[07:59:45] yo [08:16:22] morning phuedx [08:18:09] o/ joakino [08:18:19] watched the credit presentation this morning [08:18:22] well "watched" [08:18:33] had it on in the background while i was making breakfast [08:18:40] * phuedx gives a thumbs up [08:18:52] 👍 [08:19:28] i learnt about platypus and toyed with wqs, seems like it's coming together nicely finally [08:21:26] yeah [08:21:41] it's just a year after we wanted it -- which isn't a criticism [08:21:51] but a harsh reality that we didn't seem to want to face [08:26:31] yup [15:52:23] bearND: o/ any further objection to merging this: https://gerrit.wikimedia.org/r/288116 [15:57:44] dbrant: when you were working on fresco, did you run into any trouble around retrieving captcha images? [15:58:48] dbrant: no, I've changed to +1 [15:59:51] mdholloway: i don't recall any trouble. Although I also don't recall testing it particularly thoroughly ;) What's the trouble? [16:01:03] dbrant: i'm just not having any luck loading captcha images from the authmanager labs instance in the app [16:01:33] dbrant: i can do it with the Postman client so it seems to be working correctly on the server side [16:01:45] mdholloway: Fresco simply takes a url to the image... what does the url look like? [16:02:23] dbrant: example: http://authmanager.wmflabs.org/w/index.php?title=Special:Captcha/image&wpCaptchaId=357590062 [16:08:12] dbrant: the api does send back a session cookie with the response containing the url so my theory is that you need to send back the cookie to get the image from that link (which is why it works from Postman but not if you or I click that link here) but fresco for some reason isn't picking it up for this request [16:09:20] dbrant: but the app seems to be handling cookies for authmanager.wmflabs.org as intended, so i'm not sure where the problem there could lie either [16:12:00] mdholloway: i see; so we actually need to tell fresco to use the same cookie handler as the app itself. [16:12:53] dbrant: yes -- but then again, fresco is currently picking up the captcha image for the production wikipedias, so i'm not sure what the difference could be [16:14:40] mdholloway: yes, captchaId is a key for the captcha data which is stored in the session [16:14:57] I don't think there were any changes in cookie handling [16:18:54] tgr: hm, yeah, it seems likely a client-side problem [16:20:36] dbrant: i *think* Fresco uses our OkHttp client which in turn uses our SharedPreferenceCookieManager, but i'm not 100% on that [16:23:41] dbrant: anyway, maybe a next step before standup is manually set the cookie on Fresco's request for the image even though in theory the cookie manager should be handling that [16:23:47] dbrant: and see what happens [16:26:33] mdholloway: may be of use: http://frescolib.org/docs/using-other-network-layers.html [16:28:09] dbrant|brb: ah, interesting, thanks [16:34:55] mdholloway: i think there was an issue with the cookies. i put a patch in yesterday [16:35:21] https://gerrit.wikimedia.org/r/#/c/288533/ [16:35:40] niedzielski: ooh, let me take a look! [18:04:24] bmansurov /cc bearND dbrant|brb mdholloway: that vysor app you used yesterday is pretty nice. i hadn't heard of it before [18:04:54] niedzielski: cool, joaquin told me about it during the offsite [19:24:37] \o/ [19:25:00] dbrant|brb: niedzielski-afk: bearND: hooking the OkHttpClient into Fresco worked! [19:25:14] excellent [19:26:11] mdholloway: great [19:27:15] yayyy [19:55:32] dbrant: how have you folks been using https://meta.wikimedia.org/wiki/Schema:MobileWikiAppSessions so far? (i'm going to spend some cycles to look at that data and derive some general insights, but want to avoid duplicate efforts) [19:57:22] dbrant: did you also look into tagging 81 items that were actually completed in 82, with 82? [19:57:47] dbrant: since I was out I didn't catch our initial commitment [19:59:57] HaeB: I shared a doc with you, where we're using that data to track differences in latency between RestBase and MediaWiki API, and also do a first-order approximation of daily sessions. look at the tab "enwiki production app (since 2016-03-24)" [20:00:02] maxbinder: looking [20:02:01] maxbinder: there's a 1-pointer that I added to 82 [20:02:10] and that's about it [20:10:48] mdholloway: mind looking at / merging this, to make sure I don't need to make more changes before I leave: https://gerrit.wikimedia.org/r/288338 [20:11:34] dbrant: sure [20:13:21] dbrant: cool, thanks! yes i remember you sharing that daily session chart. i'm going to focus more on session lengths [20:14:59] HaeB: cool, that will be useful [20:59:36] dbrant: thanks for the update. Right now there is 6 points worth of work in 82. That's about a third of Android's typical output. Even with you out next week, is that enough? [21:08:19] maxbinder: lol dang it, i thought next week is the beginning of 83... [21:08:26] maxbinder: I'm expecting the team to start work on feeds. [21:08:50] dbrant: next week is the end of 82 (5/20) [21:09:40] dbrant: I just wanna know what was initially scoped for 82. Sounds like, in part due to the offsite confusion, 82 never actually had additional work from 81? [21:11:48] maxbinder: that sounds right. I've added the first two tasks of Feeds to 82, instead of 83. The team can pull from the product backlog if there's time to spare. [21:12:28] dbrant: OK, not a big deal, just wanted to make sure I captured the scoping so you can measure your commitment vs what was actually attempted and completed [21:12:37] dbrant: if there was nothing, das cool [21:12:48] maxbinder: yep, that's cool [21:49:08] Android CI has never been this green https://integration.wikimedia.org/ci/job/apps-android-wikipedia-periodic-test/