[00:04:51] same here. i'm anti-git review since i hit some case that blew away local changes in an undocumented way. i use the gerrit pre-commit hook but just push to the special branches myself [17:31:38] dbrant: meetin' [19:30:01] niedzielski: mdholloway: bearND: anyone mind skipping grooming today? i have a conflicting meeting :( [19:30:14] ok with me [19:30:21] fine with me [19:30:23] maxbinder: cc ^ [19:30:36] dbrant: ok [20:06:56] a quick question for moble apps types: is this user agent "Wikipedia/942 CFNetwork/808.0.2 Darwin/16.0.0" from one of our mobile apps? [20:07:16] if so, we have a bug for you :-) [20:11:22] * apergos eyes niedzielski [20:12:12] apergos: Darwin sounds Mac-ish. Maybe an Android emulator or an iOS device? [20:12:30] but not one of our? [20:12:32] *ours [20:12:35] Ni [20:12:38] grrr [20:12:39] niedzielski: [20:14:17] apergos: i'm not sure about iOS (maybe coreyfloyd or joewalsh can help) but our Android _app_ should report a UA like "WikipediaApp/... (Android ...; ...)..." [20:14:33] ok, thank you. [20:14:43] let's see if one of them can rule it in/out [20:16:36] apergos: i don't think it's the mobile site on android chrome either: https://developer.chrome.com/multidevice/user-agent [20:17:04] gtk [20:17:40] basically if it's not one of ours, we work around some brokenness, if it is one of ours, I ask someone on the team to look at a phab taask [20:17:42] task [20:22:56] apergos i think the ios folks also have a mac CI and release machine that runs tests (maybe some against the web?). i'm kind of grasping at straws :] unfortunately, i don't think it's anything android related [20:24:32] ok. main question is whether it's our client (and therefore the bug in the task is one to be fixed on our side) or not [20:24:59] if some random client in the wild is broken, then, meh... [20:30:11] apergos: the UA doesn't ring a bell. If you look what kind of request it is you might get some additional clues. (Which action API is it using? Is it using RB?) [20:31:42] it's a thumbor request, asking for a resize of 0 px [20:31:53] https://phabricator.wikimedia.org/T147648#2699558 [20:32:16] which seems clearly wrong :-D [20:34:38] apergos: The apps definitely don't do anything with thumbor yet. I learned about Thumbor only last week. [20:34:49] ok [20:35:05] well it's a request asking for 0px [20:35:08] thumbnails [20:35:40] my "thumbor" was a mistake, that was me typoing from some stuff just deployed, my bad [20:36:31] apergos: niedzielski: our user agent looks like this: WikipediaApp/5.3.0.0 (iOS 10.1.1; Phone) [20:36:39] ok [20:36:50] well that's that, and thanks for your info [20:36:57] I'll report this back and we'll proceed accordingly [20:37:07] apergos: Ah, I see. The apps do at time change thumbnail URLs [20:38:15] apergos: actually though - we do re-write thumbnail sizes inside the WebView - which could report that user agent [20:38:22] oh? [20:39:04] would you be willing to add yourself to that ticket and look around a bit to see [20:39:15] when you get some time to look, not today or anything [20:39:18] ? [20:39:40] apergos: sure thing [20:39:45] thanks! [20:44:28] I'm going back to my regular channels (-operations, -tech, -dev)... thanks again [20:53:59] the iOS app does send that user agent with thumbnail requests made by the WebView - updated the ticket with this info [21:07:57] bearND: deployment today? [21:08:54] mdholloway: yes, I'm prepping it right now [21:09:05] ok, cool [22:47:04] i'm no longer seeing persistent system notifications when zero is active (though i am seeing enter/exit interstitials). is this a regression? [22:48:04] gonna close T126997 in any case since it's probably moot now and then head off, but just thought i'd check [22:48:05] T126997: [Bug] Switching the Wikipedia language while Zero is enabled does not update the nav menu zero message - https://phabricator.wikimedia.org/T126997 [23:54:33] jdlrobson: I've updated your gerrit https://gerrit.wikimedia.org/r/#/c/320628/ and made a simply kafka consumer for you [23:55:19] there's a TODO in app.js where you will receive messages from kafka and you would be able to feed them one by one to your tranding algorithm [23:55:49] which will store results in memeory which you would be able to give out in the trending-v1 route