[04:42:20] 3MediaWiki-Core-Team, Wikimedia-General-or-Unknown: Users are unable to change account password (HTTP 503) - https://phabricator.wikimedia.org/T87305#990614 (10Sunpriat) Confirm, that problem is fixed. Thank you. [04:46:24] 3MediaWiki-Core-Team, Wikimedia-General-or-Unknown: Users are unable to change account password (HTTP 503) - https://phabricator.wikimedia.org/T87305#990615 (10hoo) 5Open>3Resolved a:3hoo >>! In T87305#990614, @Sunpriat wrote: > Confirm, that problem is fixed. The user succeeded log in. Thank you. Nice :) [05:03:46] hello [05:29:10] ori: Hey [06:15:31] Krinkle: hello [17:22:36] 3MediaWiki-Core-Team, SUL-Finalization: Bad LocalRenameUserJob stuck in jobrunner for vewikimedia - https://phabricator.wikimedia.org/T87360#991136 (10bd808) 5Open>3Resolved a:3bd808 @ori was able to help me fix this yesterday. The redis purges I had done removed the job, but the jobrunner instances were s... [18:41:18] 3VisualEditor, MediaWiki-Core-Team, VisualEditor-Performance: Parsoid performance analysis - https://phabricator.wikimedia.org/T85870#991335 (10ori) [18:46:02] Tim-away: VE perf workboard is now current and relevant: . Would you like me to give you updates periodically? [19:10:05] 3MediaWiki-Configuration, MediaWiki-Core-Team: convertExtensionToRegistration.php should write out extension.json files in a nice order - https://phabricator.wikimedia.org/T86608#991398 (10Legoktm) 5Open>3Resolved [19:59:17] ori: thanks, I will look myself [19:59:39] anomie: remind me why we have vary-revision? [20:00:21] it seems like the most evil feature in the parser, we should get rid of it one way or another [20:02:07] TimStarling: It looks like the point of it is so WikiPage::doEditUpdates() can know whether it can re-use the edit that was prepared during the pre-save checks (before a rev_id was assigned), or if it needs to re-parse things. [20:02:27] yeah, sorry, I do know that [20:02:40] it makes the page be parsed twice on save, roughly doubling the save time [20:03:06] which seems like a really high price to pay just so you can have {{REVISIONDAY}} or something work correctly [20:03:56] I was against it from the outset but brion pushed it through [20:04:29] I couldn't think of any possible reason why you would want {{REVISIONID}} so much that you would want to double the save time to get it [20:04:48] And {{PAGEID}} on creations, and self-transclusion. Although, did jackmcbarn already solve the need for the last? I forget. [20:05:15] anomie: no, vary-revision is still needed for self-transclusion [20:05:23] though i think i could remove that if i tried [20:05:27] the change you just merged was about fixing self-transclusion with vary-revision [20:07:05] The change just merged was about avoiding the self-transclusion still being wrong despite vary-revision due to slave lag [20:07:49] yeah [20:07:59] maybe the solution is to not allow self-transclusion? [20:08:17] that would break a lot of things, and that's easy to make fast [20:08:48] what are the things it would break? [20:09:12] Any template that invokes itself to show an example output in its documentation [20:09:31] ok fair enough [20:10:00] so can it just return the input to Parser::parse()? then we wouldn't need vary-revision or the patch you just merged [20:10:30] i think i can make it do something like that [20:15:39] that would be very helpful I think, jackmcbarn, thanks [20:15:47] ok [20:16:03] that's probably the top cause of vary-revision being set, right? [20:16:11] I wonder if we should make some statistics [20:16:35] i'd assume it is [20:19:21] we could optimise a few of the others if they are commonly used [20:20:15] i think pretty much everything but PAGEID and REVISIONID can be made to not need it at all [20:20:53] {{REVISIONID}} could be done by inserting a placeholder row into the revision table, with rev_page=0 [20:21:33] well, i guess we could do that, but i'm not sure i like it [20:21:59] if we were using postgresql we could allocate a revision ID without inserting a row [20:22:15] since autoincrement sequences are separate from tables [20:22:18] how big of a deal is the performance of vary-revision? [20:22:33] like I said, it doubles the page save time [20:22:41] I don't know how often it is hit [20:24:49] too hard to get such information these days [20:26:53] graphite is mostly good for freezing my browser for a minute at a time [20:27:09] not sure if there is anything useful in there [20:27:14] is there a CLI interface to it? [20:27:48] ori: ^ [21:36:40] TimStarling: yes, just a sec [21:42:04] bd808: Is there some magic in mediawiki-vagrant to have it update its copies of MediaWiki and extensions and such to master? [21:42:20] vagrant git-update [21:42:49] Bah, and it's right there in the help. Thanks! [21:42:51] does git pulls + runs update.php + runs composer [22:27:44] TimStarling: so, save times are in tungsten:/var/lib/carbon/whisper/mw/performance/save/HHVM as Whisper (*.wsp) files. Use whisper-fetch () to query from the command line. There is no data from the 14th of this month onward; I'm not sure what broke and haven't had a chance to poke (tracked in ) [22:28:14] thanks [22:34:46] More or less just arrived... [22:35:11] Are you people in the office right now? [22:38:14] hoo: Yeah [22:38:19] Daniel was too [22:38:24] is/was [22:38:30] hoo: It's still a work day :P [22:41:11] "work days"... [22:41:51] Anyway, I'll head be there in a bit [22:42:01] Are you alright getting in? [22:42:44] Huh? [22:43:38] to the office :P [22:43:40] I guess you don't have a security card etc [22:44:17] No, I don't... Will wait until someone comes out to smoke or so [22:45:05] You can just tell security where you're going etc [22:45:17] They should let you into a lift [22:45:24] Or do you want my mobile number and text me when you're here? [22:45:53] Give me your mobile number as a backup [22:46:10] Last time I had trouble with security AFAIR [22:47:00] See PM [22:48:38] Got it. Leaving in a moment. [22:48:44] woot :D [23:01:26] I will never accidentally click yes to date you. Go away. - http://ur1.ca/jioqk [23:02:20] ^d: ^^^^^^^^ [23:02:35] aww, the link is broken [23:03:03] <^d> Tweet doesn't exist? [23:03:45] hmm [23:03:48] it existed at some point! [23:04:10] there's a twitter service that keeps track of tweets deleted by UK politicians [23:10:25] 3MediaWiki-JobRunner, MediaWiki-Core-Team: Job runner tight loop when job class does not exist - https://phabricator.wikimedia.org/T87398#992051 (10aaron) [23:11:30] * Reedy goes to find hoo [23:16:28] Reedy: he's in town? [23:16:40] read up :P [23:16:54] legoktm: bah [23:20:06] DSF 384 // 1B5949D8CE9D3B1C1A5651D5C4E07E235F87F196E61304D756FFF563F38A524C00990FDF9F173ADA7E3979542FB50E0D - http://ur1.ca/jip16 [23:21:06] VOY 384 // B8B074792EBBBCD48159C7528ED635A62704FF1A6FD309C8438BEF851186565F48C25035DD9E9DE9552D59ACE45A51A2 - http://ur1.ca/jip1s [23:23:07] SALIENT 384 // BC4B667403B43E0B60B5807AF22F93AB176437676E12B4A4143B8365F18F4FCD5B3DEBF327FE19FBEC2FCC841EA0234F - http://ur1.ca/jip2u [23:43:13] If the NSA found out most online discussion and comments are made by sentient machines trying to influence humans, would they tell us? - http://ur1.ca/jipfs [23:59:18] 3MediaWiki-Core-Team, MediaWiki-API: Clean up ApiResult and ApiFormatXml, create format=json2 - https://phabricator.wikimedia.org/T76728#992343 (10Anomie) [23:59:35] 3MediaWiki-Core-Team, MediaWiki-API: API: Add jsconfigvars to action=parse - https://phabricator.wikimedia.org/T67015#992345 (10Anomie) a:3Anomie