[08:41:24] Anyone know who is responsible for RESTBase now days? [08:42:01] In particular I'm wondering if there is something equivalent to the origin parameter or JSONP? [08:48:06] "The PHP team are proposing creating a new 'dialect' of PHP called P++ to answer a variety of common criticisms of the language." – https://webopsweekly.com/link/68355/web (via webopopsweekly) [08:48:24] Krenair: Services Team was merged into Core Platform, effectively. [09:37:49] TimStarling: if you are around, I'd like to use our regularly scheduled half hour to talk about T227776 [09:37:49] T227776: Generalize ParserCache into a generic service class for large "current" page-derived data - https://phabricator.wikimedia.org/T227776 [09:38:02] we had a good discussion about it during the last techcom meeting too [13:19:48] Does MediaWiki.Commenting.FunctionComment.MissingParamName do something I'm not expecting? [13:20:12] >Instead of that, can we say "type should go before variable name" or something? [13:20:13] Aha [13:42:39] anomie: About? [13:42:47] re https://phabricator.wikimedia.org/T227662 [13:42:53] Just wondering how far you want to split things down [13:42:56] Reedy: For the moment. I have a meeting in a few. [13:43:08] In https://gerrit.wikimedia.org/r/530374 I've moved the _temp tables to their own files so far [13:43:14] I'm just wondering about the additional columns [13:43:22] Whether to move those to their own file each for completeness [13:44:44] I'm guessing we probably should, as the error was missing ar_actor etc [13:46:41] Reedy: The ideal once we have abstract schema changes would be heavily along the lines of how PostgresUpdater.php handles the field additions/changes (although e.g. lines 475 and 476 would probably be combined IMO). But that would be a huge pain for Mysql and the others with how things are currently structured, way too many tiny files in maintenance/archives. [13:47:54] With that rfc only just being approved... It being implemented (never mind backported) is a long way away [13:49:33] And this is annoying, I do agree.. but at least allows us to get a fix into master, and backport to tidy up for other people in a maintence release [13:52:35] Reedy: The alternative to many tiny files is to make sure the one tested in MysqlUpdater etc is the last one done in the file. At least then it'll try to re-run it instead of silently skipping because it thinks everything is already done. [13:53:05] But will still fail if some/half of the stuff in the middle is done [13:53:20] Which seems to be the case, that some stuff is done, not all of it [13:53:32] catch 22 [13:57:23] At least it'll fail loudly instead of quietly. [13:58:02] heh [13:58:09] and still have people asking "how do we fix this?" [13:59:33] Yeah, probably. But at least everyone not already in the know will be looking in the right place to figure out that it's a half-done change. [13:59:38] * anomie goes to a meeting now [14:32:36] I've split everything down [14:52:12] TimStarling: I amended https://gerrit.wikimedia.org/r/#/c/mediawiki/core/+/526157/ fyi