[07:42:00] I've been getting logged out automatically within the same browser session today. Did a change break something? [12:36:17] "Apple switched from Google Maps to its own in-house team for the data behind the default map app on iPhones in 2012" they mean OSM? http://www.theguardian.com/technology/2016/jan/22/google-paid-apple-1bn-to-be-default-ios-search-engine?CMP=twt_a-technology_b-gdntech [19:01:19] is there a problem with logins [19:01:22] I keep getting logged out [19:04:36] Dragonfly6-7: yes we're resetting all user's sessions: https://phabricator.wikimedia.org/T124440 but you shouldn't get logged out anymore [19:05:34] legoktm, your aware that last link in that is a dead link? [19:05:52] I just changed the Status part [19:06:10] oh right, fair enough [19:06:26] greg-g "might" have set that, but who knows? [19:13:15] Wondering if https://phabricator.wikimedia.org/T124453 ("Meta login fails") is also related to all our login issues [19:46:19] Cookies52, looks like it's trying to get errors from a server that was decommissioned in april [20:50:14] MarkTraceur, can you please explain the schedule for this? https://phabricator.wikimedia.org/T119336 [20:50:37] yannf: It'll roll out with the train, but anything using async upload-by-url is broken anyway [20:50:48] is it already live? [20:51:00] It's merged, or being merged now [20:51:06] ok [20:51:15] what do you mean by "broken anyway"? [20:51:19] It's not quite merged yet. [20:51:27] yannf: Well, I've never seen any working code that used it [20:51:33] It won't go live until Wikimedia updates it's copy of MW to a version that includes it [20:51:39] yannf: And MatmaRex claims that it never worked before [20:52:00] Unless you have a counter-example that works... [20:52:10] In which case I'll gleefully send out an announce about it [20:52:11] should I test it? [20:52:28] yannf: I'm 99% sure nobody is actually using this feature. I can check if you like. [20:52:46] Searching the job logs for async upload jobs [20:52:50] Shouldn't be too difficult [20:52:58] But I'm guessing MatmaRex did something like that already? [20:54:17] I managed to upload up to 300 MB with upload-by-url [20:54:23] Yes, but async? [20:54:45] but most films are 350-400 MB [20:54:54] so a bit short [20:55:33] yannf: If there's a timeout issue, that's separate - we can fix that any number of ways and none of them are maintaining a broken, crufty feature [20:56:02] ok [20:58:30] yannf: the code i'm removing is not in use anywhere. enabling that config variable just causes an exception when trying to upload by url. [20:58:58] yannf: upload by url still works, and async upload still works. async upload by url has never worked, as far as i know, and we definitely don't have it enabled. [21:01:51] And it's merged. I'm going to double-check that flickr upload via UW is still working on betacommons once the code updates. [21:01:57] Maybe triple-check. [21:04:30] ok, now it failed with a 324 MB file [21:08:27] but 306 MB works [21:09:05] yannf: In production? No code there has changed today. [21:09:53] ok, we could see the difference when it is [21:13:30] yannf: Unless you're passing async=true, the patch I merged won't change anything. [21:13:44] And if you are passing async=true then your uploads will all fail anyway [21:13:49] So seriously nothing will change [21:14:27] yannf: If you need to upload a bigger file, either download it and use chunked uploading, or file a task for a direct upload via SSH [21:15:42] well, the point of upload-by-url to avoid these paths [21:16:16] for most people, uploading a 300 MB is a no-go [21:17:36] ok, bye all [21:22:47] MarkTraceur: no, if you're passing async=true, nothing changes either. [21:23:19] only asyncdownload=true or whatever that parameter for async upload by url was. [21:23:47] you can still upload a file to stash with async=true, which publishes the file asynchronously [21:27:15] Oh, right. [21:27:29] But that's a pretty negligible difference [21:27:37] And it wouldn't help with big files much [21:39:11] Confirmed flickr upload still works on betacommons after code-update, pretty confident we won't have problems. [22:14:37] Still need to figure out what the ArchCom RFC meeting will be about next week: https://phabricator.wikimedia.org/E137 [22:15:51] I need to take a break right now, but I'd appreciate comments on E137 if you have thoughts. MaxSem might care about this. [22:29:53] robla, I think Tim already promised to decide upon PHP [22:30:59] OuKB: ah, right .... thank you for the reminder! [22:31:09] * robla sets about editing E137 [22:32:13] just yesterday, I discovered I kinda need a PHP 5.4 feature :P [22:32:18] like, a lot [22:33:21] OuKB: could you elaborate in T118932? [22:36:11] my a r and y keys broke, so I need to use short array syntax [23:39:32] Danny_B: Thanks for your work on the accessibility workboard, appreciate that! [23:41:25] * d3r1ck is thanking Danny_B too :)