[00:04:39] (03CR) 10jenkins-bot: [V: 04-1] If there isn't one already, append reference list to page when creating a reference [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/149117 (https://bugzilla.wikimedia.org/54906) (owner: 10Alex Monk) [00:13:33] 10VisualEditor: Inconsistent denotation on the context menus across different combinations of ways to add Cite Web - https://phabricator.wikimedia.org/T91343#1080545 (10Ryasmeen) 3NEW [00:14:26] 10VisualEditor: Inconsistent denotation on the context menus across different combinations of ways to add Cite Web - https://phabricator.wikimedia.org/T91343#1080552 (10Ryasmeen) [00:15:24] 10VisualEditor: Inconsistent denotation on the context menus across different ways to add Cite Web - https://phabricator.wikimedia.org/T91343#1080555 (10Ryasmeen) [00:18:44] 10MediaWiki-ResourceLoader, 7Technical-Debt, 7Wikimedia-log-errors: ResourceLoader: Invalid argument supplied for foreach - https://phabricator.wikimedia.org/T46907#1080557 (10Jdforrester-WMF) Does this still occur? Filed over two years ago… [00:28:50] 10VisualEditor, 10WikiEditor, 6Collaboration-Team, 10Flow, 7Design: Editing toolbar(s) for Flow (VE and/or wikitext toolbar) - https://phabricator.wikimedia.org/T78346#1080560 (10DannyH) a:3Pginer-WMF [00:29:02] 10VisualEditor: Inconsistent denotation on the context menus across different ways to add Cite Web - https://phabricator.wikimedia.org/T91343#1080561 (10Ryasmeen) [00:29:55] (03CR) 10Jforrester: [C: 032] RadioInputWidget: Add description and example [oojs/ui] - 10https://gerrit.wikimedia.org/r/193895 (owner: 10Kmenger) [00:30:13] (03CR) 10Jforrester: [C: 032] ProgressBar: Add description and example [oojs/ui] - 10https://gerrit.wikimedia.org/r/193869 (owner: 10Kmenger) [00:32:32] 10VisualEditor: Scroll bar appearing in context menu for a multi-lined / lengthy comment - https://phabricator.wikimedia.org/T91345#1080573 (10Ryasmeen) 3NEW [00:37:31] (03Merged) 10jenkins-bot: RadioInputWidget: Add description and example [oojs/ui] - 10https://gerrit.wikimedia.org/r/193895 (owner: 10Kmenger) [00:42:53] (03Merged) 10jenkins-bot: ProgressBar: Add description and example [oojs/ui] - 10https://gerrit.wikimedia.org/r/193869 (owner: 10Kmenger) [00:42:54] 10VisualEditor: List displayed too far to the left margin after deleting blank lines or slugs - https://phabricator.wikimedia.org/T91349#1080617 (10Whatamidoing-WMF) 3NEW [00:46:54] 10VisualEditor, 10WikiEditor, 6Collaboration-Team, 10Flow, 7Design: Editing toolbar(s) for Flow (VE and/or wikitext toolbar) - https://phabricator.wikimedia.org/T78346#1080631 (10DannyH) I talked to James about toolbar use in VE -- they don't currently have tracking on which buttons get used the most. Bu... [00:49:09] (03CR) 10Krinkle: "https://github.com/jshint/jshint/issues/674" [oojs/core] - 10https://gerrit.wikimedia.org/r/192969 (owner: 10Jforrester) [00:50:06] (03CR) 10Krinkle: "The only remaining warnings for oojs are for 'static'." [oojs/core] - 10https://gerrit.wikimedia.org/r/192969 (owner: 10Jforrester) [00:54:41] 10VisualEditor, 10VisualEditor-MediaWiki, 10Collaboration-Team-Sprint-R-2015-03-11, 10Flow: R1. Spike: Add relevant MW features to VE on Flow - https://phabricator.wikimedia.org/T90760#1080649 (10Mattflaschen) >>! In T90760#1078920, @matthiasmullie wrote: > @Mattflaschen If I understand it correctly, it's... [01:20:20] 10VisualEditor: Deleting the table which follows a link triggers the context menu for the link on previous line to open - https://phabricator.wikimedia.org/T91353#1080685 (10Ryasmeen) 3NEW [01:21:13] (03CR) 10Mooeypoo: [C: 032] "Works great!" [oojs/ui] - 10https://gerrit.wikimedia.org/r/193500 (owner: 10Trevor Parscal) [01:22:50] (03PS1) 10Mooeypoo: ProcessDialog: Dismiss errors on teardown [oojs/ui] - 10https://gerrit.wikimedia.org/r/194013 [01:22:59] (03Merged) 10jenkins-bot: Infer retry button action flags from symbolic name [oojs/ui] - 10https://gerrit.wikimedia.org/r/193500 (owner: 10Trevor Parscal) [01:27:10] (03PS2) 10Jforrester: ProcessDialog: Dismiss errors on teardown [oojs/ui] - 10https://gerrit.wikimedia.org/r/194013 (owner: 10Mooeypoo) [01:27:26] (03CR) 10Jforrester: [C: 031] ProcessDialog: Dismiss errors on teardown [oojs/ui] - 10https://gerrit.wikimedia.org/r/194013 (owner: 10Mooeypoo) [01:29:39] James_F: http://etherpad.wikimedia.org/p/VE-Chromium-Upstream [01:29:43] Im in contact with upstream [01:29:47] Feel free to add to the list :) [01:29:51] Krinkle: You've found someone? Awesome. [01:29:57] Krinkle: RoanKattouw_away is probably the best person. [01:30:00] Expanded into Wikimedia-wide. Kept pad title for now [01:30:08] Will be building a tracking bug [01:30:14] * James_F nods. [01:30:16] 10VisualEditor, 10WikiEditor, 6Collaboration-Team, 10Flow, 7Design: Editing toolbar(s) for Flow (VE and/or wikitext toolbar) - https://phabricator.wikimedia.org/T78346#1080709 (10Mattflaschen) Link should be exactly the same as VE once you click it, both for a consistent user experience and so we don't h... [01:30:18] (upstream that is) [01:43:34] Krinkle: How goes your list? [01:43:50] Which one :) [01:44:02] The one I wrote. :_) [01:46:14] Mostly through it. Testing the oojs-ui commits at the moment [01:46:20] * James_F nods. [01:46:21] Good. [01:46:45] CI is blocked on new slaves which is blocked on our existing instances not matching puppet and puppet not passing - zuul-cloner fails to install. [01:47:05] https://phabricator.wikimedia.org/T90984 [01:47:27] Lovely. [01:47:43] Do you need it on Precise? [01:48:29] Krinkle: do you know why OO.ui.ClippableElement.clip gets triggered when scrolling in VE? [01:49:01] ori: If you have a toolbar dropdown, dialog or inspector open, then yes. [01:49:20] nope [01:49:57] In those cases, It's called since the element in question has a certain position on the page and may need to shrink or grow for space allowing based on it moving away or toward the browser edge [01:50:04] ori: In that case, don't know [01:50:07] i'm not even mousing over any VE things, my mouse is resting on the sidebar [01:50:21] ori: Perhaps an overzellous event handler [01:50:26] i'll file a bug and investigate later [01:50:29] yeah, probably [01:50:29] cool [01:50:44] ori: Having the event handler may be reasonable, but not what it does. [01:51:00] Can you include what the handler does? E.g. assuming it's not a cheap early return [01:51:23] James_F: For CI in general and to migrate yes. for karma, no. [01:52:03] Krinkle: Could we migrate karma early and use that? [01:52:06] But for the new pool to be activated, the precise instances need to work as well. [01:52:15] Right now the new pool is depooled. [01:53:44] It was supposed to take 10 minutes to re-create the pool, but has been taking work spread over 3 days (most of that just waiting blocked and doing other stuff instead). [01:54:00] I think this is the last blocker, so I'll give hashar tomorrow to figure it out [01:54:18] 10VisualEditor, 10VisualEditor-Performance: OO.ui.ClippableElement.clip called on scroll - https://phabricator.wikimedia.org/T91355#1080751 (10ori) 3NEW [01:54:33] Otherwise I can fragment the new pool and pool half of it, but that'll require extra work that could be avoided if he fixes it first. So I'll try that first and resolve some other things tomorrow. [01:54:46] Such as disconnects with karma. And see where we're at. [01:55:00] Because evenif it magically worked now, I don't think we're ready to switch over to it just yet. [01:55:06] Yeah. [01:55:07] 10VisualEditor, 10VisualEditor-Performance: OO.ui.ClippableElement.clip called on scroll - https://phabricator.wikimedia.org/T91355#1080765 (10ori) [01:55:11] * James_F sighs. [01:58:29] what code is responsible for detecting wikitext {{ or # and reacting to it? [02:00:07] InezK: SequenceHandler or something like that. [02:00:44] InezK: EventSequencer. [02:01:17] InezK: The MW-specific ones are in ve.ui.MWSequenceRegistry.js [02:01:57] perfect, thank you [02:02:10] Happy to help. :-) [02:06:26] 10Cite: Convert Cite to use extension registration - https://phabricator.wikimedia.org/T87893#1080801 (10Legoktm) T89151 is going to make this annoying since we'll need to keep the global function around. [02:08:40] legoktm: Or just break it. [02:08:45] * James_F coughs. [02:08:53] >.> [02:09:04] legoktm: https://github.com/search?q=wfcite+%40wikimedia&type=Code&utf8=%E2%9C%93 [02:09:25] mediawiki-extensions-WYSIWYG which is dead. [02:09:34] well, that is already broken since it's not an extension function [02:09:40] Then just SearchHighlighter. [02:09:50] oh no, it does check in $wgHooks [02:09:51] sigh. [02:10:02] So fix that, and REL1_25 of Cite isn't valid for MW < REL1_25 [02:10:04] Simples. [02:10:14] (Hey says.) [02:12:20] (03CR) 10Ori.livneh: "There's a performance regression involving an expensive call to ve.ui.Toolbar.calculateOffset that I think may be related to this patch; i" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/192047 (https://phabricator.wikimedia.org/T89543) (owner: 10Krinkle) [02:18:05] ori: https://code.google.com/p/chromium/issues/detail?id=463348 / http://etherpad.wikimedia.org/p/Chromium-Upstream [02:18:08] Feel free to expand :) [02:18:31] 10VisualEditor, 10WikiEditor, 6Collaboration-Team, 10Flow, 7Design: Editing toolbar(s) for Flow (VE and/or wikitext toolbar) - https://phabricator.wikimedia.org/T78346#1080813 (10Quiddity) Based on these example threads, and our goal to rollout further at mediawikiwiki, and He7d3r's comment above, and my... [02:19:18] cool [02:22:00] 10VisualEditor: Please make the "More" and "Fewer" setting in the Insert menu be sticky - https://phabricator.wikimedia.org/T85638#1080821 (10Whatamidoing-WMF) 5declined>3Open I've had second thoughts about whether this will affect most users. Earlier, I was thinking mostly of people typing in English from... [02:33:20] 10OOjs-UI, 10UI-Standardization, 6WMF-Design, 7Design, 5Patch-For-Review: button styles differ between OOjs UI and mediawiki UI - https://phabricator.wikimedia.org/T88449#1080846 (10kaldari) I've committed changes to both MediaWiki UI and OOjs UI to bring them more into sync: https://gerrit.wikimedia.org... [02:35:38] 10OOjs-UI, 10UI-Standardization, 6WMF-Design, 7Design, 5Patch-For-Review: button styles differ between OOjs UI and mediawiki UI - https://phabricator.wikimedia.org/T88449#1080847 (10kaldari) [02:35:46] (03PS1) 10Jforrester: Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 [04:04:10] 10MediaWiki-ResourceLoader, 7Technical-Debt, 7Wikimedia-log-errors: ResourceLoader: Invalid argument supplied for foreach - https://phabricator.wikimedia.org/T46907#1080981 (10Krinkle) 5Open>3declined a:3Krinkle No matches in the last 7 days. While the code still has no check for a false return value,... [04:20:09] 10MediaWiki-ResourceLoader, 7Wikimedia-log-errors: Warning: Invalid argument supplied for foreach() in lessc.inc.php on line 507 - https://phabricator.wikimedia.org/T63706#1081005 (10Krinkle) https://github.com/wikimedia/mediawiki/blob/wmf/1.23wmf15/includes/libs/lessc.inc.php#L504-L517 ```lang=php protected... [04:20:27] 10MediaWiki-ResourceLoader, 7Wikimedia-log-errors: Warning: Invalid argument supplied for foreach() in lessc.inc.php on line 507 - https://phabricator.wikimedia.org/T63706#1081006 (10Krinkle) [04:52:26] 10VisualEditor, 10VisualEditor-Performance, 10Parsoid, 10RESTBase, and 3 others: Store & load data-mw separately - https://phabricator.wikimedia.org/T78676#1081057 (10ssastry) [06:18:08] 10VisualEditor, 10Collaboration-Team-Sprint-R-2015-03-11, 10Flow: Sometimes VE toolbar gets stuck in upper left until reflow/resize window - https://phabricator.wikimedia.org/T90762#1081157 (10Mattflaschen) Note to self, look at {1c5c13fd16be5293816487b9316a1153c5d6894b} [06:29:17] (03PS1) 10Krinkle: build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194057 [06:38:37] 10VisualEditor, 10WikiEditor, 6Collaboration-Team, 10Flow, 7Design: Editing toolbar(s) for Flow (VE and/or wikitext toolbar) - https://phabricator.wikimedia.org/T78346#1081180 (10Qgil) +1 to the idea of starting with a subset of the current VE toolbar for articles. +1 to the idea of enabling in the Flow... [06:45:30] (03CR) 10jenkins-bot: [V: 04-1] build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194057 (owner: 10Krinkle) [07:51:16] (03PS2) 10Krinkle: FieldLayout: Fix display of bulleted list [oojs/ui] - 10https://gerrit.wikimedia.org/r/193885 (owner: 10Kmenger) [07:54:35] (03PS1) 10Krinkle: build: Remove footer override from jsduck [oojs/ui] - 10https://gerrit.wikimedia.org/r/194067 [07:57:15] (03PS3) 10Krinkle: FieldLayout: Fix display of bulleted list [oojs/ui] - 10https://gerrit.wikimedia.org/r/193885 (owner: 10Kmenger) [07:57:51] (03CR) 10Krinkle: "I can't reproduce the pre-formatted rendering locally either. But this should fix it indeed. The first two spaces should be save though, a" [oojs/ui] - 10https://gerrit.wikimedia.org/r/193885 (owner: 10Kmenger) [07:58:28] (03CR) 10Krinkle: [C: 032] FieldLayout: Fix display of bulleted list [oojs/ui] - 10https://gerrit.wikimedia.org/r/193885 (owner: 10Kmenger) [08:00:32] (03Merged) 10jenkins-bot: FieldLayout: Fix display of bulleted list [oojs/ui] - 10https://gerrit.wikimedia.org/r/193885 (owner: 10Kmenger) [08:01:22] (03CR) 10Krinkle: "Fixed. https://doc.wikimedia.org/oojs-ui/master/#!/api/OO.ui.FieldLayout" [oojs/ui] - 10https://gerrit.wikimedia.org/r/193885 (owner: 10Kmenger) [08:01:39] (03PS5) 10Prtksxna: Use only two variables each for each semantic color [oojs/ui] - 10https://gerrit.wikimedia.org/r/193793 (https://phabricator.wikimedia.org/T88622) [08:07:46] (03PS6) 10Prtksxna: Use only two variables each for each semantic color [oojs/ui] - 10https://gerrit.wikimedia.org/r/193793 (https://phabricator.wikimedia.org/T88622) [08:08:48] (03CR) 10Prtksxna: "> What about icon colors defined in images.json?" [oojs/ui] - 10https://gerrit.wikimedia.org/r/193793 (https://phabricator.wikimedia.org/T88622) (owner: 10Prtksxna) [08:14:19] (03CR) 10Krinkle: "recheck" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194057 (owner: 10Krinkle) [08:15:25] (03CR) 10Krinkle: [C: 032] "Synchronised with If45d4706573215 in integration/config." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194057 (owner: 10Krinkle) [08:23:28] (03Merged) 10jenkins-bot: build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194057 (owner: 10Krinkle) [08:26:53] (03PS1) 10Krinkle: Update VE core submodule to master (0cb3b6f) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194070 [08:27:47] (03CR) 10Krinkle: [C: 032] Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [08:28:53] (03CR) 10jenkins-bot: [V: 04-1] Update VE core submodule to master (0cb3b6f) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194070 (owner: 10Krinkle) [08:32:13] (03CR) 10jenkins-bot: [V: 04-1] Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [08:33:50] (03CR) 10Krinkle: [C: 032] "False failure due to sqlite db error." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [08:35:00] (03CR) 10jenkins-bot: [V: 04-1] Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [08:37:51] (03CR) 10Krinkle: [C: 032] Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [08:39:13] (03CR) 10jenkins-bot: [V: 04-1] Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [08:43:30] 10MediaWiki-ResourceLoader, 10Continuous-Integration, 10MediaWiki-Vagrant, 10Wikidata, and 2 others: qunit test broken without explicitly setting $wgResourceLoaderMaxQueryLength - https://phabricator.wikimedia.org/T90453#1081310 (10adrianheine) >>! In T90453#1073327, @Krinkle wrote: > * The relevant softwa... [08:45:16] (03CR) 10Krinkle: [C: 032] Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [08:46:27] (03CR) 10jenkins-bot: [V: 04-1] Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [08:54:56] 10VisualEditor, 10WikiEditor, 6Collaboration-Team, 10Flow, 7Design: Editing toolbar(s) for Flow (VE and/or wikitext toolbar) - https://phabricator.wikimedia.org/T78346#1081317 (10TheDJ) if anyone is going to write 'yet another toolbar', can someone please make sure to really think stuff true ? I'm counti... [09:34:36] 10VisualEditor, 10WikiEditor, 6Collaboration-Team, 10Flow, 7Design: Editing toolbar(s) for Flow (VE and/or wikitext toolbar) - https://phabricator.wikimedia.org/T78346#1081413 (10Pginer-WMF) The idea is for the different controls to behave like they do on the VE toolbar (preferably using the same compone... [10:04:57] (03CR) 10Catrope: "IME the toolbar calculateOffset thing was always there. Maybe it was previously hitching a ride on another reflow, but I seem to recall se" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/192047 (https://phabricator.wikimedia.org/T89543) (owner: 10Krinkle) [10:15:57] 10VisualEditor, 10VisualEditor-Performance: OO.ui.ClippableElement.clip called on scroll - https://phabricator.wikimedia.org/T91355#1081520 (10Catrope) This is caused by the edit notices popup being visible. If you dismiss the popup, then do another profile while scrolling, you won't see calls to `.clip()`. I... [10:39:51] (03CR) 10Catrope: [C: 032] Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [10:41:10] (03CR) 10jenkins-bot: [V: 04-1] Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [10:51:37] Krinkle|detached: Dude https://etherpad.wikimedia.org/p/Chromium-Upstream is awesome [10:51:43] Krinkle|detached: I added a few bugs there [11:19:26] (03Abandoned) 10Catrope: Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [11:19:43] Krinkle|detached: WTF, various CI things are going crazy on https://gerrit.wikimedia.org/r/#/c/194070/1 [11:20:31] Krinkle|detached: jsduck explodes in a cloud of not knowing what jQuery.Event is, and QUnit fails with a syntax error that seems to appear consistently [11:34:48] Although I can't reproduce it locally [11:41:16] Krenair: So, Schema:Edit [11:42:31] Krenair: I'm sorry I haven't been helping you with that recently [11:52:58] (03CR) 10Catrope: Send abort event on unload (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [13:12:19] (03PS4) 10Catrope: Send abort event on unload [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [13:14:23] hello [13:14:27] can someone help me? [13:14:35] (03CR) 10jenkins-bot: [V: 04-1] Send abort event on unload [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [13:42:49] (03PS2) 10Gerrit Patch Uploader: Move result of thumb from i18n to config [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/187338 [13:42:52] (03CR) 10Gerrit Patch Uploader: "This commit was uploaded using the Gerrit Patch Uploader [1]." [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/187338 (owner: 10Gerrit Patch Uploader) [13:50:46] RoanKattouw_away: checking into it now [15:01:55] (03PS10) 10Mvolz: Add ability to set requesting User-Agent [services/citoid] - 10https://gerrit.wikimedia.org/r/193134 (https://phabricator.wikimedia.org/T89757) [15:04:18] (03CR) 10Mobrovac: [C: 032] Add ability to set requesting User-Agent [services/citoid] - 10https://gerrit.wikimedia.org/r/193134 (https://phabricator.wikimedia.org/T89757) (owner: 10Mvolz) [15:09:23] (03PS3) 10Mvolz: [WIP] Add mocha tests [services/citoid] - 10https://gerrit.wikimedia.org/r/193580 (https://phabricator.wikimedia.org/T91168) [15:14:22] Krinkle: https://integration.wikimedia.org/ci/job/mwext-VisualEditor-qunit/14136/console has the same problem [15:14:38] Completely different commit [15:18:45] (03PS4) 10Mvolz: Add mocha tests [services/citoid] - 10https://gerrit.wikimedia.org/r/193580 (https://phabricator.wikimedia.org/T91168) [15:20:13] (03PS5) 10Mvolz: Add mocha tests [services/citoid] - 10https://gerrit.wikimedia.org/r/193580 (https://phabricator.wikimedia.org/T91168) [15:20:47] (03PS6) 10Mvolz: Add mocha tests [services/citoid] - 10https://gerrit.wikimedia.org/r/193580 (https://phabricator.wikimedia.org/T91168) [15:21:38] (03CR) 10Mobrovac: [C: 04-1] "Why have both tests/index.js and test/index.js ? No need for tests/index.js. Remove it, please." [services/citoid] - 10https://gerrit.wikimedia.org/r/193580 (https://phabricator.wikimedia.org/T91168) (owner: 10Mvolz) [15:23:46] (03PS7) 10Mvolz: Add mocha tests [services/citoid] - 10https://gerrit.wikimedia.org/r/193580 (https://phabricator.wikimedia.org/T91168) [15:24:06] (03CR) 10Mvolz: "rename that went awry!" [services/citoid] - 10https://gerrit.wikimedia.org/r/193580 (https://phabricator.wikimedia.org/T91168) (owner: 10Mvolz) [15:25:29] (03CR) 10Catrope: [C: 032] Send abort event on unload [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [15:25:59] (03PS5) 10Catrope: Send abort event on unload [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [15:26:06] (03CR) 10Catrope: [C: 032] Send abort event on unload [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [15:27:26] (03CR) 10jenkins-bot: [V: 04-1] Send abort event on unload [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [15:29:36] RoanKattouw, hey [15:31:18] Hey [15:31:49] (03CR) 10Catrope: "Jenkins won't let me merge anything in VE-MW right now :(" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [15:32:58] (03CR) 10Mobrovac: [C: 032] Add mocha tests [services/citoid] - 10https://gerrit.wikimedia.org/r/193580 (https://phabricator.wikimedia.org/T91168) (owner: 10Mvolz) [15:40:21] (03PS3) 10Catrope: Reduce complexity around onbeforeunload handler [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/177107 [15:41:18] (03CR) 10Catrope: "I can't reproduce the behavior mentioned in PS1. When I make a change and then click Recent Changes, I am prompted with the native Stay on" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/177107 (owner: 10Catrope) [15:41:44] RoanKattouw, the issue I ran into is that the save failures will need to be stored in the same way [15:42:15] Oh save failures in the wikitext editor you mean? [15:42:16] which means both VE (on the client) and WE (in the server hooks) will need to be able to detect all the same events [15:42:56] Yeah [15:43:01] That's going to be difficult for some things [15:43:33] RoanKattouw: mw-ve failing is this https://phabricator.wikimedia.org/T91399 [15:43:50] [2015-02-19 12:51:56] RoanKattouw: the whole of ve.init.mw.trackSubscriber.js screams "I am annoyed that I have to do this, so I'll try to get it over with as fast as possible". getDefaultTiming is weird. [15:44:07] Ori was right, I was annoyed that I had to it, and I did try to get it over with [15:44:14] I guess what I'm saying is, I feel your pain :| [15:44:53] (03CR) 10jenkins-bot: [V: 04-1] Reduce complexity around onbeforeunload handler [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/177107 (owner: 10Catrope) [15:50:02] Morning. [15:51:58] RoanKattouw, I wonder if we can do savefailure type detection on the server for VE or something [15:52:15] then we'd not have to rewrite the same logic to work in WikiEditor [15:52:47] Hmm [15:52:52] That sounds sensible [15:52:59] It would be pretty easy to put in ApiVisualEditor.php [15:53:28] Or even in one hook that attaches to all edits, if we wanted [15:53:30] Well, you'd re-write the JS to work in PHP, and then copy the PHP. [15:53:34] Still sounds like one re-write. [15:53:47] Yeah but we wouldn't have duplication that way [15:58:01] … except for the copy-and-pasting. [15:58:29] Or are you suggesting that WikiEditorHooks.php would call into ApiVisualEditor.php? If so, ick. [16:01:55] Yeah I'm not sure how to structure that yet [16:02:14] But it should be much easier to figure out how to put it in one place if it's at least implemented in the same language [16:02:21] (03CR) 10Catrope: [C: 04-1] [WIP] Schema:Edit instrumentation (035 comments) [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) (owner: 10Alex Monk) [16:02:24] (03CR) 10Jforrester: [C: 032] build: Remove footer override from jsduck [oojs/ui] - 10https://gerrit.wikimedia.org/r/194067 (owner: 10Krinkle) [16:03:17] Krenair: That instrumentation commit ---^^ mostly looks good. I commented on a few things inline but the general approach of passing around a session ID through a hidden form field makes sense to me [16:03:31] Krenair: Is there anything major that's missing from that commit still? Or anything you have questions about? [16:03:55] abort.type and saveFailure.type need to be addressed before it can go in [16:05:25] (03Restored) 10Krinkle: Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [16:05:40] (03CR) 10Krinkle: [C: 032] Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [16:06:07] (03CR) 10Krinkle: "T91399 is fixed." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [16:07:30] (03Abandoned) 10Catrope: Update VE core submodule to master (2bd526b) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194029 (owner: 10Jforrester) [16:07:43] (03CR) 10Catrope: [C: 032] Update VE core submodule to master (0cb3b6f) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194070 (owner: 10Krinkle) [16:07:48] RoanKattouw: :-( [16:08:20] James_F: The one I abandoned is a strict subset of the one I just +2e [16:08:21] d [16:08:33] RoanKattouw: YEah, but it's not working yet [16:08:33] RoanKattouw: I rather wanted to make Timo's confusing patch a smaller change. [16:08:39] Urgh, right [16:08:45] That's the one jshint was freaking out on [16:08:46] Alright [16:08:47] And yeah, it needs to wait. [16:08:49] I'll swap them then [16:08:51] I'll fix it now though [16:08:52] Well, "wait" [16:08:54] 1 se [16:08:55] c [16:09:06] It's in VE core's master already, so it can't really wait that long [16:09:34] RoanKattouw: And self-merged by Krinkle, no less. ;-) [16:09:34] (03PS2) 10Krinkle: Update VE core submodule to master (0cb3b6f) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194070 [16:09:58] James_F: Had to take load off of prod slaves. [16:09:59] (03CR) 10Catrope: [C: 032] Update VE core submodule to master (0cb3b6f) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194070 (owner: 10Krinkle) [16:10:08] So forged ahead with migrating all jsduck to labs slaves [16:10:47] It was previously tested in unicodejs as a small experiment. But then left. [16:11:18] Krinkle: Aha. OK. [16:11:33] Krinkle: I know, I know, just complaining. :-) [16:11:44] No it's a good point [16:11:47] It hurt to do it :) [16:24:06] (03CR) 10Jforrester: "recheck" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [16:24:20] (03CR) 10Jforrester: [C: 032] Update VE core submodule to master (0cb3b6f) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194070 (owner: 10Krinkle) [16:28:13] Krinkle: So what's the plan for repos which have JS and other languages too with documentation needs? .docs/* -> .jsduck/* rather closes the door on that being the common directory [16:29:06] (03CR) 10Jforrester: "Still failing. :-(" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194070 (owner: 10Krinkle) [16:29:09] (03CR) 10jenkins-bot: [V: 04-1] Update VE core submodule to master (0cb3b6f) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194070 (owner: 10Krinkle) [16:32:59] James_F: https://gerrit.wikimedia.org/r/#/c/194125/ hasn't merged yet [16:33:02] the revert [16:33:42] It would've been was zuul not restarted in the middle. [16:33:45] Unlucky :) [16:35:42] Krinkle: Helpful. [16:40:26] (03CR) 10Krinkle: "recheck" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [16:42:12] RoanKattouw: You'd be surprised how productive one is with classic songs of Kinderen voor Kinderen in the background. [16:42:44] haha [16:42:57] Op een onbewoond ei-hei-land... [16:43:26] Krinkle: Children for Children? [16:43:33] I have a mug in my kitchen cabinet at home (SF) with the logo of Tweehek Recreatie on it [16:43:48] They ran (and still run) youth camps out in Drenthe [16:43:52] I went there in, like, 2002? [16:43:53] Krinkle: Nursery rhymes? [16:44:06] Somehow I still have the mug ~13 years later [16:44:24] Those camps is where I picked up most of the KvK songs I know [16:44:41] .. Ik heb zo wa-waw-waw-waw-waanzinnig gedroomd... [16:44:56] And of course the gooische G sound [16:45:03] and the R [16:45:14] There was so much fuss about the R [16:45:49] they went from requiring children to have it and not allow non-western entries, to the opposite [16:46:02] Because it was convenient and less travel distance [16:46:16] (03CR) 10Jforrester: [C: 032] "Per Roan." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [16:46:25] Are you saying the modern KvK stuff is all eastern accents? [16:46:34] hehe [16:46:39] (03PS3) 10Mvolz: [WIP] Add ability to request AcceptLanguage [services/citoid] - 10https://gerrit.wikimedia.org/r/193154 [16:46:39] It's more a mix [16:46:45] I do know that most of the stuff I heard is upper-class western accents [16:46:52] Though I don't know the recent stuff well [16:47:03] I mostly heard the first 10 years. Had 'em on CD when I was younger. [16:47:52] It's also interesting how they changed the opening lyrics. [16:48:31] (03Merged) 10jenkins-bot: Send abort event on unload [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/190144 (owner: 10Alex Monk) [16:48:43] Turns out saying that children below the equator often grow up to be beggers, wasn't the best way to summarise the issues. And they also stopped being a charity later. [16:49:24] Krinkle: Wow. What? [16:49:45] haha [16:49:48] James_F: Yeah, the opening theme basically has two lines of lyrics [16:49:51] "Children for children" [16:49:52] Yeah that is insensitive [16:50:06] (03CR) 10Jforrester: [C: 032] "…" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194070 (owner: 10Krinkle) [16:50:11] "a child below the equator often grows up to be a begger" [16:50:18] Oh right [16:50:20] evenaar, bedelaar [16:50:23] Yeah [16:51:55] RoanKattouw: I do miss Willem Nijholt's presentation. He was great at that stuff. So sober. [16:52:05] the first edition is on youtube :) [16:52:24] (03Merged) 10jenkins-bot: Update VE core submodule to master (0cb3b6f) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194070 (owner: 10Krinkle) [16:53:00] \o/ [17:31:17] Hmm [17:31:24] Krinkle: What JS interpreter does ecmabot-wm use? [17:31:32] RoanKattouw: v8 [17:31:39] 1.5year old version. [17:32:02] https://github.com/Krinkle/oftn-bot [17:32:28] It's a fork of ##javascript irc bot, oftn-bot. I diverged from upstream and unable to upgrade basically. [17:32:42] OK, let's try this [17:33:11] >>> a = new Array(120000); a.splice(300,0,'x'); a.splice(260,1); a.indexOf('x'); [17:33:11] RoanKattouw: (number) 299 [17:33:21] Krinkle: In Safari 8, that returns 300 [17:33:49] Looks like this v8 still matches latest Chrome canary in this regard [17:33:54] also 299 [17:34:12] and 300 in Safari 8, yeah. [17:34:25] and 299 in latest WebKit nightly [17:34:38] My command line jsc returns 299 as well [17:34:40] Oh good [17:34:44] That means they've fixed it already [17:34:48] Which uses JavaScriptCore engine, just like Safari. [17:34:52] But a year or so ahead [17:35:04] Yeah [17:35:25] The length of the array appears to matter [17:35:50] RoanKattouw: So, what does this thing do [17:36:15] (03CR) 10Jforrester: [C: 032] "…" [oojs/ui] - 10https://gerrit.wikimedia.org/r/194067 (owner: 10Krinkle) [17:36:19] Create sparse array, insert 'x' at offset 300, remove an element (a gap, really) at offset 260, assert that 'x' has now shifted to offset 299 [17:36:41] When removing gaps from a sparse array, Safari 8 removes a gap at the end rather than at the requested offset [17:37:15] (03CR) 10Jforrester: [C: 032] core: Improve class related unit tests [oojs/core] - 10https://gerrit.wikimedia.org/r/193153 (owner: 10Krinkle) [17:37:22] So before the splice, a = [ undefinedx300, 'x', undefinedx119699 ] and afterwards it should be a = [ undefinedx299, 'x', undefinedx119699 ] but instead it's a = [ undefinedx300, 'x', undefinedx119698 ] [17:38:19] Surprisingly, this happens only if the length is >= 100000 [17:38:28] You would expect a power of 2 as a threshold value [17:38:42] Maybe there's a different implementation that's used for sparse arrays of length >= 100k [17:38:56] (03CR) 10Jforrester: "recheck" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/149117 (https://bugzilla.wikimedia.org/54906) (owner: 10Alex Monk) [17:39:16] (03CR) 10Jforrester: "recheck" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/177107 (owner: 10Catrope) [17:39:41] (03CR) 10Jforrester: [C: 04-1] "It doesn't." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/191037 (https://phabricator.wikimedia.org/T88148) (owner: 10Esanders) [17:39:49] And the index of 'x' has to be >= 8 :O [17:40:32] So a = new Array(n); a.splice(i,0,'x'); a.splice(0,1); a.indexOf('x'); is supposed to return i-1, but returns i if n>=100000 && i>=8 [17:40:56] I gotta eat now, but when I get back I'll document this ridiculous bug and add a workaround for it [17:41:01] Reusing the one for Opera [17:41:19] (03Merged) 10jenkins-bot: core: Improve class related unit tests [oojs/core] - 10https://gerrit.wikimedia.org/r/193153 (owner: 10Krinkle) [17:41:50] edsanders: RoanKattouw: when VE is initialized, there is a call to ve.ui.Toolbar.prototype.calculateOffset, to calculate the left and right offsets of the toolbar. But the offsets should be the same as div#content, right? So could we compute the offsets earlier, when we're still waiting for the API response? [17:42:58] (03Merged) 10jenkins-bot: build: Remove footer override from jsduck [oojs/ui] - 10https://gerrit.wikimedia.org/r/194067 (owner: 10Krinkle) [17:46:39] ori, sounds reasonable [17:47:00] the only edge case I can thing of is if switching to VE changes whether or not you have a scroll bar [17:47:04] *think [17:54:29] RoanKattouw_away: haha :D [17:56:52] (03PS1) 10Mvolz: Fix: itemType was missing in URLs with scrape err [services/citoid] - 10https://gerrit.wikimedia.org/r/194143 [17:59:39] (03CR) 10Krinkle: [C: 032] build: Add disconnect tolerance to karma config [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/193843 (https://phabricator.wikimedia.org/T90673) (owner: 10Jforrester) [17:59:48] Thanks Krinkle. [18:01:02] (03CR) 10Bartosz Dziewoński: "No idea. Probably from some early spec." [oojs/ui] - 10https://gerrit.wikimedia.org/r/193793 (https://phabricator.wikimedia.org/T88622) (owner: 10Prtksxna) [18:02:21] (03CR) 10Mobrovac: [C: 032] Fix: itemType was missing in URLs with scrape err [services/citoid] - 10https://gerrit.wikimedia.org/r/194143 (owner: 10Mvolz) [18:02:40] (03Merged) 10jenkins-bot: Fix: itemType was missing in URLs with scrape err [services/citoid] - 10https://gerrit.wikimedia.org/r/194143 (owner: 10Mvolz) [18:05:01] edsanders, ori: Also consider the case where the toolbar is floating [18:05:15] It could be initially floating if the user scrolled down while loading, or used a keyboard shortcut to load [18:05:31] RoanKattouw_away: but the offset computation on initialization is just for left/right iirc [18:05:53] Oh OK [18:06:27] James_F: RoanKattouw_away: Hm.. did oojs-ui demos live examples work in the past? It's failing as it's trying to load oojs.js and dist/, makes sense, but I can imagine it worked in the past [18:06:34] docs that is [18:06:36] (03PS1) 10Esanders: Fix typo in direction calculation [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194148 [18:06:45] https://doc.wikimedia.org/oojs-ui/master/#!/api/OO.ui.Dialog [18:07:16] Krinkle: They worked locally [18:07:24] I mean I had to fix it twice, but they did work [18:08:17] (03PS3) 10Bartosz Dziewoński: tests: Run JS/PHP tests for widgets with required parameters, too [oojs/ui] - 10https://gerrit.wikimedia.org/r/193114 [18:08:29] (03PS6) 10Bartosz Dziewoński: tests: Add infusion tests [oojs/ui] - 10https://gerrit.wikimedia.org/r/193123 [18:08:56] (03CR) 10Catrope: [C: 031] "lol. What did this break? Deferring to David for final approval." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194148 (owner: 10Esanders) [18:12:20] RoanKattouw_away: Yeah. [18:12:52] (03CR) 10Alex Monk: [WIP] Schema:Edit instrumentation (031 comment) [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) (owner: 10Alex Monk) [18:12:59] I'm not sure if they ever worked on doc.wm.o though [18:14:31] (03CR) 10jenkins-bot: [V: 04-1] tests: Add infusion tests [oojs/ui] - 10https://gerrit.wikimedia.org/r/193123 (owner: 10Bartosz Dziewoński) [18:15:00] (03CR) 10Bartosz Dziewoński: [C: 032] Use only two variables each for each semantic color [oojs/ui] - 10https://gerrit.wikimedia.org/r/193793 (https://phabricator.wikimedia.org/T88622) (owner: 10Prtksxna) [18:17:30] (03PS9) 10Alex Monk: [WIP] Schema:Edit instrumentation [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) [18:17:40] (03Merged) 10jenkins-bot: Use only two variables each for each semantic color [oojs/ui] - 10https://gerrit.wikimedia.org/r/193793 (https://phabricator.wikimedia.org/T88622) (owner: 10Prtksxna) [18:18:26] (03PS10) 10Alex Monk: [WIP] Schema:Edit instrumentation [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) [18:22:43] (03PS1) 10Kaldari: Syncing some button styles with MediaWiki UI [oojs/ui] - 10https://gerrit.wikimedia.org/r/194155 (https://phabricator.wikimedia.org/T88449) [18:22:48] (03CR) 10jenkins-bot: [V: 04-1] Syncing some button styles with MediaWiki UI [oojs/ui] - 10https://gerrit.wikimedia.org/r/194155 (https://phabricator.wikimedia.org/T88449) (owner: 10Kaldari) [18:32:57] (03Merged) 10jenkins-bot: build: Add disconnect tolerance to karma config [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/193843 (https://phabricator.wikimedia.org/T90673) (owner: 10Jforrester) [18:33:59] (03PS2) 10Kaldari: Syncing some button styles with MediaWiki UI [oojs/ui] - 10https://gerrit.wikimedia.org/r/194155 (https://phabricator.wikimedia.org/T88449) [18:37:25] (03PS3) 10Kaldari: Syncing some button styles with MediaWiki UI [oojs/ui] - 10https://gerrit.wikimedia.org/r/194155 (https://phabricator.wikimedia.org/T88449) [18:39:09] (03CR) 10jenkins-bot: [V: 04-1] Syncing some button styles with MediaWiki UI [oojs/ui] - 10https://gerrit.wikimedia.org/r/194155 (https://phabricator.wikimedia.org/T88449) (owner: 10Kaldari) [18:39:55] (03CR) 10Kaldari: "recheck" [oojs/ui] - 10https://gerrit.wikimedia.org/r/194155 (https://phabricator.wikimedia.org/T88449) (owner: 10Kaldari) [18:41:09] (03CR) 10jenkins-bot: [V: 04-1] Syncing some button styles with MediaWiki UI [oojs/ui] - 10https://gerrit.wikimedia.org/r/194155 (https://phabricator.wikimedia.org/T88449) (owner: 10Kaldari) [18:42:16] (03PS1) 10Mvolz: Fix: Change itemType of 520 citation to 'webpage' [services/citoid] - 10https://gerrit.wikimedia.org/r/194159 [18:42:55] (03CR) 10Mvolz: "Just one of those days... ugh." [services/citoid] - 10https://gerrit.wikimedia.org/r/194159 (owner: 10Mvolz) [18:52:22] (03CR) 10Mobrovac: [C: 032] Fix: Change itemType of 520 citation to 'webpage' [services/citoid] - 10https://gerrit.wikimedia.org/r/194159 (owner: 10Mvolz) [18:52:40] (03Merged) 10jenkins-bot: Fix: Change itemType of 520 citation to 'webpage' [services/citoid] - 10https://gerrit.wikimedia.org/r/194159 (owner: 10Mvolz) [18:55:12] (03PS1) 10Mooeypoo: Work with the new mw.Feedback and add UserAgent checkbox [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194162 [18:58:54] (03PS2) 10Jforrester: Work with the new mw.Feedback and add UserAgent checkbox [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194162 (owner: 10Mooeypoo) [18:59:13] (03CR) 10Jforrester: [C: 032] Work with the new mw.Feedback and add UserAgent checkbox [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194162 (owner: 10Mooeypoo) [19:00:09] Argh [19:00:20] // This polyfill only been discovered to be necessary on Opera [19:00:22] // and it seems to handle up to 1048575 function parameters. [19:00:22] RoanKattouw, do you know how to work with WikimediaMessages extensions ? [19:00:25] Well guess what Safari doesn't like [19:00:47] Opera? [19:01:38] haha [19:01:42] Well possibly that too [19:01:51] But it also doesn't like large numbers of function parameters [19:02:22] Fortunately there's some code that zz_sucheta wrote as a cleanup thing when she first started on our team and that saves my butt here [19:02:24] (03Merged) 10jenkins-bot: Work with the new mw.Feedback and add UserAgent checkbox [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194162 (owner: 10Mooeypoo) [19:02:45] RoanKattouw: Nice. :-) [19:03:16] ve.batchPush, in particular [19:05:54] mooeypoo: James_F: https://gerrit.wikimedia.org/r/194164 :o [19:06:21] MatmaRex: Meh. What about Firefox users? [19:06:23] ;-) [19:06:29] * James_F is unhelpful. [19:06:32] James_F: well, they're in deep [19:06:43] but i'm going to try fixing that today, maybe [19:06:47] MatmaRex: Good. [19:07:08] James_F: i was also thinking that the feedback dialog could really use minRows in addition to maxRows, for the text input widget. [19:07:24] MatmaRex: Meh. It's just a feedback box. :-) [19:07:25] since the field kind of doesn't look like it'll take long text. [19:07:31] but yeah, i don't care enough [19:07:44] :-) [19:09:42] MatmaRex, I've been reading a bit about this -- can we use a mutation observer ? [19:10:22] According to mozilla hack blog, it works, and there are polyfills for IE 8-10 [19:11:33] MatmaRex, https://hacks.mozilla.org/2012/05/dom-mutationobserver-reacting-to-dom-changes-without-killing-browser-performance/ [19:14:05] mooeypoo: (thinking) [19:14:19] MatmaRex, this can be a fallback [19:14:35] so, if DOMNodeInsertedIntoDocument doesn't exist, use that [19:14:41] ? [19:16:09] mooeypoo: there was a patch proposed for it, even - https://gerrit.wikimedia.org/r/#/c/176681/ - but the problem for our use case is that we would have to observe document root, and thus pay the vague unspecified performance penalty on the entire document [19:16:19] because we're not watching for changes inside a node [19:16:36] we're watching for changes "outside" a node - we want to know when a node is attached, not when something is attached to a node [19:17:01] the old deprecated mutation events just happen to provide this exact functionality, with firefox not supporting it [19:17:38] we can use the mutation events in the same way as mutationobserver, by attaching some handlers to the root node, and possibly pay the same unspecified penalty [19:18:21] i don't think that anybody has actually tested how big the penalty is for us, but it's allegedly not tiny [19:19:09] MatmaRex, I started working on a proof of concept when I saw it, but I don't yet have much, I was mostly researching [19:19:57] MatmaRex, why do we have to observe document root? Create an observer in the TextInputWidget and have it expand the input [19:20:21] mooeypoo: okay, but what would the observer observe? [19:20:28] we don't need to listen to the entire page, just specific items, it can be internal, and we can also destroy/disconnect them [19:20:31] the input itself [19:20:44] only changes inside the element. we don't care about them [19:20:50] we care about element's parent changing [19:20:54] $element's* [19:20:56] hmm [19:21:03] MatmaRex, not according to this bug report... ? [19:21:09] Krinkle: are you good with wikitext changes or do you feel something is blocking? cc milimetric [19:21:17] the entire problem is that inputs don't expand to the text size, no? [19:21:21] that's inside the input [19:21:34] okay, zoom out [19:22:16] the big picture is that we want to detect when the computed width or font-size or line-height etc. of the changes, to update our height of it. [19:22:19] Okay, how often do we even listen to 'when element is attached' events at all? We seem to frown on listening to those in general anyways [19:22:23] there are very few occasions when we do [19:22:37] exactly one, in fact, in textInputWidget [19:22:41] yes [19:23:02] and those affect the text input itself or its wrapper, both of which are in TextInputWidget [19:23:23] these computed things can change when a node is attached somewhere in the DOM, thanks to CSS and cascading and stuff [19:23:38] MatmaRex, but they'll affect the input's wrapper too, then [19:23:49] Krenair: [20:21] nuria Krinkle: are you good with wikitext changes or do you feel something is blocking? cc milimetric [19:23:53] whatever affects the input, affects its wrapper... ? [19:23:54] nuria: (I think you meant Krenair ) [19:23:59] well, whatever computed [19:24:10] yeah, but if it's inherited from parent, there is no mutation event fired, or no change for MutationObserver to notice [19:24:25] so we were trying to detect the attachment [19:24:29] but, what you said made me htink [19:24:31] ayyyy RoanKattouw you are soo righttt [19:24:42] hm [19:24:46] we could provide our own unused parent node for $element [19:24:52] Krenair: are you good with wikitext changes or do you feel something is blocking? cc milimetric [19:24:52] (I have to run for a bit, will you be here in 20 minutes? ) [19:24:57] liek, this.$parent = $('
').append(this.$element) [19:25:01] and not attach it anywhere [19:25:08] and add the MutationObserver to the $parent [19:25:11] ooh [19:25:19] and when it fires, well, it means that $element was reattached elsewhere [19:25:26] interesting [19:25:42] so we add new MutationObserver on the new parent (or create new fake parent if there's no parent) [19:25:50] that could actually perhaps work [19:25:56] interesting [19:26:11] It should be a fallback anyways [19:26:26] the current event is probably preferable if it exists [19:26:36] the current method we use (mutation events) is allegedly deprecated [19:26:39] but yeah, either [19:26:44] whichever works [19:26:47] * mooeypoo nods [19:26:57] some browsers support only one, some both [19:26:59] MatmaRex, are you going to work on it? [19:27:16] or do you want me to attempt? [19:27:25] yeah, i will i think [19:27:29] cool [19:27:52] ok sounds good [19:27:55] i'll be back in a bit [19:28:09] wikibugs down? :( [19:29:23] RoanKattouw, I think compareDocumentOrder is being used wrongly too [19:29:36] the documentation isn't very clear [19:31:11] edsanders: Yeah that's why I asked David to look [19:31:50] it uses the same signature as Array.sort would expect [19:32:01] i.e. f(a,b) = a-b [19:32:11] On the one hand, it's just a typo fix, but on the other hand, I'm a bit scared of that code [19:32:19] so it's positive if a comes after b [19:33:01] in his code he writes direction = compare(before, after) [19:33:39] which would give you a direction of '1' if before is after after [19:33:58] i.e. you moved 'left' (for some meaning of 'left') [19:43:14] nuria, wikitext changes? [19:43:20] Krenair: yes [19:43:26] edits to the schema? [19:45:38] Krenair: i was thinking more of your changeset to instrument wikitext [19:46:07] I thought about changing the schema, but didn't know what to change [19:48:07] Krenair: are schema changes blocking you from merging your wikitext patch? [19:48:14] Possibly. [19:48:27] Krenair: doesn't seem that it would be the case though..how so? [19:48:35] cc milimetric [19:48:37] We need to work out how we're going to gather the same information about save failures in both VE and the wikitext editor. [19:49:55] Krenair: information is not going to be identical [19:50:34] Krenair: so -given that the schema is skewed heavily towards client side- Wikitext will fill in the bits it can fit [19:50:44] Krenair: rest will be unfilled [19:51:05] Krenair: that is a good starting point [19:51:51] So you're happy with just leaving those type fields empty? [19:52:01] I like saying "increasingly smaller" [19:53:20] increasingly smaller is totally fine [19:53:22] Krenair: there is no other choice, is there? as many of the schema codes only apply to VE (like "responseUnknown") [19:53:47] Well it might be possible to make them generate the same type data [19:53:52] I don't know [19:54:15] Krenair: not really, you are in the server not in the client [19:54:30] James_F, is leaving that blank going to be okay with you? [19:54:35] Krenair: in an ideal scenario the schema should be "generic" and not assume client side [19:54:43] Krenair: but we are not in an ideal scenario [19:54:54] Krenair: Leaving which ones? [19:55:19] The same ones that I've been talking to you about for the past two weeks. [19:55:32] Krenair: when you say "it might be possible", we have to consider time. So if we can do it with a lot of effort, it's better to leave them empty for now, then iterate later if we solve it [19:56:37] Krenair: Sure. About 300 people have talked to me about things to do with VisualEditor in the past few weeks. If you think you need my sign-off, write it down in a single list. :-) [19:56:58] guys - let's be straight here [19:57:08] let me lead this, I feel that it's been somewhat unproductive so far [19:57:10] James_F: you are in the middle of the hotness.... [19:57:19] so - recap [19:57:21] :-) [19:57:34] James_F, the problematic fields are action.saveFailure.type and action.abort.type. [19:57:35] Visual Editor logs a few failure types, visualized here: https://edit-analysis.wmflabs.org/adhoc.html#ve-failures.tsv [19:57:48] I have been focusing on the saveFailure part [19:58:23] The Wikitext instrumentation that Alex is working on is stalled on the fact that he can't emit the same exact set of failure types from the server [19:58:41] Yeah. [19:58:44] so far, he can only get "conflict", "pageDeleted", and "responseUnknown" [19:59:01] and our advice has been: that's ok for now, at least if we deploy it we'll start getting some data [19:59:04] What about hook-related ones? [19:59:21] what's an example of that? [19:59:38] AbuseFilter, CAPTCHA, SpamBlacklist, TitleBlacklist. [19:59:45] You mean trying to detect which extension has prevented it from being saved? [19:59:47] Those are our top 4 I think. [19:59:49] gotcha. Krenair: is that possible? [19:59:51] Yeah. [20:00:02] James_F: But IIRC we only capture three [20:00:03] (Top 4 after "responseUnknown".) [20:00:16] So maybe the fourth is all the unknowns? [20:00:19] RoanKattouw: Oh, do we not cover TitleBlacklist yet? [20:00:23] RoanKattouw is right, only 3: https://meta.wikimedia.org/wiki/Schema:Edit [20:00:27] * James_F nods. [20:00:33] OK, that might be something to fix. :-) [20:00:36] Anyway, sure. [20:00:39] James_F: Maybe TitleBlacklist is where those unknowns come from [20:00:46] I will take another look to see if I can special-case those extensions. [20:00:55] I suggested we might want to add a free text field for failure info [20:00:56] RoanKattouw: that'd be great to find out [20:01:01] So we can figure out what the unknowns are [20:01:04] Now and in the future [20:01:28] a free text field would be a little harder to analyze, but sounds like a great idea [20:01:57] RoanKattouw: that sounds to me too much like client side error logging [20:02:18] So instead of responseUnknown we'd record the actual response in action.saveFailure.type? [20:02:22] RoanKattouw: so not so fond of that, one thing is analyzing user behaviour, other application exceptions [20:02:24] It's currently an enum. [20:02:32] if we had a free-form field we could understand better where the errors meet the real exception cases, nuria, because now it's tangled in "responseUnknown" [20:02:42] nuria: We have no idea whether these are application exceptions. [20:02:51] * milimetric nods [20:03:06] nuria: The same kind of information could be logged on the server too. The distinction is application exceptions vs user behavior, the client vs server side distinction is irrelevant [20:03:16] James_F / RoanKattouw: we can make a second field, leave the enum alone [20:03:23] Yeah make a second field [20:03:27] milimetric: I'd prefer that. [20:03:36] leave "responseUnknown" for now, and just detail it in the new field [20:03:39] (Yay, another schema revision.) [20:03:45] Yeah.. [20:03:50] but if you think Title blacklist is causing some, do fix that as well [20:03:50] James_F: The distinction between what the user sees and "what error made your application fail" is not irrelevant [20:04:02] My change is already out of date because of the addition of mediawiki.version [20:04:11] James_F: if all your 5 errors end up displaying a white page , when it comes to a funnel is teh same "state" [20:04:22] Krenair: you can target the old schema revision, you don't have to emit to the latest version [20:04:32] but if it's not too hard, you can add the version field too [20:04:33] up to you [20:05:05] nuria: We're not talking about a white page, though. [20:05:36] nuria: We're talking about (in VE) a message saying "Sorry, we couldn't save your edit at this time because: xxx", or (in WT) a similar message. [20:05:37] agreed, nuria, the problem here is they are having a problem understanding what exactly to log. Maybe a second schema would help RoanKattouw? Like an "Edit_Debug"? [20:06:00] James_F: understood, you should use this schema for " user-funnel states" [20:06:02] nuria: /Possibly/ a few % of these are actually catastrophic everything-broke client-crashed situations. [20:06:28] Most of the errors are caused by the user though [20:06:30] nuria: Mostly these are "The DB is locked" or "That word is banned" or "You're not logged in" or whatever. [20:06:34] Like SpamBlacklist and TitleBlacklist [20:06:35] nuria: These are user errors. [20:06:41] nuria: Or, well, "errors". [20:06:47] yeah, what do we do about read only errors in VE? [20:06:52] James_F: but not "responseUnknown" [20:07:03] Krenair: It's a fatal recoverable save state. [20:07:04] Well responseUnknown could be either [20:07:07] nuria: User actions which the system is telling them was wrong. [20:07:12] nuria: No. [20:07:14] that's what they're saying, that perhaps TitleBlacklist is responseUnknown's best customer [20:07:25] It could be an extension we didn't account for, or a catastrophic failure we didn't account for [20:07:25] nuria: responseUnknown == "VisualEditor's developers haven't coded this one in yet". [20:07:59] Right. that is info for the developer RoanKattouw not to understand user behaviour [20:08:05] NO! [20:08:08] so, James_F, did we have a decision for Krenair so he can finish up the first version of his work? [20:08:20] I am not saying is not useful, i am saying mixing both is going to lead to confusion [20:08:22] I'm going to take another look to see what I can pull from the status [20:08:44] Krenair: you're looking into detecting the others from interactions with extensions via hooks, right? [20:08:48] but yes, milimetric is right, let's give Krenair the green light to go on so we can get some data to compare [20:09:05] Will check to see what I can do about other extensions [20:09:17] If I can get a string error message (or a message key), I'm going to make a field for it in the schema. [20:09:18] ok, so this discussion is good here. Krenair: poke us once you've checked [20:09:20] milimetric: Yes. (a) Krenair to finish first version as is. (b) Adjust the schema to record the responseUnknown failure state so we can work out what they are (WHO?). (c) Krenair to add WT extension detection as a second patch. (d) ??? [20:09:44] James_F: (e) Add TitleBlacklist detection to VE [20:10:02] ok, great [20:10:10] and to the enum in the schema [20:10:11] RoanKattouw: Eh. I'd rather we record responseUnkowns first rather than striking out randomly at extensions without data. [20:10:17] thanks guys [20:10:20] Sure, that's fine too [20:10:46] RoanKattouw: https://www.mediawiki.org/wiki/Category:APIEditBeforeSave_extensions [20:12:31] PageContentSave would be better, I guess. [20:12:51] RoanKattouw: https://www.mediawiki.org/wiki/Category:PageContentSave_extensions [20:14:49] TrevorParscal can I haz review? IT's what we talked about yesterday with dismissing OO.ui.Error in ProcessDialog teardown https://gerrit.wikimedia.org/r/#/c/194013/ [20:15:31] (03CR) 10Trevor Parscal: [C: 032] ProcessDialog: Dismiss errors on teardown [oojs/ui] - 10https://gerrit.wikimedia.org/r/194013 (owner: 10Mooeypoo) [20:16:00] wow that was quick [20:16:16] \o/ [20:22:58] (03Merged) 10jenkins-bot: ProcessDialog: Dismiss errors on teardown [oojs/ui] - 10https://gerrit.wikimedia.org/r/194013 (owner: 10Mooeypoo) [20:45:49] if ( editApi && editApi.info && editApi.info.indexOf( 'Hit AbuseFilter:' ) === 0 && editApi.warning ) { [20:45:49] this.emit( 'saveErrorAbuseFilter', editApi ); [20:45:58] (03CR) 10Kaldari: "recheck" [oojs/ui] - 10https://gerrit.wikimedia.org/r/194155 (https://phabricator.wikimedia.org/T88449) (owner: 10Kaldari) [20:46:00] RoanKattouw, yeah... [20:46:21] Well crap [20:46:27] Ain't that some fine code [20:46:34] This is not the sort of check I wanted to be adding to wikieditor for extensionAbuseFilter [20:46:58] perhaps $status->getErrorsArray()[0][0] === 'abusefilter-disallowed' [20:48:33] Great, now crossbrowsertest.com isn't working [20:49:50] RoanKattouw: It doesn't rain but it pours… [20:52:38] I got it working again [20:52:49] I'm guessing it might not like é in URLs [20:56:37] Ha. Weak. [20:56:39] James_F: 5 megs of html? Maybe use VisualEditor instead. ;-) [20:57:44] Krinkle: That crashes too. [20:58:25] RoanKattouw: Ha, reminds me. last year the iOS simulator that browserstack hosted converted 'é' to 'e' and '´'. Which is impressive, since that is actually its own character, it's not a grapheme cluster. So something is trying to be smart, splitting it and then failing to put it back to together again. [20:59:02] Status object for a ConfirmEdit error... [20:59:11] Krenair: Bad? [20:59:15] Seems to only have error message keys for 'hookaborted'. [20:59:29] So yes, responseUnknown [21:00:24] Oh. Interesting. [21:00:28] That could explain a lot. [21:00:47] what could it explain? [21:01:20] A lot of our random responseUnknown counts. Isn't ConfirmEdit quite common? [21:01:45] No, VE handles this based on the API response, which includes a captcha key [21:02:07] it becomes extensionCaptcha (who decided that?) [21:02:56] Ah. [21:03:04] Me? Roan? [21:06:28] (03PS4) 10Krinkle: ve.init.sa.Target: Add jsduck @demo for simplest editor [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/183508 [21:08:15] (03CR) 10Krinkle: "Works-ish. Though messages are still missing." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/183508 (owner: 10Krinkle) [21:08:27] Krinkle: Should we merge and do messages later? [21:08:44] Krinkle: Without messages it still gives an idea of how it looks… [21:08:50] I'd rather not. It's a show case that Target had a regression and is not usable for standalone right now. [21:09:04] It'll give a bad example. [21:09:07] But it works fine in standalone… [21:09:10] Nobody should use this code, not until/unless it works. [21:09:12] It doesn't [21:09:20] Our standalone demo is not standalone [21:09:27] Hmm. [21:09:29] it's sa.Target + a million lines of code to make it work [21:09:40] Yes, we call those lines of code "VisualEditor". [21:09:50] No I mean demos/** [21:09:51] ;-) [21:10:16] Oh, that stuff. Eh. [21:10:17] messages is just the first visible failure. the problem is larger. [21:10:25] Then mark your code as WIP. [21:10:29] Though easier to solve I think RoanKattouw had an idea for why it was failing. [21:10:31] Either it's good to merge or it isn't. [21:10:43] (03PS5) 10Krinkle: [WIP] ve.init.sa.Target: Add jsduck @demo for simplest editor [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/183508 [21:10:50] Thanks. [21:10:52] Just using it to try the new build. [21:11:39] Oh wow. That's why it's crashing [21:11:44] We're copying node_modules [21:11:52] Why on Earth are we doing that? [21:11:52] rsync --recursive --delete-after --force ./lib/ ./docs/lib/ [21:11:52] in ve-mw [21:11:55] Oh, yeah, there was a thing that used to work but now it doesn't work anymore [21:12:00] For standalone messages [21:12:00] lib/ve/* is a lot. [21:12:06] hehe [21:12:08] One of Ed's demo refactors broke it and I haven't fixed it yet [21:12:12] Aha. [21:12:16] RoanKattouw: Not a priority. :-) [21:12:35] Noooope [21:12:53] RoanKattouw: Yeah, I'm more worried about our integrity than the demo itself. Consider @example an integration test. as well as a dev UX test. I expect the patch to be in src/ not demos/ or docs/ [21:12:56] RoanKattouw: How's the Safari testing going? [21:15:15] I finally got the local proxy thing to work again [21:15:29] So now I just need to confirm that the code that I already wrote works [21:15:38] And not crash Safari's dev console this time [21:17:09] ugh, rsync is so annoying. And more over that OSX has its own version [21:17:30] OK my code works [21:17:38] $ rsync --recursive --force lib/ve/src docs/lib/ve/src -> Assuming the target directory does not exist, this will still create doc/lib/ve/src/src (two src) [21:17:49] WHY! [21:19:37] Source should have trailing slash but not dest. that's basically how to make it work. Arrrgh [21:19:37] Krinkle: Because. [21:20:05] The trailing slash symbolising you want to copy the contents. Fine. Except that's not how mv and cp work. [21:20:07] Yeah the slash-sensitivity stuff in rsync sucks [21:20:12] It's documented, but still [21:20:21] No I think cp does work that way too [21:20:24] RoanKattouw, isset( $editPage->getArticle()->getPage()->ConfirmEdit_ActivateCaptcha ) ? [21:20:26] It always confusesme [21:20:27] :D [21:20:33] Krenair: pfft [21:20:34] RoanKattouw: cp is worse in some way. It depends on whether the target exists. [21:20:39] Krenair: Nothing in the Status object that gives any indication? [21:20:39] same for mv [21:20:42] Oh right [21:20:44] J [21:20:47] I don't think so [21:21:03] Krinkle: Where did you say that splice bug was fixed? Some recent version of WebKit? [21:21:18] RoanKattouw: Somewhere between WebKit of Safari 8 and current nightly, yes. [21:22:10] Unless there is a different reason that Safari and WebKit behave differently. Which is not unimaginable considering Safari is a synchronised port of WebKit. [21:22:23] Right [21:22:26] Although jsc tends to be a straight copy to release [21:22:35] Do you happen to have access to a Safari nightly? [21:22:38] RoanKattouw, https://github.com/wikimedia/mediawiki-extensions-ConfirmEdit/blob/master/Captcha.php#L596 [21:22:44] If not, no big deal [21:22:46] RoanKattouw: yeah, it's a download from webkit.org [21:22:49] RoanKattouw: That's how I tested it [21:22:53] Oh OK [21:22:53] Krinkle: Currently nightly of what? Safari 8? Something that is going to be in Safari 8? Something that is going to be in Safari N? Something that is likely to be in Safari N? [21:22:59] Ah. [21:23:02] Safari does not have nightlies. [21:23:14] That was why I was confused. :-) [21:23:19] Although "WebKit Nightly for Mac" is essentially the Safari build. [21:23:29] It's even titled Safari in the menu bar [21:23:34] HA. [21:23:40] Not like Chromium/Chrome. [21:23:47] Rught. [21:23:49] i [21:23:51] To make a clean build of pure WebKit you'd have to do some rigging. [21:24:43] Anyway, Safari tends to have some propietary stuff in WebCore (the layout engine in WebKit, like Blink is in Chromium). But JSC tends to be a straight copy. [21:24:51] So if it's fixed in nightly, I'm pretty sure it'll end up in Safari as-is. [21:24:52] Krinkle: I'll just say "fixed in WebKit nightly" [21:24:56] Yeah [21:25:03] It's just a commit summary, it doesn't have to be researched to death :) [21:27:58] (03PS1) 10Jforrester: VisualEditor.php: Warn people to use extension.json instead [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194192 [21:29:07] (03CR) 10Legoktm: [C: 032] VisualEditor.php: Warn people to use extension.json instead [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194192 (owner: 10Jforrester) [21:29:33] :-( https://doc.wikimedia.org/VisualEditor/master/lib/ve/.git [21:29:55] "gitdir: /srv/ssd/jenkins-slave/workspace/mwext-VisualEditor-jsduck-publish/.git/modules/lib/ve " [21:29:58] Lovely. [21:30:02] (03CR) 10jenkins-bot: [V: 04-1] VisualEditor.php: Warn people to use extension.json instead [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194192 (owner: 10Jforrester) [21:30:16] Ha. [21:30:19] legoktm: Your comment sucks. [21:30:20] Ha lucky, yeah. so because it's a submodule and git doens't do symlinks, it's not the entire repo [21:30:27] wat [21:30:42] lol [21:30:48] legoktm: 104 characters long > 100 character length hard limit > 80 character length guideline. [21:30:59] legoktm: Coding Standards. They're a thing. :-P [21:31:12] wider monitors, they're a thing! [21:31:42] (03PS11) 10Alex Monk: [WIP] Schema:Edit instrumentation [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) [21:31:54] legoktm: :-) [21:32:07] (03CR) 10jenkins-bot: [V: 04-1] [WIP] Schema:Edit instrumentation [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) (owner: 10Alex Monk) [21:32:20] (03PS2) 10Jforrester: VisualEditor.php: Warn people to use extension.json instead [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194192 [21:32:35] (03PS1) 10Catrope: Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194194 (https://phabricator.wikimedia.org/T76998) [21:33:44] RoanKattouw: I take it this doesn't magically fix https://phabricator.wikimedia.org/T76715 too? [21:34:13] No [21:34:16] That's a UI issues [21:34:18] *issue [21:34:27] I will look into that first thing tomorrow [21:34:36] Right now I'm going to sleep [21:35:33] RoanKattouw: Excuses. [21:35:57] (03CR) 10Legoktm: [C: 032] VisualEditor.php: Warn people to use extension.json instead [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194192 (owner: 10Jforrester) [21:36:14] RoanKattouw, nuria, milimetric: https://meta.wikimedia.org/w/index.php?title=Schema:Edit&diff=11448361&oldid=11319708 [21:36:21] does that look ok? [21:38:18] (03Merged) 10jenkins-bot: VisualEditor.php: Warn people to use extension.json instead [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194192 (owner: 10Jforrester) [21:40:23] (03PS1) 10Bartosz Dziewoński: TextInputWidget: Use MutationObserver for #onElementAttach support [oojs/ui] - 10https://gerrit.wikimedia.org/r/194197 (https://phabricator.wikimedia.org/T64084) [21:41:17] Krenair: Looks sensible to me [21:41:32] Krenair: Did we want to add anything to the enum whilst we're at it? [21:41:38] (cc whomever) [21:41:41] pow. [21:41:47] MatmaRex: Nice. [21:41:50] TrevorParscal: Hey, we should chat. [21:42:07] i should probably try it out with VE. [21:42:09] James_F: ok, pm [21:42:11] MatmaRex: Probably. [21:45:55] (it does work.) [21:46:12] MatmaRex: Nice. [21:46:21] magic [21:47:25] (03PS2) 10Bartosz Dziewoński: Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194194 (https://phabricator.wikimedia.org/T76998) (owner: 10Catrope) [21:48:24] (03CR) 10Bartosz Dziewoński: Work around Array.prototype.splice being broken in Safari (031 comment) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194194 (https://phabricator.wikimedia.org/T76998) (owner: 10Catrope) [21:49:52] (03CR) 10Catrope: Work around Array.prototype.splice being broken in Safari (031 comment) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194194 (https://phabricator.wikimedia.org/T76998) (owner: 10Catrope) [21:51:10] RoanKattouw: Won't surprise you but Safari 7 is also affected. [21:51:25] I misreported that it was broken for me in Safari 8, I'm running 7. 8 requires the OSX 10.10 which I don't have yet. [21:51:34] (03PS3) 10Catrope: Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194194 (https://phabricator.wikimedia.org/T76998) [21:51:55] I went through a lot of crossbrowsertesting pain to test in Safari 8 so I didn't test any other versions [21:52:03] Yeah. [21:52:05] But yeah I'm assuming all versions < 8 are also affected [21:52:17] RoanKattouw: Hm.. anything particular that made it hard? [21:52:29] Did you use in-browser or vnc? [21:52:47] In-browser, so slow [21:52:57] Also typing in the console sometimes stops working and you have to click out and back in [21:53:04] Also launching sessions with a URL containing é doesn't work [21:53:27] Also $wgServer = "//127.0.0.1" doesn't work, you have to use "/local" which of course isn't set up to work for me locally so it breaks my wiki in my normal browsers [21:53:30] Hm.. I only use crossbrowsertesting when I want to use VNC (e.g. to use keyboard commands, or zooming) [21:53:51] Also the Safari console hangs when dumping large sparse arrays (and reproducing the bug required array sizes over 100k) [21:53:53] For anything else I tent to use saucelabs and browserstack, in-browser. Quite fast. And quick tunneling setup so you don't need to change any wiki settings [21:54:10] Also there appears to be no key that maps to the command key, so you can't do Cmd+F or Cmd+P or such useful things [21:54:18] browserstack now also has a chrome extension to require even less setup [21:54:29] Ah, yeah. For that you'd need to do VNC. [21:54:32] where is wikibugs [21:54:35] legoktm? [21:54:38] which is only supported by crossbrowsertesting [21:55:15] Krenair: it's having issues. [21:55:19] RoanKattouw: Hm.. Can you check how this relates to es5-shim? [21:55:26] Ahm, wat [21:55:29] Why es5-shim? [21:55:35] Array.prototype.splice is ES3 [21:55:37] RoanKattouw: Well, they have extensive testing. [21:55:43] To correct browser bugs. [21:55:46] Which apparently doesn't prevent TWO browsers from implementing it wrong [21:55:47] Not just based on if(obj.method) [21:56:04] did the window fix from a couple of days ago in ooui got released to mw ooui ? [21:56:08] Well, it's also part of es5 naturally. [21:56:19] I'm curious if loading es5-shim fixes it [21:56:23] Krinkle: I doubt they have a test for this or for the Opera bug, they're both quite obscure [21:56:32] Could you look into that maybe? [21:56:36] RoanKattouw: I'd be surprised if they don't have it. [21:56:39] But yeah, I'll check [21:56:41] You know that project and those people, and you're presumably not about to fall asleep [21:56:51] The Safari one requires a sparse array of over 100k elements [21:56:54] hehe [21:57:16] RoanKattouw: The test you have there is good right? [21:57:55] Krinkle: they don't have fixes for this, i checked some time ago [21:58:55] (03PS12) 10Alex Monk: [WIP] Schema:Edit instrumentation [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) [21:59:53] yeah, it's broken even with es5-shim loaded [21:59:55] I'll file upstream though [22:00:09] but yeah, it'd be nice not to load the es5-shim in Safari 8 [22:00:15] considering it's supposed to support ES5 [22:01:37] Krinkle: Yeah both the Safari and Opera tests in ve.utils.js expose the respective issues [22:02:31] For Safari I know the parameters pretty well (array has to be sparse, has to have length >=100k, the 'x' has to be at offset >= 8), for Opera I don't know them at all other than that that specific example fails [22:04:54] Hi. I am encountering a problem which also seems to be described at this link: https://www.mediawiki.org/wiki/Extension_talk:VisualEditor#Progress_bar_keeps_on_forever_when_.22Edit.22_54260. A few days ago, we had VisualEditor working. I did a database refresh without changing other parameters (I did perform the database update) and now the progress bar at the upper right hand corner hangs indefinitely. I hav [22:05:02] we had discussions about this back then, if you have logs :) [22:05:26] Pickle-san: You got cut off after "indefinitely. I hav" [22:05:49] RoanKattouw: OK. Can I cite you for that on the Github issue? [22:05:51] "For Safari I know the parameters pretty well (array has to be sparse, has to have length >=100k, the 'x' has to be at offset >= 8), for Opera I don't know them at all other than that that specific example fails" [22:05:51] Krenair: re: the change adding saveFailure.message: that's ok with me, check with James and Roan if they want anything else in the schema in the next week or so, and otherwise go ahead and use your new version. [22:06:15] Krinkle: Yes [22:06:35] RoanKattouw: Sorry. Parsoid log indicates the page is getting parsed. Common.js is blanked out. [22:06:51] milimetric, Krenair: An additional enum value for TitleBlacklist [22:06:57] Or whichever one the missing extension is [22:07:11] RoanKattouw: splicing into a sparse array with >=256 elements after the 256th element removes things that were afterwards [22:07:14] James listed four of them and we only have three in the enum [22:07:19] MatmaRex: If you merge, I'll reference the commit on the upstream report :) [22:07:45] (03CR) 10Bartosz Dziewoński: [C: 032] Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194194 (https://phabricator.wikimedia.org/T76998) (owner: 10Catrope) [22:08:01] MatmaRex: You mean splicing into a spare array whose length is >=256 at index > 256 removes all pre-existing array elements whose index is >= 256? [22:08:02] RoanKattouw / Krenair: yes, it was TitleBlacklist [22:08:43] James_F: https://gerrit.wikimedia.org/r/194194 is ready for your cherry-picking pleasure, MatmaRex +2ed it [22:08:54] You guys are on your own for any SWATting though, I am going to sleep [22:09:07] RoanKattouw: it seems more complicated than that [22:09:43] (03Merged) 10jenkins-bot: Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194194 (https://phabricator.wikimedia.org/T76998) (owner: 10Catrope) [22:09:45] RoanKattouw: Awesome, thanks. [22:09:48] RoanKattouw, milimetric: So the wikitext editor won't even get to that error [22:09:52] Yeah, I've never noticed a sensible pattern other than "a[256]" disappears if you splice right after it [22:10:05] (titleblacklist) [22:10:10] Krenair: sure, but it'd be useful for Roan [22:10:20] and it's best if we don't have *too* many schemas [22:10:30] Krenair: It's probably possible in theory though? [22:10:38] because you can't get a creation form for a blacklisted title [22:10:49] let alone attempt to save [22:10:58] sure, maybe useful for ve [22:10:58] Or is there something else that stops me from submitting stuff to ?title=F-you&action=edit ? [22:11:01] Krenair: is that an internal error message or a message given to the user? just list what type of message it is [22:11:02] OK, right [22:11:03] (03PS1) 10Jforrester: Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194204 [22:11:07] Well also maybe VE should detect this early [22:11:13] Rather than when you attempt to create [22:11:23] Krenair: What happens when the title is blacklisted while you're already editing? [22:11:30] I mean something must be happening somewhere [22:11:37] But maybe not in that code path [22:11:43] nuria, given to the user [22:12:26] nuria: This is an error like "You are not allowed to create a page whose name contains the word 'f***'" [22:12:55] (03PS1) 10Jforrester: Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194205 [22:13:03] (03CR) 10Jforrester: [C: 032] Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194205 (owner: 10Jforrester) [22:13:05] (03CR) 10jenkins-bot: [V: 04-1] Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194204 (owner: 10Jforrester) [22:13:21] RoanKattouw, actually in this case VE displays two notices [22:13:25] (03CR) 10Jforrester: [C: 032] Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194204 (owner: 10Jforrester) [22:13:28] Argh. [22:13:32] RoanKattouw: Ok, it's an error message for the user to see then [22:13:32] Krinkle! [22:13:35] * James_F sighs. [22:13:59] James_F: backport my patch :P [22:14:11] Let me draft a simplified version [22:14:11] Krinkle: You do it. [22:14:15] OK. [22:14:21] * James_F goes for meeting. [22:14:28] "Warning: This page can only be edited by administrators and other users with the tboverride right because it matches the following title blacklist entry: HookTest" [22:15:00] and "You do not have permission to create pages, for the following reason: Mediawiki_dev has restricted the ability to create new pages. [22:15:00] You can go back and edit an existing page, or log in or create an account." (this does not appear when trying to create non-blacklisted pages) [22:15:00] (03CR) 10jenkins-bot: [V: 04-1] Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194205 (owner: 10Jforrester) [22:15:25] anyway I'll add it to the schema RoanKattouw [22:15:28] (03CR) 10jenkins-bot: [V: 04-1] Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194204 (owner: 10Jforrester) [22:15:28] Krinkle, James_F|Away: Feel free to ignore Jenkins (especially jsduck) in wmf branches [22:16:21] https://github.com/es-shims/es5-shim/issues/295 [22:17:01] RoanKattouw: I've been out all morning/afternoon PST. Anything you need from me before you head off for the night? [22:17:06] Hm.. wmf branch in VisualEditor core? [22:17:10] Ah, right. [22:17:16] not ve-mw [22:20:26] (03PS1) 10Krinkle: build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194211 [22:20:28] (03CR) 10jenkins-bot: [V: 04-1] build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194211 (owner: 10Krinkle) [22:20:32] ori: For the next week or so I am pivoting to stability issues instead of performance issues. Trevor and James may have told you [22:20:43] yes [22:21:05] If you are doing that too, then see if you can help mvolz and Alexandros et al with Citoid/Zotero things [22:21:21] May require async communication and waiting till tomorrow because it's pretty late here in Europe [22:21:38] (03PS2) 10Krinkle: build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194211 [22:21:42] (03CR) 10jenkins-bot: [V: 04-1] build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194211 (owner: 10Krinkle) [22:21:54] RoanKattouw: yeah, I chatted with Alexandros yesterday. He has a patch staged and I'm going to be reviewing it. [22:21:54] Intriguing. My VE session was hung for an hour. I hit F5, it hung for a moment, then it finally entered into edit mode. [22:22:02] what? How can it reply that fast. It's too fast. The patch isn't available yet. [22:22:02] If you do not have an immediate way to be useful there, then another thing you could work on is using iteration instead of hashing in IndexValueStore [22:22:04] mvolz: Anything I can do to help you with Citoid/Zotero stuff? [22:22:22] We talked about that earlier, and maybe I'm blind and you already wrote a patch that I didn't see, but if not, you could write one [22:22:34] Ha, .gitreview is wrong [22:23:08] RoanKattouw: I wrote one last week, but it's not faster. I know what I have to do with it, though, so I'm good. [22:23:16] Such a patch would also more or less obviate your GeneratedContentNode WIP patches, or at least the perf implications thereof (weaning GCN off of IVStore is still a good idea for other reasons) [22:23:17] (03Abandoned) 10Krinkle: build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194211 (owner: 10Krinkle) [22:23:39] ori: OK, I just wanted to make sure that didn't get left by the wayside [22:24:03] (03PS1) 10Krinkle: build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194212 [22:24:04] Yep. [22:24:34] I was going to ask Krinkle to take on some of the perf tasks I had in my sight [22:25:00] T89928 and T91248 are straightforward browser detection bugs, but for those we should really bring jquery.client back into VE core [22:25:12] That seems like something that would be up Krinkle's alley [22:25:44] File a bug or assign to me. I'm gonna forget. [22:25:48] Sounds interesting. [22:25:56] I'll file a bug that blocks those two [22:26:20] Basically I need jquery.client back in VE core (it was removed at some point because it was unused) so I can put in two trivial browser sniffing things [22:27:01] s/so I can/so someone (maybe Ori) can/ [22:27:07] (03Abandoned) 10Mooeypoo: Reject promises with OOUI error handling [extensions/Citoid] - 10https://gerrit.wikimedia.org/r/191123 (owner: 10Mooeypoo) [22:28:01] RoanKattouw: relevant: https://phabricator.wikimedia.org/T89928 [22:28:43] ori: As I said, [23:24] RoanKattouw T89928 and T91248 are straightforward browser detection bugs, but for those we should really bring jquery.client back into VE core [22:28:45] :) [22:28:56] oh right [22:29:54] Ugh, I can't add "tasks X and Y block Z" to Z, I have to edit X and Y separately to say Z blocks them [22:29:58] Thanks Phabricator [22:30:32] Krinkle: https://phabricator.wikimedia.org/T91460 [22:34:51] ori: If you see David around later today (he's not always on IRC, maybe ping him on Hangouts in a few hours when it's not stupid early in HK), ask him to fix up https://gerrit.wikimedia.org/r/#/c/193787/ [22:34:55] (03CR) 10Jforrester: [C: 032] build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194212 (owner: 10Krinkle) [22:35:04] (03CR) 10Jforrester: "Needs picking to wmf18 too." [VisualEditor/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194212 (owner: 10Krinkle) [22:35:08] i.e. find a way to do what I intended there, which is to prevent setupSlugs from being called twice for every node during initt [22:35:33] RoanKattouw: will do [22:35:39] OK [22:35:45] That's all I have for now [22:35:48] RoanKattouw, the phabricator issue you ran into is https://phabricator.wikimedia.org/T33 [22:36:19] Those were all the perf bugs that were on my mind, apart from https://gerrit.wikimedia.org/r/#/c/193026/ needing more work but I'll talk to Krinkle about that one tomorrow) [22:37:13] ori: Oh and you might derive some entertainment value from what I worked on today: https://gerrit.wikimedia.org/r/194194 [22:37:17] (03PS3) 10Trevor Parscal: Make annotation inspector getMode method trim [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/193971 [22:38:38] (03Merged) 10jenkins-bot: build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194212 (owner: 10Krinkle) [22:39:00] (03PS2) 10Jforrester: Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194205 [22:39:12] (03CR) 10Jforrester: [C: 032] Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194205 (owner: 10Jforrester) [22:39:15] James_F|Away: Ugh, it doenst apply cleanly to wmf18? [22:39:16] weird [22:40:45] (03PS1) 10Krinkle: build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194215 [22:40:47] (03CR) 10jenkins-bot: [V: 04-1] build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194215 (owner: 10Krinkle) [22:40:48] Ah --external changed [22:41:09] (03Abandoned) 10Krinkle: build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194215 (owner: 10Krinkle) [22:41:11] (03Merged) 10jenkins-bot: Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194205 (owner: 10Jforrester) [22:41:49] (03CR) 10Catrope: [C: 032] Make annotation inspector getMode method trim [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/193971 (owner: 10Trevor Parscal) [22:41:54] (03PS1) 10Krinkle: build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194216 [22:42:10] (03CR) 10Krinkle: [C: 032] "Backport." [VisualEditor/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194216 (owner: 10Krinkle) [22:45:15] (03Merged) 10jenkins-bot: Make annotation inspector getMode method trim [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/193971 (owner: 10Trevor Parscal) [22:45:17] (03Merged) 10jenkins-bot: build: Rename JSDuck config standard jsduck.json [VisualEditor/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194216 (owner: 10Krinkle) [22:45:20] edsanders: you 'round? [22:47:02] i'm trying to get to the bottom of https://phabricator.wikimedia.org/T91307 and it appears that https://gerrit.wikimedia.org/r/#/c/193077/ and https://gerrit.wikimedia.org/r/#/c/192738/ don't get on [22:47:07] could use some help [22:48:31] what's going on, is when going from "closed -> inspector" in DesktopContext, because this.lastClosedSelection is empty (getRanges() -> []) it takes the "don't reposition" route [22:52:43] hmmm... this is only a problem in mw... [22:57:36] 10VisualEditor: Polish: Does Page title make sense in formatter control? - https://phabricator.wikimedia.org/T91462#1084823 (10Jaredzimmerman-WMF) 3NEW [22:58:44] (03PS2) 10Krinkle: Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194204 (owner: 10Jforrester) [23:00:43] 10MediaWiki-ResourceLoader, 10MediaWiki-Vagrant, 5Patch-For-Review, 7Performance: Possible performance degradation in lessc - Local vagrant super slow - https://phabricator.wikimedia.org/T78733#1084876 (10bd808) We looked at this for a bit today and may be narrowing in on some parts of the problem. Resouce... [23:01:41] 10MediaWiki-ResourceLoader, 10MediaWiki-Vagrant, 5Patch-For-Review: ResourceLoader slow under MediaWiki-Vagrant with VirtualBox - https://phabricator.wikimedia.org/T78733#1084903 (10bd808) [23:02:01] 10MediaWiki-ResourceLoader, 10MediaWiki-Vagrant, 5Patch-For-Review: ResourceLoader slow under MediaWiki-Vagrant with VirtualBox - https://phabricator.wikimedia.org/T78733#852085 (10bd808) a:3bd808 [23:02:42] 10VisualEditor, 10VisualEditor-MediaWiki-References: [Regression wmf13] Creating a citation without filling in a required field throws "Cannot read property 'createElement' of null" - https://phabricator.wikimedia.org/T78626#1084934 (10Etonkovidova) Uncaught TypeError: Cannot read property 'elements' of nullV... [23:04:32] 10MediaWiki-ResourceLoader, 10MediaWiki-Vagrant, 5Patch-For-Review: ResourceLoader slow under MediaWiki-Vagrant with VirtualBox - https://phabricator.wikimedia.org/T78733#1084956 (10Jdlrobson) Magnitudes faster with that patch 12 seconds -> 4 seconds. Still room for improvement though. cc @rmoen @jhernandez... [23:05:02] (03CR) 10Jforrester: [C: 032] Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194204 (owner: 10Jforrester) [23:06:09] (03PS7) 10Bartosz Dziewoński: tests: Add infusion tests [oojs/ui] - 10https://gerrit.wikimedia.org/r/193123 [23:06:31] Bleh. [23:07:12] * MatmaRex looms [23:07:30] (03Merged) 10jenkins-bot: Work around Array.prototype.splice being broken in Safari [VisualEditor/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194204 (owner: 10Jforrester) [23:08:51] (03CR) 10jenkins-bot: [V: 04-1] tests: Add infusion tests [oojs/ui] - 10https://gerrit.wikimedia.org/r/193123 (owner: 10Bartosz Dziewoński) [23:11:11] (03PS1) 10Jforrester: Update VE core submodule for cherry-pick (81d12a6) [extensions/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194219 [23:11:24] (03PS1) 10Jforrester: Update VE core submodule for cherry-pick (26845f2) [extensions/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194220 [23:11:48] Krenair: Could you get ^^^ those two SWATted at 16:00 (midnight)? Sorry to ask, but I have a meeting then. :-( [23:13:00] James_F, Krinkle: Why the jsduck config change? [23:13:20] Krenair: Because Krinkle made a breaking change to CI for jsduck. [23:13:28] ok [23:13:33] Krenair: If we wanted it to ever merge under CI we needed to change the file name. [23:14:06] I could have maintained a separate pipeline and blacklisted wmf<20 from the main one and whitelist the other for >20 but this was easier. [23:14:26] TrevorParscal, hey [23:14:32] hey [23:14:35] Another reason for moving more state like these into the repo instead of in Jenkins where they change at once. [23:15:45] still stuck on the context thing? [23:17:23] (03CR) 10Alex Monk: [C: 032] Update VE core submodule for cherry-pick (81d12a6) [extensions/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194219 (owner: 10Jforrester) [23:17:28] (03CR) 10Alex Monk: [C: 032] Update VE core submodule for cherry-pick (26845f2) [extensions/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194220 (owner: 10Jforrester) [23:17:31] Krenair: You rock. [23:17:38] :) [23:19:56] (03Merged) 10jenkins-bot: Update VE core submodule for cherry-pick (81d12a6) [extensions/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194219 (owner: 10Jforrester) [23:19:58] (03Merged) 10jenkins-bot: Update VE core submodule for cherry-pick (26845f2) [extensions/VisualEditor] (wmf/1.25wmf18) - 10https://gerrit.wikimedia.org/r/194220 (owner: 10Jforrester) [23:22:28] Hi. I had a couple questions about TextInputWidget. Anyone have a moment to help answer them? If not, I'll just commit what I have so far and go back to the confusing parts later :) [23:23:01] hi kirstenMA [23:23:25] James_F: copying/pasting as well as undo/re-paste very quickly scrolls me back up (exposing the logo on top left) and then back to where it was. I've got a video capture and can file a bug if needed. [23:23:38] Krinkle: Already filed. [23:23:50] Krinkle: If you have a fix that'd be awesome, of course. ;-) [23:24:39] James_F: https://phabricator.wikimedia.org/T74390? [23:25:20] My text didn't contain a tab and it does scroll back down after a split seond [23:25:29] just fyi [23:25:40] Krinkle: Yeah. [23:25:56] Hmm. [23:25:57] http://cl.ly/0V1c2B2N0v0C [23:25:58] * James_F looks further. [23:26:02] hi kirstenMA, what's up? [23:26:57] I had a couple questions about textinput methods. [23:27:06] 10VisualEditor, 10VisualEditor-DataModel, 7Browser-Support-Apple-Safari, 5WMF-deploy-2015-02-18_(1.25wmf18), and 2 others: Categories incorrectly moved up, possibly due to improperly closed table - https://phabricator.wikimedia.org/T72373#1084981 (10Jdforrester-WMF) 5Open>3Resolved a:3Catrope [23:27:07] 10VisualEditor, 10VisualEditor-EditingTools, 10VisualEditor-MediaWiki-Plugins, 10VisualEditor-Tables, 7Tracking: Support tables (tracking) - https://phabricator.wikimedia.org/T41596#1084984 (10Jdforrester-WMF) [23:27:45] * MatmaRex listens [23:27:55] 10VisualEditor, 10VisualEditor-DataModel, 7Browser-Support-Apple-Safari, 5WMF-deploy-2015-02-18_(1.25wmf18), and 2 others: Categories incorrectly moved up, possibly due to improperly closed table - https://phabricator.wikimedia.org/T72373#751578 (10Jdforrester-WMF) I'm provisionally declaring that this was... [23:27:58] thanks, MatmaRex [23:28:09] 10VisualEditor, 10VisualEditor-ContentEditable, 10VisualEditor-ContentLanguage: After selecting table cell by single clicking, typing with any IME does not work - https://phabricator.wikimedia.org/T86589#1084989 (10Inez) This is a quick and dirty solution: http://pastebin.com/PQr3uZ5S. At this moment it copi... [23:28:17] I am confused about the difference between setLabelPosition and positionLabel [23:28:53] And also a note that appears under setReadOnly, which says 'This should probably change the widget's appearance and prevent it from being used.' [23:30:08] 10VisualEditor: Use consistent icons for same/similar functions between editors - https://phabricator.wikimedia.org/T91467#1085008 (10Jaredzimmerman-WMF) 3NEW [23:30:38] I'm not sure who that instruction is for/if its a call to action. [23:30:41] yeah. hmm [23:31:25] as for setReadOnly(), that was probably intended for people implementing hypothetical subclasses; there's a similar comment on OO.ui.Widget.prototype.setDisabled [23:31:27] 10VisualEditor: Polish: Does Page title make sense in formatter control? - https://phabricator.wikimedia.org/T91462#1085016 (10Jdforrester-WMF) p:5Triage>3Low It's not a negotiable functionality (unless you want to argue we should remove it from MediaWiki entirely in which case, this is not the right venue).... [23:31:29] I'm also a little confused by select() method, because it is not an option? But maybe this is for another purpose? [23:31:43] in fact, that's the exact same comment, so that was probably copy-pasted… [23:31:55] since there are no subclasses of TextInputWidget, that's not very useful there [23:33:06] ah, okay [23:33:17] sounds like maybe we can remove it from here. [23:34:36] yeah, probably [23:34:37] setLabelPosition() is a setter like setLabel() or setIcon() or such; it calls updatePosition() can also be called on its own (in case something caused the label to be mispositioned? not sure); and that calls positionLabel(), which is private method actually implementing the positioning calculations [23:34:50] updatePosition() which can* [23:34:58] 10VisualEditor: Polish: Does Page title make sense in formatter control? - https://phabricator.wikimedia.org/T91462#1085029 (10Jaredzimmerman-WMF) {F54376} we already have a pattern for lesser used items, elsewhere, perhaps all the non-heading related styles belong "below the fold" It would be great if we had us... [23:35:00] 10VisualEditor: Put jquery.client back in VE core - https://phabricator.wikimedia.org/T91460#1085030 (10Jdforrester-WMF) p:5Triage>3Normal [23:35:10] 10VisualEditor: Put jquery.client back in VE core - https://phabricator.wikimedia.org/T91460#1084669 (10Jdforrester-WMF) p:5Normal>3High [23:35:30] great. thank you. that clears up a lot. [23:36:00] Oh my [23:36:09] and select() selects the entire text of the textarea, this is just a thin wrapper over the jQuery method [23:36:13] I've got myself an XHR- loop on mw.org just now when saving an edit [23:36:19] It keeps making more post requests [23:36:25] (which itself is a thin wrapper over native method) [23:36:28] badtoken [23:36:35] Infinite loop [23:36:44] 10VisualEditor, 10VisualEditor-EditingTools, 10VisualEditor-MediaWiki-References: Context menu does not update the change immediately unless you re-open it - https://phabricator.wikimedia.org/T91336#1085053 (10Jdforrester-WMF) p:5Triage>3Normal a:3TrevorParscal [23:37:30] Did I take too long to edit? [23:37:39] Refreshing shows my session is still in tact [23:37:48] at least our refrsh token system is broken [23:38:07] 10VisualEditor, 10VisualEditor-EditingTools: There should be help links in every context - dialog boxes, inspectors, etc. - https://phabricator.wikimedia.org/T53798#1085059 (10Jdforrester-WMF) a:5TrevorParscal>3None [23:38:08] 10VisualEditor, 10VisualEditor-EditingTools: VisualEditor: Link inspector inside dialog is clipped against dialog boundary, which makes it too small - https://phabricator.wikimedia.org/T74121#1085060 (10Jdforrester-WMF) a:5TrevorParscal>3None [23:38:09] 10VisualEditor, 10VisualEditor-EditingTools: VisualEditor: Empty transclusion editor is empty, needs some kind of reassuring message - https://phabricator.wikimedia.org/T52281#1085061 (10Jdforrester-WMF) a:5TrevorParscal>3None [23:38:11] 10VisualEditor, 10VisualEditor-EditingTools, 7Browser-Support-Firefox: VisualEditor: In Firefox, the context menu's background sometimes does not appear for items inside a media caption (but the icons appear still) - https://phabricator.wikimedia.org/T66332#1085062 (10Jdforrester-WMF) a:5TrevorParscal>3No... [23:38:47] (03PS1) 10Esanders: Only store linear selections as last context position [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194227 (https://phabricator.wikimedia.org/T91307) [23:39:23] 10VisualEditor: Polish: Does Page title make sense in formatter control? - https://phabricator.wikimedia.org/T91462#1085070 (10Jdforrester-WMF) >>! In T91462#1085029, @Jaredzimmerman-WMF wrote: > {F54376} > we already have a pattern for lesser used items, elsewhere, perhaps all the non-heading related styles bel... [23:39:31] edsanders: Nice. [23:39:46] 10VisualEditor, 10VisualEditor-ContentEditable, 10VisualEditor-EditingTools, 5Patch-For-Review: [Regression pre-wmf20] Link inspector opening in wrong position when selection is at the start of a line - https://phabricator.wikimedia.org/T91307#1085071 (10Jdforrester-WMF) a:5TrevorParscal>3Esanders [23:39:55] TrevorParscal: Did edsanders steal your thunder? :-) [23:40:07] he's helping me [23:40:13] Ah. [23:40:14] Cool. [23:40:21] TrevorParscal: Also, https://phabricator.wikimedia.org/maniphest/query/Bl8AwbZzeyEh/#R [23:41:26] TrevorParscal: James_F: hey you guys, what's the status/priority of the toolbar tweaks we talked about? [23:41:29] 10VisualEditor, 10VisualEditor-MediaWiki: New internal links on /w/index.php?… pages are created relative to /w/ not /wiki/ in the CE and so fail to go anywhere - https://phabricator.wikimedia.org/T85622#1085074 (10Jdforrester-WMF) [23:41:31] 10VisualEditor, 10VisualEditor-Links, 5§ VisualEditor Q3 Blockers: Clicking on the link from the new preview interface link dialogue opens a "404-page not found"page instead of opening the article - https://phabricator.wikimedia.org/T91291#1085073 (10Jdforrester-WMF) [23:41:34] MatmaRex: Which ones? [23:41:39] (should i poke at them more or do something more useful for now?) [23:41:49] https://gerrit.wikimedia.org/r/193484 and https://gerrit.wikimedia.org/r/193282 and such [23:41:52] MatmaRex: Right now I can think of 10+ toolbar changes that are proposed/doing/partially done. [23:42:14] https://gerrit.wikimedia.org/r/#/c/193484/ is to go in next week's OOjs UI release (0.9.0) [23:42:25] James_F: and https://phabricator.wikimedia.org/T52227#1079767 [23:42:27] https://gerrit.wikimedia.org/r/#/c/193282/ possibly too? I dunno. [23:42:56] And understood. [23:43:16] (that one needs some more work) [23:43:23] 10VisualEditor: Search field in Category dialog should be above currently added labels rather than below. - https://phabricator.wikimedia.org/T91469#1085078 (10Jaredzimmerman-WMF) 3NEW [23:44:11] (03PS2) 10Trevor Parscal: Only store linear selections as last context position [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194227 (https://phabricator.wikimedia.org/T91307) (owner: 10Esanders) [23:45:50] i need to go to the post office with my family to submit our passports [23:45:54] running out of time on that [23:46:24] (03CR) 10jenkins-bot: [V: 04-1] Only store linear selections as last context position [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194227 (https://phabricator.wikimedia.org/T91307) (owner: 10Esanders) [23:46:38] 10VisualEditor: Search field in Category dialog should be above currently added labels rather than below. - https://phabricator.wikimedia.org/T91469#1085090 (10Jdforrester-WMF) Really? Why? This adds one to the end. It's in exactly the place it then gets "created". I agree that it should be in a better place, bu... [23:46:44] (03PS3) 10Trevor Parscal: Only store linear selections as last context position [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194227 (https://phabricator.wikimedia.org/T91307) (owner: 10Esanders) [23:47:00] 10VisualEditor: Polish: Does Page title make sense in formatter control? - https://phabricator.wikimedia.org/T91462#1085095 (10Jaredzimmerman-WMF) I'm not sure that users would think of these very differently, and reusing existing patterns will surely help with user understanding. [23:47:26] (03PS4) 10Trevor Parscal: Only store linear selections as last context position [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194227 (https://phabricator.wikimedia.org/T91307) (owner: 10Esanders) [23:47:33] 10TemplateData, 7Wikimedia-log-errors: Invalid foreach and undefined property in TemplateData - https://phabricator.wikimedia.org/T91470#1085096 (10Chad) 3NEW [23:47:52] 10VisualEditor, 10VisualEditor-Links, 5§ VisualEditor Q3 Blockers: New internal links on /w/index.php?… pages are created relative to /w/ not /wiki/ in the context menu and so fail to go anywhere useful - https://phabricator.wikimedia.org/T91291#1085103 (10Jdforrester-WMF) [23:48:22] 10TemplateData, 7Wikimedia-log-errors: Invalid foreach and undefined property in TemplateData - https://phabricator.wikimedia.org/T91470#1085107 (10Jdforrester-WMF) p:5Triage>3High a:3Mooeypoo [23:49:11] edsanders: If you fixed https://phabricator.wikimedia.org/T85622 it'd probably fix https://phabricator.wikimedia.org/T91291 for Trevor if you have 20 seconds to look… [23:51:23] Hehe, we're doing API queries in the link inspector for queries starting with a # [23:51:43] it actually yields results. Interesting [23:54:15] 10VisualEditor: Search field in Category dialog should be above currently added labels rather than below. - https://phabricator.wikimedia.org/T91469#1085119 (10Jaredzimmerman-WMF) I don't understand how search is the wrong metaphor here, you're searching through available categories to add, the fact that they ar... [23:59:19] (03PS5) 10Esanders: Only store selections which yield a valid context position [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194227 (https://phabricator.wikimedia.org/T91307)