[00:00:21] (03PS2) 10Mooeypoo: Add optional text to deprecated property [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/194748 (https://phabricator.wikimedia.org/T90734) [00:00:37] 10MediaWiki-Page-editing: Implement (part of) Wikipedia's edit notice system in MediaWiki core - https://phabricator.wikimedia.org/T91715#1094284 (10Krinkle) 3NEW [00:01:00] James_F: actually, this might work ^ [00:01:58] (03Merged) 10jenkins-bot: Follow-up Id9e657f12: Don't add moreTextStyle to the same group as it breaks downstream [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194757 (owner: 10Jforrester) [00:02:04] (03PS1) 10Jforrester: build: Update jscs to latest [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194760 [00:03:26] Krinkle: Would the class hiding work for our needs? Wouldn't it retain the $.getVisibleText issue? [00:04:12] (03CR) 10Jforrester: [C: 031] "Product +1." [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/194748 (https://phabricator.wikimedia.org/T90734) (owner: 10Mooeypoo) [00:05:48] Krinkle: "Wikipedia could continue to hide this message for non-sysops using the same class hiding.” I mean. [00:08:43] 10MediaWiki-Page-editing: Implement (part of) the English Wikipedia's edit notice system in MediaWiki core - https://phabricator.wikimedia.org/T91715#1094315 (10Krenair) [00:09:24] 10VisualEditor, 10VisualEditor-MediaWiki, 5§ VisualEditor Q3 Blockers: VisualEditor got into endless visualeditoredit API loop when saving the page - https://phabricator.wikimedia.org/T91158#1094317 (10matmarex) a:3matmarex Easy to reproduce, at least. Just make the edit token that VE stores invalid (`ve.i... [00:09:43] 10VisualEditor, 10VisualEditor-MediaWiki: VisualEditor: Button to initiate UploadWizard and insert image on completion - https://phabricator.wikimedia.org/T40030#1094328 (10Jdforrester-WMF) [00:09:52] 10VisualEditor, 10VisualEditor-CopyPaste: VisualEditor: Upload-by-insertion of images via copy-paste - https://phabricator.wikimedia.org/T39932#1094330 (10Jdforrester-WMF) [00:14:15] 10VisualEditor, 10VisualEditor-MediaWiki, 5§ VisualEditor Q3 Blockers: [Regression] VisualEditor no long able to rescue itself from badtoken error state on save - https://phabricator.wikimedia.org/T91158#1094343 (10Jdforrester-WMF) p:5High>3Unbreak! [00:14:32] 10VisualEditor, 10VisualEditor-MediaWiki, 5§ VisualEditor Q3 Blockers: [Regression] VisualEditor no long able to rescue itself from badtoken error state on save - https://phabricator.wikimedia.org/T91158#1075415 (10Jdforrester-WMF) [00:15:38] James_F: No, because it wouldn't be output as an edit notice [00:15:42] but just as junk output on EditPage.php [00:15:57] Hence the separate inteface message [00:16:03] *not* inside the edit notice [00:16:24] Krinkle: Ah, right. [00:16:33] Krinkle: And we'd have special treatment for it in VE. [00:16:41] ugh. the serialize-then-save code path doesn't handle badtoken at all D: [00:16:55] the just-save path tries to, but fails [00:17:10] James_F: VE extracts them from EditPage/API. Which would work as expected. [00:17:24] They wouldn't be in the editnotices array returned by the server. [00:17:48] 10VisualEditor, 10VisualEditor-MediaWiki: VisualEditor: Initiate file upload by drag-and-drop - https://phabricator.wikimedia.org/T40031#1094349 (10Jdforrester-WMF) [00:17:49] 10VisualEditor, 10VisualEditor-MediaWiki: VisualEditor: Button to initiate UploadWizard and insert image on completion - https://phabricator.wikimedia.org/T40030#1094351 (10Jdforrester-WMF) [00:18:10] VE would then be free to re-implement that part of EditPage if it wants to by loading that interface message and displaying it somewhere. [00:18:22] Just like we do with all other editpage stuff [00:18:25] data driven [00:18:36] * James_F nods. [00:18:55] Krinkle: Please, though, have them be an API that both EditPage and VE and whatever loads. [00:19:06] Rather than scraped out of EditPage like the checkboxes. [00:19:44] :-( [00:19:58] (03PS1) 10Jforrester: Update VE core submodule to master (300b89a) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194763 [00:20:49] Krinkle: OK, do you want to go ahead with T91715 then? I like it. [00:22:55] Krinkle: If you say yes I'll assign and make it a Q3 task, so… [00:23:08] (03CR) 10Kaldari: "Also, the label isn't centered vertically." [oojs/ui] - 10https://gerrit.wikimedia.org/r/194753 (owner: 10Nirzar) [00:23:53] 10VisualEditor: IE11- Inserting Template repositions the cursor to the middle of the page - https://phabricator.wikimedia.org/T91718#1094354 (10Etonkovidova) 3NEW [00:24:08] 10VisualEditor, 10VisualEditor-MediaWiki, 5§ VisualEditor Q3 Blockers: [Regression] VisualEditor no long able to rescue itself from badtoken error state on save - https://phabricator.wikimedia.org/T91158#1094364 (10matmarex) At a quick glance, we have several different code paths towards saving (thanks to sp... [00:25:33] 10MediaWiki-Page-editing, 6Editing, 5§ VisualEditor Q3 Blockers: Implement (part of) the English Wikipedia's edit notice system in MediaWiki core - https://phabricator.wikimedia.org/T91715#1094365 (10Jdforrester-WMF) p:5Triage>3High [00:25:47] 10MediaWiki-Page-editing, 6Editing, 5§ VisualEditor Q3 Blockers: Implement (part of) the English Wikipedia's edit notice system in MediaWiki core - https://phabricator.wikimedia.org/T91715#1094284 (10Jdforrester-WMF) [00:25:49] 10VisualEditor, 10VisualEditor-Performance, 5§ VisualEditor Q3 Blockers: Eliminate call to $.getVisibleText() in ve.init.mw.Target.prototype.onNoticesReady - https://phabricator.wikimedia.org/T87412#1094367 (10Jdforrester-WMF) [00:25:56] 10MediaWiki-Page-editing, 6Editing, 5§ VisualEditor Q3 Blockers: Implement (part of) the English Wikipedia's edit notice system in MediaWiki core - https://phabricator.wikimedia.org/T91715#1094369 (10Jdforrester-WMF) a:3Krinkle [00:29:38] (03PS1) 10Bartosz Dziewoński: ve.init.mw.ViewPageTarget: Use mw.Api#postWithToken [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194765 [00:32:30] (03PS1) 10Jforrester: build: Bump various devDependencies [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194766 [00:34:44] 10VisualEditor, 10VisualEditor-EditingTools: VisualEditor: Pressing Return at the end of the document throws Error: Offset could not be translated to a DOM element and offset: 675 - https://phabricator.wikimedia.org/T66567#1094378 (10Etonkovidova) Re-checked the page and the test scenario - no errors displaye... [00:34:56] 10VisualEditor, 10VisualEditor-EditingTools: VisualEditor: Pressing Return at the end of the document throws Error: Offset could not be translated to a DOM element and offset: 675 - https://phabricator.wikimedia.org/T66567#1094379 (10Etonkovidova) 5Open>3Resolved [00:35:30] James_F: Okay. Until when is Q3? [00:35:50] Krinkle: Realistically you'd need to finish this within two weeks' time, ideally sooner. [00:36:09] Krinkle: Q3 ends at the end of March. To achieve the deployment, switch over and removal of the slow code in VE… [00:36:30] Krinkle: As they say, "Move Quickly And Fix Stuff". :-) [00:41:00] 10MediaWiki-Page-editing, 6Editing, 3VisualEditor 2014/15Q3 blockers: Implement (part of) the English Wikipedia's edit notice system in MediaWiki core - https://phabricator.wikimedia.org/T91715#1094383 (10Jdforrester-WMF) [00:42:13] 10OOjs-UI, 10UI-Standardization, 5Patch-For-Review: ButtonWidget (frameless…)‎ to be styled like quiet buttons/normal anchors - https://phabricator.wikimedia.org/T88622#1094384 (10violetto) The icons next to the label is still colored on normal state instead of #555. [00:51:10] Krinkle: Does that sound OK? [00:54:36] James_F: Hm.. I'll give it a shot early next week [00:54:50] (03CR) 10Krinkle: [C: 032] build: Bump various devDependencies [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194766 (owner: 10Jforrester) [00:54:59] Krinkle: Great. :-) [00:56:35] 10VisualEditor, 10WikiEditor, 6Collaboration-Team, 10Flow, 7Design: Editing toolbar(s) for Flow (VE and/or wikitext toolbar) - https://phabricator.wikimedia.org/T78346#1094396 (10DannyH) We talked about this in the design meeting this morning, and decided on a plan for v1. We need to get a simple v1 of... [00:57:56] 10VisualEditor, 10WikiEditor, 6Collaboration-Team, 10Flow, 7Design: Editing toolbar(s) for Flow (VE and/or wikitext toolbar) - https://phabricator.wikimedia.org/T78346#1094397 (10DannyH) Pau, I think the only mockup that we still need is the mentions modal... Am I forgetting anything? [00:59:30] (03PS3) 10Mooeypoo: Add optional text to deprecated property [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/194748 (https://phabricator.wikimedia.org/T90734) [01:03:53] (03CR) 10jenkins-bot: [V: 04-1] build: Bump various devDependencies [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194766 (owner: 10Jforrester) [01:05:54] * James_F glares at npm [01:06:07] (03CR) 10Jforrester: [C: 032] "Krinkle said merge, damn it." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194766 (owner: 10Jforrester) [01:08:13] (03Merged) 10jenkins-bot: build: Bump various devDependencies [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194766 (owner: 10Jforrester) [01:16:27] 10VisualEditor, 10VisualEditor-EditingTools, 7Design, 7Epic, and 2 others: Cite: Auto-filled references based on destination URL / ISBN / etc. - https://phabricator.wikimedia.org/T62768#1094506 (10KHammerstein) [01:26:34] 10MediaWiki-ResourceLoader, 10MediaWiki-extensions-Sentry, 6Multimedia: Use source urls in mw.loader.store - https://phabricator.wikimedia.org/T90524#1094530 (10Tgr) [01:37:56] 10VisualEditor: [Regression wmf20] Cursor traverse through the table unexpectedly with right/left arrows and throws following error in the console "Uncaught TypeError: Cannot read property 'start' of null" - https://phabricator.wikimedia.org/T91726#1094596 (10Ryasmeen) 3NEW [01:42:19] 10VisualEditor: [Regression wmf20] Cursor traverses through the table unexpectedly with right/left arrow keys and throws following error in the console "Uncaught TypeError: Cannot read property 'start' of null" - https://phabricator.wikimedia.org/T91726#1094625 (10Ryasmeen) [01:44:08] 10VisualEditor: [Regression pre-wmf21] The Row/Column indicator for table is not aligned with the cell properly - https://phabricator.wikimedia.org/T91727#1094630 (10Ryasmeen) 3NEW [01:49:18] (03PS2) 10Bartosz Dziewoński: ve.init.mw.ViewPageTarget: Use mw.Api#postWithToken [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194765 [01:49:19] (03PS1) 10Bartosz Dziewoński: ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194783 (https://phabricator.wikimedia.org/T91158) [01:49:21] (03PS1) 10Bartosz Dziewoński: ve.ui.MWLinkTargetInputWidget: Fix documentation [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194784 [01:50:30] James_F: https://gerrit.wikimedia.org/r/194783 would be good to merge and deploy soon. not really 5-pm-on-friday-critical, but critical nonetheless. [01:51:39] (03CR) 10jenkins-bot: [V: 04-1] ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194783 (https://phabricator.wikimedia.org/T91158) (owner: 10Bartosz Dziewoński) [01:52:10] oh come on. [01:52:28] meh, unrelated. [01:54:22] (03CR) 10GWicke: "@Alex, the goal is indeed to move this logic into core, and make VRS easily accessible, possibly from RequestContext. I do expect that to " [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194338 (https://phabricator.wikimedia.org/T89066) (owner: 10Mobrovac) [01:54:57] Bash is the new Gulp [01:54:58] http://blog.keithcirkel.co.uk/how-to-use-npm-as-a-build-tool/ [02:55:57] 10VisualEditor, 10VisualEditor-EditingTools, 7Design, 7Epic, and 2 others: Cite: Auto-filled references based on destination URL / ISBN / etc. - https://phabricator.wikimedia.org/T62768#1094741 (10Eloquence) Can we project a "week of" deployment ETA? [03:12:02] (03CR) 10Legoktm: "check experimental" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194456 (https://phabricator.wikimedia.org/T90943) (owner: 10Legoktm) [03:44:55] 10TemplateData: No warning about invalid JSON in existing block, odd behavior. - https://phabricator.wikimedia.org/T91730#1094756 (10Spage) 3NEW a:3Mooeypoo [04:41:00] 10VisualEditor, 10MediaWiki-extensions-EasyTimeline, 10Possible-Tech-Projects: EasyTimeline: in-browser visual editing - https://phabricator.wikimedia.org/T29157#1094809 (10NiharikaKohli) Is this same as T29156? [04:44:23] (03PS1) 10Krinkle: jsduck: Set --processes=0 to fix warnings-exit-nonzero [unicodejs] - 10https://gerrit.wikimedia.org/r/194812 [04:44:59] (03PS1) 10Krinkle: jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/core] - 10https://gerrit.wikimedia.org/r/194813 [04:45:22] (03PS1) 10Krinkle: jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/ui] - 10https://gerrit.wikimedia.org/r/194814 [04:45:52] (03PS1) 10Krinkle: jsduck: Set --processes=0 to fix warnings-exit-nonzero [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194815 [04:49:44] (03CR) 10jenkins-bot: [V: 04-1] jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/ui] - 10https://gerrit.wikimedia.org/r/194814 (owner: 10Krinkle) [04:54:42] (03PS1) 10Krinkle: jsduck: Add MouseEvent and KeyboardEvent to externals [oojs/ui] - 10https://gerrit.wikimedia.org/r/194817 [05:44:48] 10VisualEditor, 10GlobalUserPage: GlobalUserPage links are red in VisualEditor - https://phabricator.wikimedia.org/T91731#1094866 (10Eloquence) 3NEW [06:09:26] 10VisualEditor, 10GlobalUserPage: GlobalUserPage links are red in VisualEditor - https://phabricator.wikimedia.org/T91731#1094890 (10Krinkle) The link completion is an interesting point. Something for GlobalUserPage to perhaps tap into OpenSearch in a way that currently Special pages are able to generate subpa... [06:09:45] 10VisualEditor, 10GlobalUserPage: GlobalUserPage links are red in VisualEditor - https://phabricator.wikimedia.org/T91731#1094892 (10Krinkle) [06:11:21] 10VisualEditor, 10VisualEditor-MediaWiki, 10GlobalUserPage, 10MediaWiki-extensions-EducationProgram: VisualEditor redlink detection should obey TitleIsAlwaysKnown hook (so that EducationProgram pages and GlobalUserPage links render correctly) - https://phabricator.wikimedia.org/T72123#1094896 (10Krinkle) [06:11:50] o/ [06:24:40] (03CR) 10Krinkle: "recheck" [unicodejs] - 10https://gerrit.wikimedia.org/r/194812 (owner: 10Krinkle) [06:25:36] (03CR) 10Krinkle: "recheck" [oojs/core] - 10https://gerrit.wikimedia.org/r/194813 (owner: 10Krinkle) [06:25:39] (03CR) 10Krinkle: "recheck" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194815 (owner: 10Krinkle) [06:25:45] (03CR) 10Krinkle: "recheck" [oojs/ui] - 10https://gerrit.wikimedia.org/r/194817 (owner: 10Krinkle) [06:26:34] (03CR) 10jenkins-bot: [V: 04-1] jsduck: Set --processes=0 to fix warnings-exit-nonzero [unicodejs] - 10https://gerrit.wikimedia.org/r/194812 (owner: 10Krinkle) [11:11:05] (03PS3) 10Catrope: ve.init.mw.ViewPageTarget: Use mw.Api#postWithToken for setting hidebetawelcome pref [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194765 (owner: 10Bartosz Dziewoński) [11:11:34] (03CR) 10Catrope: "PS3: Edited commit summary to clarify that this does not apply to saving (which would break the identity change warnings)" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194765 (owner: 10Bartosz Dziewoński) [11:12:16] Krinkle|detached: Re TargetLoader, if you're confused about what any of the changes are for, ask me [11:12:30] I probably had good reasons for making them, I just haven't yet gotten around to documenting why I made them [11:13:12] (03CR) 10Catrope: [C: 032] ve.init.mw.ViewPageTarget: Use mw.Api#postWithToken for setting hidebetawelcome pref [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194765 (owner: 10Bartosz Dziewoński) [11:13:17] (03CR) 10jenkins-bot: [V: 04-1] ve.init.mw.ViewPageTarget: Use mw.Api#postWithToken for setting hidebetawelcome pref [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194765 (owner: 10Bartosz Dziewoński) [11:14:18] (03CR) 10Catrope: [C: 032] ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194783 (https://phabricator.wikimedia.org/T91158) (owner: 10Bartosz Dziewoński) [11:14:35] (03PS2) 10Catrope: ve.ui.MWLinkTargetInputWidget: Fix documentation [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194784 (owner: 10Bartosz Dziewoński) [11:15:19] (03CR) 10Catrope: [C: 032] "Why is there even a documentation comment at all? Can't this just @inheritdoc from LookupElement?" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194784 (owner: 10Bartosz Dziewoński) [11:16:26] (03CR) 10jenkins-bot: [V: 04-1] ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194783 (https://phabricator.wikimedia.org/T91158) (owner: 10Bartosz Dziewoński) [11:16:28] (03CR) 10jenkins-bot: [V: 04-1] ve.init.mw.ViewPageTarget: Use mw.Api#postWithToken for setting hidebetawelcome pref [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194765 (owner: 10Bartosz Dziewoński) [11:16:30] (03CR) 10jenkins-bot: [V: 04-1] ve.ui.MWLinkTargetInputWidget: Fix documentation [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194784 (owner: 10Bartosz Dziewoński) [11:17:11] wtf [11:17:16] Thanks Jenkins [11:17:49] 11:14:39 Query: INSERT OR IGNORE INTO user_groups (ug_user,ug_group) VALUES ('1','sysop') [11:17:50] 11:14:39 Function: User::addGroup/single-row [11:17:52] 11:14:39 Error: 10 disk I/O error [11:17:54] Cool story [11:18:12] (03CR) 10Catrope: [C: 032] ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194783 (https://phabricator.wikimedia.org/T91158) (owner: 10Bartosz Dziewoński) [11:20:16] (03Merged) 10jenkins-bot: ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194783 (https://phabricator.wikimedia.org/T91158) (owner: 10Bartosz Dziewoński) [11:21:54] (03Merged) 10jenkins-bot: ve.init.mw.ViewPageTarget: Use mw.Api#postWithToken for setting hidebetawelcome pref [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194765 (owner: 10Bartosz Dziewoński) [11:21:56] (03Merged) 10jenkins-bot: ve.ui.MWLinkTargetInputWidget: Fix documentation [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194784 (owner: 10Bartosz Dziewoński) [12:38:29] (03CR) 10Alexandros Kosiaris: [C: 031] "@Mvolz so we got https://gerrit.wikimedia.org/r/#/admin/projects/mediawiki/services/zotero/translators which is a local copy to facilitate" [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/194548 (https://phabricator.wikimedia.org/T89872) (owner: 10Alexandros Kosiaris) [12:52:06] (03CR) 10Mobrovac: [C: 031] "+1 re zotero changes. I propose to include them right away in our local repo(s) (after testing them, ofc), but also to report them upstrea" [services/citoid/deploy] - 10https://gerrit.wikimedia.org/r/194548 (https://phabricator.wikimedia.org/T89872) (owner: 10Alexandros Kosiaris) [13:28:04] I'm having trouble getting the parsoid service to run properly [13:28:27] Installed it on Debian Wheezy with the parsoid and nodejs repos [13:28:37] TypeError: Cannot read property 'lang' of undefined << in the parsoid.log when trying to edit a page [13:28:53] Any suggestions? [14:23:40] 10VisualEditor, 10VisualEditor-ContentEditable: VisualEditor: Sometimes it's somehow possible for VE/Parsoid to insert a
at the start of a heading - https://phabricator.wikimedia.org/T53444#1095533 (10Elitre) 5declined>3Open Providing steps. 1) Start with [[ https://test.wikipedia.org/w/index.php?tit... [14:34:20] 10VisualEditor, 10MediaWiki-General-or-Unknown: The guest warning cannot be closed - https://phabricator.wikimedia.org/T91757#1095543 (10Nnemo) 3NEW [14:37:05] 10VisualEditor: VisualEditor: Coordinates template is displaying in the wrong place - https://phabricator.wikimedia.org/T67648#1095554 (10Elitre) I tried as well, and what happens for me in FF and Chrome is that the template moves completely to the right and there is no margin between the E and the scrollbar. [14:45:36] 10VisualEditor: VisualEditor: adding section for references will suppress the Related Articles one - https://phabricator.wikimedia.org/T55699#1095563 (10Elitre) Yup, can't reproduce this one. [14:58:54] 10VisualEditor, 10VisualEditor-EditingTools: Give exact match when looking for a specific file - https://phabricator.wikimedia.org/T64917#1095585 (10Elitre) Works for me as well. (Shouldn't this be Invalid or something else?) [15:13:50] 10VisualEditor, 10VisualEditor-MediaWiki: At nowiki the nowiki-tag shows up in some edits done with VisualEditor - https://phabricator.wikimedia.org/T74442#1095602 (10Elitre) [15:14:13] 10VisualEditor, 10VisualEditor-MediaWiki: At nowiki the nowiki-tag shows up in some edits done with VisualEditor - https://phabricator.wikimedia.org/T74442#749624 (10Elitre) [15:23:32] RoanKattouw: In general if you get false negatives on Jenkins, however brief or anything, a bug report, irc ping or email would be nice. A short excerpt and link to build would be great. I'm trying to document the known issues because we're good at forgetting. [15:24:08] RoanKattouw: Re targetLoader, let me phrase it this way to remove any ambiguity: What is the main thing the commit tries to achieve? It seems to move part of the init sequence to a second module that is also loaded on every page. [15:24:40] It tries to fire the API request for the Parsoid HTML immediately when the user clicks edit [15:25:07] Rather than first waiting for the RL request for VPT (which depends on all of VE core these days) [15:25:29] It also loads all the RL modules at once, rather than in two stages (currently VPT loads the remaining modules) [15:30:53] 10VisualEditor, 10MediaWiki-General-or-Unknown: "You are not logged in." warning cannot be closed in Safari on iOS - https://phabricator.wikimedia.org/T91757#1095618 (10Aklapper) [15:31:04] 10VisualEditor: "You are not logged in." warning cannot be closed in Safari on iOS - https://phabricator.wikimedia.org/T91757#1095543 (10Aklapper) [15:32:00] 10VisualEditor: "You are not logged in." warning cannot be closed in Safari on iOS - https://phabricator.wikimedia.org/T91757#1095543 (10Aklapper) Cannot reproduce with Firefox 36 on desktop... If Safari on iOS had some developer tools I'd be curious to see the JavaScript error console output when clicking. [15:32:25] Krinkle: So the main goal of the commit is to avoid the waterfall effect where we do something like mw.loader.using( 'ext.ve.vpt' ).then( function () { return $.when( apiRequest(), mw.loader.using( other modules ) ); } ).then( build DM and CE ); [15:32:44] Instead I want to do $.when( apiRequest(), mw.loader.using( everything we need ) ).then( build DM and CE ); [15:32:52] If that makes any sense [15:34:55] Although now I'm thinking we probably need to run some code in VPT as soon as it loads, so it can do DOM manipulation and stuff, without first waiting for the API request to finish [15:46:11] (03PS1) 10Jforrester: ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194869 (https://phabricator.wikimedia.org/T91158) [15:46:21] (03PS1) 10Jforrester: ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] (wmf/1.25wmf20) - 10https://gerrit.wikimedia.org/r/194870 (https://phabricator.wikimedia.org/T91158) [15:46:49] 10VisualEditor, 10VisualEditor-MediaWiki, 5Patch-For-Review, 3VisualEditor 2014/15Q3 blockers, and 3 others: [Regression] VisualEditor no long able to rescue itself from badtoken error state on save - https://phabricator.wikimedia.org/T91158#1095640 (10Jdforrester-WMF) [15:48:30] Heya. [15:49:20] RoanKattouw: Yeah, that makes sense. [15:49:31] RoanKattouw: Perhaps we can split some of it up in a second or even third commit. [15:50:48] RoanKattouw: I assume the reason the not-init init module (targetLoader) is in the bottom queue so that it starts loading earlier and becomes part of the existing load request, too. [15:51:00] (03CR) 10Bartosz Dziewoński: [C: 032] jsduck: Add MouseEvent and KeyboardEvent to externals [oojs/ui] - 10https://gerrit.wikimedia.org/r/194817 (owner: 10Krinkle) [15:51:12] It will still need using() in the init code in case it's slow, but it'll be faster yeah [15:52:04] Yeah that's why [15:52:10] And yes it does use using() in init init [15:52:26] I would love to split this along whatever lines you think are sensible [15:53:00] Although I think this might be one of those situations where we have to do the entire thing and then go backwards to split it up [15:53:13] (03Merged) 10jenkins-bot: jsduck: Add MouseEvent and KeyboardEvent to externals [oojs/ui] - 10https://gerrit.wikimedia.org/r/194817 (owner: 10Krinkle) [15:53:15] For instance, loading VPT late will require a lot of what activate() does to be factored differently [15:53:22] (03PS2) 10Krinkle: jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/ui] - 10https://gerrit.wikimedia.org/r/194814 [15:53:24] That could be done in a separate commit beforehand [15:53:38] RoanKattouw: I dont mind which goes first. As long as we can figure one out that actually can go first. [15:53:39] But you have to make the TargetLoader changes first before you find out you need to do that :) [15:54:17] Yeah that's what I'm saying, the one that can go first will probably end up being something we figure out near the end [15:54:20] So maybe start with moving some of VPT into targetLoader and add it to the bottom queue. Nothing much would change in init-init, except that its using() will now hook on the load that is already complete/started. [15:54:25] But that's OK, commits do not have to be in order of discovery [15:54:28] (03CR) 10Bartosz Dziewoński: [C: 032] jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/core] - 10https://gerrit.wikimedia.org/r/194813 (owner: 10Krinkle) [15:54:30] (03CR) 10Bartosz Dziewoński: [C: 032] jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/ui] - 10https://gerrit.wikimedia.org/r/194814 (owner: 10Krinkle) [15:54:32] (03CR) 10Bartosz Dziewoński: [C: 032] jsduck: Set --processes=0 to fix warnings-exit-nonzero [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194815 (owner: 10Krinkle) [15:54:37] (03CR) 10jenkins-bot: [V: 04-1] jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/ui] - 10https://gerrit.wikimedia.org/r/194814 (owner: 10Krinkle) [15:54:39] Hmm right [15:54:48] Stuff like setting .ve-activating could move there [15:55:27] (03CR) 10jenkins-bot: [V: 04-1] jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/core] - 10https://gerrit.wikimedia.org/r/194813 (owner: 10Krinkle) [15:57:13] (03Merged) 10jenkins-bot: jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/ui] - 10https://gerrit.wikimedia.org/r/194814 (owner: 10Krinkle) [15:57:33] (03Merged) 10jenkins-bot: jsduck: Set --processes=0 to fix warnings-exit-nonzero [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194815 (owner: 10Krinkle) [16:14:36] 10MediaWiki-ResourceLoader, 10Continuous-Integration, 10MediaWiki-Unit-tests, 5Patch-For-Review: Fix "DatabaseSqlite::replace/single-row NOT NULL constraint failed" for md_module table - https://phabricator.wikimedia.org/T91567#1095657 (10Krinkle) 5Open>3Resolved [16:15:13] RoanKattouw: Yeah [16:16:39] RoanKattouw: moving addClass ve-activating could be done in that same first commit. Probably in activateTarget? [16:16:53] (03CR) 10Krinkle: "recheck" [unicodejs] - 10https://gerrit.wikimedia.org/r/194812 (owner: 10Krinkle) [16:19:21] Yeah probably [16:22:11] Gah. [16:22:17] OK, now back online for good, I hope. [16:22:25] * James_F offers a token sacrifice to OIT. [16:24:08] RoanKattouw: Could you +2 https://gerrit.wikimedia.org/r/#/q/I103cf888d339b44e3fd4fe2376edf5e37ce4157f,n,z cherry-picks please? greg-g's said it's OK to deploy now. [16:24:27] (03CR) 10Catrope: [C: 032] ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194869 (https://phabricator.wikimedia.org/T91158) (owner: 10Jforrester) [16:24:30] Ta. [16:24:39] (03CR) 10Catrope: [C: 032] ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] (wmf/1.25wmf20) - 10https://gerrit.wikimedia.org/r/194870 (https://phabricator.wikimedia.org/T91158) (owner: 10Jforrester) [16:24:46] James_F: Who's deploying it? Me or someone else? [16:25:02] RoanKattouw: I was going to ask Krenair if he's around. [16:25:16] RoanKattouw: You need to focus on getting me my lovely new TargetModel thing. [16:25:20] Or whatever it's called. [16:25:24] (hint hint) [16:26:43] (03Merged) 10jenkins-bot: ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] (wmf/1.25wmf19) - 10https://gerrit.wikimedia.org/r/194869 (https://phabricator.wikimedia.org/T91158) (owner: 10Jforrester) [16:26:49] Hi James_F [16:26:55] (03Merged) 10jenkins-bot: ve.init.mw.Target: Don't go into infinite recursion on API errors [extensions/VisualEditor] (wmf/1.25wmf20) - 10https://gerrit.wikimedia.org/r/194870 (https://phabricator.wikimedia.org/T91158) (owner: 10Jforrester) [16:27:13] James_F: I'm making transactions exception-safe [16:27:21] RoanKattouw: That's good too. [16:27:26] I finally found a good strategy to do it [16:27:29] We don't have a swat right now because it's friday [16:27:32] Krenair: Want to do a quick deploy? Or are you busy with other things? [16:27:32] Inspired by code I wrote three years ago [16:27:38] Krenair: Yeah, I know. [16:27:39] is this unbreak now, so we need to deploy it asap? [16:27:40] And I'm halfway through porting the code to it [16:27:44] Krenair: Yup. [16:27:47] alright [16:27:48] RoanKattouw: Nice. [16:28:02] greg-g, ^ fyi [16:28:04] Krenair: It's infinite retry loops for API requests that encounter token failures [16:28:12] Fallout from the mw.Api change apparently [16:28:38] * Krenair curses submodule updates [16:29:55] (03CR) 10Jforrester: Add composer entry point for phpcs and linting (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194456 (https://phabricator.wikimedia.org/T90943) (owner: 10Legoktm) [16:30:20] Hey edsanders [16:31:44] Krenair: yep, ty [16:31:46] 10VisualEditor: Category was duplicated and given different sort keys - https://phabricator.wikimedia.org/T89618#1095690 (10Jdforrester-WMF) a:5Jdforrester-WMF>3Catrope [16:32:13] 10VisualEditor, 10VisualEditor-MediaWiki, 5Patch-For-Review, 3VisualEditor 2014/15Q3 blockers, and 3 others: [Regression] VisualEditor no long able to rescue itself from badtoken error state on save - https://phabricator.wikimedia.org/T91158#1095694 (10Jdforrester-WMF) 5Open>3Resolved [16:36:16] (03CR) 10Jforrester: [C: 032] jsduck: Set --processes=0 to fix warnings-exit-nonzero [unicodejs] - 10https://gerrit.wikimedia.org/r/194812 (owner: 10Krinkle) [16:36:38] (03CR) 10Jforrester: [C: 032] jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/core] - 10https://gerrit.wikimedia.org/r/194813 (owner: 10Krinkle) [16:41:06] (03Merged) 10jenkins-bot: jsduck: Set --processes=0 to fix warnings-exit-nonzero [unicodejs] - 10https://gerrit.wikimedia.org/r/194812 (owner: 10Krinkle) [16:41:08] (03CR) 10jenkins-bot: [V: 04-1] jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/core] - 10https://gerrit.wikimedia.org/r/194813 (owner: 10Krinkle) [16:42:26] (03CR) 10Legoktm: Add composer entry point for phpcs and linting (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194456 (https://phabricator.wikimedia.org/T90943) (owner: 10Legoktm) [16:43:05] (03CR) 10Jforrester: Add composer entry point for phpcs and linting (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194456 (https://phabricator.wikimedia.org/T90943) (owner: 10Legoktm) [16:44:12] Looks like saucelabs is down [16:44:19] or unavailable via their tunnel at least [16:44:23] Fun. [16:44:36] (03PS2) 10Legoktm: Add composer entry point for phpcs and linting [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194456 (https://phabricator.wikimedia.org/T90943) [16:48:42] (03CR) 10Jforrester: [C: 032] Add composer entry point for phpcs and linting [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194456 (https://phabricator.wikimedia.org/T90943) (owner: 10Legoktm) [16:51:22] (03Merged) 10jenkins-bot: Add composer entry point for phpcs and linting [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194456 (https://phabricator.wikimedia.org/T90943) (owner: 10Legoktm) [16:52:07] Krinkle, we have a TemplateData php-vs-javscript json'ability issue. Hoorah... So apparently S found a template with TemplateData that has multiline description to one of the parameters. The PHP extension finds it valid (it displays it in the table and allows to save) but the Javascript json parse has no idea what to do with it, and presents the dialog as empty. We are probably going to have to find how many cases of these exist in p [16:52:07] roduction to try and fix them. Is this possible? Finding specific types of TemplateData? [16:52:59] mooeypoo: Why is that a JSON issue? [16:53:09] RoanKattouw, what do you mean? [16:53:13] Does it not parse as valid JSON according to JSON.parse()? [16:53:21] seems not [16:53:34] RoanKattouw, https://www.mediawiki.org/w/index.php?title=Template:RFC/doc&action=edit [16:53:36] ^^ [16:53:42] RoanKattouw: \n characters as opposed to actual \ns aren't valid in JSON strings. [16:53:46] (I think?) [16:53:47] Look at the description of 'status' [16:54:17] Oh wow [16:54:26] And PHP's JSON functions allow this? [16:54:29] RoanKattouw, so I had an idea, but I'm not sure it'll work, I'm testing this [16:54:30] I hope they at least don't produce it [16:54:41] RoanKattouw, it seems so -- if you look at the actual page in read mode, the table is there [16:55:08] Hmm but don't we like normalize those pages? [16:55:17] apparently not [16:55:21] so I can *start* normalizing [16:55:43] php > echo json_encode(array('foo'=>"bar\nbaz")); [16:55:44] {"foo":"bar\nbaz"} [16:55:46] RoanKattouw, or I can cheat... I have an idea. It seems like if you ask the API for templateData you get it normalized even if it looks like this [16:55:56] RoanKattouw, http://www.mediawiki.org/w/api.php?action=templatedata&titles=Template:RFC [16:56:12] maybe instead of having TemplateData dialog parse from the text of the page, I should call the api for the templatedata [16:56:21] It will still be normalized on *save* [16:56:26] but at least it won't die on read [16:57:08] Hmm [16:57:12] But that's kind of evil [16:57:16] yes. [16:57:21] I feel like we should be normalizing on save [16:57:24] How are we not already doing this? [16:57:25] We should [16:57:30] (03PS1) 10Legoktm: jenkins job validation, do not submit [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194893 [16:57:30] Normalizing with pretty-print [16:57:35] RoanKattouw, the php doesn't seem to [16:58:24] So, I can start doing that from now onwards, *but* (a) we'll have unhappy users who see dirty diff for their templatedata and (b) it will only work on save, so this won't solve the issue of loading an existing templatedata with this behavior into the javascript [16:58:24] I swear we implemented this *somewhere* [16:58:27] But maybe it was Gadgets 2.0 [16:58:42] RoanKattouw, it allows you to save, which means the php sees no issues with it [16:58:52] if you're missing a comma somewhere, ironically, it would stop you [16:58:58] with an "invalid json" message [16:59:01] Yeah I verified that PHP parses JSON with newlines just fine, even though it's invalid [16:59:03] There's a V8 bug for this. Marked Wontfix per JSON spec. [16:59:09] So it validates but it doesn't normalize [16:59:12] PHP is in violation here and we should reject those from being saved server-side. [16:59:13] yeah [16:59:14] (03Abandoned) 10Legoktm: jenkins job validation, do not submit [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194893 (owner: 10Legoktm) [16:59:16] I think that eventually, we should normalize [16:59:18] I can add normalization [16:59:23] at this point, we seem to have to [16:59:24] But for now the API thing you mentioned works as a workaround [16:59:33] otherwise the editor can't do anything with it [16:59:51] RoanKattouw, even still, if i take the info from the api, it will still be normalized somewhat on *save* [16:59:57] because the javascript is the one that outputs the json [17:00:03] so dirty diffs, etc [17:00:10] This is not on many pages right? We can make a breaking change (or rather, removing support for something undocumented) that will make newlines break. [17:00:15] James_F, suggested we look for how many templatedata strings we have on production [17:00:16] TDG won't have to support it [17:00:18] not sure we can do it [17:00:25] It [17:00:28] It [17:00:49] mooeypoo: Won't it massively dirty-diff anyway because the source is often pretty-printed and the result from the JS won't be? [17:00:51] Krinkle, it doesnt have to break, we can "just" normalize it [17:01:01] This is why normalization to a pretty-printed format is nice [17:01:02] RoanKattouw, the result from the JS is pretty printed too [17:01:07] Oh? How? [17:01:11] sometimes a little differently, but mostly okay [17:01:11] mooeypoo: I don't want people inserting newlines and spreading the bug that that will work. [17:01:14] I didn't realize you could do that [17:01:33] >> JSON.stringify( {a:123,b:[123]}, null, 4); [17:01:37] RoanKattouw, it's a flag [17:01:40] Krinkle: Do you remember the JSON normalization stuff I did for Gadgets 2.0? [17:01:59] RoanKattouw: That's because we were using $.toJSON, not JSON.stringify [17:02:04] >> JSON.stringify( {a:123,b:[123]}, null, 4); [17:02:04] Krinkle: (string) '{\n "a": 123,\n "b": [\n 123\n ]\n}' [17:02:07] RoanKattouw, JSON.stringify( result, null, '\t' ); [17:02:07] You could save any old JSON and it would run json_encode( json_decode( $whatYouTyped ), JSON_PRETTY_PRINT ); and save that instead [17:02:29] RoanKattouw: Yeah, but that's for dedicated pages [17:02:31] that's what I use in the output code [17:02:38] Do we want to make a PST? [17:02:38] Krinkle: Crap right of course [17:02:40] This is a tag [17:02:42] lol no [17:02:47] Please no [17:02:49] :D [17:02:52] O.O [17:02:58] We /want/ to make it its own namespace. [17:03:04] (Pre save transformation, like subst:) [17:03:09] (and ~~~~_ [17:03:15] At which point we can use JsonContentHandler, which hopefully has PST normalization [17:03:25] Well, even if it does we don't care. [17:04:56] I think a PST is the only sane way. [17:05:18] I wonder if php has a way to PROPERLY validate json [17:07:10] James_F: If there's a relatively painless way to make a PST thing then I'd be all for it, but I'm concerned there probably isn't one [17:07:28] *then I'm all for it [17:07:48] RoanKattouw: Just swapping CRs for \ns inside ""s? [17:09:03] hm, the php part of validating is weird [17:09:03] James_F: Well no if you have a good way to PST the contents of a tag, you should go all the way and do $contents = json_encode( json_decode( $contents ), JSON_PRETTY_PRINT ); [17:09:08] James_F: RoanKattouw: edsanders: The tracker bug for "Wikipedia" on chromium is now working and with cascading notifications, so star it if you're interested [17:09:16] Krinkle: Link? [17:09:17] Krinkle: Link? [17:09:18] Ha. [17:09:26] https://code.google.com/p/chromium/issues/detail?id=463348 [17:09:34] RoanKattouw: That'd be very neat, actually. If that works? [17:09:37] And Paul Irish took over ownership :) [17:09:42] Krinkle: Nice. [17:09:58] Krinkle: Did you ever look at the ones I added to https://etherpad.wikimedia.org/p/Chromium-Upstream ? [17:10:04] Yep. [17:10:07] Doing it now [17:10:07] The Etherpad suggests they haven't been added to the tracking bug yet [17:10:34] Tracker bug "for Wikipedia"? [17:11:17] 10VisualEditor: VE feedback page on mediawiki.org: replace LiquidThreads with Flow or with wikitext - https://phabricator.wikimedia.org/T91768#1095868 (10Elitre) 3NEW [17:11:43] 10VisualEditor: VE feedback page on mediawiki.org: replace LiquidThreads with Flow or with wikitext - https://phabricator.wikimedia.org/T91768#1095877 (10Jdforrester-WMF) Flow or nothing. [17:12:22] 10VisualEditor, 10Flow: Convert LiquidThreads to Flow for the mw:VisualEditor/Feedback page - https://phabricator.wikimedia.org/T91768#1095878 (10Jdforrester-WMF) [17:13:14] RoanKattouw: I want to make sure not to include things that may be fixed or no longer affect us. [17:13:23] Those three (e.g. https://code.google.com/p/chromium/issues/detail?id=304826 ) are still live ? [17:14:59] Krinkle: Let me check [17:15:54] The filter one is definitely still live, and was already duped to something else [17:16:37] The selection color in lists one is still live [17:17:55] The 
 serialization bug is also still live
[17:18:12] 	 And as I said, is likely to encounter resistance
[17:18:36] 	 Yeah
[17:18:52] 	 RoanKattouw: Maybe we'll fix it the same way we fixed innerText
[17:18:58] 	 by standardising to textContent
[17:19:01] 	 The reload bug is also live
[17:19:09] 	 e.g. Node.htmlContent
[17:19:13] 	 hah, maybe
[17:19:32] 	 Which was affected by a very similar bug and is also introduced by MSIE
[17:19:34] 	 Well Opera didn't seem to care that they went from having a working version to a broken one
[17:19:38] 	 When they moved to Blink
[17:20:03] 	 I'm generally not very optimistic about fixing pre serialization but if that ever did happen it would be very nice
[17:22:37] 	 Krinkle, I'm a little confused about the TemplateData php code, specifically the onPageContentSave
[17:24:15] 	 Krinkle, when I try to get a json string with multilines through json_decode() it fails, with an error that it's not a valid json
[17:24:49] 	 but the php extension seems to tkae another way of validating the json -- and that seems to state the json's okay. From the comment I understand that this is the way that MW api is validating/working with the JSON
[17:25:45] 	 Krinkle, But the API also normalizes it, it seems ... So I don't understand how ->isOk() returns true for those cases in onPageContentSave
[17:25:46] 	 mooeypoo: json_decoding() a multiline string worked for me
[17:26:24] 	 uh json_decode() with the multiline didn't work for me
[17:26:29] 	 10VisualEditor: [Regression pre-wmf21] Tool-Format menu does not display all options - https://phabricator.wikimedia.org/T91772#1095930 (10Etonkovidova) 3NEW
[17:26:44] 	 php > $json = "{\"foo\":\"bar\nbaz\"}";
[17:26:46] 	 php > var_dump(json_decode($json));
[17:26:47] 	 object(stdClass)#1 (1) {...}
[17:28:59] 	 RoanKattouw_away, add a \n before bar
[17:29:47] 	 10TemplateData: TemplateData: Store TemplateData in its own namespace ("Template data:") with a JSON content type and associated to the Template: namespace - https://phabricator.wikimedia.org/T56140#1095938 (10Jdforrester-WMF)
[17:30:53] 	 err.
[17:36:43] 	 mooeypoo: worked?
[17:36:43] 	 10OOjs-UI, 10UI-Standardization, 7Design: PopupButtonWidget (frameless, with popup head)‎ shows at hover instead of onClick? - https://phabricator.wikimedia.org/T88630#1095960 (10Jaredzimmerman-WMF) Hover seems to make sense for things like this as long as we also support click (for touch enabled devices)
[17:38:24] 	 Hm.. pressing Shift when having text selected scroll be back up to the top of the selection. Is that intentional / known bug?
[17:38:53] 	 annoying when trying to expand a selection or pressing shift for another reason and results in a scroll.
[17:42:01] 	 yeah it works
[17:42:04] 	 though it shouldn't
[17:42:07] 	 it's not a valid json
[17:55:18] 	 oh,crap, I just realized my workaround isn't really going to work
[17:55:59] 	 Krinkle, okay, so I could workaround the issue *for now* by asking for the API version of the templatedata string as a fallback -- so, if the javascript can't parse the json it is given, ask the API for one, because the API seems to produce a normalized json.
[17:56:45] 	 Krinkle, *but* that would mean that if the user edited the string by hand pre-save, and then loaded the editor, they'll get the saved (and therefore not current) version of the template data.
[17:56:50] 	 so... not too good.
[17:56:59] 	 I'm going to try and at least get an error message going. 
[17:57:20] 	 James_F: Do we have an online url to ve-standalone demo?
[17:57:59] 	 Krinkle: https://en.wikipedia.org/w/extensions/VisualEditor/lib/ve/demos/ve/desktop.html
[17:58:02] 	 hilarious, i know
[17:58:27] 	 Interesting
[17:58:37] 	 doesn't work on beta labs, alas.
[17:59:23] 	 WE don't require any build script for that?
[17:59:43] 	 (03PS1) 10Kmenger: Dialog: Mark private methods and add description of methods and configs [oojs/ui] - 10https://gerrit.wikimedia.org/r/194906 
[18:00:10] 	 10VisualEditor: [Regression pre-wmf21] Tool-Format menu does not display all options - https://phabricator.wikimedia.org/T91772#1096028 (10matmarex) a:3Jdforrester-WMF Fixed in https://gerrit.wikimedia.org/r/#/c/194757/ in VE, need to pull it to VE-MW.
[18:00:21] 	 nope
[18:00:45] 	 Krinkle: we do for desktop-dist.html
[18:00:52] 	 Right
[18:02:37] * Krenair pokes James_F 
[18:03:40] 	 (03PS2) 10Bartosz Dziewoński: jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/core] - 10https://gerrit.wikimedia.org/r/194813 (owner: 10Krinkle)
[18:03:51] 	 (03CR) 10Bartosz Dziewoński: [C: 032] jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/core] - 10https://gerrit.wikimedia.org/r/194813 (owner: 10Krinkle)
[18:04:37] 	 (03CR) 10jenkins-bot: [V: 04-1] jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/core] - 10https://gerrit.wikimedia.org/r/194813 (owner: 10Krinkle)
[18:06:28] 	 mooeypoo: You mean {"foo":\n"bar"} or something?
[18:10:46] 	 10OOjs-UI, 5Patch-For-Review: Toolbars are messed up, especially in MediaWiki theme - https://phabricator.wikimedia.org/T91135#1096069 (10matmarex) 5Open>3Resolved
[18:12:45] 	 TrevorP|Away: https://phabricator.wikimedia.org/T91152 is going to be "fun"
[18:14:56] 	 10VisualEditor, 6Mobile-Web: Editor switcher styling regression - https://phabricator.wikimedia.org/T91283#1096096 (10KLans_WMF)
[18:15:20] 	 10VisualEditor, 6Mobile-Web: Editor switcher styling regression - https://phabricator.wikimedia.org/T91283#1078925 (10KLans_WMF) @khammerstein can you take a look at this?
[18:16:10] 	 (03PS1) 10Legoktm: Add composer entry point for phpcs and linting [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/194910 (https://phabricator.wikimedia.org/T90943) 
[18:16:33] 	 (03CR) 10Legoktm: "check experimental" [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/194910 (https://phabricator.wikimedia.org/T90943) (owner: 10Legoktm)
[18:16:40] 	 10OOjs-UI, 5OOjs-UI-0.9.0, 5WMF-deploy-2015-03-11_(1.25wmf21): Toolbars are messed up, especially in MediaWiki theme - https://phabricator.wikimedia.org/T91135#1096106 (10Jdforrester-WMF)
[18:19:25] 	 Krenair: ?
[18:19:40] 	 We were waiting for you to join :p
[18:19:57] 	 Krenair: We were waiting for people to GTFO of the room. :-)
[18:28:41] 	 MatmaRex: https://wikimedia.github.io/VisualEditor/
[18:29:08] 	 Seems to work.
[18:29:19] 	 Pre-compiles npm-install and includes dist/
[18:29:40] 	 We can add this to dev.wm.o later and demos in docs etc.
[18:29:49] 	 But this was 1 checkbox
[18:29:54] 	 to enable :)
[18:29:55] 	 :o
[18:32:09] 	 10VisualEditor: Help the HTML Editing Task Force shape "Editing on the Web" (aka future of contentEditable) - https://phabricator.wikimedia.org/T91781#1096192 (10Krinkle) 3NEW
[18:32:19] 	 #whatwg channel just told me we should work with them to fix content editable.
[18:32:25] 	 They're doing a rewrite, a 2.0 if you will.
[18:32:47] 	 I told them I'd pass it on, but I'm spread too thin to do it myself right now
[18:36:07] 	 Krinkle: Eh.
[18:36:12] 	 Krinkle: I've been following the list for months.
[18:36:18] 	 Oh?
[18:36:23] 	 Krinkle: They seem to think that replacing CE is the only way.
[18:36:34] 	 Which is a bit defeatist, and means support for Chrome 55+ only.
[18:36:39] 	 Which sucks as a user outcome.
[18:36:49] 	 They're not necessarily /wrong/.
[18:37:01] 	 Whatever they come up with, even if it changes existing CE will be future versions of any browser. 
[18:37:02] 	 Though we've done quite well with CE, it's been very hard and there are still some issues.
[18:37:24] 	 They're not focussed on best practices at using CE.
[18:37:32] 	 They're not focussed on work-arounds for common issues.
[18:37:47] 	 They're not focussed on narrowing down a few minor bugs in the browsers that block wider adoption.
[18:37:52] 	 10VisualEditor, 10Flow: Convert LiquidThreads to Flow for the mw:VisualEditor/Feedback page - https://phabricator.wikimedia.org/T91768#1096267 (10DannyH)
[18:37:53] 	 In other words, not really our thing.
[18:38:31] 	 10VisualEditor, 10Flow, 10LiquidThreads, 3Collaboration-Team-Sprint-R-2015-03-11: R6. LQT conversion test for Mediawiki.org VE/Feedback - https://phabricator.wikimedia.org/T90025#1096272 (10Jdforrester-WMF)
[18:38:46] 	 Maybe I'm wrong. :-)
[18:38:57] 	 James_F: Have you posted anything?
[18:39:17] 	 Krinkle: I posted a "hey, I'm James from VisualEditor" message.
[18:39:56] 	 Krinkle: 2014-07-11.
[18:40:49] 	 10VisualEditor, 10Flow: Convert LiquidThreads to Flow for the mw:VisualEditor/Feedback page - https://phabricator.wikimedia.org/T91768#1096279 (10DannyH) Turns out we are currently working on that right now! We did a test conversion of VE/Feedback yesterday, and found one issue that needs to get fixed. ErikB i...
[18:40:59] 	 James_F: Well, I don't know how flexible they are but I do know that it's gonna be what we'll be stuck with. We can either work with them or against them.
[18:41:14] 	 Krinkle: Or ignore them.
[18:41:25] 	 Krinkle: http://lists.w3.org/Archives/Public/public-editing-tf/2014Jul/0015.html
[18:41:27] 	 If we can't make them see it our way, we can at least try and get a few things into the spec within their larger view.
[18:41:32] 	 Krinkle: A lot of W3 TFs go nowhere.
[18:41:59] 	 Oh FFS
[18:42:00] 	 Krinkle: And I didn't think it a valuable use of e.g. 10 hours of Roan's or your time a week chasing stardust.
[18:42:03] 	 RoanKattouw: ?
[18:42:13] 	 I'm making personal edits on eowiki, as you do
[18:42:18] 	 And they have an old version of WikEd
[18:42:26] 	 With a broken Function.prototype.bind polyfill
[18:42:28] 	 RoanKattouw: Don't use WikEd. Simples.
[18:42:40] 	 Which to add insult to injury overwrites the native implementation with a broken one
[18:42:47] 	 Fun.
[18:42:57] 	 Oh no it's Lupo's thing, whatever
[18:45:02] 	 Krinkle, RoanKattouw (and everyone, really): Do you have wishlist items you'd like RelEng to schedule to make our lives easier? I'm making a list…
[18:45:10] 	 https://etherpad.wikimedia.org/p/4pKxno7DAI
[18:59:32] 	 James_F: Is T90678 about unit or integration tests
[19:00:12] 	 MatmaRex: how come https://en.wikipedia.org/w/extensions/VisualEditor/lib/ve/demos/ve/desktop.html works but https://en.wikipedia.org/w/vendor/oojs/oojs-ui/demos/index.html
[19:00:22] 	 doesn't ?  (No big deal, just curious)
[19:00:25] 	 10VisualEditor, 10MediaWiki-General-or-Unknown, 10RESTBase, 6Services, and 3 others: Use RESTBase from the MediaWiki Virtual Rest Service - https://phabricator.wikimedia.org/T89066#1096453 (10greg)
[19:00:29] 	 10VisualEditor, 6Mobile-Web: Editor switcher styling regression - https://phabricator.wikimedia.org/T91283#1096454 (10KHammerstein) Thanks @Jdlrobson Lots of things in mobile VE are visually broken or failing user tests. I would love to work on them, but not sure which team has the capacity to fix these bugs....
[19:00:44] 	 spagewmf: Because we don't expose /w
[19:00:49] 	 Contents of /w is a whitelisted subset
[19:00:58] 	 It's not the real script directory
[19:01:12] 	 https://github.com/wikimedia/operations-mediawiki-config/tree/master/w
[19:01:20] 	 Which is in purpose, and vendor should not be exposed there.
[19:01:21] 	 10VisualEditor, 10VisualEditor-MediaWiki: VisualEditor: "Unknown error" on saving on officewiki - https://phabricator.wikimedia.org/T70235#1096456 (10Catrope) >>! In T70235#719892, @Krenair wrote: > So that last srwikibooks entry appears to be in an unused file... Roan, can you release and upstream the LAPI.js...
[19:01:51] 	 spagewmf: Also remember that /w is global, so it's not the correct wmf branch for that wiki 
[19:02:02] 	 Krinkle: Both, but it's written about browser tests.
[19:02:49] 	 Krinkle: thanks!  I suppose /w/extensions has to be exposed to access extension resources in some mode, maybe RL debug?
[19:03:04] 	 spagewmf: Nah, nothing is for debug.
[19:03:17] 	 spagewmf: wgExtensionAssetsPath needs to public.
[19:03:18] 	 RL debug is on bits.wm.o I think
[19:03:25] 	 Hi, I'm looking for some help troubleshooting a new VE installation.
[19:03:28] 	 10VisualEditor, 10MediaWiki-General-or-Unknown, 10RESTBase, 6Services, and 3 others: Use RESTBase from the MediaWiki Virtual Rest Service - https://phabricator.wikimedia.org/T89066#1096459 (10GWicke) @mobrovac, I'll amend your patch to address some of the comments today.  I hope that we can get this in ear...
[19:03:28] 	 spagewmf: But its set to the one on bits.
[19:03:55] 	 spagewmf: w/extensions might not have to be exposed. It's there from pre-RL. Where extensions loaded scripts via script tags.
[19:04:00] 	 spagewmf: It has php disabled however
[19:07:51] 	 10VisualEditor, 10VisualEditor-Performance, 10MediaWiki-General-or-Unknown, 10MediaWiki-Interface, and 8 others: Use Parsoid HTML for read-only views too - https://phabricator.wikimedia.org/T55784#1096461 (10Jdforrester-WMF)
[19:11:06] 	 RoanKattouw, just to verify, inside a dialog's getSetupProcess I have an async function with .then() and then both success and failure methods. In the failure function, I return OO.ui.Error() -- that *should* trigger the error reporting of the dialog, shouldn't it? 
[19:11:13] 	 it doesn't, and I'm trying to figure out why
[19:11:24] 	 No
[19:11:31] 	 Error reporting doesn't work in the setupProcess
[19:11:34] 	 Only in ActionProcesses
[19:11:38] 	 but --
[19:11:42] 	 oh poop.
[19:11:46] 	 That sucks.
[19:12:01] 	 What does it mean to fail to set up?
[19:12:05] 	 RoanKattouw, well, the templatedata dialog reads the current templatedata json string and tries to parse it
[19:12:11] 	 Hmm right
[19:12:12] 	 then it goes "wait, I can't parse this."
[19:12:24] 	 Up until now I had a custom div with an error at the top of the dialog
[19:12:32] 	 but in THIS case, it could really be good to display an actual error
[19:12:57] 	 mostly because the user has to notice that we had an error that meant the dialog is working with information from scratch
[19:13:13] 	 I guess I can *call* an action process when I set up
[19:13:17] 	 and have it succeed or fail there
[19:13:24] 	 but that's really backwards :\
[19:13:41] 	 RoanKattouw, will copy that Function#bind fix to other wikis
[19:13:48] 	 10VisualEditor, 10VisualEditor-Tables, 3Roadmap: Be able to re-order the columns or rows of a table by dragging to to another position - https://phabricator.wikimedia.org/T88694#1096476 (10Jdforrester-WMF)
[19:14:18] 	 RoanKattouw, how horrible is it if I go this.actions.executeAction( 'readModel' ) from getSetupProcess so I can trigger the error if needed? :\
[19:14:25] 	 mooeypoo: Ask Trevor
[19:14:34] 	 TrevorParscal, ^^
[19:14:35] 	 mooeypoo: And tell him your use case for a fail-able setup process
[19:14:48] 	 the second option is to validate before the dialog comes up
[19:14:51] 	 10VisualEditor, 10VisualEditor-Tables, 3Roadmap: Be able to re-order the columns or rows of a table by dragging them to another position in VisualEditor - https://phabricator.wikimedia.org/T88694#1096493 (10Jdforrester-WMF)
[19:14:57] 	 and display  message dialog if things are bad
[19:15:04] 	 but that's unnecessarily complex, I think.
[19:15:12] 	 TrevorParscal, I need error handling in getSetupProcess! :)
[19:15:52] 	 mooeypoo: thanks for working on this. Gimme $5 and I'll fix the bad templatedata in {{RFC/doc}} and mark the bug "Lowest" priority :)
[19:15:54] 	 James_F: review https://gerrit.wikimedia.org/r/194910 ? :)
[19:16:14] 	 (03PS1) 10Kmenger: ProcessDialog: Add description and example and mark private methods [oojs/ui] - 10https://gerrit.wikimedia.org/r/194923 
[19:16:15] 	 spagewmf, put it on my tab
[19:16:16] 	 :D
[19:16:36] 	 mooeypoo: hmm
[19:16:41] 	 spagewmf, but yeah, the main issue here is that the php json parsing seems to be different than the javascript one
[19:16:42] 	 error handling in getSetupProcess
[19:16:52] 	 and by "different" i mean it sucks. It accepts bad json formats :(
[19:17:09] 	 so, if you return an error in one of the OOjs UI steps in getSetupProcess, I suppose that just fails in a some silent way?
[19:18:04] 	 TrevorParscal, seems so yeah
[19:18:36] 	 TrevorParscal: got any thoughts on https://phabricator.wikimedia.org/T91152 before i do it and break everything?
[19:18:39] 	 TrevorParscal, the TemplateData dialog, on setup, reads the templatedata string and parses it. But if there's an error in parsing, the dialog opens *empty* letting the user re-add stuff. I'd like to be able to splash the user with "Wait! you have an error in your json!" before I let them get frustrated with empty dialog.
[19:19:24] 	 mooeypoo: Why don't you just get the JSON from that API thing that does always return something valid, as you said before?
[19:19:40] 	 RoanKattouw, because that doesn't account for changes the user just made to the string
[19:19:52] 	 Ahm
[19:20:00] 	 I'm not too familiar with the TD editor
[19:20:06] 	 so let's say you start a new templatedata, you used the dialog, applied -- you have a string in the textbox. Now you reopen the dialog.
[19:20:08] 	 But why is the user able to edit raw JSON?
[19:20:16] 	 Oh because you're in an edit page with a textbox
[19:20:22] 	 mooeypoo: hmm, you could probably have the API parse the article text, and extract the 'templatedata' property from API output?
[19:20:24] 	 OK well in that case the user could have entered any old crap in there
[19:20:26] 	 Because the result is pasted into the edit input
[19:20:31] 	 yes
[19:20:36] 	 The user could vandalize the JSON and then load your code
[19:20:38] 	 hence why I am using more defensive coding in there
[19:20:40] 	 yep
[19:20:42] 	 they could.
[19:20:42] 	 So this isn't necessarily related to the newline issue
[19:20:56] 	 Not necessarily, no -- but in any other case, the PHP side won't let the user *save*
[19:21:04] 	 so you won't see it in random pages already in 
[19:21:29] 	 so if I mess up the json, and then hit "save" you'll get "Invalid JSON" message and your page won't be saved until you fix it
[19:22:00] 	 the main issue is that the php json parser doesn't care about newlines, so we have pages in production (potentially) that have non-parse'able templatedata strings for the javascript editor
[19:22:02] 	 that makes sense?
[19:23:05] 	 RoanKattouw, the dialog "protects" itself fine, it tries to read the json and if it can't, it displays a message. I just want to make that message properly big and in-the-user's-face
[19:23:20] 	 and preferably, let the user decide "yes, start from scratch" or "cancel" and fix their own templatedata
[19:23:39] 	 Yeah OK
[19:24:02] 	 I was just trying to understand things here, and figured out that the reason you're having this problem isn't just because of the newline thing
[19:24:07] 	 It's also because the user may do silly things
[19:24:49] 	 mooeypoo: the question is, will that error be recoverable or retryable?
[19:25:08] 	 I think what you are talking about there is a different kind of error, we've discussed this before, it's a warning or a notice or something
[19:25:28] 	 TrevorParscal, so, my next question to you would've been whether I can change the text of the 'retry' button
[19:25:43] 	 because in this case it won't retry, really, it will just let you redo the templatedata string
[19:26:07] 	 i think it's a bad idea to hack the current error system
[19:26:15] 	 that makes sense? I think that's sensible user behavior; the dialog tries to parse -- if it fails, you have the option either to get out of it and manually fix what is broken or start from scratch.
[19:26:20] 	 we would be better off adding a notice system
[19:26:31] 	 Hmm
[19:26:53] 	 esp. since we could notify the user asynchronously (using a popup button widget with a warning icon that opens the popup automatically or something)
[19:26:55] 	 Hmm couldn't you factor this out of the dialog?
[19:27:13] 	 we don't HAVE to block the user on it
[19:27:13] 	 it's just a notice
[19:27:13] 	 I could, but it's a little elaborate
[19:27:18] 	 Have the dialog's setup take an object rather than a string
[19:27:18] 	 Then there has to be some code that parses the JSON in the box and feeds it to the dialog setup
[19:27:46] 	 RoanKattouw, so, the other idea I had was to do that -- to parse outside the dialog, and then if there's an issue, open a MessageDialog
[19:27:50] 	 If that code encounters a problem, it could launch a MessageDialog asking the user what to do, and then launch the TemplateDataDialog with obj={} 
[19:27:51] 	 I could do that.
[19:27:55] 	 yeah
[19:28:16] 	 It's just a bit more elaborate; I was thinking a notice or something would be easier. But I can totally work with an intermediate dialog
[19:28:18] 	 I don't think we need to worry about existing production pages using illegal new lines. 
[19:28:22] 	 Users can fix that.
[19:28:30] 	 mooeypoo: what Roan is saying makes some sense, since the way they fix is is outside the dialog
[19:28:32] 	 right?
[19:28:51] 	 Krinkle, it's not just the illegal new lines, though -- we need to account for *any* case where the templatedata string is invalid, including a case where the user edited it manually first (even by mistake) and opened the dialog.
[19:28:55] 	 I wouldn't normally, but this seems quite rare.
[19:29:01] 	 TrevorParscal, yeah
[19:29:20] 	 mooeypoo: Yeah, but it has (or shoudl have) that already for when the user edits the wikitext first in the editor and makes it invalid.
[19:29:27] 	 So the user goes back in and fixes it.
[19:29:46] 	 Krinkle, yes, but the message it displays at the top of the dialog is insufficient imho
[19:29:47] 	 Or you could convert new lines to \\n maybe and help them in that regard.
[19:29:53] 	 mooeypoo: What does it say?
[19:30:13] 	 But from now on the PHP side should reject new lines. 
[19:30:33] 	 it says "Bad JSON format. Either correct it, or delete the current  tags and try again." and presents an empty dialog
[19:30:38] 	 the issue is that it presents an empty dialog
[19:30:56] 	 Users may miss that message/banner
[19:31:22] 	 I think it might be better user experience to make a message dialog with the same text and then let the user decide f they want to continue to the editor -- but then create a whole new templatedata or to cancel and go fix it
[19:32:02] 	 Yeah, they should consciously decide to init the dialog or go back to the text editor
[19:32:17] 	 Krinkle, yeah, I can normalize myself -- but I am also worried about other sort of errors since the string is completely editable even if it shouldn't be edited.
[19:34:56] 	 10VisualEditor, 10VisualEditor-MediaWiki: VisualEditor: "Unknown error" on saving on officewiki - https://phabricator.wikimedia.org/T70235#1096566 (10Krenair) I fixed it up in the remaining places: ```commonswiki         MediaWiki:LAPI.js dewikiversity       MediaWiki:LAPI.js enwiki              MediaWiki:LAPI...
[19:35:10] 	 10VisualEditor: Change MediaWiki:Visualeditor-feedback-tool to say "Leave feedback about this software" - https://phabricator.wikimedia.org/T91804#1096567 (10Whatamidoing-WMF) 3NEW
[19:35:40] 	 10VisualEditor: "You are not logged in." warning cannot be closed in Safari on iOS - https://phabricator.wikimedia.org/T91757#1096576 (10Etonkovidova) Confirmed on iPad iOS7.1 mini with Mobile Safari7
[19:35:52] 	 mooeypoo: I think the point is that you should check for valid input first, show a separate message dialog if it's a problem, and then continue to the normal dialog if all is going well
[19:36:07] 	 one of the options in the message dialog could be "continue anyway"
[19:36:16] 	 you can have as many options as you like in a message dialog
[19:36:22] 	 TrevorParscal, yeah. That sounds reasonable.
[19:36:29] 	 and you don't have to subclass the message dialog, so it should be simple
[19:36:47] 	 yeah I used that for the feedback dialog, it should be easy
[19:37:13] 	 the harder part is to switch the entire initialization out of the dialog, but that's not really hard either, just requires a bit of reorganization.
[19:37:15] 	 sounds good
[19:37:20] 	 10VisualEditor, 10MediaWiki-General-or-Unknown, 10RESTBase, 6Services, and 3 others: VisualEditor performance: Use RESTBase from the MediaWiki Virtual Rest Service - https://phabricator.wikimedia.org/T89066#1096578 (10Eloquence)
[19:38:17] 	 mooeypoo: i think this will be a better solution than shoehorning it into the error system
[19:39:30] 	 TrevorParscal, yeah 
[19:39:32] 	 I agree
[19:39:56] 	 TrevorParscal, it's also going to be a better solution in general for the templatedata dialog itself. Have all the validation done outside of it is a good idea
[19:41:41] 	 cool
[19:50:26] 	 10VisualEditor, 6Collaboration-Team, 10Flow: Flow support for VisualEditor's mediawiki.feedback form - https://phabricator.wikimedia.org/T91805#1096593 (10Quiddity) 3NEW
[19:50:57] 	 10VisualEditor, 6Collaboration-Team, 10Flow: Flow support for VisualEditor's mediawiki.feedback form - https://phabricator.wikimedia.org/T91805#1096605 (10Jdforrester-WMF) This is just {T59989}.
[19:52:19] 	 (03CR) 10Jforrester: [C: 032] Add composer entry point for phpcs and linting [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/194910 (https://phabricator.wikimedia.org/T90943) (owner: 10Legoktm)
[19:57:09] 	 (03Merged) 10jenkins-bot: Add composer entry point for phpcs and linting [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/194910 (https://phabricator.wikimedia.org/T90943) (owner: 10Legoktm)
[19:57:35] 	 10VisualEditor, 10MediaWiki-General-or-Unknown, 10RESTBase, 6Services, and 3 others: VisualEditor performance: Use RESTBase from the MediaWiki Virtual Rest Service - https://phabricator.wikimedia.org/T89066#1096628 (10Jdforrester-WMF) This is about all Parsoid HTML consumers of the MediaWiki core service,...
[19:58:34] 	 Krenair: Quick merge of https://gerrit.wikimedia.org/r/#/c/194760/ and https://gerrit.wikimedia.org/r/#/c/194763/ ? :-)
[20:00:14] 	 (03CR) 10Jforrester: [C: 031] Schema:Edit instrumentation [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/191221 (https://phabricator.wikimedia.org/T88027) (owner: 10Alex Monk)
[20:00:29] 	 James_F, I have no idea what we're fixing in either of these
[20:00:49] 	 mooeypoo: ok, so I'm installing citoid, but https://www.mediawiki.org/wiki/Citoid#Run_the_server the first step in this section makes no sense
[20:01:07] 	 10VisualEditor: "You are not logged in." warning cannot be closed in Safari on iOS - https://phabricator.wikimedia.org/T91757#1096633 (10Etonkovidova) p:5Triage>3Low
[20:01:09] 	 or, i guess it does
[20:01:17] 	 crap, so there are 2 servers?
[20:01:23] 	 Krenair: One of them is to stop npm whining about out-of-date-ness; the other is a pull-through of a fix to unbreak the styling section of the toolbar in VE-MW.
[20:02:30] 	 where was it whining about that?
[20:03:31] 	 Krenair: `npm outdated`.
[20:04:10] 	 Krenair: `npm outdated --depth=0` actually unless you want to cry.
[20:04:28] 	 TrevorParscal: Yes
[20:04:36] 	 TrevorParscal: Zotero / Translation server, and the Citoid server
[20:05:46] 	 dude, i'm like 5 installs deep now, not only do I have to install the translation server, but I have to install a specific version of XULRunner and that has incompatibility with my mac, so I have to install macports, and then install libidl
[20:06:02] 	 this makes me sad
[20:06:08] 	 RoanKattouw: You have anything further to add to https://etherpad.wikimedia.org/p/4pKxno7DAI ?
[20:06:49] 	 I think my install of VE is broken :/
[20:06:57] 	 Krenair: Oops.
[20:07:06] 	 TrevorParscal: Yeah, it's a mess.
[20:07:52] 	 Uncaught TypeError: Expecting a function in instanceof check, but got undefined
[20:08:04] 	 oh, need to update core I guess
[20:08:24] 	 TrevorParscal: Yeah isn't citoid fun
[20:08:35] 	 TrevorParscal: No wonder I did such a shoddy job of puppetizing it
[20:08:57] 	 Alexandros was unhappy with what I'd done but he slowly became more understanding and forgiving as he got lost in those weeds
[20:09:07] 	 :-)
[20:09:09] 	 I mean what I did was still horrible, but at least he forgave me for it :D
[20:09:18] * James_F grins.
[20:09:32] 	 RoanKattouw: *cough* https://etherpad.wikimedia.org/p/4pKxno7DAI *cough* … sending in 5 minutes.
[20:09:53] 	 Yeah looking
[20:10:25] 	 I don't know of any others
[20:10:46] 	 That's fine. :-)
[20:10:51] 	 (03CR) 10Bartosz Dziewoński: [C: 032] Update VE core submodule to master (300b89a) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194763 (owner: 10Jforrester)
[20:10:56] 	 Thanks MatmaRex.
[20:12:06] 	 TrevorParscal, I've installed the deployment server, so I didn't go through all the 'from scratch' steps (I probably should at some point)
[20:12:16] 	 but yeah, you always have to run 2 servers, the translation server and the citoid server
[20:12:20] 	 (03CR) 10Alex Monk: [C: 032] Update VE core submodule to master (300b89a) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194763 (owner: 10Jforrester)
[20:12:26] 	 ...
[20:12:42] 	 (03PS1) 10Legoktm: jenkins job validation, do not submit [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/194940 
[20:12:58] 	 James_F, is `npm outdated` supposed to return a blank line to me?
[20:13:08] 	 (03Merged) 10jenkins-bot: Update VE core submodule to master (300b89a) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/194763 (owner: 10Jforrester)
[20:13:10] 	 Krenair: If it's up-to-date, yes.
[20:13:21] 	 Krenair: Observe the output on master for comparison.
[20:13:28] 	 I am observing the output on master.
[20:13:36] 	 Hmm.
[20:13:37] 	 alex@alex-laptop:~/Development/MediaWiki/extensions/VisualEditor (master)$ npm outdated
[20:13:37] 	 alex@alex-laptop:~/Development/MediaWiki/extensions/VisualEditor (master)$ 
[20:13:49] 	 (but with a blank line in the middle)
[20:14:09] 	 $ npm outdated --depth=0
[20:14:09] 	 Package     Current  Wanted  Latest  Location
[20:14:09] 	 grunt-jscs    1.2.0   1.2.0   1.5.0  grunt-jscs
[20:14:29] 	 (03CR) 10Legoktm: "recheck" [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/194940 (owner: 10Legoktm)
[20:14:47] 	 (03Abandoned) 10Legoktm: jenkins job validation, do not submit [extensions/TemplateData] - 10https://gerrit.wikimedia.org/r/194940 (owner: 10Legoktm)
[20:14:49] 	 10VisualEditor: Extra text format icon appearing inside the text formatting menu, clicking on which does not expand to show more options but closed the menu - https://phabricator.wikimedia.org/T91809#1096679 (10Ryasmeen) 3NEW
[20:14:51] 	 James_F, do I need to run any particular command to set it up?
[20:15:03] 	 Krenair: npm install first, maybe?
[20:15:14] 	 10VisualEditor: Extra text format icon is appearing inside the text formatting menu, clicking on which does not expand to show more options but closes the menu - https://phabricator.wikimedia.org/T91809#1096686 (10Ryasmeen)
[20:15:34] 	 That new task ryasmeen posted sounds like the bug that was just fixed
[20:15:44] 	 10VisualEditor, 10VisualEditor-MediaWiki: [Regression pre-wmf21] Extra text format icon is appearing inside the text formatting menu, clicking on which does not expand to show more options but closes the menu - https://phabricator.wikimedia.org/T91809#1096688 (10Jdforrester-WMF) a:3Jdforrester-WMF
[20:15:56] 	 10VisualEditor, 10VisualEditor-MediaWiki: [Regression pre-wmf21] Extra text format icon is appearing inside the text formatting menu, clicking on which does not expand to show more options but closes the menu - https://phabricator.wikimedia.org/T91809#1096679 (10Jdforrester-WMF) Fixed by https://gerrit.wikimed...
[20:16:02] 	 10VisualEditor, 10VisualEditor-MediaWiki: [Regression pre-wmf21] Extra text format icon is appearing inside the text formatting menu, clicking on which does not expand to show more options but closes the menu - https://phabricator.wikimedia.org/T91809#1096692 (10Jdforrester-WMF) 5Open>3Resolved
[20:16:11] 	 oh cool! :)
[20:16:16] 	 Yup.
[20:16:22] 	 Fast, aren't we?
[20:16:34] 	 hehehe yeah you are!
[20:18:04] 	 James_F, tried that, no luck
[20:18:22] 	 Krenair: Hmm. Did you run with ` --depth=0`?
[20:19:33] 	 both with and without
[20:19:43] 	 Hmm.
[20:19:45] 	 Odd.
[20:19:54] 	 Maybe Windows npm doesn't work very well?
[20:20:02] 	 i'm getting errors when I hit the parsoid service
[20:20:07] 	 You're running windows?
[20:20:20] 	 No, but aren't you?
[20:20:24] 	 No.
[20:20:36] 	 Oh. OK, well, I don't know. What version of npm are you running?
[20:20:38] 	 I very rarely run windows for work.
[20:20:43] 	 i notice the instructions say node v0.10, i'm using v0.11.14 - is that a known issue?
[20:21:05] 	 1.3.10
[20:23:16] 	 RoanKattouw: http://pastebin.com/uwrF1w4W
[20:23:23] 	 any ideas?
[20:23:24] 	 Krenair: Oh. I'm runnning 2.1.14.
[20:23:31] 	 Hm.
[20:23:34] 	 Krenair: Maybe updating npm would help?
[20:23:39] 	 Probably.
[20:23:51] 	 "string is not a function"? WTF?
[20:24:26] 	 TrevorParscal: I've never seen that before
[20:25:12] 	 Oh, I understand what's going on
[20:25:23] 	 It's doing callback(...); where callback is a string
[20:25:38] 	 That's in the HTTP 520 code path
[20:25:45] 	 "Unable to scrape resource"
[20:26:30] 	 TrevorParscal: Try git revert f59613eeb332ebac11a0832d2dcfc79d8ca5777d
[20:26:45] 	 Looking at the diff of Marielle's latest change it seems very plausible that that broke it
[20:26:46] 	 James_F, seems to be the latest version available from the ubuntu repositories
[20:26:59] 	 Krenair: Oh dear.
[20:27:43] 	 RoanKattouw: yes, seems to have fixed it
[20:28:00] 	 Yeah her commit is broken
[20:28:09] 	 She changed a function signature but didn't quite change it everywhere
[20:28:35] 	 Krenair: npm 2.0.0 was released on 2014-09-12; master is 2.7.1.
[20:28:46] 	 (03CR) 10Catrope: "FIXME: this causes the following error: http://pastebin.com/uwrF1w4W" [services/citoid] - 10https://gerrit.wikimedia.org/r/193154 (owner: 10Mvolz)
[20:28:52] 	 That's probably why
[20:30:36] 	 10VisualEditor: [Regression pre-wmf21] Tool-Format menu does not display all options - https://phabricator.wikimedia.org/T91772#1096726 (10matmarex)
[20:30:37] 	 10VisualEditor, 10VisualEditor-MediaWiki: [Regression pre-wmf21] Extra text format icon is appearing inside the text formatting menu, clicking on which does not expand to show more options but closes the menu - https://phabricator.wikimedia.org/T91809#1096725 (10matmarex)
[20:31:13] 	 10VisualEditor, 10VisualEditor-MediaWiki: [Regression pre-wmf21] Tool-Format menu does not display all options - https://phabricator.wikimedia.org/T91772#1096728 (10Jdforrester-WMF) 5Open>3Resolved
[20:31:17] 	 10VisualEditor, 10VisualEditor-MediaWiki: [Regression pre-wmf21] Tool-Format menu does not display all options - https://phabricator.wikimedia.org/T91772#1095930 (10Jdforrester-WMF) Still fixed.
[20:31:33] 	 10VisualEditor, 10VisualEditor-MediaWiki: [Regression pre-wmf21] Tool-Format menu does not display all options - https://phabricator.wikimedia.org/T91772#1096732 (10matmarex) I still see the issue. Either it didn't work or beta is not updating.
[20:32:36] 	 10VisualEditor: "You are not logged in." warning cannot be closed in Safari on touch devices - https://phabricator.wikimedia.org/T91757#1096735 (10matmarex)
[20:33:18] 	 10VisualEditor: "You are not logged in." warning cannot be closed in Safari on touch devices - https://phabricator.wikimedia.org/T91757#1095543 (10matmarex) I can reproduce this on a Windows 8.1 tablet.  >>! In T91757#1096576, @Etonkovidova wrote: > Note: another issue (possibly, not related) - all drop down men...
[20:36:03] 	 10OOjs-UI, 10VisualEditor: Pop-up "You are not logged in." warning cannot be closed via the "X" on touch devices - https://phabricator.wikimedia.org/T91757#1096742 (10Jdforrester-WMF) p:5Low>3Normal
[20:37:15] 	 (03PS1) 10Kmenger: MessageDialog: Add description, example, and mark private methods [oojs/ui] - 10https://gerrit.wikimedia.org/r/194942 
[20:40:54] 	 mooeypoo: ok, i have citoid running an everything, but I'm seeing some issue in the inspector, like CSS might not be loading properly?
[20:41:00] 	 the results are not styled properly
[20:41:30] 	 mh let me try again
[20:41:36] 	 it was working a couple of days ago, but let me verify
[20:42:01] 	 http://cl.ly/image/2u3e1w2G0g0z
[20:42:47] 	 O.o eww
[20:42:53] 	 let me check
[20:44:02] 	 what the hell
[20:44:40] 	 the icon is supposed to be on top
[20:45:07] 	 brb
[20:45:20] 	 did we change the way icons are attached in option widgets lately in ooui?
[20:45:37] 	 ed changed some icon CSS
[20:45:37] 	 I blame Ed.
[20:45:44] 	 :-D
[20:45:44] 	 ha
[20:45:59] 	 but probably not option widgets itself, unless you're doing something bad to them! :>
[20:46:01] 	 well it was "top: 1.5em" and now it should be top: -2em
[20:46:10] 	 meh
[20:46:21] 	 MatmaRex: TextInputWidget got icon: and indicator: elements removed in 0.9.0 or something.
[20:46:41] 	 the associated events only. and that one is my doing.
[20:46:53] 	 Oh, OK.
[20:47:07] 	 Hmm. Bad form if we had a breaking change without announcing. :-(
[20:47:32] 	 it wasn't *really* breaking
[20:47:34] 	 ok this is frustrating. The location of the icon changes with the size of the widget.
[20:47:41] 	 (as in, no one was using that)
[20:47:59] 	 Apparently they were. :-)
[20:48:34] 	 oh, the icon/indicator background size thing
[20:48:39] 	 mehhhh
[20:48:42] 	 so a small reference seems to push the icon up, so top: -2em puts it waaaay up the widget
[20:49:04] 	  a large reference puts it way down, so it needs top: -2em to work and be at the top
[20:49:10] 	 meh
[20:49:18] 	 mooeypoo: i think it now tries to center the icon, or something
[20:49:27] 	 MatmaRex, how do I get rid of it
[20:49:28] 	 :P
[20:49:34] 	 or to match the size of the container
[20:49:38] 	 background-size: 24px 24px;
[20:49:48] 	 background-position: …uhh, 'top left' probably?
[20:50:10] 	 10VisualEditor, 10VisualEditor-MediaWiki: [Regression pre-wmf21] Tool-Format menu does not display all options - https://phabricator.wikimedia.org/T91772#1096758 (10matmarex) (Works now.)
[20:51:14] 	 there y'are, icon
[20:51:24] 	 meh it's still moving around 
[20:52:07] 	 TrevorParscal: i need second opinion on https://phabricator.wikimedia.org/T91152 before i spend tons of time un-up-screwing the situation
[20:54:09] 	 TrevorParscal, MatmaRex okay, I'm going to eat, I'll brb, but the reason the cite inspector looks like crap is because the icon keeps moving... I changed "top" to -2em and it fixed the icon position for larger citations, but broke it for smaller ones (like in TrevorParscal's image) but then 1.2em which works for Trevor's example fails in the larger ones. I am stumped. 
[20:54:16] 	 I added sizing to it, too. It doesn't work.
[20:54:17] 	 :(
[20:54:19] 	 brb
[21:05:30] 	 (03CR) 10Krinkle: [C: 032] MessageDialog: Add description, example, and mark private methods [oojs/ui] - 10https://gerrit.wikimedia.org/r/194942 (owner: 10Kmenger)
[21:07:30] 	 (03Merged) 10jenkins-bot: MessageDialog: Add description, example, and mark private methods [oojs/ui] - 10https://gerrit.wikimedia.org/r/194942 (owner: 10Kmenger)
[21:09:32] 	 (03PS1) 10Krinkle: build: Remove obsolete 'build' task from grunt-doc [oojs/ui] - 10https://gerrit.wikimedia.org/r/194946 
[21:10:15] 	 10VisualEditor: Add a "Done" button or a way to slide up the special character menu, right now its not apparent that clicking on the menu icon again will close that dialog - https://phabricator.wikimedia.org/T91814#1096820 (10Ryasmeen)
[21:13:01] 	 Do we have a page explaining how story points are allocated?
[21:13:39] 	 MatmaRex: The doc_parser functions as a way to validate @ keywords
[21:14:04] 	 My go-to example to test if jsduck error returns exit code 1 properly is by making a bad @keywrd, but that now fails earlier in the doc parser
[21:14:07] 	 in oojs-ui :D
[21:14:19] 	 Krinkle: if by "failing with an obscure error message" you mean "validate", yes. ;)
[21:14:53] 	 Krinkle: use @see, docparser supports it but jsduck doesn't.
[21:15:17] 	 Yeah, >> /Users/krinkle/Development/oojs-ui/bin/docparser.rb:156:in `block (2 levels) in parse_file': unrecognized keyword: absrc (RuntimeError) 
[21:15:38] 	 oh, that's actually less obscure that i suspected
[21:16:03] 	 most doc syntax errors because it to complain that something was nil :(
[21:16:06] 	 cause it*
[21:27:43] 	 (03PS1) 10Krinkle: build: Move pre/post 'doc' task into package.json [oojs/ui] - 10https://gerrit.wikimedia.org/r/194956 
[21:31:28] 	 Hi. Can someone suggest a place where I can get help troubleshooting a new install of VE?
[21:31:32] 	 10VisualEditor: Bumping up/down CE while opening/closing a special character makes it hard to find where the cursor position was for editing - https://phabricator.wikimedia.org/T91818#1096854 (10Ryasmeen) 3NEW
[21:31:47] 	 10VisualEditor: Bumping up/down CE while opening/closing a special character menu makes it hard to find where the cursor position was for editing - https://phabricator.wikimedia.org/T91818#1096861 (10Ryasmeen)
[21:35:08] 	 10VisualEditor: Bumping up/down CE while opening/closing a special character menu makes it hard to find where the cursor position was for editing - https://phabricator.wikimedia.org/T91818#1096864 (10Ryasmeen) However, moving down CE makes sense only when the cursor is at some position where the dialog might ove...
[21:37:22] 	 kirstenMA: James_F|Away: OOjs UI @example now works in production 
[21:37:22] 	 https://doc.wikimedia.org/oojs-ui/master/#!/api/OO.ui.Dialog
[21:37:53] 	 Awesome!
[21:38:26] 	 10VisualEditor: Bumping up/down CE while opening/closing a special character menu makes it hard to find where the cursor position was for editing - https://phabricator.wikimedia.org/T91818#1096867 (10Ryasmeen)
[21:38:48] 	 kirstenMA: Do you run it locally as well? If so, what do you use?
[21:39:16] 	 Run @example? 
[21:39:29] 	 kirstenMA: Yeah, do you install jsduck locally to preview it?
[21:39:32] 	 Yes
[21:39:39] 	 What command do you use to run it
[21:39:42] 	 I have jsduck locally.
[21:39:48] 	 10VisualEditor, 10VisualEditor-Tables, 3Roadmap: Be able to re-order the columns or rows of a table by dragging them to another position in VisualEditor - https://phabricator.wikimedia.org/T88694#1096872 (10Eloquence) When is this expected to land?
[21:39:55] 	 I've been using 'grunt doc'
[21:40:07] 	 10VisualEditor: Bumping up/down CE while opening/closing special character dialog makes it hard to find where the cursor position was for editing - https://phabricator.wikimedia.org/T91818#1096873 (10Ryasmeen)
[21:40:09] 	 kirstenMA: Okay. That's going to change in a few minutes.
[21:40:32] 	 okay. thanks for the heads up
[21:40:34] 	 kirstenMA: It'll now be 'npm run-script doc'. This already works now, but 'grunt doc' will stop working after today.
[21:40:56] 	 got it. 
[21:41:17] 	 mookie: nowhere else than here. what is your question?
[21:42:15] 	 Hey thanks. I've got VE installed on a private wiki
[21:42:48] 	 Edit tab appears, I can click on it, and I see the animated progress bar, but WYSIWYG editor never appears
[21:42:58] 	 I can see that Parsoid is working on the server
[21:43:19] 	 API call is going through, with full content returned and parsed
[21:43:46] 	 I set the cookie forwarding for Parsoid
[22:00:43] 	 mooeypoo: status of that patch?
[22:01:32] 	 TrevorParscal, I am at a loss. Something changed with the icons where now they are moving with the element, which means that setting a position for them (top: 1.2em) which worked before doesn't work now
[22:01:42] 	 mookie: Do any errors appear in your browser console?
[22:01:53] 	 for small citations (like in your picture) we actually need -2em now
[22:01:57] 	 (The console that appears when you right-click the page and choose "inspect element")
[22:02:09] 	 and for "regular" ones (that parse completely) it seems more like +1em
[22:02:18] 	 TrevorParscal, somewhere, I'm missing something, but I have no idea what since the icons changed
[22:02:20] 	 MatmaRex, help
[22:02:21] 	 mooeypoo: Is this because of background-size: cover; perhaps?
[22:02:27] 	 mooeypoo: I'm pretty sure there's more wrong with that patch than the icon though
[22:02:40] 	 also, NEVER use background position to position an icon plz
[22:02:49] 	 it's super limited and causes all sorts of issues
[22:02:55] 	 Yeah background-position should almost always be center center
[22:02:56] 	 TrevorParscal, it isn't the background position, its the position of the icon SPAN itself
[22:02:59] 	 not to mention is doesn't work properly in RTL
[22:03:04] 	 mooeypoo: ok
[22:03:05] 	 good
[22:03:10] 	 but the span is moving now
[22:03:16] 	 sorry... I didn't mean to accuse you of being evil
[22:03:29] 	 It's okay, TrevorParscal, evil in RTL is live
[22:03:29] 	 mooeypoo: well, are you seeing the same thing my picture showed?
[22:03:37] 	 TrevorParscal, I am
[22:03:44] 	 but if you try for another link, the effect is different
[22:03:48] 	 mooeypoo: which patch is it, and is it live somewhere?
[22:03:55] 	 TrevorParscal, try to lookup this link: http://www.nytimes.com/2015/03/06/arts/design/review-bjork-unfurled-in-many-guises-at-moma.html?ref=arts
[22:04:05] 	 MatmaRex: be warned, you have to get citoid going first
[22:04:11] 	 i can't check it out because either gerrit or my connection is fubar again
[22:04:14] 	 MatmaRex, https://gerrit.wikimedia.org/r/#/c/190973/
[22:04:22] 	 and yes, citoid is needed :\
[22:04:29] 	 I can do a hangout and share my screen...
[22:04:57] 	 mooeypoo: http://cl.ly/image/452A3C462z2h
[22:05:01] 	 doesn't seem any better
[22:05:07] 	 mooeypoo: i recommend http://localtunnel.me/
[22:05:13] 	 TrevorParscal, it isn't, but the icon isn't in the same place either
[22:05:24] 	 TrevorParscal, so when i fix it to the previous image, i get it screwed up for the new one
[22:05:29] 	 i'm doing something wrong with the css
[22:05:44] 	 wait, why is this empty?
[22:06:27] 	 "Empty citation (help)" is the rendering
[22:06:37] 	 and News is HUGE
[22:06:41] 	 and the icon is misplaced
[22:06:48] 	 10VisualEditor, 7I18n: Regression: Using Unicode Combining diacritics duplicates/replaces/adds unwanted characters - https://phabricator.wikimedia.org/T85093#1096937 (10Etonkovidova) p:5Triage>3Low
[22:06:48] 	 so, you have a few issues there perhaps?
[22:07:34] 	 TrevorParscal, I added background-size: 24px 24px to the icon like MatmaRex suggested, and then corrected the "top" value to -2em. I see this in my reference: http://postimg.org/image/4p7j617mp/
[22:07:36] 	 which looks good
[22:07:48] 	 but then when I take something that won't parse, I see this: http://postimg.org/image/uwlhti24x/
[22:08:22] 	 TrevorParscal, I don't see the news as that huge?
[22:08:41] 	 RoanKattouw: Yes: "Uncaught ReferenceError: importArticles is not defined" but that appears before the user clicks the Edit tab
[22:08:44] 	 they're about the size of the icon, which I thoght was more or less what it should be? I can make it smaller, but I don't know if see what you see
[22:08:47] 	 mooeypoo: it seems like it should be smaller, maybe just bold, see the latest design work
[22:08:55] 	 mooeypoo: make your wiki available via localtunnel or something like it, and i'll debug
[22:09:10] 	 MatmaRex, localtunnel?
[22:09:22] 	 mooeypoo: [23:05]	MatmaRex	mooeypoo: i recommend http://localtunnel.me/
[22:09:26] 	 oh
[22:09:29] 	 missed that
[22:10:53] 	 TrevorParscal, hm, I thought I did go by the design work, I'll take another look. Was it changed or was I just blind...?
[22:11:07] 	 hmm
[22:12:54] 	 mooeypoo: the design looks like it should be similar size to the button text (1em, bold) and the rendering should be smaller than that
[22:13:08] 	 you have 1.5em on the title
[22:13:32] 	 I do, yeah. For some reason I remember talking about this as if we're "mocking" a fieldset label/title
[22:13:33] 	 hm
[22:13:36] 	 I'll fix that
[22:13:36] 	 10VisualEditor: Cursor easily gets placed in the line behind the special character menu and adds characters there - https://phabricator.wikimedia.org/T91823#1096955 (10Ryasmeen) 3NEW
[22:13:42] 	 MatmaRex, https://wuvfwkzuqz.localtunnel.me
[22:13:43] 	 mooeypoo: no worries
[22:14:32] 	 MatmaRex, okay that's not working, hang on
[22:15:24] 	 MatmaRex, check if this works: https://tknyvqlugn.localtunnel.me/wiki
[22:16:18] 	 omg, having a cold is so crap
[22:16:28] 	 https://tknyvqlugn.localtunnel.me/wiki/index.php/Main_Page seems to kind of work
[22:16:29] 	 mooeypoo: nope, your wiki redirects to 'localhost' :)
[22:16:37] 	 oh, hah
[22:16:37] 	 But I think you might have $wgServer = 'localhost' issues
[22:16:43] 	 i need to lie down for a bit, I'll be back in a bit when my head stops feeling like it's literally exploding
[22:16:44] 	 eh
[22:16:46] 	 brb
[22:16:50] 	 yeah, that's easy to fix though
[22:16:58] 	 TrevorParscal, feel better!
[22:17:12] 	 Cause it tries to hit http://localhost/wiki/load.php
[22:17:34] 	 10VisualEditor: [Regression pre-wmf21] The Insert option remains enabled in the menu when the table cell is selected for no reason - https://phabricator.wikimedia.org/T91826#1096986 (10Ryasmeen) 3NEW
[22:17:37] 	 mooeypoo: $wgServer = "http://" . ( isset( $_SERVER['HTTP_HOST'] ) ? $_SERVER['HTTP_HOST'] : 'localhost' );
[22:17:37] 	 k, well, what do I change it to seeing as I don't have my own dns?
[22:17:43] 	 ah
[22:18:19] 	 MatmaRex, try now
[22:18:47] 	 yep, that works. okay.
[22:19:41] 	 Cite -> Autofill from URL -> use this url as an example: http://www.nytimes.com/2015/03/06/arts/design/review-bjork-unfurled-in-many-guises-at-moma.html?ref=arts
[22:19:55] 	 MatmaRex, and you can also use "blabla" or something as a lookup example of a failed fetch
[22:19:57] 	 mooeypoo: your wiki is using debug mode, it's still loading for me ;)
[22:20:07] 	 MatmaRex, want me to remove debug mode? 
[22:20:25] 	 nah, just gonna wait, it's going to be quicker the next time
[22:20:26] 	 (you can also do debug=false)
[22:20:32] 	 oh yeah, i guess
[22:20:44] 	 Oh hah I guess you can
[22:20:46] 	 I made the title smaller anyways, but the icon is weird
[22:21:00] 	 mooeypoo: Thank you for teaching me about a feature I didn't realize existed in code I wrote over 4 years ago :D
[22:21:07] 	 :D
[22:21:13] 	 anytime, RoanKattouw \o/
[22:21:20] 	 10VisualEditor, 10VisualEditor-EditingTools: VisualEditor: The hand icon remains active after opening the Media Settings dialog and Reference dialog which selects some part inside the dialog upon double clicking - https://phabricator.wikimedia.org/T65398#1097008 (10Etonkovidova) Re-checked - non-reproducible a...
[22:21:30] 	 10VisualEditor, 10VisualEditor-EditingTools: VisualEditor: The hand icon remains active after opening the Media Settings dialog and Reference dialog which selects some part inside the dialog upon double clicking - https://phabricator.wikimedia.org/T65398#1097009 (10Etonkovidova) 5Open>3Resolved
[22:21:50] 	 also MatmaRex this is the design https://phabricator.wikimedia.org/T62768
[22:22:00] 	 so see, the icon should be aligned with the title "News" or "Journal" or whatever else
[22:22:36] 	 mooeypoo: bah, it's sending API queries to localhost:1970
[22:22:53] 	 oh
[22:22:54] 	 crap
[22:23:03] 	 (citoid api)
[22:23:03] 	 hang on let me use the same trick in citoid too
[22:23:47] 	 since it's client-side, then you can probably make a second localtunnel, and point wgCitoidServiceUrl to that
[22:24:07] 	 MatmaRex, crap, the citoid config is js. Let me check again how to get the server address from url in js
[22:24:34] 	 10VisualEditor: The option for Header cell/Content cell in a table does not retain the check mark next to it , that appears initially while selecting one of them - https://phabricator.wikimedia.org/T91827#1097018 (10Ryasmeen) 3NEW
[22:25:11] 	 mooeypoo: it's probably simpler to do it on my side. what is the output of the api query supposed to be?
[22:25:32] 	 for http://localhost:1970/api?action=query&format=mediawiki&search=http%3A%2F%2Fwww%2Enytimes%2Ecom%2F2015%2F03%2F06%2Farts%2Fdesign%2Freview-bjork-unfurled-in-many-guises-at-moma%2Ehtml%3Fref%3Darts
[22:26:37] 	 MatmaRex, http://pastebin.com/5Vadw50f
[22:27:19] 	 MatmaRex, but if you try it now, it should work. I switched the localhost in citoid service
[22:28:35] 	 oh, okay
[22:30:54] 	 "I can't tell why this is here. I wonder if any tests break if I comment this out"
[22:30:54] * RoanKattouw is having fun with TransactionProcessor
[22:30:54] 	 None of them :O
[22:30:54] 	 One of my gravest concerns was in a piece of code I appear to be able to just delete
[22:30:54] 	 edsanders: I have a transaction question for ya
[22:30:55] 	 Say I have 
Foo
 and I process the following transaction: retain 1, startAnnotating bold, retain 1, replace ['o'] with ['x'], retain 1, stopAnnotating bold, retain 1
[22:31:19] 	 edsanders: What is the result? A) [ { type: 'paragraph' }, ['F', bold], ['x', bold], ['o', bold], { type: '/paragraph' } ]   B) [ { type: 'paragraph' }, ['F', bold], 'x', ['o', bold], { type: '/paragraph' } ]   C) Why are you nesting annotation and replace operations, that's cray
[22:31:22] 	 *crazy
[22:31:38] 	 hmm, still using localhost:1970 for me?
[22:35:16] 	 O.o
[22:35:18] 	 damnit
[22:36:10] 	 oh
[22:36:11] 	 haha
[22:41:16] 	 MatmaRex, okay, it's not working
[22:41:35] 	 (03PS2) 10Nirzar: Button styles between OOJS and MW [oojs/ui] - 10https://gerrit.wikimedia.org/r/194753 
[22:41:38] 	 i don't know how to set it up to work for that :\ i gave it the temporary host and it's still not working
[22:43:26] 	 10VisualEditor: [Regression wmf 20] Table gets broken after changing the cell type for a merge cell operation - https://phabricator.wikimedia.org/T91831#1097091 (10Ryasmeen) 3NEW
[22:44:07] 	 10VisualEditor: [Regression wmf 20] Table gets broken while changing the cell type after a merge cell operation - https://phabricator.wikimedia.org/T91831#1097099 (10Ryasmeen) p:5Triage>3High
[22:46:55] 	 mooeypoo: okay, i got it
[22:47:20] 	 10VisualEditor: [Regression wmf 20] Table gets broken while changing the cell type after a merge cell operation - https://phabricator.wikimedia.org/T91831#1097114 (10Catrope) a:3Esanders
[22:48:00] 	 mooeypoo: hmm, so
[22:48:31] 	 TrevorParscal, is there an event that a messageDialog emits for a primary action? the 'safe' action dismisses, which is fine, but without extending a MessageDialog how can I respond to the primary action?
[22:48:40] 	 10VisualEditor: [Regression pre-wmf20] The transition while switching between Readmode and VE is not smooth, the whole block of content repeats as a flicker below the actual content - https://phabricator.wikimedia.org/T91442#1097117 (10Etonkovidova) I saw the issue on FF35, Chrome40, and Safari7. And I noticed a...
[22:49:08] 	 mooeypoo: is looking like this okay? Access-Control-Allow-Origin
[22:49:09] 	 bah
[22:49:28] 	 like this: http://i.imgur.com/BdLCvSk.png
[22:49:59] 	 MatmaRex, no
[22:50:11] 	 MatmaRex, https://phabricator.wikimedia.org/T62768
[22:50:20] 	 look at the first image
[22:50:41] 	 see how in the "Confirms citation" the title is near the icon, and the citation is underneath
[22:50:43] 	 10VisualEditor, 10VisualEditor-MediaWiki, 10VisualEditor-Performance, 3VisualEditor 2014/15Q3 blockers: Load and display the welcome dialog straight after the toolbar, without waiting for the content area - https://phabricator.wikimedia.org/T90454#1097125 (10Krenair) I had a go at this, but it seems to con...
[22:50:49] 	 mooeypoo: in general, i think the problem is that 
 is inserted directly into OptionWidget, and it doesn't like that
[22:50:58] 	 it used to.
[22:51:09] 	 MatmaRex, I have to add it, the contents inside it is generated
[22:51:12] 	 and hence, replaced.
[22:51:18] 	 i have to have SOMETHING stable wrapping it
[22:51:35] 	 but in any case, that's inside the label. And it used to work fine. Why is the icon jumping around now?
[22:51:54] 	 it's not inside the label
[22:52:04] 	 which is why it goes bonkers
[22:52:45] 	 10VisualEditor: [Regression pre-wmf20] The transition while switching between Readmode and VE is not smooth, the whole block of content repeats as a flicker below the actual content - https://phabricator.wikimedia.org/T91442#1097136 (10Ryasmeen) @Aklapper:    This is not a browser specific issue , happens in bot...
[22:52:50] 	 ergh
[22:53:09] 	 MatmaRex, right, sorry, the title is the label, and the generated contents are appended to the option widget
[22:53:34] 	 it doesn't really fit into SelectWidget and OptionWidget constraints, especially if you're going to have more buttons there
[22:53:35] 	 Why would that make it go bonkers at all?
[22:53:43] 	 10VisualEditor, 10VisualEditor-EditingTools: VisualEditor:  What you see is not the Babel box that you get - https://phabricator.wikimedia.org/T64745#1097138 (10Etonkovidova) Re-checked it in production - no difference in displaying {{#babel:en|de-1|es-1|vo-0}} in VE and in Read mode.
[22:53:52] 	 10VisualEditor, 10VisualEditor-EditingTools: VisualEditor:  What you see is not the Babel box that you get - https://phabricator.wikimedia.org/T64745#1097139 (10Etonkovidova) 5Open>3Resolved
[22:53:58] 	 but okay, whatever
[22:54:02] 	 for the immediate issue
[22:54:09] 	 mooeypoo: the icon is being centred vertically.
[22:54:29] 	 because `.oo-ui-decoratedOptionWidget .oo-ui-iconElement-icon` has height: 100%
[22:54:42] 	 Can we not override that?
[22:54:57] 	 sure we can
[22:55:03] 	 instead of the 'top' elsewhere
[22:55:15] 	 (oh, that's not elsewhere)
[22:55:22] 	 so just instead of the 'top'
[22:56:03] 	 okay, height: what? auto ? 24px ?
[22:56:30] 	 height: 2.5em; seems to work well
[22:56:37] 	 this won't affect the size of the icon
[22:56:48] 	 just the vertical positioning
[22:57:13] 	 24px worked, but i'll use 2.5em
[22:57:47] 	 mooeypoo: at least 1.875em, or 24px indeed, which is the size of the icon
[22:57:49] 	 sweet
[22:57:59] 	 if you have that, you can also control it with 'top' sensibly
[22:58:35] 	 :)
[22:58:45] 	 yes!
[22:58:46] 	 now it works
[22:58:47] 	 thanks!
[23:06:58] 	 (03CR) 10Jforrester: [C: 032] build: Remove obsolete 'build' task from grunt-doc [oojs/ui] - 10https://gerrit.wikimedia.org/r/194946 (owner: 10Krinkle)
[23:07:46] 	 (03CR) 10Jforrester: [C: 032] build: Move pre/post 'doc' task into package.json [oojs/ui] - 10https://gerrit.wikimedia.org/r/194956 (owner: 10Krinkle)
[23:09:00] 	 James_F, I wonder if we can just move the LQT feeback page to an archive for now, and flow-ify the current one
[23:09:01] 	 ori: Could I ask you to review https://gerrit.wikimedia.org/r/#/c/191221/ and merge if you are happy with it?
[23:09:27] 	 Eloquence: The conversion script is pretty painless. The main issue is that neither LQT nor Flow work with ?section=new requests from the API.
[23:09:44] 	 Eloquence: So mediawiki.feedback responses go to the wrong place.
[23:10:03] 	 *nod* and you don't want to at least get rid of LQT for now?
[23:10:16] 	 James_F: reviewing
[23:10:22] 	 I'm happier with LQT than a wikitext talk page.
[23:10:30] 	 But Flow >> LQT of course. :-)
[23:10:35] 	 ori: Thanks!
[23:11:33] 	 well, we're not going to implement section=new for LQT, it's broken now, so unless we have a patchset in a few days for Flow to do the magic ...
[23:11:40] 	 I'd say let's use plain vanilla talk til we can get that sorted?
[23:11:56] 	 broken as in the mw.org feedback gets misrouted which is not nice 
[23:11:57] 	 Retrenching on UX feels like the wrong answer.
[23:12:03] 	 But yeah.
[23:12:17] 	 I'm sure Flow would work on it if you asked them. ;-)
[23:12:22] 	 true enough
[23:12:26] 	 do we have a T# for that yet?
[23:12:31] 	 Yes.
[23:12:33] * James_F finds it.
[23:12:50] 	 https://phabricator.wikimedia.org/T59989
[23:13:48] 	 (03Merged) 10jenkins-bot: build: Remove obsolete 'build' task from grunt-doc [oojs/ui] - 10https://gerrit.wikimedia.org/r/194946 (owner: 10Krinkle)
[23:13:52] 	 ok, will ask danny to estimate it and then we can decide
[23:13:56] 	 Thanks.
[23:13:57] 	 let's just not leave it in a broken state for too long
[23:14:10] * James_F nods.
[23:14:42] 	 10VisualEditor, 6Collaboration-Team, 10Flow, 10MediaWiki-General-or-Unknown: Flow support for the mediawiki.feedback form - https://phabricator.wikimedia.org/T91805#1097198 (10Jdforrester-WMF)
[23:14:58] 	 10VisualEditor, 6Collaboration-Team, 10Flow, 10MediaWiki-General-or-Unknown: Flow support for the mediawiki.feedback form - https://phabricator.wikimedia.org/T91805#1096593 (10Jdforrester-WMF) Also, nothing VE-specific; this is MediaWiki core's code that Flow doesn't work with. :-)
[23:15:03] 	 And https://phabricator.wikimedia.org/T91805 is the ticket for converting the page.
[23:15:10] 	 Err, no, for supporting the tool.
[23:15:19] 	 Which is mostly a dupe, but eg.
[23:16:15] 	 (03Merged) 10jenkins-bot: build: Move pre/post 'doc' task into package.json [oojs/ui] - 10https://gerrit.wikimedia.org/r/194956 (owner: 10Krinkle)
[23:17:41] 	 RoanKattouw: Any feedback for me?
[23:19:29] 	 (03CR) 10Jforrester: [C: 032] "Let's see if SauceLabs is now working…" [oojs/core] - 10https://gerrit.wikimedia.org/r/194813 (owner: 10Krinkle)
[23:20:18] 	 mookie: Sorry I got distracted
[23:20:36] 	 mookie: If you load the page with ?debug=true, and keep the browser console open while you click the edit button, do any errors appear?
[23:20:38] 	 (03CR) 10jenkins-bot: [V: 04-1] jsduck: Set --processes=0 to fix warnings-exit-nonzero [oojs/core] - 10https://gerrit.wikimedia.org/r/194813 (owner: 10Krinkle)
[23:20:44] 	 RoanKattouw: np I see you're busy
[23:20:45] 	 That error you reported is concerning but probably not why VE is failing
[23:20:54] 	 10VisualEditor, 10VisualEditor-EditingTools, 7Design, 7Epic, and 2 others: Cite: Auto-filled references based on destination URL / ISBN / etc. - https://phabricator.wikimedia.org/T62768#1097222 (10Jdforrester-WMF) >>! In T62768#1094741, @Eloquence wrote: > Can we project a "week of" deployment ETA?  No. Ev...
[23:24:01] 	 RoanKattouw: No, no other errors with debug=true
[23:24:30] 	 Hm, weird
[23:26:01] 	 RoanKattouw: I don't see any errors in the server log (it's IIS, ugh)
[23:28:15] 	 (03PS1) 10Bartosz Dziewoński: Remove remnants of window isolation [oojs/ui] - 10https://gerrit.wikimedia.org/r/194974 
[23:28:17] 	 (03PS1) 10Bartosz Dziewoński: demo: Simplify @media styles [oojs/ui] - 10https://gerrit.wikimedia.org/r/194975 
[23:28:19] 	 (03PS1) 10Bartosz Dziewoński: PanelLayout: Add 'framed' config option [oojs/ui] - 10https://gerrit.wikimedia.org/r/194976 (https://phabricator.wikimedia.org/T91153) 
[23:29:52] 	 10VisualEditor, 10VisualEditor-ContentEditable: Display bug for character formatting:  characters are bolded until you press enter at the end of a line - https://phabricator.wikimedia.org/T91672#1097233 (10Jdforrester-WMF)
[23:30:10] 	 (03CR) 10Catrope: [C: 032] Remove remnants of window isolation [oojs/ui] - 10https://gerrit.wikimedia.org/r/194974 (owner: 10Bartosz Dziewoński)
[23:30:41] 	 (03CR) 10Jforrester: [C: 032] Dialog: Mark private methods and add description of methods and configs [oojs/ui] - 10https://gerrit.wikimedia.org/r/194906 (owner: 10Kmenger)
[23:31:06] 	 RoanKattouw: Lack of focus much? ;-P
[23:31:52] 	 (03CR) 10Jforrester: [C: 032] ProcessDialog: Add description and example and mark private methods [oojs/ui] - 10https://gerrit.wikimedia.org/r/194923 (owner: 10Kmenger)
[23:32:40] 	 (03CR) 10Jforrester: [C: 032] demo: Simplify @media styles [oojs/ui] - 10https://gerrit.wikimedia.org/r/194975 (owner: 10Bartosz Dziewoński)
[23:34:44] 	 James_F: can you get that slacker trevor to look at https://phabricator.wikimedia.org/T91152 , read the description and confirm that i have the right idea?
[23:35:18] 	 because fixing that is going to be a lot of pain, and i'd rather not start if we decide not to
[23:35:19] 	 MatmaRex: :-)
[23:35:53] 	 MatmaRex: I'm chasing him for code review for you first, though.
[23:35:54] 	 James_F: also. should OOUI dialogs in Vector skin use 0.8em for everything, or 0.875em (Vector's body font size)?
[23:36:00] 	 (03CR) 10jenkins-bot: [V: 04-1] Remove remnants of window isolation [oojs/ui] - 10https://gerrit.wikimedia.org/r/194974 (owner: 10Bartosz Dziewoński)
[23:36:03] 	 (03CR) 10jenkins-bot: [V: 04-1] demo: Simplify @media styles [oojs/ui] - 10https://gerrit.wikimedia.org/r/194975 (owner: 10Bartosz Dziewoński)
[23:36:06] 	 that is a question we also need to answer
[23:36:09] 	 MatmaRex: Meh.
[23:36:14] 	 or, in other words
[23:36:28] 	 should the theme control the font size of widgets, or should it use "natural" size?
[23:37:51] 	 (03Merged) 10jenkins-bot: Dialog: Mark private methods and add description of methods and configs [oojs/ui] - 10https://gerrit.wikimedia.org/r/194906 (owner: 10Kmenger)
[23:38:06] 	 Move to rem and live it up?
[23:38:30] 	 10VisualEditor: Help the HTML Editing Task Force shape "Editing on the Web" (aka future of contentEditable) - https://phabricator.wikimedia.org/T91781#1097248 (10Sideshowbarker) I work for the W3C and know Krinkle from #whatwg on freenode and would be glad to help out in any way I can here with facilitating coor...
[23:39:02] 	 (03PS15) 10Mooeypoo: Refactor Citoid extension as an inspector [extensions/Citoid] - 10https://gerrit.wikimedia.org/r/190973 (https://phabricator.wikimedia.org/T88152) 
[23:39:53] 	 MatmaRex, do you know some about MessageDialog ?
[23:40:12] 	 some
[23:40:18] 	 Trevor mentioned that I can use it for a sort of yes/no dialog without extending it. But I don't see how I can respond to the affirmative action
[23:40:40] 	 I don't see an event that's being triggered if the user hits 'yes' or whatever label i give it
[23:40:45] 	 the dismiss just dismisses, which is fine
[23:41:03] 	 but I don't see a way to do the actual action *except* for extending the dialog... am I missing something ?
[23:41:37] 	 10VisualEditor, 3Editing Department 2014/15Q4 blockers: Help the HTML Editing Task Force shape "Editing on the Web" (aka future of contentEditable) - https://phabricator.wikimedia.org/T91781#1097250 (10Jdforrester-WMF) a:3Catrope
[23:41:39] 	 mooeypoo: second-to-last example on https://www.mediawiki.org/wiki/OOjs_UI/Windows/Message_Dialogs seems to be this
[23:42:01] 	 you set some actions, and wait for it to close with the given action
[23:42:10] 	 (we actually have docs with examples now! woo)
[23:42:43] 	 oooh
[23:42:51] 	 hm, some bit of nesting promises in there
[23:42:59] 	 but yay woohoo
[23:43:10] 	 MatmaRex, thanks!
[23:43:33] 	 (03PS1) 10Catrope: Factor attribute setting code into ElementLinearData [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194977 
[23:43:35] 	 (03PS1) 10Catrope: Move ve.dm.Document#spliceMetadata to dm.MetaLinearData [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194978 
[23:43:37] 	 (03PS1) 10Catrope: Remove unused variables in dm.TransactionProcessor#applyAnnotations [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194979 
[23:43:39] 	 (03PS1) 10Catrope: Remove unnecessary support for nesting annotation and replace operations [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194980 
[23:43:41] 	 (03PS1) 10Catrope: Queue linear model modifications in TransactionProcessor [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194981 
[23:43:43] 	 (03Merged) 10jenkins-bot: ProcessDialog: Add description and example and mark private methods [oojs/ui] - 10https://gerrit.wikimedia.org/r/194923 (owner: 10Kmenger)
[23:43:45] 	 (03CR) 10Jforrester: [C: 032] "…" [oojs/ui] - 10https://gerrit.wikimedia.org/r/194974 (owner: 10Bartosz Dziewoński)
[23:45:19] 	 (03CR) 10jenkins-bot: [V: 04-1] Remove unnecessary support for nesting annotation and replace operations [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194980 (owner: 10Catrope)
[23:45:29] 	 (03CR) 10jenkins-bot: [V: 04-1] Remove unused variables in dm.TransactionProcessor#applyAnnotations [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194979 (owner: 10Catrope)
[23:45:31] 	 (03CR) 10jenkins-bot: [V: 04-1] Queue linear model modifications in TransactionProcessor [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194981 (owner: 10Catrope)
[23:45:32] 	 RoanKattouw: Ouch.
[23:45:45] 	 wat
[23:45:51] 	 jscs error.
[23:45:54] 	 ugh
[23:46:00] 	 At the bottom of the stack.
[23:46:19] 	 James_F: Karma has a canary release they'd like us to try.
[23:46:29] 	 Krinkle: For mw-ext?
[23:46:43] 	 James_F: well, in general regarding disconnects and race conditions
[23:46:56] 	 mw-ext is where we see it 100% of the time, though?
[23:47:01] 	 Or is that different issues as well?
[23:47:27] 	 I'm gonna do a concurrency stresstest  later tonight to see how it behaves. There is no single repo where it happens every time, but when I ran it 20x in parallel with mwext it always failed.
[23:47:41] 	 Hmm. OK.
[23:47:55] 	 I'll be around over the weekend if you need a rubber duck to +2 changes. ;-)
[23:47:55] 	 basically like this: https://phabricator.wikimedia.org/T90673#1073556
[23:48:10] 	 Fun.
[23:49:55] 	 http://i.imgur.com/MHlRBOc.png ugh
[23:50:11] 	 why do we have some of these randomly attached to the prototype, and some to the constructor
[23:51:01] 	 and they're all bound to something else anyway
[23:51:21] 	 (03Merged) 10jenkins-bot: Remove remnants of window isolation [oojs/ui] - 10https://gerrit.wikimedia.org/r/194974 (owner: 10Bartosz Dziewoński)
[23:51:22] 	 (03Merged) 10jenkins-bot: demo: Simplify @media styles [oojs/ui] - 10https://gerrit.wikimedia.org/r/194975 (owner: 10Bartosz Dziewoński)
[23:51:50] 	 is there a reason for this, or just organic growth and hate-fueled code?
[23:52:23] 	 MatmaRex: mwTarget is dark and full of terrors.
[23:52:48] 	 MatmaRex: There's a tech debt ticket to fix it. Every time I mention it to RoanKattouw he says "oh gods".
[23:53:00] 	 MatmaRex: However, he is in the middle of a big refactor in that area.
[23:53:02] 	 it = my screenshot, or mw.Target?
[23:53:16] 	 (03PS1) 10Kmenger: TextInputMenuSelectWidget: Add description and mark private methods [oojs/ui] - 10https://gerrit.wikimedia.org/r/194983 
[23:53:18] 	 it is full or terrors all right
[23:53:31] 	 Well, mostly to fix target, init and platform's relationships with each other, as I understand it.
[23:53:32] 	 i am just considering fixing the FIXME i added yesterday when unbreaking badtoken handling
[23:53:52] 	 but if it's being rewritten, then i guess i shouldn't waste time
[23:54:15] 	 Oh that garbage
[23:54:21] 	 (you see all of these methods taking 'jqXHR'? they actually don't anymore)
[23:54:29] 	 MatmaRex: Please do fix the jqXHR mess
[23:54:31] 	 (as in, it's not being passed)
[23:54:41] 	 (so status and data get gerbled)
[23:54:44] 	 That's particularly harmful as we saw
[23:54:44] 	 garbled
[23:54:56] 	 RoanKattouw: Will you be OK to rebase on top of it?
[23:55:07] 	 RoanKattouw: Or do you want MatmaRex to do his patch on top of yours?
[23:55:10] 	 The non-prototype mess really needs to be fixed but is nontrivial because of naming conflicts with the subclass (yay!)
[23:55:17] 	 Fun.
[23:55:17] 	 James_F: I can rebase on top of him
[23:55:21] 	 eurgh
[23:55:23] 	 RoanKattouw: "Can". Sure.
[23:55:33] 	 And nonsensical communication via events on top of tha
[23:55:35] 	 RoanKattouw: But… efficacy and efficiency…
[23:55:42] 	 oh god, the events
[23:56:16] 	 ve.init.mw.Target.onFoo  is really a method because it's invoked with .call( this ), and then it does this.emit( 'foo' ) and VPT has done this.connect( this, { foo: 'onFoo' } ); so ve.init.mw.ViewPageTarget.prototype.onFoo is called
[23:56:40] 	 And that insanity is how almost everything in mw.Target works
[23:56:42] 	 "works"
[23:56:59] 	 James_F: jqXHR stuff won't affect TargetLoader much
[23:57:03] 	 OK.
[23:57:37] 	 RoanKattouw: Can you take the 10 minutes to write the commit summary for that change so Kr*/etc. can review and merge, or does it need more?
[23:57:48] 	 (03CR) 10Jforrester: [C: 032] TextInputMenuSelectWidget: Add description and mark private methods [oojs/ui] - 10https://gerrit.wikimedia.org/r/194983 (owner: 10Kmenger)
[23:58:05] 	 James_F: It needs work
[23:58:11] 	 I talked to Krinkle about it this morning
[23:58:18] 	 Right now it works but that's mostly an illusion
[23:58:26] 	 RoanKattouw: Lots of work? Something you can finish in a day or two?
[23:58:26] 	 Because I avoided solving all the hard problems
[23:58:38] 	 (03CR) 10jenkins-bot: [V: 04-1] Move ve.dm.Document#spliceMetadata to dm.MetaLinearData [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194978 (owner: 10Catrope)
[23:58:38] * James_F is trying to get a feel for how quickly it will land.
[23:58:41] 	 Somewhere in between those two
[23:58:48] 	 Fun.
[23:59:09] 	 I might be able to do it in a day or two but I'm afraid it'd be more like three or four
[23:59:40] 	 :-(
[23:59:46] 	 So, Tuesday–Friday? ;-)
[23:59:51] 	 Yeah hopefully