[00:00:19] niedzielski: ok, now it worked, after the other patches were merged [00:00:38] r4r fo realz [00:01:09] bearND: thanks, hopefully a one time weirdness. i'm checking out the patch now [00:07:02] Hi mobile! Just to say, the new CN code just deployed everywhere. Looks good so far! [00:08:04] bearND: ok, looks like we're all good :) [00:11:07] niedzielski: \o/ thanks! [00:12:46] bearND: np! thank you for this much needed network strategy switcher and the dev settings code is a real nice touch [11:18:24] hullo [11:45:59] hallo [13:40:52] bmansurov: yt? [13:42:06] yep [14:19:41] niedzielski: fyi, I've commented on https://gerrit.wikimedia.org/r/#/c/235395/ [14:20:54] bearND: thanks, i saw that yesterday but didn't get to following up on it. will do [14:40:33] woah [14:40:41] bmansurov: segfaults during builds :o [14:47:27] bmansurov: i've raised a bug for the segfaul [14:47:29] *t [15:23:47] bearND: hey! i'm rebasing some of my other jenkins patches and had a question on our lint discussion. since gerrit can merge (not just fast forward) on a +2, don't we want to run lint in this case? [15:27:11] niedzielski: yes, that's a good point. [15:28:52] bearND: with this in mind, are you ok with collapsing the lint job into the tests job? [15:29:24] niedzielski: what's the advantage of collapsing gerrit check tasks? [15:32:00] bearND: hm, i guess nothing too great. i think the only advantages are a very modest speed improvement and we can have one jenkins job builder template [15:33:09] niedzielski: ok. It's up to you. I'm fine either way. [15:33:09] bearND: i was thinking that since it's gradle-ized, it could be nice to just run the lot of tests at once: /gradlew -q clean lintAlphaDebug checkstyle testAllAlphaRelease [15:35:21] bearND: ok cool, thanks [16:02:02] dbrant: i can repro the proguard issue locally. i will make the change and push up if that's cool [16:08:37] I will let you know when I see dbrant around here [16:08:37] @notify dbrant [16:31:09] niedzielski: would you check out https://gerrit.wikimedia.org/r/#/c/235678/ and merge if ok? [16:31:38] niedzielski: mdholloway|eat : I want to get the outstanding patches merged before the deployment [16:31:40] bearND: sure thing [16:37:32] hey joakino [16:38:09] joakino: can we skip design review today? it overlaps metrics and trying to catch up on work too [16:38:37] joakino: i've been in touch with jon r about max-width and surveys [16:40:52] bearND: on it [16:48:08] rmoen: my bus appears to be running late, so don't wait for me if I'm late to standup [16:54:36] jdlrobson: noted [16:55:16] \o phuedx [16:55:18] thanks for all the sign off [16:55:20] appreciated [16:55:26] kaity: joakino is on hols [16:55:38] kaity: so i guess you're request is honoured [16:55:47] haha thanks jdlrobson [16:56:32] wish our browser tests didn't take 10 minutes to run :( [16:56:41] ^ new problems [16:59:39] bearND: should i get on the hangout now or wait until your patches are in? [17:00:01] mdholloway: let's wait until the patches are in [17:00:08] bearND: ok cool [17:00:55] rmoen bmansurov ping: [17:02:41] 1sec [17:03:24] jdlrobson: RE: https://phabricator.wikimedia.org/T107395 - it's blocked on mobile removing use of that old directory. [17:03:41] (and either using the new images or copying the old ones into the new directory) [17:05:57] jdlrobson: bmansurov jhobs rmoen phuedx sorry missign standup: had some startegy stuff run over [17:06:18] kristenlans: no worries [17:11:00] mdholloway: niedzielski: can we do https://gerrit.wikimedia.org/r/#/c/235678/ for now, and address the outstanding comments after the deploy? It would take me too long to get this working [17:11:21] bearND: you got it [17:11:30] sure, doing a separate patch is fine with me (i was about to suggest it actually) [17:12:31] bearND: i'll write up a phab task [17:23:27] okie poke [17:23:32] AWAY! [17:27:48] later [17:30:38] mdholloway: https://gerrit.wikimedia.org/r/#/c/235649/ should be ready [17:31:07] bearND: ok [17:34:27] mbinder: lost you guys [17:37:44] mdholloway: niedzielski: can https://gerrit.wikimedia.org/r/#/c/235649/ get a +2? [17:38:39] mdholloway: going to the hangout [17:38:57] bearND: sure thing [17:38:58] bearND: omw. [17:39:14] niedzielski: looks like your comment on that patch was not a blocker, mind if I +2 it? [17:39:28] mdholloway: definitely not a blocker. go for it [17:39:58] thanks guys [17:53:30] posted a patch to Gather that i think will fix the merge issues [18:02:06] jdlrobson: looking [18:19:09] Hm.. MobileFrontend is blocking the whole gate with its 20min-slow selenium job [18:19:31] maybe we can rename its 'npm' and 'php-composer' jobs to have a prefix so that they aren't associated with the meidawiki/extension/* gate. [18:19:58] mdholloway|brb: I've updated https://www.mediawiki.org/wiki/Wikimedia_Apps/Team/RESTBase_services_for_apps/Ubuntu_setup_notes#Update_the_deployment_repo [18:25:16] Krinkle: That would be nice. [18:31:55] bearND: cool, thx for the heads-up [18:41:19] jdlrobson: barry is famous at metrics meetign! [18:44:23] bearND: btw i was able to connect to tin without any trouble [18:47:33] mdholloway: great. We're going to try again after the MW train [18:47:40] bearND: cool [18:48:29] kristenlans: i can see ;-) [18:48:37] allowed me to send a perfectly timed email [18:49:21] Krinkle: the 20 min slow selenium job is worth it as it makes sure we ship good code. A small cost. [18:49:50] I disagree. It's not worth everyone elses time to wait 20 minutes for one repo to finish their tests. [18:50:03] well the problem is something broke somewhere [18:50:08] and that's leading to congestion. [18:50:20] lots of patches getting rebased at same time. [18:50:44] we could probably change how the bucketing works (rather than first come first serve have some mechanism to limit the number of requests per extension) [18:52:33] mdholloway: niedzielski: got a new patch for you: https://gerrit.wikimedia.org/r/235835 [18:52:56] bearND: taking a look [18:53:38] checking [18:54:31] bearND: btw, are there any patches you don't want us revising while you're away if changes are needed to merge? [18:56:15] niedzielski: no. mdholloway should be free to merge what he seems is ready. I'll check the git log when I get back. [18:56:31] bearND: gr8! :) [18:56:46] s/seems is/sees as/ [18:57:16] niedzielski: for the app code you're da man now [18:57:46] bearND: i'll do my best! [19:02:10] bearND: for the rb patch you just linked, the diff is a little funky. the only functional changes in mwapi.js are apiGet() / checkResponseStatus(), right? [19:03:54] niedzielski: those are the main changes. The other changes are additions of the logger parameter. [19:05:01] the main reason this looks funky is because I moved the error handling function to the top of the file. For apiGet, I extracted checkResponseStatus into it's own function. [19:07:51] bearND: niedzielski: are there cases where the API will return a status other than 200 in the 2xx range, except in error? [19:08:45] mdholloway: not sure. So far I've only seen 200. But theoretically, any status code in the 2xx range is success [19:09:21] well, I should rephrase that, since with mwapi it also returns a 200 in error cases. [19:09:21] bearND: that was my only comment on the patch, so thought i'd just ask here. just asking b/c from marko's content here it sounded like it would be either 200 or a problem. [19:09:34] marko's comment earlier, that is [19:09:50] mdholloway: yeah, i definitely can't speak to mw api but http://httpstatus.es/ says 2xx = success [19:10:38] mdholloway: no, the problem he sees is that 200 doesn't automatically mean that it all went fine. we'd have to parse the body for error tags [19:10:48] mdholloway: I'll probably leave https://phabricator.wikimedia.org/T111407 for you or others btw. [19:11:34] bearND: right. i'm fine leaving the check for 200-299 [19:11:47] bearND|lunch: sounds good about T111407 [19:32:17] mdholloway: are you ready to try again? [19:32:24] bearND: sure [19:33:01] bearND: do you want to just call me directly? [19:33:09] mdholloway: ok [19:37:37] hey jdlrobson I noticed the title font in beta has been changed to libertine [19:39:07] jdlrobson: do you know whats happening with that? [19:50:15] jdlrobson: [19:50:33] I am little confused now what should I do?? [19:57:16] kaity: it has? [19:57:21] adisha: hey [19:58:04] adisha: so from my perspective I don't like the addition and removal of mw-ui-input - is there any reason to do that? [19:58:22] apart from that it looks great! :) [19:59:18] I would suggest removing the class mw-ui-input from https://gerrit.wikimedia.org/r/#/c/234852/12/includes/skins/MinervaTemplate.php and the addition/removal of class in https://gerrit.wikimedia.org/r/#/c/234852/12/resources/mobile.search/init.js [19:59:43] Ok.. [20:00:11] I added It is showing errors in npm too.. [20:00:19] How to solve them [20:01:27] https://integration.wikimedia.org/ci/job/npm/22775/console [20:02:27] adisha: so to replicate those locally.. [20:02:28] jdlrobson: yea, i can't find any phab tickets around it [20:02:32] run: > npm install [20:02:35] > grunt test [20:02:38] jdlrobson: are you seeing it? [20:02:40] it should complain about the same things [20:02:43] kaity: looking now [20:03:02] kaity: it's always been that way.. font-family: "Linux Libertine",Georgia,Times,serif; [20:03:14] Linux Libertine is the first choice. I don't have that font installed though [20:03:32] jdlrobson: hmm i wonder if more browsers just started supporting it [20:03:40] jdlrobson: I never saw it before [20:03:51] kaity very possible [20:04:42] dr0ptp4kt: is android planning happening? [20:05:07] kaity: yup [20:05:17] i've asked them to start hangout [20:07:34] jdlrobson: [20:08:04] adisha: hey :) did the commands work? [20:08:05] I have to remove the mw-ui-input ? else is fine?> [20:08:18] I will try it now [20:08:41] adisha: yes it's not a good idea to change classes that relate to presentation with js imo [20:08:46] great! [20:08:52] Ok [20:18:31] jdlrobson: I have one last doubt [20:18:43] adisha: ask away :) [20:19:16] Should I remove the portion that says for only 320px placeholder will be added [20:19:22] or I should keep it [20:19:30] I chose 320px [20:20:09] As I thought In more than 320px search placeholder will not cutoff [20:20:25] Is it a valid reason to chose 320px.. [20:21:33] * jdlrobson is thinking [20:22:10] adisha: how does it look at 320px with js disabled? can you post a screensht? [20:22:29] brb relocating to where the coffee is [20:22:29] yes I will do it..wait.. [20:23:58] http://198.199.96.81/core/index.php/Main_Page [20:24:21] jdlrobson: You may wish to check the live change [20:24:33] And I am posing a screenshot to [20:24:42] adisha: be careful sharing your private wiki in public you never know who might be looking :) [20:25:22] I will delete it after bug solving..And yes I will be careful next time [20:25:29] adisha: i think you should show the placeholder in js regardless of width [20:25:44] adisha: ngrok is a really useful tool btw https://ngrok.com/ [20:25:55] adisha: so i would suggest dropping the 320px test [20:26:26] ok I will drop the 320px idea.. [20:26:30] jdlrobson: is the font stack different for beta and stable? [20:26:41] kaity: nope [20:27:49] jdlrobson: I'm still seeing Georgia in stable and Libertine in beta [20:28:01] jdlrobson: both desktop chrome and ios safari [20:34:38] jdlrobson: Please check it I have removed the class and amended the changes [20:34:58] kaity: i'm not sure why that would be happening.. :/ what page are you seeing this on [20:35:01] adisha: okay will do! [20:35:25] ok.. thanks.. [20:37:22] jdlrobson: its weird when any article page first loads I see georgia, but then as more of the page loads it changes to libertine [20:37:36] logged in? anon? [20:38:23] jdlrobson: happening in both [20:38:29] jdlrobson: you're not seeing it? [20:38:52] kaity: which page? can i have a url? [20:39:03] jdlrobson: https://en.m.wikipedia.org/wiki/Dispersion_of_the_Apostles [20:39:35] mm looks like we are using js to load the font now [20:39:57] jdlrobson: in stable: https://www.dropbox.com/s/x4beeul8w82k99s/Screenshot%202015-09-03%2016.39.20.png?dl=0 [20:40:01] guess something changed [20:40:09] in beta: https://www.dropbox.com/s/uau7sur7kg90bwa/Screenshot%202015-09-03%2016.39.08.png?dl=0 [20:40:52] i have no idea why the font is loading but someone seems to be loading the font [20:44:35] kaity: maybe ULS related? [20:44:47] jdlrobson: whats that? [20:44:54] universal language selector. [20:45:00] I think something is installing the fonts [20:45:08] i haven't got time to look in detail to why though [20:47:19] adisha: doesn't seem to be working for me [20:48:04] adisha: are you seeing the placeholder with JavaScript disabled? [21:00:14] yes.. [21:01:10] Without JS you will only see defauld placeholder ,i.e. 'Search' [21:02:00] jdlrobson: You think I should remove that search also [21:02:51] just set placeholder to empty string adisha [21:03:10] ok [21:04:44] ok, signing off for now. Will check back later tonight [21:06:50] bearND|afk dbrant_ mdholloway: i've been seeing a weird issue on master wherer the page's "read in other languages" comes up empty for articles that should have translations, like Solid and Fungi. i don't know what the repro steps are yet but please watch out for it [21:06:56] If I remove placeholder text and set it to ' ' jdlrobson it is difficult to identify where is search input and looks something like this https://drive.google.com/file/d/0B16C1BWsjffGRXBuY0tqelktSms/view?usp=sharing [21:07:17] adisha: yeh don't worry about that. I can fix that in a follow up. [21:07:24] (or a patch beforehand) [21:07:25] brb [21:07:27] Ok [21:07:54] I will remove the default searchInut variable too. [21:12:23] niedzielski: interesting. thanks for the heads-up [21:13:30] bearND|afk dbrant_ mdholloway: er, actually. maybe this is the issue? here's the language link look up for molecule on the arabic site: https://ar.m.wikipedia.org/w/api.php?action=query&format=json&prop=langlinks&titles=%D8%A7%D9%84%D8%AC%D8%B2%D9%8A%D8%A6%D8%A7%D8%AA&lllimit=500&continue= [21:13:58] bearND|afk dbrant_ mdholloway: and the same look up on english site returning a bunch of results? https://en.m.wikipedia.org/w/api.php?action=query&format=json&prop=langlinks&titles=Molecule&lllimit=500&continue= [21:14:26] bearND|afk dbrant_ mdholloway: i'm not sure if that's correct or not [21:15:22] niedzielski: that's interesting in itself, anyway. [21:15:45] bearND|afk dbrant_ mdholloway: i guess i'll open a ticket for it and we can dig in there [21:17:43] niedzielski: i still get a bunch of results for "read in other languages" from the arabic molecule entry with arabic as the app language (via search bar) [21:18:07] niedzielski: but yeah, worth opening a ticket [21:18:35] mdholloway: i don't think that list is filtered [21:18:59] mdholloway: i think that's any language the app supports [21:19:52] niedzielski: i'm not following. [21:20:09] niedzielski: i mean, you're right, but what filters are you applying? [21:22:28] jdlrobson: I removed the placeholder i.e. set it to ' ' for js disabled browsers [21:24:32] mdholloway: sorry my pc crashed. well, all the input drivers crashed so i couldn't interact with it anyway [21:25:03] niedzielski: d'oh! [21:25:45] mdholloway: what i mean to say is that the app langauges are prebaked whereas each article has a subset of languages its available in. we request that subset with those urls i mentioned [21:27:31] niedzielski: i assumed (without digging into the code) that those are the queries performed when we tap "Read in other languages" from a given article -- is that not the case? [21:28:13] cannot... unsee!!! https://usercontent.irccloud-cdn.com/file/BkO39Ixw/Screen%20Shot%202015-09-03%20at%202.27.34%20PM.png [21:28:31] niedzielski: right, the list of app languages is pre-generated -- maybe i was unclear [21:28:34] mdholloway: yeah that's my understanding too but my reason for bringing it up is that i don't think the search language does any of that [21:29:10] mdholloway: search languages is always all languages where as article language is a subset that's dependent on that request coming back with good data [21:32:18] niedzielski: right. yeah, i definitely was misleading. what i meant to say was, I changed the app language to arabic via the search bar (same effect as doing it via the settings), searched for 'molecule' (which worked b/c the searches are in all languages), reached جزيء, tried "Read in other languages" and got a full set of results [21:34:55] niedzielski: were you seeing the lack of results in english or other languages? [21:35:29] niedzielski: (maybe i should use a different term from "results") -- lack of other language options, that is? [21:35:32] mdholloway: ah, weird. so i just tried again and it worked fine. the titles parameter is different though: titles=جزيء vs titles=الجزيئات [21:35:42] (latter doesn't work) [21:35:56] maybe copy and paste error on my part? [21:36:38] mdholloway: mh, i don't think i've seen it when on an english article and don't read any arabic dialects unfortunately [21:37:45] niedzielski: me neither, i only knew i was on the right article because the "Read in other languages" entries had titles like "Molekuul" :) [21:39:02] niedzielski: now that i think about it, i seem to remember having the results of langlinks queries vary (though never having no results where there should have been results) [21:39:43] mdholloway: oh wow. so you might get 100 languages one time and 10 the next for the same article? [21:40:23] niedzielski: more like 15 one time and 12 the next, but yeah (hopefully i'm not misremembering) [21:40:45] mdholloway: hm, any idea how long ago this was? [21:42:07] niedzielski: I haven't really looked at that functionality since March/April, so around then [21:43:30] mdholloway: ok thanks [21:47:24] mdholloway: so according to google translate, the one title is plural and the other is not [21:47:32] molecule vs molecules [21:48:12] niedzielski: ah -- where did you find the plural one? [21:48:19] niedzielski: i think that's the one you were using [21:49:21] mdholloway: idk it's what was in the logcat for the request issued [21:55:23] hello bearND|afk ... [21:57:15] nuria: i think bearND|afk has left for the day (and the next couple of weeks) [21:58:11] mdholloway: i see. anyone here knows about the state of: https://phabricator.wikimedia.org/T109383 [21:58:28] mdholloway: to differentiate between previews and pageviews in android? [21:59:17] dr0ptp4kt maybe? [21:59:44] nuria: dr0ptp4kt is a good bet. [22:00:09] nuria: or dbrant [22:00:12] maybe [22:00:24] dbrant hello? [22:18:55] Jenkins is complaining about some Gather schema on mediawiki/corre [22:18:57] ERROR: '[GatherClicks] Missing or empty schema' [22:20:43] https://integration.wikimedia.org/ci/job/mediawiki-extensions-qunit/9976/console [22:21:04] I don't think it's what's causing Jenkins to fail but would still be nice to clean up [22:55:55] nuria: mdholloway|afk i know nothing. what now? [22:57:14] oh, that: bearND|afk has been working on java patches with milimetric overseeing. dbrant will step in if necessary when bearND|afk is ooo; and i'm generally familiar with the problem space (i was in the meetings) [22:57:30] nuria: mdholloway|afk ^ [22:57:57] https://phabricator.wikimedia.org/T109383 < bearND|afk dbrant what i'm trying to speak to with nuria and mdholloway|afk [23:53:54] hi nuria, dr0ptp4kt, re: the code [23:54:38] we're doing that thing where two people meet in the hallway and they dance, both going in the same direction :) [23:55:28] so, who's gonna write the code? nuria can do it if bearND|afk is uncertain, which is what it sounded like in the last patch comment [23:56:12] if so, nuria, we can catch up tomorrow. dr0ptp4kt, we'll just assume that we're writing it if we don't hear back by tomorrow [23:58:52] o/