[13:12:37] MatmaRex: Yes, I would be unhappy if you merged 186757. Why is #3 a no-go? It's more work, sure, but it's also more correct. [14:16:36] 6MediaWiki-API-Team, 10MediaWiki-API, 5Patch-For-Review: PHP Notice: Undefined index: callback in includes/api/ApiFormatJson.php - https://phabricator.wikimedia.org/T96479#1220588 (10Anomie) [14:59:40] legoktm: Re 204968, want to fix where I made the same error in ApiFormatPhp in the same patch? [14:59:56] legoktm: Never mind, someone else jumped on +2ing it. :/ [15:01:37] I'll submit another patch [15:02:06] 6MediaWiki-API-Team, 10MediaWiki-API, 5Patch-For-Review: PHP Notice: Undefined index: callback in includes/api/ApiFormatJson.php - https://phabricator.wikimedia.org/T96479#1220738 (10Anomie) 5Open>3Resolved a:3Anomie [15:04:32] anomie: https://gerrit.wikimedia.org/r/205257 [15:47:07] legoktm: how's sul going? [15:47:58] greg-g: good! we're on nowiki right now [15:48:33] legoktm: it's alphabetical for small wikis, yes? [15:49:11] were all done with small and medium wikis (exception of mediawikiwiki), right now it's going through https://noc.wikimedia.org/conf/highlight.php?file=large.dblist except dewiki + enwiki were skipped and are going to be last [15:49:26] ahhhhhhh [15:49:32] sweet sweet [15:49:47] legoktm: glad it's going well, did you have to spend much time this weekend on it? [15:50:30] * bd808 votes for legoktm to get a week off in the location of his choosing when SUL is done [15:50:34] not actively, I was just sitting next to my laptop watching it scroll by :P [15:50:58] bd808: my vacation starts on wednesday :P [15:51:13] nice [15:51:37] legoktm: Did you start on a patch for that deprecation in merge-plugin yet? [15:51:52] I can if you don't already have it in progress [15:52:10] I haven't, I'm still testing it to make sure nothing is actually broken [15:52:38] our tests are failing :/ [15:53:49] bd808: http://fpaste.org/213415/42954521/raw/ no functionality appears to be broken though [15:54:14] k [15:55:14] That's with the latest HEAD? [15:55:21] for composer I mean [15:55:31] no that's with alpha10 [15:55:51] well, I think we're testing with composer master though? [15:56:10] "composer/composer": "1.0.*@dev", [15:57:33] *nod* master is 45 commits ahead of a10 [15:58:56] * bd808 updates to master locally to messa around [15:59:52] Composer\Script\CommandEvent is certainly used in a lot of places still for being deprecated [16:00:14] :/ [16:00:27] github is making breaking API changes today (https://developer.github.com/changes/2015-02-20-migration-period-removing-authorizations-token/) which is why they released a new version [16:03:19] ah. they made CommandEvent an empty child of the Event class [16:07:09] legoktm: *grumble* the errors aren't there with HEAD -- https://dpaste.de/FoEQ [16:08:07] :| [16:08:08] sigh. [16:08:08] afais functionality wise everything is fine with alpha10 [16:16:21] 6MediaWiki-API-Team, 10MediaWiki-API, 5MW-1.25-release: Backport API format changes and follow-ups to REL1_25 branch - https://phabricator.wikimedia.org/T96570#1220914 (10Legoktm) 3NEW [16:16:36] anomie: ^ did I miss any patches? [16:18:34] legoktm: Maybe https://gerrit.wikimedia.org/r/#/c/205278/ that I just submitted, if that seems like something important enough to backport. [16:19:17] legoktm: https://gerrit.wikimedia.org/r/#/c/191103/ would be nice if it can get merged to master in time. [16:20:35] hmm, I don't like falling back to jsonfm [16:21:06] Less than throwing an unhandled UsageException? [16:21:37] I'd rather have something like {'error':"Unknown error [hash]"} and log the exception [16:22:16] sending a bunch of HTML to a user when they're not expecting it doesn't seem like a good idea to me [16:23:31] most clients should be at least checking for 'error' and stuff [16:25:33] Why throw json at them when they may not have requested that either? [16:27:10] 6MediaWiki-API-Team, 10MediaWiki-API, 5MW-1.25-release: Backport API format changes and follow-ups to REL1_25 branch - https://phabricator.wikimedia.org/T96570#1220947 (10Anomie) [16:29:30] 6MediaWiki-API-Team, 10MediaWiki-API, 5MW-1.25-release: Backport API format changes and follow-ups to REL1_25 branch - https://phabricator.wikimedia.org/T96570#1220962 (10Anomie) These might be nice too: * {61233fc84724833dc487ecc55e27c1b39459befc} * {589258a7331282dedfb7743270bb3be40a3b8c45} * {9a504d18fb13... [16:33:43] 6MediaWiki-API-Team, 10MediaWiki-API, 5MW-1.25-release: Backport API format changes and follow-ups to REL1_25 branch - https://phabricator.wikimedia.org/T96570#1220984 (10Anomie) [16:37:38] 6MediaWiki-API-Team, 10MediaWiki-API, 5MW-1.25-release: Backport API format changes and follow-ups to REL1_25 branch - https://phabricator.wikimedia.org/T96570#1221010 (10Anomie) [16:48:03] anomie: :( [16:49:00] anomie: re https://gerrit.wikimedia.org/r/#/c/186757/ , having the ids in page output and making them configurable was part of the point of the indicators. it's presumably better than having inline css. [16:49:16] anomie: can you mwgrep for 'mw-indicator', actually? i wonder if people are making use of this functionality [16:51:22] MatmaRex: https://phabricator.wikimedia.org/P536 [16:54:18] MatmaRex: Or they could do ... to use global CSS. [16:54:42] yeah, but that would kind of suck, no? [16:55:06] if we can provide the same or better functionality from the software [16:56:01] Depends how much it's actually used. The problem with "the same or better functionality from the software" is that there's apparently no good place to transform the raw id to something CSS likes. [16:56:03] code like above would generate nested structures and probably cause people some pain if they try shifting the indicator's position with CSS positioning [16:56:24] we could also just stop escaping ids everywhere [16:56:37] which i would say would be entirely reasonable [16:56:46] you can escape ids in CSS, too, if you really need to [16:57:00] (but you should just not choose names that are hard to represent) [16:57:16] we only escape the id because we escape ids everywhere else in mediawiki. [16:57:33] (for no good reason; of it there is a good reason, i don't know it) [16:59:18] MatmaRex: In part because HTML5 is newer than MediaWiki, and in part because some escaping (e.g. quotes) is necessary to avoid injection. [17:00:58] anomie: I didn't mean use json specifically, just output that in whatever format they had requested [17:02:19] anomie: okay, but HTML entity escaping and the weird things we do to ids are entirely independent [17:04:26] ugh: The first fallback is to try that. But if that fails too, the second fallback is jsonfm. Then the third fallback is unhandled exception. [17:13:47] marktraceur: Is there a reason T89765 isn't closed yet? [17:14:46] tgr: ^ (it looks like everything is done, just not sure if you guys are changing anything else) [17:16:44] csteipp: T85856 now reminds me of "Warning: Coffee is hot" ;) [17:17:03] anomie: I know right :) [17:17:10] But really, some people need that warning... [17:25:14] mmm, pretty sure all wiki users understand all pages are visible [17:25:22] at least wp users [17:25:52] it's like "users can post private insformation on their user pages" [17:26:56] legoktm: https://gerrit.wikimedia.org/r/205293 doc tweaks [18:34:24] 6MediaWiki-API-Team, 10MediaWiki-API: Update extensions (again) for ApiResult rewrite - https://phabricator.wikimedia.org/T96595#1221561 (10Anomie) 3NEW a:3Anomie [18:35:31] 6MediaWiki-API-Team, 10MediaWiki-API: Add wfDeprecated() to deprecated ApiResult methods - https://phabricator.wikimedia.org/T96596#1221585 (10Anomie) 3NEW a:3Anomie [18:35:41] 6MediaWiki-API-Team, 10MediaWiki-API: Update extensions (again) for ApiResult rewrite - https://phabricator.wikimedia.org/T96595#1221595 (10Anomie) [18:35:42] 6MediaWiki-API-Team, 10MediaWiki-API: Add wfDeprecated() to deprecated ApiResult methods - https://phabricator.wikimedia.org/T96596#1221594 (10Anomie) [18:37:44] 6MediaWiki-API-Team, 10MediaWiki-API: Add wfDeprecated() to deprecated ApiResult methods - https://phabricator.wikimedia.org/T96596#1221585 (10Anomie) [18:43:10] 6MediaWiki-API-Team, 10MediaWiki-API: Add wfDeprecated() to deprecated ApiResult methods - https://phabricator.wikimedia.org/T96596#1221655 (10Anomie) [18:46:09] 6MediaWiki-API-Team, 10MediaWiki-API, 5MW-1.26-release: Add wfDeprecated() to deprecated ApiResult methods - https://phabricator.wikimedia.org/T96596#1221703 (10Legoktm) [19:09:42] 6MediaWiki-API-Team, 10MediaWiki-API, 5MW-1.25-release: Backport API format changes and follow-ups to REL1_25 branch - https://phabricator.wikimedia.org/T96570#1221868 (10Anomie) Assuming they get merged, [[https://gerrit.wikimedia.org/r/#/q/status:open+topic:api-cleanup-PS25,n,z|these 28 extension patches]]... [19:15:55] anomie: https://gerrit.wikimedia.org/r/#/c/205311/ [19:19:32] legoktm: https://gerrit.wikimedia.org/r/#/c/205344/ [19:20:26] :D [19:23:32] randomcat: https://gerrit.wikimedia.org/r/#/c/205345/ [19:24:54] legoktm: http://memegenerator.net/instance/61465941 [19:25:25] ori: needs rebase [19:26:12] haha :) [19:28:16] legoktm: https://gerrit.wikimedia.org/r/#/c/36330/ [19:30:47] randomcat: rebased [19:30:56] Krenair: does User::fetchAllRemoteGroups() still need caching if we're no longer using the API? [19:31:48] Oh I think the release notes change is obsolete now [19:32:50] * randomcat can't find that method [19:33:12] randomcat, https://gerrit.wikimedia.org/r/#/c/36330/46/includes/User.php [19:33:17] legoktm, I don't think so [19:33:46] ahh, thats a patch [19:33:52] * randomcat never saw that [19:36:02] Krenair: commented. [19:36:40] thanks [19:36:41] will do [19:38:06] It's my oldest patch in gerrit, been there since 2012 [19:38:06] :| [19:42:33] anomie: https://gerrit.wikimedia.org/r/#/c/205351/ fixes Push tests [19:47:41] * anomie was lazy and didn't want to figure that out [19:51:06] legoktm, IIRC those release notes were only relevant when it was using the api [19:51:39] ok [19:52:09] anomie: ok, just flow is left and I poked ebernhardson about that earlier [19:52:34] legoktm: You are awesome. [19:53:52] :) [19:54:16] backporting all of these is going to be fun [19:54:45] legoktm: so... I can't recreate your test failures for composer-merge-plugin even with the same version of composer -- https://dpaste.de/hN7h [19:54:57] but we are using different php versions [19:55:04] bd808: that's...weird. :/ [19:55:33] I'm on 5.6.7 [19:55:48] I was running with 5.5.19 there [19:56:52] Somewhere I have a VM with a bunch of different php versions installed. I'll try to dig that up I guess [19:56:54] and you ran composer update beforehand right? [19:57:09] oh... nope [19:57:33] * bd808 does that now [19:58:38] legoktm: fails the same way now [19:58:42] that's nicer [19:58:48] I think the tests run against what version of composer is in require-dev, not whatever you're using [19:58:51] yay! :/ [19:59:00] 6MediaWiki-API-Team, 10MediaWiki-API, 5Patch-For-Review, 7Wikimedia-log-errors: Call to a member function getConfig() on a non-object (NULL) - https://phabricator.wikimedia.org/T96583#1222197 (10Anomie) a:3Anomie [20:00:51] randomcat, oh I just saw your comment [20:00:53] Which class? [20:01:46] speaking of backports, can i poke you guys about a REL1_25 commit i want merged? https://gerrit.wikimedia.org/r/#/c/204929/ (not actually a backport) [20:01:56] if not, who should be poked? [20:03:50] +2'd [20:04:56] thanks [20:20:12] 6MediaWiki-API-Team, 10MediaWiki-API, 5Patch-For-Review: Update extensions (again) for ApiResult rewrite - https://phabricator.wikimedia.org/T96595#1222253 (10Anomie) 5Open>3Resolved [20:20:14] 6MediaWiki-API-Team, 10MediaWiki-API, 5MW-1.26-release, 5Patch-For-Review: Add wfDeprecated() to deprecated ApiResult methods - https://phabricator.wikimedia.org/T96596#1222254 (10Anomie) [21:21:35] legoktm: any time to look at https://gerrit.wikimedia.org/r/#/c/200325/ ? [21:23:18] maybe, I have some API patches to look at first [21:53:54] legoktm: I figured out the merge-plugin test problem. They made a change to the InstallerEvent constructor that stuck a new param in the middle of the existing signature. :/ [22:23:08] legoktm: https://github.com/wikimedia/composer-merge-plugin/pull/23 [22:40:06] bd808: woot, merged [22:40:39] kewl. poking at some others now since I'm there already [22:49:24] bd808: is there a way we can require a minimum version of composer? [22:50:15] hmmm... only if they bump the plugin api I think [22:51:20] https://getcomposer.org/doc/articles/plugins.md#plugin-package [22:53:21] I'd be willing to bet that if you asked the Composer folks they would say that supporting the latest tagged version is "good enough" [22:53:56] since there hasn't really ever been a stable release yet [22:53:58] legoktm: I'll accept daily facts and figures when you get to it, please :) [22:55:33] Keegan: Don't distract legoktm with real work while I'm distracting him with other things. ;p [22:56:40] I said when he gets to it! [22:56:52] Subtle! [22:56:58] And please. [22:57:26] Look, I've heard y'all have been having trouble putting the covers on your TPS reports... [22:57:40] * bd808 moves his desk back into the corner [22:58:08] Keegan: we're about halfway through ruwiki right now, 1,854,813 renames done so far, ~962,595 left [22:58:32] Whee. Thanks [23:01:25] TimStarling: idea for your next event / family gathering / birthday: http://www.tankedup.com.au/our-promotional-vehicles/game-over-armoured-personnel-carrier-for-hire/ [23:02:44] the gold exterior is the best part, I think [23:03:44] shiny [23:06:43] bd808: https://github.com/composer/composer/issues/3955 [23:13:23] https://github.com/composer/composer/issues/3938 oh god [23:13:54] only a ps2? any respectible tank OR limo should have a ps3 at least. really by now a ps4 seems mandatory. [23:48:24] legoktm: y u no want beeeeep? [23:48:58] I'm not sure I want to know the guy who thinks that would be a usability improvement