[00:36:10] bmansurov: What is the Commons category media viewer? [00:36:52] kaldari: it's a link to commons category where you can see more images i think [00:41:58] bmansurov: Is there a URL where I can see it? I've never even heard of it. [00:42:39] kaldari: https://en.m.wikipedia.org/wiki/Brad_Pitt?mobileaction=alpha [00:42:46] kaldari: click on the wikidata infobox [00:42:52] you'll see the link at the bottom [00:43:30] thanks, I see it now [00:43:38] np [07:22:00] phuedx, I figured out what they were giving us at WWW and calling "coffee". It's nescafe. I found some at my AirBNB in copenhagen since I needed some caffeine and it tastes the same. Like burnt marshmallows and devoid of any flavor I identify as coffee. [07:22:16] *tastes the same as the stuff at WWW. [08:13:54] good morning [08:47:40] hey halfak [08:47:45] bleh -- also good catch [08:47:54] halfak: you're in copenhagen? [08:47:58] hey hashar [08:50:28] phuedx: good morning :) [08:55:41] phuedx, yup. Our here for a workshop. [08:55:44] *out [08:56:12] o/ hashar [09:30:12] good morning! [10:11:04] hey joakino [10:11:50] hey phuedx ! [10:12:25] phuedx: hey i'm testing out locally http://www.mediawiki.org/wiki/Extension:GettingStarted but getting no related results, do you know how can I feed it those "categories" that it needs? [10:13:55] i was about to send an email asking [10:14:12] ah, one sec [10:14:19] i'll dig out the relevant config [10:18:55] joakino: https://github.com/wikimedia/operations-mediawiki-config/commit/23f2da85c2511c72ad8ec7a64b70b8fd58951ea4 [10:19:04] ^ take a look at that, and i'll try and explain it here [10:19:20] on it [10:19:31] joakino: first enable it, obvs [10:19:56] second: map the 'copyedit' task to a category [10:20:05] phuedx: i guess vagrant did some crap for me, not sure how much [10:20:28] make sure you make the entry like: 'copyedit' => array( 'Category:Title without spaces' ) [10:21:43] then you'll need to run: cd /path/to/GettingStarted/ && php maintenance/populate_categories.php [10:21:55] which grabs the entries in the category and shunts 'em into redis [10:22:08] phuedx: should I do 'morelike' instead of 'copyedit' since that's what I'm querying? [10:22:35] joakino: sure [10:22:55] yeah -- the api will validate that the task exists and then do a thing [10:23:10] so as long as you're configuring "morelike" and requesting the "morelike" task then that's cool [10:24:15] oki, gonna try it out [10:28:27] phuedx: why is there no morelike config on that file? it doesn't make much sense that I would have to list hundreds of categories on the config file, right? [10:33:32] joakino: i don't understand the second part of yer question [10:33:58] joakino: because we were only ever interested in getting users to edit pages during the experiment [10:34:02] hence "copyedit" [10:36:21] phuedx: i understand, but the "morelike" option on the api works, that's the one I'm using per matt's suggestion, see http://en.wikipedia.org/wiki/Special:ApiSandbox#action=query&prop=pageimages&format=json&piprop=thumbnail&pithumbsize=200&pilimit=3&generator=gettingstartedgetpages&ggsgptaskname=morelike&ggsgpexcludedtitle=Programming&ggsgpcount=3 [10:36:41] that's the one I'm interested in getting working locally, but I'm not sure how [10:37:35] yeah, don't use copyedit, use morelike [10:37:58] i was agreeing with you in general, but the example config i showed you was just configuring copyedit instead :/ [10:48:02] phuedx: so you mean that in some config there is a huge array of categories mapped for morelike? manually? [10:50:20] joakino: sorry, mibad [10:50:24] i forgot how they were mapped [10:50:30] morelike means "use cirrussearch" [10:50:46] i don't think you need any other config than the enabled flag [10:50:48] so it should work without special config? [10:51:12] yup, sorry [10:51:16] i'd sorta forgotten [10:51:21] it was only february last year :/ [10:52:45] phuedx: no worries [10:53:58] i'm going to manually enable cirrussearch in vagrant, as it doesn't seem enabled by default [10:54:55] huh [10:54:56] :/ [10:57:17] magic http://media.giphy.com/media/WjKv3iNnPZpsI/giphy.gif [11:05:54] so funny https://media1.giphy.com/media/9x2VKqvE7n71K/200.gif [12:20:40] mhurd: Please respond on your talk page, this is going nowhere. https://www.wikidata.org/w/index.php?title=Q830&action=history [12:21:28] According to our policy, a description on a Wikidata entry is a short phrase designed to disambiguate items with the same or similar labels. https://www.wikidata.org/wiki/Help:Description [12:25:51] sjoerddebruin: so you're overloading "description" to be something other than a description? that seems... odd. especially given wikidata itself uses the descriptions when it presents search results - as does both the iOS and android wikipedia apps, and i believe both desktop and visual editor plan to do the same... [12:26:32] In the beginning, there were no plans to use the description outside Wikidata if I can remember. [12:26:48] sjoerddebruin: no plans to use wikidata outside wikidata??? [12:26:57] "the description" [12:27:02] that's data [12:28:01] sounds like you'll need an "internal disabiguating description" if that's really what you're trying to do [12:28:18] That was the original designation for the description field. [12:29:18] sjoerddebruin: what do you intend to call the "description" then? [12:30:19] I don't know, but you should start a discussion instead of starting a small edit war on a item. [12:30:43] not warring at all. just entering a description into a description field [12:30:50] changing* [12:31:25] the old description makes so little sense it's been passed around as a joke :( [12:31:25] Four times. [12:35:27] All I wanted to say, it's really weird for a other user to see someone editing every time without looking at their talk page. [12:41:00] sjoerddebruin: i didn't consider entering a good description into a description field would be controversial. [14:45:06] gonna take a break for lunch [15:31:51] back [15:35:41] mdholloway: do you have any idea how the android app navigates to or checks for main pages? [15:35:51] i.e. go to main page for $wiki or is $title a main page? [15:42:42] bgerstle: to check whether the page is a main page, we check for "mainpage" in "pageprops" in the api response [15:44:02] gotchya [15:44:25] mdholloway: don't you also navigate to the user's wiki (based on language settings) on startup? [15:44:32] how do you navigate to main page at that point? [15:45:28] Yeah, looking at that now, want to double-check the class the request was constructed in appears to have been refactored out of existence :/ [15:45:40] *want to double-check but... [15:56:14] bgerstle: ok, so there's an automatically generated file containing a list of main pages for each language: [15:56:25] bgerstle: https://github.com/wikimedia/apps-android-wikipedia/blob/b9f1c6d817ec8ac95a520feb277a315a52e333f8/wikipedia/src/main/java/org/wikipedia/staticdata/MainPageNameData.java [15:57:18] bgerstle: so we just grab the values from that in constructing the URL to request in PageActivity.displayMainPage() [15:59:04] mdholloway: yeah, we're doing the same thing on iOS [15:59:21] and crashing on startup if the user's language doesn't have a key/value pair in ou rlist [15:59:22] our list* [15:59:37] i'm talking w/ API people in #wikimedia-dev [15:59:44] apparently you can query the site to find out it's main page [15:59:54] curl 'http://en.wikipedia.org/w/api.php?action=query&meta=siteinfo&format=json' [15:59:59] mdholloway: ^ [16:00:18] mdholloway: the remaining problem is "how do we know the site URL?" [16:00:31] do you know what standard Android uses for language codes? [16:00:58] yeah, i was just looking at this yesterday, let me find you a link [16:01:26] bgerstle: https://developer.android.com/reference/java/util/Locale.html [16:01:45] mdholloway: apparently we use ISO-639-{1,2} [16:01:58] seems you do as well [16:02:11] greg-g: my connection is bad [16:02:12] bgerstle: yep [16:02:57] bgerstle: i haven't heard of crashes for that reason (though of course that doesn't mean it isn't happening) [16:03:12] mdholloway: yeah, i found out about in a roundabout way [16:03:27] and only now that iOS *finally* has crash reports from Apple, can i actually see the crash being reported [16:03:31] of course there aren't ANY alerts [16:04:00] i just saw an OTRS ticket from a user i had previously talked to that said "and other users on Norwegian app store seem to have having problems as well [with crashing on start-up]" [16:04:56] so i fired up the app w/ the simulator set to Lang => Borsk & Region => Norway, and 💥 [16:05:31] mdholloway: do you guys at least have handling for cases where the list doesn't contain the specified lang? [16:07:18] i'm doing some testing now. given how our app behaves elsewhere, it's more likely that we include the language in the list, so it doesn't crash (but is effectively unreadable). there are certain languages (e.g., cherokee) that no version of android supports but can still appear in, for instance, our "read in other languages" list [16:07:43] bgerstle: it's actually a tech debt backlog task i was looking at yesterday [16:09:21] bgerstle: but no, then it couldn't be the case that it would be the system language anyway. hmm. [16:09:34] coreyfloyd: bgerstle anything for https://phabricator.wikimedia.org/tag/roadmap/ ? i'm on the regular meeting for this [16:09:44] phuedx: ^ anything from you? [16:09:45] bgerstle: scratch that, i was responding to myself. [16:09:53] bgerstle: ha. [16:09:55] dr0ptp4kt_i: we're working on 4.1.4 release. just started testing w/ TSG yesterday [16:10:07] mdholloway: hehe np [16:11:13] dr0ptp4kt_i: working on re-enabling mobilewebsearch logging that won't destroy eventlogging [16:11:17] eom [16:13:32] bgerstle: i wonder if we use the same python script to generate our list of languages. generate_wiki_languages.py? [16:13:34] bgerstle: bgerstle phuedx got cards that can be tagged? which week should they be dropped into? [16:13:42] (with the #release project) [16:13:51] ^ wasn't aware of that [16:13:54] will tag now [16:14:12] mdholloway: no idea, ours is a JSON object w/ key/value pairs [16:14:12] greg-g: i'm trying to re-connect to the hangout [16:14:27] bgerstle: no, ours looks android-specific [16:15:26] dr0ptp4kt_i: T99788 [16:15:33] should be merged soonish [16:15:43] dr0ptp4kt_i: actually there isn't, we nixed the last release because it got blocked [16:15:47] bgerstle: and draws from the list i mentioned earlier with languages android doesn't support on it. so the user would end up with a main page with blank space or empty boxes or the like instead of meaningful script, but the app wouldn't crash. [16:15:53] dr0ptp4kt_i: let me create a new task for 4.1.4 (130) [16:16:12] mdholloway: well at least it doesn't crash :-) [16:16:47] phuedx and bgerstle - thanks, if you would please also drag the card into the appropriate week, bonus points. i can drag it if you tell me which week, too [16:18:12] dr0ptp4kt_i: done [16:18:35] dr0ptp4kt_i: ok, will let you know when it's up [16:18:50] dr0ptp4kt_i: unfortunately, i need to re-assess the situation due to a potential blocker [16:21:16] mdholloway: not sure what the priority is for you, but i'm going to fix this now on iOS [16:22:06] i'll probably set up some fallback logic which should handle the cases as best we can. namely: use standard codes, on error, point users at at the site matrix and ask them to specify the language code for their wiki automatically [16:22:08] site matrix: https://meta.wikimedia.org/wiki/Special:SiteMatrix [16:22:36] bgerstle: worth writing up a phab task and looking into, for sure [16:23:01] bgerstle: oh, nice [16:23:36] s/automatically/manually/ [16:24:04] mdholloway: the idea is that standard codes should "just work" but there are edge cases where wikis were created before standard codes exist and we haven't yet moved them or set up a redirect [16:24:37] bgerstle: that makes sense, do you know any examples offhand? [16:24:49] would also be great to slap some event logging on the error case, so that we can measure which wikis are failing when access is attempted through standard codes [16:25:21] mdholloway: according to anomie: "In at least one case the non-standard code we picked (alswiki for Swiss German) now belongs to a different language (Tosk)." [16:25:28] so, there's also that [16:25:32] :-D [16:26:39] but yeah, in that case users should be able to "figure it out" and specify the site/domain manually [16:27:11] but, in the vast majority of cases, using ${ISO_lang_code}.wikipedia.org should work. just need to handle edge cases [16:27:37] bgerstle: there's at least one situation where our current behavior is to default to English if a language code isn't found in the list, just to prevent a crash. but yeah, better to let the user specify [16:28:04] yeah... gotta love kicking cans down roads [16:29:51] coreyfloyd: bgerstle remind me, did event logging for Schema:Search get implemented on ios yet? [16:30:28] dr0ptp4kt_i: i believe so. i have to double check to see when it was added [16:30:44] dr0ptp4kt_i: slated for release in 4.1.4 [16:30:51] bgerstle: was that coreyfloyd ? [16:30:55] dr0ptp4kt_i: yep [16:31:06] bgerstle: thanks [16:37:36] jdlrobson: yo [16:37:44] quieter in here [16:50:42] mhurd: You around? [16:51:03] Could you please respond to messages left on your talk page? See https://www.wikidata.org/wiki/User_talk:Montehurd [16:53:18] multichill: hi, i'm another iOS developer at WMF [16:53:43] just so you know, he probably made that edit as a test for a hack he did at the WMF Lyon hackathon this past weekend [16:54:17] I thought so too but than noticed that his latest edits are today [16:54:22] bgerstle: no that's not related [16:54:25] oh ok [16:54:36] carry on, then :-) [16:54:36] See https://www.wikidata.org/wiki/Help:Description [16:55:04] mhurd: I wasn't sure if you were seeing the page talk ping because you were doing mobile edits [16:55:14] mhurd, actually, this makes me think... if we start letting people edit Wikidata stuff from the app, we should support "Talk" [16:55:15] multichill: I'm afk - scroll up to a couple of hours ago to see my input on the matter [16:55:39] mhurd: You should respond on https://www.wikidata.org/wiki/User_talk:Montehurd [16:55:42] maybe not "raw" talk pages, but something on top of it [16:56:25] multichill: I will when I have time and inclination and not before [16:56:28] or, write a separate app just for "talk" a la FB messenger. though i don't necessarily agree w/ the separate app approach, could be useful for people to get messages from non-wikipedia wikis [16:57:16] bgerstle: ya def something to consider. [16:57:44] bgerstle: multichill grabbing food - back later this evening if I can steal away :) [16:57:46] mhurd might want to change your IRC nick so we know you're AFK [16:57:57] mhurd don't sweat it, enjoy your travels! [16:58:13] bgerstle: ya I'm too lazy about that lately [16:58:15] are you still in milan? [16:58:29] mhurd: FYI you can just type "/nick mhurd_afk" in IRC cloud [16:58:43] I feel some passive aggressiveness in your messages mhurd. Don't shoot the messenger. Wikidata is a community project. If someone brings up something on your talk, just respond to it before you continue editing. [16:59:04] multichill: I felt like you were demanding I do something [16:59:35] Nope [17:00:07] multichill: well that's how it read. Next time I'll give you the benefit the doubt :) [17:00:23] multichill: I get cranky when I'm hungry [17:00:32] That's the default mode you know, asume good faith ;-) [17:00:35] mhurd: you should have a snickers [17:00:58] multichill: well aware. My bad. [17:01:14] bgerstle: lol [17:02:12] mhurd: https://www.wikidata.org/wiki/Wikidata:Administrators%27_noticeboard#User:Montehurd FYI [17:02:28] bgerstle: Coming back to your point. Yes, talk is important [17:03:05] Maybe focus on flow support? [17:03:28] multichill: perhaps... might be difficult given the unstructured nature of conversations on talk pages [17:03:42] but, such is the nature of all things wiki [17:04:03] actually, i'm not even sure what flow is [17:04:10] Dude! [17:04:12] bit of a wiki novice, myself [17:04:26] Didn't you go to the hackathon like mhurd? [17:04:31] multichill: i know :-( [17:04:40] yeah [17:04:57] did i miss something? [17:05:16] bgerstle: Short version: New interaction system that is more structured and has an api [17:05:19] Long version at https://www.mediawiki.org/wiki/Flow [17:05:40] At the hackathon someone made a commandline version of it using the api, that was pretty cool [17:05:48] bgerstle: ya flow API may be the way to go. It's a big proj though [17:05:50] multichill: http://data2.whicdn.com/images/48904992/large.jpg [17:06:47] Problem is that flow won't be enabled everywhere for ages so you still end up needing to support the old system for a long time [17:07:39] again, such is the nature.. [17:08:54] i do like the idea of a standalone app... but ultimately is a product decision. we might not do anything for a while since we're focused on "reading" use cases atm. that said, it might still be valuable to have a talk/flow client [17:09:53] multichill: either way, thanks for the tip. i'll try to ping you if anything happens on this front [17:10:00] gotta focus on some other stuff now. cheers o/ [17:10:20] Me too, cooking! :-) [17:10:28] Nice talking to you and have fun at the WMF [17:10:44] thanks, enjoy the foodz [17:10:48] And do have a look at https://tools.wmflabs.org/wikidata-game/# some time bgerstle [17:11:12] hrmmm [17:11:24] will do [17:11:28] * bgerstle bookmarks it [17:20:39] phuedx: bmansurov yo [17:21:04] so the human generating a release works as a temporary measure but whilst we do it we basically need two people to do it [17:21:17] one person to write the release patch and the other to merge it asap before another patch gets merged. [17:22:12] jdlrobso_: agree, or we should look at how VE team is doing it [17:22:15] or core [17:23:05] yeh so hence why i felt we should have a chat about this [17:23:33] * phuedx wonders whether or not he should stop gerrit accepting patches for a while [17:23:44] there are 29 (ish) open patches [17:23:50] VE use package.json for oojs ui [17:24:54] not sure if they do releases for VisualEditor and if they do how [17:24:58] i mean oojs ui release notes [17:25:18] i get emails from them regularly [17:27:10] bmansurov: I write them. :-) [17:28:03] James_F: thanks, they are awesome [17:29:19] James_F: is your release process documented anywhere? MF could use some ideas. [17:30:29] bmansurov: oojs ui is a node module though [17:30:34] so it becomes a lot easier.. :) [17:30:52] add 'em to https://phabricator.wikimedia.org/T100296 when you find out [17:31:44] given that we're not actually sure what we want those release notes for, is it worth generating 'em now? [17:32:00] the unease is coming from the fact that we don't know what our process looks like [17:32:51] feels like a sync might be in order [17:33:10] phuedx: FWIW we've been creating phab "release tickets" which are templated w/ a checklist of stuff we need to do for each release: https://phabricator.wikimedia.org/T99019 [17:34:01] bgerstle: that's nice -- the team could create one for every sprint [17:34:14] phuedx: yes, we can wait. [17:34:27] phuedx: that's essentially what we do.. i'd almost prefer for us to have the sprint "projects" *be* the release tickets [17:34:40] since we try to release stuff at the end of every sprint [17:34:40] bgerstle: you're reaching [17:34:49] it's not a perfect world [17:34:50] :P [17:34:55] phuedx: NOT UNTIL WE MAKE IT PERFECT [17:34:59] can't [17:35:10] entropy [17:35:15] it's destined to be chaotic [17:35:17] maybe we can settle for an excellent one by striving for it ;-P [17:35:18] just accept it [17:35:24] let the despair wash over you [17:35:32] make a release anyway [17:35:51] yolo [17:35:57] although, having release tickets allows you to handle weird edge cases like what we're in now on iOS [17:36:06] sprint 56 release was blocked, so we're essentially releasing 56 & 57 [17:36:31] can you have tickets blocked by a project...? [17:36:46] i.e. create "Release 4.1.4" and have it blocked by all tasks in "Sprint 56 & 57" projects [17:37:02] i'll try it and let you know how it goes :-P [17:37:23] all of this gets back to Definition of Done, etc... [17:37:35] must have more red tape! proceduralize all the things [17:37:46] i'm ranting, must be time for lunch [17:37:47] o/ [17:38:59] YOLO indeed [17:40:08] jdlrobso_, joakino, bmansurov: i'd encourage y'all to add thoughts to https://phabricator.wikimedia.org/T100296 [17:40:38] and things that you find painful [17:40:57] i've tried to capture what i feel is painful in the description, but if it's not clear then i can rewrite it [17:41:24] this sprint isn't about pushing patches, it's about refining/tidying up [17:42:20] +1^, but can we get the patches that are already pushed merged? [17:42:52] all 29 of 'em? [17:43:58] i see only 11 [17:44:10] make gerrit [17:44:14] bmansurov: https://gerrit.wikimedia.org/r/#/q/status:open+project:mediawiki/extensions/MobileFrontend,n,z says different [17:44:21] -1 counts [17:44:32] also only 2 more sprints to go before http://www.imdb.com/title/tt0029808 [17:45:43] i just have a patch that needs to get merged before the code is pushed to prod. https://gerrit.wikimedia.org/r/#/c/214113/ [17:46:38] jdlrobson: ^ you've been hot on the discussion (and it's been a hot discussion) so could you follow up with bmansurov before pushing anything new [17:46:51] (you also seem to have a grasp on this caching malarky) [17:47:17] phuedx: oh shoot new i forgot something [17:47:28] ta [17:47:48] bmansurov: i'm still extremely confused how a html change can fix this though [17:48:20] * jdlrobson quickly checks before leaving [17:48:35] jdlrobson: sorry for making you extremely confused. It's another issue as stated in the commit message: "Also fix nested s." [17:48:58] bmansurov: maybe confused is the wrong word.. basically i suspect your patch isn't right.. [17:49:15] jdlrobson: i would appreciate a constructive review [17:50:19] i'm being constructive. I'm checking out the patch now and testing it but these issues can only be fixed by css or javascript as the html as i've explained is cached [17:59:53] so bmansurov it works but the text is not truncated on cached pages as i suspected. Adding #mw-mf-last-modified a.truncated-texte { display: block; } will fix that for cached pages. I'm also seeing links within a link. [18:00:02] it fixes the padding and the background though [18:00:25] (checked with ./dev-scripts/cachedpage.sh 67f521b0cbd91ad5d71d5b5e1cba5648e0d76c4c 'Benedict_Cumberpatch' ) [18:00:41] i have to leave now as i'm late for dinner plans but i'll sync up tomorrow afternoon [18:00:57] jdlrobson: thanks [18:30:02] JonKatz: If you couold look at https://phabricator.wikimedia.org/T100297 today, that would be awesome :) [18:37:06] okie poke: dinner [18:37:59] sudo su [18:38:08] err, wrong window:P [18:39:21] MaxSem: what'd i tell you about trying to use sudo in IRC cloud? :-P [23:29:30] I think something is broken with infoboxes on beta: http://en.m.wikipedia.beta.wmflabs.org/wiki/Animal vs https://en.m.wikipedia.org/wiki/Animal [23:30:02] not sure if it's been like that for a while or a recent development