[00:35:42] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements: Calculate ORES score thresholds for 7 more wikis - https://phabricator.wikimedia.org/T151970#2833743 (10jmatazzoni) [00:36:04] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements: Calculate ORES score thresholds for 7 more wikis - https://phabricator.wikimedia.org/T151970#2833743 (10jmatazzoni) Aaron, did I describe the task properly? And may I assign it to you? [01:01:17] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements, 06Revision-Scoring-As-A-Service, 10rsaas-editquality: Implement new precision-based test stats for editquality models - https://phabricator.wikimedia.org/T151970#2833819 (10Halfak) [01:01:28] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements, 06Revision-Scoring-As-A-Service, 10rsaas-editquality: Implement new precision-based test stats for editquality models - https://phabricator.wikimedia.org/T151970#2833743 (10Halfak) a:03Halfak [01:13:54] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements, 06Revision-Scoring-As-A-Service, 10rsaas-editquality: Implement new precision-based test stats for editquality models - https://phabricator.wikimedia.org/T151970#2833839 (10Halfak) @jmatazzoni may I suggest that we use t... [01:22:32] (03CR) 10Catrope: [C: 032] Rename language codes sr-ec and sr-el to sr-cyrl and sr-latn [extensions/Thanks] - 10https://gerrit.wikimedia.org/r/323684 (https://phabricator.wikimedia.org/T117845) (owner: 10Fomafix) [01:40:49] 06Collaboration-Team-Triage, 10MediaWiki-extensions-PageCuration, 03Community-Tech-Sprint: Update PageTriage speedy deletion tag list to match Twinkle's - https://phabricator.wikimedia.org/T150628#2833883 (10kaldari) [01:41:45] 06Collaboration-Team-Triage, 10MediaWiki-extensions-PageCuration, 03Community-Tech-Sprint: Update PageTriage speedy deletion tag list to match Twinkle's - https://phabricator.wikimedia.org/T150628#2791502 (10kaldari) @Kudpung, @TTO: I've restored the entire list on this ticket per Kudpung. Feel free to discu... [01:42:31] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page, 10Collaboration-Community-Engagement, 07Documentation: Improve Filters for Special:Recent Changes documentation page on mediawiki.org - https://phabricator.wikimedia.org/T151477#2833886 (10jmatazzoni) [01:43:29] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Notifications, 10Thanks: Move Thanks icons to Thanks repo - https://phabricator.wikimedia.org/T151768#2833888 (10jmatazzoni) [01:44:54] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page: Inform users when a URL cannot be adapted to the new filter system for Recent Changes - https://phabricator.wikimedia.org/T151873#2833892 (10jmatazzoni) [01:46:01] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Thanks: ApiRevThankIntegrationTest broken - https://phabricator.wikimedia.org/T151878#2833894 (10jmatazzoni) a:03Catrope [01:50:40] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page, 10Collaboration-Community-Engagement: Integrate a feedback page link in Recent Changes Beta filters - https://phabricator.wikimedia.org/T150959#2833905 (10jmatazzoni) a:03Pginer-WMF [01:52:27] 10Collab-Notifications-Page, 06Collaboration-Team-Triage, 07Mobile, 13Patch-For-Review, and 2 others: [mobile] Special:Notifications the counter does not display the count - https://phabricator.wikimedia.org/T150999#2833909 (10jmatazzoni) [10:47:33] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page: Design a user option for the beta page describing the new RC page filters and enabling users to opt in - https://phabricator.wikimedia.org/T144457#2834455 (10Pginer-WMF) From the feedback we got, the popup version se... [10:58:19] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page: Design a user option for the beta page describing the new RC page filters and enabling users to opt in - https://phabricator.wikimedia.org/T144457#2834512 (10Pginer-WMF) [11:31:48] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page, 07Design: Design ways to facilitate repetitive use for the new Recent Changes filters - https://phabricator.wikimedia.org/T151994#2834562 (10Pginer-WMF) [11:57:20] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page: Explain how the 'IP contribs' filter works (so we can fix it) - https://phabricator.wikimedia.org/T151799#2834640 (10Pginer-WMF) Communities seem to add links to the top of the Recent Changes page to facilitate the a... [14:23:18] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page: Explain how the 'IP contribs' filter works (so we can fix it) - https://phabricator.wikimedia.org/T151799#2834986 (10Catrope) >>! In T151799#2834640, @Pginer-WMF wrote: > It would be good to know: (a) how are those f... [14:30:33] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page, 07Design: Design ways to facilitate repetitive use for the new Recent Changes filters - https://phabricator.wikimedia.org/T151994#2834999 (10Pginer-WMF) [14:34:28] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page, 07Design: Design ways to facilitate repetitive use for the new Recent Changes filters - https://phabricator.wikimedia.org/T151994#2835013 (10Pginer-WMF) [14:45:07] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page, 07Design: Design ways to facilitate repetitive use for the new Recent Changes filters - https://phabricator.wikimedia.org/T151994#2835038 (10Trizek-WMF) Do you plan to investigate more, or can we just add that to t... [14:48:05] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page: Explain how the 'IP contribs' filter works (so we can fix it) - https://phabricator.wikimedia.org/T151799#2835047 (10Trizek-WMF) >>! In T151799#2834986, @Catrope wrote: >>>! In T151799#2834640, @Pginer-WMF wrote: >>... [14:51:55] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page: Explain how the 'IP contribs' filter works (so we can fix it) - https://phabricator.wikimedia.org/T151799#2835051 (10Pginer-WMF) Thanks for the input @Catrope and @Trizek-WMF. I captured both wiki-specific and user-s... [15:16:09] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page, 07Design: Design ways to facilitate repetitive use for the new Recent Changes filters - https://phabricator.wikimedia.org/T151994#2835087 (10Pginer-WMF) >>! In T151994#2835038, @Trizek-WMF wrote: > Do you plan to i... [15:21:32] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page: Explain how the 'IP contribs' filter works (so we can fix it) - https://phabricator.wikimedia.org/T151799#2835091 (10Etonkovidova) Thank @Catrope for the explanation: >As for what this link actually does, it's "hide... [16:37:40] (03PS10) 10Etonkovidova: WIP RSpec test for Mention notificationon [extensions/Echo] - 10https://gerrit.wikimedia.org/r/313221 (https://phabricator.wikimedia.org/T146916) [16:52:12] (03Draft2) 10Aftab: Echo.alias for Bengali (bn) [extensions/Echo] - 10https://gerrit.wikimedia.org/r/324491 [16:52:46] (03PS3) 10Aftab: Echo.alias for Bengali (bn) [extensions/Echo] - 10https://gerrit.wikimedia.org/r/324491 (https://phabricator.wikimedia.org/T151927) [16:55:49] (03CR) 10jenkins-bot: [V: 04-1] WIP RSpec test for Mention notificationon [extensions/Echo] - 10https://gerrit.wikimedia.org/r/313221 (https://phabricator.wikimedia.org/T146916) (owner: 10Etonkovidova) [16:57:55] (03PS11) 10Etonkovidova: WIP RSpec test for Mention notificationon [extensions/Echo] - 10https://gerrit.wikimedia.org/r/313221 (https://phabricator.wikimedia.org/T146916) [17:03:27] (03CR) 10jenkins-bot: [V: 04-1] WIP RSpec test for Mention notificationon [extensions/Echo] - 10https://gerrit.wikimedia.org/r/313221 (https://phabricator.wikimedia.org/T146916) (owner: 10Etonkovidova) [18:17:41] 10Collab-Notifications-Page, 06Collaboration-Team-Triage, 13Patch-For-Review: Make mobile-friendly version of the Notifications page left nav - https://phabricator.wikimedia.org/T139525#2434810 (10KMuthu) {F4908179} I did not make changes to the notifications header because the header was consistently center... [18:23:30] kmuthu, hey [18:23:58] mooeypoo, hi [18:24:17] kmuthu, I'm making my way to the office, but I have to stop at the post office first, so I figured I'll suggest a task you can do today while I code-review mobile phillip [18:24:30] Sure [18:24:43] Make the wikipedia library notification thing (the fix you worked on before) into its own extension [18:25:28] so just copy the code into an extension in your local vagrant/mediawiki/extensions and - for now - you'll have to hook it manually (which I'll show you when I get there) [18:25:35] but we'll at least have the files/folder set up [18:25:37] sounds good? [18:25:46] you can also use mwstew as a basis for the files, btw [18:25:48] it should work [18:26:07] Sure, I can do that [18:26:25] http://tools.wmflabs.org/mwstew/ [18:26:33] ^ this can help you with a boilerplate for an extension [18:27:04] you don't really need any of the "fancy" stuff like special page or the hooks in that list, but at least you'll have a base folder [18:27:05] Thanks ! [18:27:11] we can always add/remove later [18:27:28] sure [18:27:39] kmuthu, also, this is a good resource: https://www.mediawiki.org/wiki/Manual:Extensions [18:28:13] you can use it to "install" the extension manually in your local machine [18:28:37] Great, thanks ! [18:29:17] I'll be there in a bit! [18:29:21] see you soon [19:43:51] * RoanKattouw waves [19:44:24] I'm on a plane flying through a storm, but the wifi still works :) [19:46:22] RoanKattouw: great! Btw, do you know anything about the new option in Preferences - Recent Changes- Revision scoring? The new (and the third) option is displayed as [19:46:36] In prod?! [19:46:42] RoanKattouw:yup [19:46:47] I know a third option was added [19:46:55] I think I even merged that change [19:47:02] Didn't realize he forgot to add i18n :/ [19:47:06] Fixing [19:47:19] RoanKattouw: ahh :) [19:47:48] No he added it, wtf [19:47:55] + "ores-damaging-softest": "Lowest (flags least edits possible)", [19:49:53] etonkovidova: Hmm which wiki did you see this on? [19:51:26] Can confirm on Wikidata [19:52:25] RoanKattouw: enwiki [19:52:55] huh [19:53:02] * RoanKattouw checks Special:Version to verify he's not crazy [19:53:41] WTF is going on [19:54:02] This is a new feature in wmf4, how did this get half-deployed to wmf3 wikis [20:01:01] Aha I see it [20:03:18] etonkovidova: Turns out someone was a little too trigger-happy with https://gerrit.wikimedia.org/r/#/c/324540/1 . I complained there. [20:03:57] RoanKattouw: thanks! :) [20:04:13] etonkovidova: Also you are super fast, this change was deployed only 12 minutes before you pinged me [20:04:26] 11:34am Pacific [20:05:23] RoanKattouw: hahaha - I should say that I monitor the production all the time, but ... it was just a coincidence [20:05:47] I think you've broken Rummana's record AFAIK [20:06:34] which was something like 20 minutes [20:08:50] RoanKattouw: heh [20:09:20] That was a bit different though, it was 20 mins after merge not 20 mins after deployment [20:09:39] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-ReviewStream, 06Editing-Analysis: Research what anti-vandalism tools are most used/most productive? - https://phabricator.wikimedia.org/T152037#2836179 (10jmatazzoni) [20:09:47] I merged a patch changing X and shortly after she goes "why is X broken" [20:10:19] :) [20:59:51] (03CR) 10Catrope: [C: 032] Echo.alias for Bengali (bn) [extensions/Echo] - 10https://gerrit.wikimedia.org/r/324491 (https://phabricator.wikimedia.org/T151927) (owner: 10Aftab) [21:05:43] 06Collaboration-Team-Triage, 10Edit-Review-Improvements, 10MediaWiki-extensions-ORES, 06Revision-Scoring-As-A-Service: Post about ORES review tool including ERI filters - https://phabricator.wikimedia.org/T150858#2836400 (10Halfak) 05Open>03Resolved [21:08:36] (03Merged) 10jenkins-bot: Echo.alias for Bengali (bn) [extensions/Echo] - 10https://gerrit.wikimedia.org/r/324491 (https://phabricator.wikimedia.org/T151927) (owner: 10Aftab) [21:18:16] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page, 13Patch-For-Review: Implement functionality for RC page 'Contribution Quality' filters (ORES) - https://phabricator.wikimedia.org/T149734#2836461 (10SBisson) As I work on this I'm wondering what should be done abou... [21:19:18] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page, 13Patch-For-Review: Implement functionality for RC page 'Contribution Quality' filters (ORES) - https://phabricator.wikimedia.org/T149734#2836463 (10jmatazzoni) [21:22:32] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Flow, 13Patch-For-Review, 05WMF-deploy-2016-11-29_(1.29.0-wmf.4), 07Wikimedia-log-errors: Flow fatal: Default format of 'topic-title-html' or 'topic-title-wikitext' can only be used to format to... - https://phabricator.wikimedia.org/T149827#2764373 [21:32:20] quiddity: Thanks for the email [21:32:44] quiddity: Sorry I dropped the ball on what I promised to do yesterday, I'll try to fit it in today after the extension.json wrangling I'm doing today [21:33:45] * RoanKattouw stops speaking in mystery-speak [21:34:48] * quiddity pauses on way out door for lunch. Thank *you* for being generally awesome, RoanKattouw :) No rush on the 'promise', as long as we can get a decision before EOD Friday. :) [21:35:28] "No rush, just get it done in the next 48h" [21:35:29] :P [21:38:15] ... [21:59:25] 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page, 13Patch-For-Review: Implement functionality for RC page 'Contribution Quality' filters (ORES) - https://phabricator.wikimedia.org/T149734#2836567 (10jmatazzoni) Thanks for asking Stephane. Re. the Watchlist and Spe... [22:10:53] (03PS54) 10Catrope: Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [22:11:26] James_F: ---^^ [22:28:38] RoanKattouw: How much testing did you do? (Too late now. ;-)) [22:28:46] (03CR) 10Jforrester: [C: 032] Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [22:29:29] (03CR) 10Paladox: "Thankyou." [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [22:29:35] Basic testing + locked a topic to verify I didn't screw up the $wgFlowActions stuff [22:29:48] * James_F nods. [22:30:47] RoanKattouw: Now Notifications. [22:31:03] (03CR) 10Paladox: "@Jforrester and @Catrope should we use manifest version 1 instead of version 2?" [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [22:31:58] Ugh the extension.json generation script is broken and puts all sorts of garbage in the autoload list [22:32:06] Classes from core and from other extensions [22:32:35] Less good. [22:35:27] James_F: Do you know what "manifest version 1" vs version 2 means? [22:36:07] RoanKattouw: Yeah, ignore it. Future-facing issue. [22:36:15] Lego reverted half the code. [22:36:44] RoanKattouw: script is broken? [22:37:11] legoktm: Yeah. Ran into it when converting both Flow and Echo [22:37:17] RoanKattouw: remove the "global $wgAutoloadClasses;" from autoload.php in Echo, that shouldn't be there anyways [22:37:25] Hmm OK [22:37:43] I've already cleaned up after it, but I'll try that if it happens again when I convert PageTriage [22:38:10] the conversion script relies upon those global variables being required in function scope so they become local variables, but if you make it an explicit global weird shit happens [22:38:23] Ah. [22:39:02] (03Merged) 10jenkins-bot: Add extension.json, empty PHP entry point [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) [22:39:40] (if you read the script you'll find out how, well, terrible it is. I finally found a legit usage for php's $$var feature) [22:41:16] * James_F looks shocked, shocked I tell you. [22:43:08] hah [22:43:19] So, what is the best way to keep documentation for $wg vars? [22:43:33] Both outward facing config vars and semi-internal ones [22:43:35] Echo has a lot of them [22:43:52] README or on-wiki or you can use the hacky @doc comments, see https://www.mediawiki.org/wiki/Manual:Extension_registration#Retaining_documentation [22:43:53] Move to wiki. [22:44:07] Extension:Echo#Configuration or whatever. [22:44:10] It's what we did for VE. [22:45:43] Oh we already have a lot of stuff at https://www.mediawiki.org/wiki/Extension:Echo#Configuration [22:46:23] Then duplicating it in repo and on-wiki means at least one of them is likely to be wrong. :-) [22:46:33] Bitrot and all that. [22:47:14] legoktm: There's a task about manifest_version:2 isn't there? Or am I forgetting? [22:48:00] https://phabricator.wikimedia.org/tag/mediawiki-configuration/ there's a column, no specific task I guess [22:48:03] gotta run, bbl [22:49:17] Kk. [22:56:40] Ugh I hate 
 support in VE
[22:56:48] 	 you can't create multi-line 
s
[22:57:08] 	 (I'm finding this out editing that bitrotten-as-you-said page)
[23:09:33] 	 OK, WTF
[23:09:37] 	 What am I missing about extension registration here
[23:10:20] 	 I have "config": { ... "EchoConfig": { "foo": { "bar": { "baz": 3 } } } say, in extension.json
[23:10:25] 	 I run wfLoadExtension( 'Echo' )
[23:10:34] 	 $wgEchoConfig is now... undefined :/
[23:11:01] 	 And setting $wgEchoConfig['foo']['bar']['baz'] = 5; works. but causes the EchoConfig setting in extension.json to be ignored completely
[23:11:23] 	 How do I allow people to override individual keys of arrays in LocalSettings.php
[23:11:32] 	 (or in MWV's settings.d in this case)
[23:11:38] 	 (03CR) 10Jforrester: "> version 2 breaks mw 1.28 and 1.27 compat." [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) 
[23:11:47] 	 RoanKattouw: something with _merge_strategy
[23:12:00] 	 (03CR) 10Paladox: "Ok." [extensions/Flow] - 10https://gerrit.wikimedia.org/r/213837 (https://phabricator.wikimedia.org/T87916) (owner: 10Paladox) 
[23:12:06] 	 06Collaboration-Team-Triage, 10Flow, 13Patch-For-Review: Convert Flow to use extension registration - https://phabricator.wikimedia.org/T87916#2836739 (10Jdforrester-WMF) 05Open>03Resolved a:03Paladox
[23:12:44] 	 it's probably documented somewhere. i'm not sure where, but this is the keyword to search for :D
[23:14:11] 	 Fond it
[23:14:18] 	 It's not documented at all, but I found the code
[23:14:46] 	 RoanKattouw: also, wfLoadExtension() only queues it to be loaded. the actual loading happens somewhere later
[23:14:57] 	 Hmm right
[23:14:57] 	 so if you did that in a REPL or something, you won't see the globals
[23:15:04] 	 That's why defined constants don't work either
[23:15:08] 	 ExtensionRegistry::getInstance()->loadFromQueue();
[23:15:12] 	 No I'm in a settings.d file
[23:15:21] 	 (it's done at the top of Setup.php)
[23:16:36] 	 Oh, here we go, the default merge strategy is array_mrege
[23:16:42] 	 OK, so I only need to override this for nested arrays
[23:17:29] 	 But the array_merge_recursive strategy is backwards :(
[23:17:31] 	 WTF
[23:17:38] 	 This why code needs documentaiton
[23:17:48] 	 RoanKattouw, if it's backwards, shouldn't it be FTW
[23:18:08] 	 'array_merge' means array_merge( $extensionJsonValue, $globalValue ) , so the global one wins
[23:18:18] 	 'array_merge_recursive' means array_merge_recursive( $globalValue, $extensionJsonValue )
[23:18:35] 	 I'm sure there's a reason / use case for this, but it's super unclear
[23:20:21] 	 Looks like 'array_plus_2d' will work for me in this case, but not in the general caes
[23:23:08] * ebernhardson wrote an arrayMergeRecursive in some class because the actual array_merge_recursive is crap
[23:23:45] 	 TIL about array_replace_recursive
[23:24:24] 	 oh interesting, that sounds like the change i needed to array_merge_recursive. good to know
[23:24:49] 	 Whoa TIL what array_merge_recursive actually does and it's horrifying
[23:26:07] 	 And it makes me question whether it's the right choice for the config settings that it's the merge strategy for
[23:27:05] 	 06Collaboration-Team-Triage (Collab-Team-Q2-Oct-Dec-2016), 10Edit-Review-Improvements-RC-Page, 13Patch-For-Review: Implement functionality for RC page 'Contribution Quality' filters (ORES) - https://phabricator.wikimedia.org/T149734#2836804 (10Halfak) @Ladsgroup, see above.    My sense is that dropping the r...
[23:32:20] 	 AARGH and _merge_strategy is a key in the *associative array itself*. Then what is the wrapper in extension.json for!
[23:32:53] 	 It's "EchoBlah": { "value": { stuff here } } so I thought it would be "EchoBlah": { "value": { stuff here }, "_merge_strategy": "..." }
[23:33:21] 	 But no, _merge_strategy goes inside the value
[23:35:08] 	 It looks like the schema and the implementation disagree on this!!
[23:36:15] 	 James_F: Who should I talk to about ---^^ apart from legoktm ?
[23:39:29] 	 RoanKattouw: Umm. Reedy and me maybe.
[23:39:41] 	 OK, so
[23:39:54] 	 _merge_strategy appears to be implemented inconsistently and uselessly
[23:39:59] 	 Yes.
[23:39:59] 	 At least for extensions
[23:40:26] 	 We've gone back and forth about it for different parts of extension.json.
[23:40:40] 	 See e.g. the users
[23:40:45] 	 Bah.
[23:40:58] 	 See e.g. the uses in VE-MW.
[23:42:16] 	 Looking
[23:42:43] 	 Oh that's a v1 schema?
[23:42:49] 	 I see
[23:43:00] 	 v2 has a "value": wrapper (didn't realize that was v2 only)
[23:43:21] 	 So I was expecting _merge_strategy to be a sibling of "value" in v2
[23:43:59] 	 Anyway, what I need is a mirrored version of the array_replace_recursive strategy
[23:44:19] 	 And I'm wondering if anything even uses that strategy currently
[23:45:41] 	 Aha, JsonConfig uses it
[23:46:55] 	 And there doesn't seem to be a good reason for why it is in the order that it is, since the default value is {}
[23:50:02] 	 So either I need to change that, or I need to introduce array_plus_recursive, which is like array_plus_2d but recurses all the way
[23:50:45] 	 Yeah.
[23:50:59] 	 It's a mess.
[23:57:39] 	 James_F: https://gerrit.wikimedia.org/r/324628 illustrates my problem
[23:58:00] 	 I'll write a change that adds a new merge strategy that does what I need, and make the Echo extension.json patch depend on it
[23:58:05] 	 That should get someone's attention hopefully
[23:58:27] 	 I also don't quite understand how we haven't already encountered this issue
[23:59:00] 	 Did we really manage to convert all but 4 WMF-deployed extensions and not have any doubly nested config overides?
[23:59:21] 	 (I guess a lot of these cases are handled correctly by array_plus_2d)