, with full specs:) [15:42:34] http://thedailywtf.com/Articles/Invasion-of-the-Consultants.aspx [15:43:02] * SleepyPanda clicks link [15:45:41] MaxSem: that... was fun [15:51:20] SleepyPanda, do you know Python? [15:52:17] MaxSem: yes, though haven't used it in a few months. Why? [15:52:19] hello brion [15:52:28] hey [15:52:33] yo [15:52:48] brion: how are you getting info of description, etc ({{Information}}) from the pages? I was proposing to MaxSem that we build an API for it, like GeoData :D [15:52:56] SleepyPanda, what would you use for XML parsing in 2.6? [15:53:07] MaxSem: lxml [15:53:28] i was going to pull the description page as html and extract the "machine-readable" stuff, haven't gotten there yet tho [15:53:58] brion: yeah, I'm going to do that too. Except that MaxSem pointed out that they can take up to a couple of seconds to generate, and then we've network transfer, and then the extraction... [15:54:07] we can do that for now, but would be nice to have an API of sorts. [15:54:29] SleepyPanda, where's it in http://docs.python.org/2.6/modindex.html#cap-X ? [15:54:40] MaxSem: it isn't part of standard library [15:54:52] then I can't use it [15:55:18] or rather I don't want that PITA [15:55:30] MaxSem: well, then you're left with the xml module [15:55:55] not much choice there no? [15:56:28] MaxSem: pulldom or sax if you are dealing with large docs, dom otherwise [15:56:39] (I personally prefer pulldom, sax is nuts) [15:56:40] yeah i vaguely remember some work towards a proper api for fetching commons image metadata, but can't find anything on it now. probably was just some experiments on toolserver [15:57:12] brion: but we have a MaxSem! [15:57:23] :) [15:57:26] brion: also are you setting UserAgent properly? I just set it today... [15:57:34] brion: and you should do a pass over my new commits too. [15:57:42] i'm not setting it yet no [15:57:43] ok [15:57:57] xml.dom looks like JS without jQuery :] [15:58:37] MaxSem: indeed. lxml has great selector support, so you can use it exactly like jquery :P [15:59:33] MaxSem: there is http://lxml.de/cssselect.html and also xpath. [15:59:46] MaxSem: lxml has a fairly up to date debian / ubuntu / fedora package, if that helps [15:59:55] heh [15:59:58] and it is highly possible that it is already installed wherever you're running it [16:00:19] the trick is that it's for a lucid server [16:01:42] MaxSem: should still have it installable, I think [16:01:54] nah, pita [16:02:15] javascript without jquery then :P [16:03:33] document.verboseMethodName('something') [16:03:42] nooooooooo the dommmmmm [16:04:02] brion: document.verboseMethodName('somethinglong').verboseMethodName2('somethingelse') [16:04:03] and so on [16:04:04] maybe I should just fucking port it to PHP:P [16:04:17] now now... [16:04:23] okay food time! [16:04:24] brb [16:04:27] brion: and CR! :) [16:04:34] whee [16:06:34] why capitalize the model? [16:08:39] breakfast time…. [16:34:25] mmmm french toast [16:35:01] or as we like to say in objective-c, FrenchToast *toast = [[FrenchToast alloc] initWithEggs:eggArray]; [16:36:37] ughh [16:36:45] it's scary:P [16:36:55] [[[[it's just a few brackets]]]] [16:39:52] (((much like)((((((LISP))))))) [16:40:01] {{or|wikitext}} [16:40:22] it's like lisp and smalltalk had an illegitimate baby but had it surrogate-carried by C [16:41:26] did Jobbs just pick the sickest language at the time? [16:42:44] in the late 80s this shit was pretty cutting-edge [16:43:00] complete with manual memory management at the time [17:05:25] zz_SleepyPanda: where are you ? [17:28:47] hey brion [17:28:54] yo [17:29:03] brion: how are you? :) [17:29:11] doing well :) [17:29:31] i posted a reply on the commons image thread. [17:29:58] how is it going with the app? [17:30:34] coming along, mostly doing internals changes (logging etc) [17:30:54] brion: need anything from me [17:32:21] thx for the reply & the screens! [17:32:36] brion: no problem! :D [17:32:41] not blocking on anything from you right now, but i think there's some things in the queue [17:33:32] https://www.mediawiki.org/wiki/Apps/Commons#Design <- we'll be adding more stuff to the image detail pages (categories, list of pages where it's used). be awesome to have some designs on that [17:34:01] brion: re tagging with cats. I was thinking we could tag with pages instead. [17:34:09] and pick up the category from the page's commonscat [17:34:10] thoughts/ [17:34:12] ? [17:34:55] hrm [17:35:05] tagging which what? [17:35:11] brion: as in, *instead* of categories [17:35:33] brion: have you seen our awesome Wikimedia Foundation Roadmap? https://trello.com/board/roadmap/51143ff97d138cd6200086ad :D [17:36:10] rofl [17:36:35] that should keep us busy until the end of fiscal year [17:37:15] YuviPanda: not sure i understand. like tagging 'this was uploaded with android version 1 blah blah'? [17:37:18] is a tough ask as is, I think [17:37:20] that's done through the template? [17:37:22] brion: oh no. [17:37:36] brion: as in, instead of showing people list of categories, we show them list of pages [17:37:45] so you won't add Category:Taj Mahal [17:37:53] but you'll pick the Taj Mahal page from autocomplete [17:38:10] and we'll find out that commonscat for Taj Mahal is Category:Taj Mahal [17:38:11] and add that [17:38:20] ah i see [17:38:23] for picking the categories [17:38:23] the advantage is for things like the Ellora Caves [17:38:31] there are I think about 20 different categories [17:38:33] but only one page [17:38:41] so we know what the 'main' category is [17:38:45] how do we look this up? [17:38:46] instead of figuring out which of the sub cats [17:38:51] brion: so... we can't :P [17:38:55] html parsing again [17:38:56] haha [17:39:10] but it will be I think a much better user experience [17:39:16] because list of pages is usually much more well defined [17:39:18] than categories [17:39:43] so this would pull from english? or from other languages too? [17:39:52] we need wikidata to be done ;) [17:41:19] brion: heh, yes [17:41:26] I think commonscat would be a niiccee thing for WikiData [17:43:06] let's stick that in the backlog and we'll implement it if we have time and can figure it out :) [17:43:21] using open search suggestions on category namespace should be good enough for a first pass [17:43:24] and much simpler :) [17:43:40] brion: lazy! :P [17:44:03] http://en.wikipedia.org/wiki/YAGNI [17:44:26] hehe :P [17:44:37] brion: but we should seriously consider moving {{Information}} parsing to an API [17:44:45] oh god yes [17:44:48] I don't think WikiData will cover that [17:45:01] well i think it probably should [17:45:04] but it may take ages :) [17:45:23] brion: true. [17:45:39] brion: do you know that wikidata covers GeoCoordinates too? [17:45:47] but I doubt they've searh like how GeoData does [17:46:00] yeah in theory those should integrate at some point [17:46:04] * brion hopes theory is ncie [17:46:06] *nice [17:46:08] :) [17:46:12] oh THAT one autocorrect doesn't catch? [17:46:16] :D [17:46:16] haha [17:46:29] I turned autocorrect off since it kept complaining about typing 'reddit' [17:46:34] YuviPanda: brion as part of the post upload workflows, can we do share in the third iteration? [17:46:58] notnarayan: like sharing the link of an uploaded file? that'd be sweet [17:47:00] MaxSem: do we have to talk to and convince tfinc to let you do {{Information}} parsing API? [17:47:25] brion: notnarayan yup. I think we can do a fair amount of hings in 'post upload' [17:47:35] brion: yeah, sharing the link to a social network perhaps? [17:47:47] yeah nice [17:47:59] shall i add it to iteration 3? [17:48:01] notnarayan: note that we can't do automatic share like Instgram (because for that we'll need to actually integrate into them, rather than doing just intents - and privacy policy, etc) [17:48:46] but we can sure have a share button on the detail view [17:48:55] YuviPanda: not like instagram. share an image on the detail view [17:48:56] :D [17:49:04] notnarayan: isn't that already in this iteration? [17:49:09] I remember seeing a share icon [17:49:39] well, i did put it in for android. :) not for iOS yet. :) that was un panned [17:49:43] *planned [17:51:02] ah, right. a feature android gets for free but iOS will have to work on :P [17:51:08] there's a framework in iOS 6 for sharing that's *almost* generalizable like intents, but hardcodes which services it supports. [17:51:09] YuviPanda: :D [17:51:11] brion: but they added something like that, no? [17:51:12] yeah [17:51:14] i remember that [17:51:21] still that'll cover the twitter/facebook/copy url/email stuff [17:51:30] brion: good enough. [17:51:33] yeah, good enough I think. [17:51:35] yup [17:52:30] yeah i think i'm dropping iOS 5 support. just won't be worth reimplementing these things [17:52:53] brion: lucky you :) but is the marketshare low enough? [17:53:43] i don't have figures but it's waaaaaaay better than android version uptake [17:53:49] that is true. [17:54:00] I'll drop 2.3 support when it gets unde 10% [17:54:02] *under [17:54:07] so that'll be... years? :) [17:54:11] heh [17:54:18] brion: notnarayan I think we should also spam the design list more. paper trial, etc. [17:54:22] trail [17:54:23] not trial [17:54:27] well you've already got intents and a grid view [17:54:50] YOU ARE ON PAPER TRIAL FOR OVERUSING PAPER [17:55:08] :D [17:55:43] hmm ideally we'd get some useful data out of http://stats.wikimedia.org/wikimedia/squids/SquidReportClients.htm [17:56:15] YuviPanda: i have been doing that, shall do it more often i guess. [17:56:23] doesn't seem to have a good version breakdown though oh well [17:56:43] brion: for Android I looked at usage stats for our android app [17:58:32] * brion stumbles around stat1 hoping not to break anything [17:58:57] brion: yeah, I'm sure you have way more rights than I do :D [18:00:42] rmoen|afk: you cycle to the office? [18:01:14] brion: standup! :) [18:06:28] [Commons-iOS] atdt opened pull request #29: Omit 'isValid' (master...master) http://git.io/1MiE0Q [18:07:25] [Commons-iOS] brion pushed 2 new commits to master: http://git.io/2Zm6XA [18:07:25] Commons-iOS/master 7022d19 Ori Livneh: Omit optional 'isValid' field... [18:07:25] Commons-iOS/master 95cbe18 Brion Vibber: Merge pull request #29 from atdt/master... [18:07:42] ori-l: thx :D [18:07:54] :) [18:08:57] ori-l: now you're an iOS developer too :) [18:09:00] i was hearing a weird background noise on the standup hangout and finally realized it's my cat snoring [18:09:08] brion: cats... snore? [18:09:14] YuviPanda: Yeah boss [18:09:17] they do! it's high pitched and cute [18:09:54] rmoen: ah, nice. I just heard a familiar voice running through the standup video, so pinged :) [18:10:24] YuviPanda: SENIOR iOS developer [18:10:45] ori-l: apologies sir, I did not mean to insult you by calling you merely an iOS developer [18:10:49] hehe [18:10:59] hey munaf! Saw the threads notnarayan opened on design-l? [18:13:12] New patchset: Zfilipin; "Updated Ruby gems" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/50032 [18:13:33] aha http://stats.wikimedia.org/wikimedia/squids/SquidReportOperatingSystems.htm [18:13:49] that's got the figures i want under "Breakdown per OS version, mobile" [18:14:00] :D [18:14:48] ori-l: it made your day to fix on iOS, but not on Android? [18:14:50] ios 5+5.1 together are about half our hits [18:14:50] * YuviPanda pouts [18:14:58] hehe [18:15:16] half of our iOS hits, that is. [18:15:22] the other half being mostly 6 [18:15:26] YuviPanda: made my way to stumble across it! I knew you were working on a lib for android but brion was quiet as a mouse [18:15:30] * my day [18:15:46] * brion is a ninja [18:15:56] isNinja: true [18:15:58] * ori-l runs. [18:16:16] hmm i'm not sure i trust these numbers [18:16:30] "iPad OS 1.X" has almost as many hits as "iPhone OS 6.0" [18:16:44] there's no such thing as iPad OS, not to mention a 1.X [18:16:51] hop into #wikimedia-analytics and harass people :D [18:17:13] ori-l-away: :P [18:17:27] tfinc: also you need to put me on the staff and contractors page. [18:18:04] YuviPanda: did HR never do that? i'll make sure you get added [18:18:17] oh gosh those figures were generated in november too [18:18:28] tfinc: nope. shankar neither. [18:18:34] so that's like…. 50% upgrade uptake within a few weeks of iOS 6 coming out :) [18:18:50] brion: they might have newer numbers now. If not they can surely run a small script to find out :) [18:18:51] so poke them [18:19:40] yep [18:19:41] :D [18:22:49] YuviPanda: thanks for the heads up. i mailed rachel who will take care of it [18:23:07] tfinc: thanks :) [18:24:13] brion: i didn't know you were doing the meetup tomorrow. nice [18:24:51] yep. it was gonna be last week but they rescheduled it on account of a) lots of people couldn't make it and b) valentine's day [18:24:53] tfinc: yeah, me neither :P [18:28:10] tfinc: ping. [18:28:17] hmm ? [18:28:49] tfinc: me and brion were wondering if we could steal MaxSem to write an API for parsing {{Information}} [18:29:14] YuviPanda: how would this API for parsing be used ? [18:29:17] He says it would take a couple of weeks. [18:29:22] and what can't we do now that this would provide [18:29:32] tfinc: so currently we will have to grab the HTML for the page and parse the data out of it manually [18:29:46] i will timeout in 1min for the wed manager meeting [18:29:54] but i'm happy to discuss over email [18:29:54] tfinc: on average that is about 3s before you can see anything on an image's details other than a title [18:30:09] yeah if we move the processing to the server we'll have to move less data over the wire [18:30:18] yeah, plus we can query more at a time [18:30:19] plus we can upgrade it to use wikidata when available later [18:30:27] brion: start an email thread? [18:30:31] will do [18:30:38] brion: sweet :D [18:30:48] brion: I'm sure Mobile Web would also find uses for it :) [18:30:54] yep [18:31:02] YuviPanda, sorry I couldn't reply to your ping cause I was in a meeting [18:31:24] np MaxSem. are you still in one? [18:31:26] no [18:32:05] MaxSem: so the API would have to be able to return whatever info is in the page (documented on https://commons.wikimedia.org/wiki/Commons:Machine-readable_data) [18:32:13] yup [18:32:28] I don't know if it needs to be queryable by anything other than file name [18:32:32] brion: ^ thoughts? [18:32:55] typing up a link to that and brief info right now :) [18:33:05] ah sweet :) [18:33:10] MaxSem: so brion is moving it to email :) [18:39:38] sent to mobile-tech [18:40:52] MaxSem: according to rob, the osm servers are all online (though no OS) [18:41:02] cool [18:47:06] brion: for this community testing event how do we get the Commons iOS version out to folx who would like to help testing? [18:47:28] jorm: see my mail about a commons testing list ? [18:47:39] just getting to it now. [18:47:51] or, rather, i'm reading my mailing list messages now. [18:47:57] i got your one about my names yesterday. [18:48:15] that's gonna be more of a "massage people" thing, i think. but yuvi was already on the ringleader. [18:48:28] jorm: indeed. [18:48:58] jorm: ragesoss has also been testing them for a long while. [18:49:28] sage is your guy who is "prolific photographer, uploads everything" [18:49:37] but he is also "good camera" [18:50:22] jorm: anyone else I should be poking? [18:50:34] odder takes a passing look now and then, mostly because he doesn't have a good enough device. [18:50:44] uhm, guillaum. [18:50:45] hmm there's a blizzard warning in tucson... [18:50:45] I suppose I should ping multichill. [18:50:51] jcmish: i'll need to get people to sign up ahead of time to build the keys for an iOS beta [18:50:57] i'll send an email by end of day with more info [18:51:03] YuviPanda: What's up? [18:51:12] will see if we can do it with test flight instead of manually [18:51:29] hello multichill. We were looking for commons people who would be interested in testing the Android / iOS uploading apps we're building :) [18:52:05] but I totally forgot which device you had when we last met... [18:53:05] A broken Iphone :P [18:53:31] oh, *now* i remember :D [18:53:35] multichill: you still don't have that fixed? [18:53:44] Got a new one [18:53:49] broke it again? :P [18:53:59] multichill: a new iPhone, you mean? [18:54:00] No no [18:54:08] Yeah [18:54:27] brion: k cool thanks [18:54:30] brion: ^ test subject :P [18:54:49] multichill: upgraded to iOS 6? [18:54:58] Not yet [18:55:24] currently iOS 6 only, trying to figure out whether it's worth the effort to add back-compat :( [18:55:38] (there's google maps and a jailbreak for iOS 6 now btw ;) [18:55:58] multichill: do you know any other heavy commons user who uses a recentish android phone I can bug? [19:00:27] jcmish_: also the Android app has in built crash reporting. if it crashes a dialog box will pop up asking people if they want to email me with a crash report. [19:00:36] jcmish_: I can change it to a common email address (mobile-feedback-l perhaps) [19:00:48] oh nice [19:01:21] I think it would be better on a common address [19:03:18] jcmish_: mobile-feedback-l? [19:03:25] are you subscribed to that list? [19:03:30] hehe looking through all the listst :D [19:03:36]