[00:42:44] 3VisualEditor / 3Editing Tools: Page Options Dialog width is not wide enough - 10https://bugzilla.wikimedia.org/70316 (10josephsieh) 3UNCO p:3Unprio s:3normal a:3None On http://en.wikipedia.beta.wmflabs.org/wiki/Cite_VisualEditor_Test?veaction=edit, when you click on the "Page Options" icon, the pop-... [00:43:27] 3VisualEditor / 3Editing Tools: Page Options Dialog width is not wide enough - 10https://bugzilla.wikimedia.org/70316#c1 (10josephsieh) Created attachment 16347 --> https://bugzilla.wikimedia.org/attachment.cgi?id=16347&action=edit Screen shot [00:53:43] 3VisualEditor: Switch to source editing dialog too tall - 10https://bugzilla.wikimedia.org/70317 (10josephsieh) 3UNCO p:3Unprio s:3normal a:3None Created attachment 16348 --> https://bugzilla.wikimedia.org/attachment.cgi?id=16348&action=edit screen shot of "Switch to source editing" dialog too tall i... [01:23:48] (03CR) 10Alex Monk: "OOUI's window manager only allows you to have one window open at a time. Sounds like we'd have to close the template/citation dialog and t" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/156559 (https://bugzilla.wikimedia.org/50747) (owner: 10Alex Monk) [02:39:26] 3VisualEditor / 3Editing Tools: VisualEditor: Stop users from deleting parameters in the template dialog if they're marked as "required" in TemplateData - 10https://bugzilla.wikimedia.org/60358#c8 (10Helder) (In reply to Ricordisamoa from comment #5) > ...a set of fields > may be mutually exclusive (and yet... [02:49:43] 3OOjs UI: Flipping in the build pipeline broken - 10https://bugzilla.wikimedia.org/70323 (10Rainer Rillke @commons.wikimedia) 3NEW p:3Unprio s:3normal a:3None Locally and on http://tools.wmflabs.org/oojs-ui/ RTL styling version appears to be not RTL. Despite , the following CSS... [02:57:43] 3VisualEditor: VisualEditor should respect [[MediaWiki:Noarticletext]] - 10https://bugzilla.wikimedia.org/70322 (10MaxBioHazard) 5UNCO>3NEW a:3None [09:33:59] 3VisualEditor / 3MediaWiki integration: [Regression] VisualEditor: anoneditwarning notice displays unparsed links with $1 and $2 - 10https://bugzilla.wikimedia.org/70334 (10Krinkle) 3NEW p:3Unprio s:3normal a:3None Created attachment 16349 --> https://bugzilla.wikimedia.org/attachment.cgi?id=16349&... [09:44:26] 3VisualEditor / 3Editing Tools: Give exact match when looking for a specific file - 10https://bugzilla.wikimedia.org/62917#c1 (10Elitre) I believe that this request is a duplicate, in that it depends on the results from the search not being perfect. If you type "descending into cave" in the media dialog on e... [09:46:27] 3VisualEditor / 3Editing Tools: VisualEditor: Be able to specify a file name in the media or image dialog, even if search can't find the file - 10https://bugzilla.wikimedia.org/66046#c3 (10Elitre) I disagree with your comment about search results working fine now. I added some findings to https://bugzilla.w... [14:36:13] 3VisualEditor / 3ContentEditable: Creating numbered lists with images resets the numbering back to 1 for each list item - 10https://bugzilla.wikimedia.org/69205#c1 (10Max Anderson) Update 9/3/2014: I've found another work around. This one is slightly more convenient but its still a workaround. 1)The user c... [14:53:29] 3VisualEditor / 3ContentEditable: VisualEditor: Link, Formula and Text Styles stay selected while editing a comment. - 10https://bugzilla.wikimedia.org/70343 (10Ritu Swain) 3UNCO p:3Unprio s:3normal a:3None Created attachment 16352 --> https://bugzilla.wikimedia.org/attachment.cgi?id=16352&action=e... [14:53:59] 3VisualEditor / 3ContentEditable: VisualEditor: Link, Formula and Text Styles stay selected while editing a comment. - 10https://bugzilla.wikimedia.org/70343#c1 (10Ritu Swain) Created attachment 16353 --> https://bugzilla.wikimedia.org/attachment.cgi?id=16353&action=edit screenshot2 [14:54:26] 3VisualEditor / 3ContentEditable: VisualEditor: Link, Formula and Text Styles stay selected while editing a comment. - 10https://bugzilla.wikimedia.org/70343 (10Ritu Swain) s:5normal>3minor [15:09:29] 3VisualEditor / 3ContentEditable: VisualEditor: Cursor jumps to the start of the line, if link and Language style text present. - 10https://bugzilla.wikimedia.org/70344 (10Ritu Swain) 3UNCO p:3Unprio s:3minor a:3None Created attachment 16354 --> https://bugzilla.wikimedia.org/attachment.cgi?id=1635... [15:09:42] 3VisualEditor / 3ContentEditable: VisualEditor: Cursor jumps to the start of the line, if link and Language style text present. - 10https://bugzilla.wikimedia.org/70344 (10Ritu Swain) [15:13:56] 3VisualEditor / 3ContentEditable: VisualEditor: Cursor jumps to the start of the line, if link and Language style text present. - 10https://bugzilla.wikimedia.org/70344#c1 (10Ritu Swain) so if i want to edit some text present in between a link and the language style text, i cannot do it. [15:14:41] 3VisualEditor / 3ContentEditable: VisualEditor: Cursor jumps to the start of the line, if link and Language style text present. - 10https://bugzilla.wikimedia.org/70344#c2 (10Ritu Swain) Environment- test2 [15:15:27] 3VisualEditor / 3ContentEditable: VisualEditor: Link, Formula and Text Styles stay selected while editing a comment. - 10https://bugzilla.wikimedia.org/70343#c2 (10Ritu Swain) Environment-test2 [15:35:44] (03PS1) 10Krinkle: Prefer String#slice over substr() or substring() [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/158110 [16:03:02] (03CR) 10Esanders: Position the paste target just below the top of the screen (031 comment) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/157876 (owner: 10Catrope) [16:20:23] (03CR) 10Esanders: "If you look at usage of this function we almost always want relative offsets." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/157868 (https://bugzilla.wikimedia.org/69122) (owner: 10Catrope) [16:43:05] (03CR) 10Esanders: [C: 032] Prefer String#slice over substr() or substring() [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/158110 (owner: 10Krinkle) [16:43:50] (03CR) 10Esanders: [C: 032] Reorder CSS rules in ve.ce.Surface.css to make sense [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/157880 (owner: 10Catrope) [16:45:38] (03Merged) 10jenkins-bot: Prefer String#slice over substr() or substring() [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/158110 (owner: 10Krinkle) [16:46:40] (03CR) 10Krinkle: "The typing presumably is focus' built-in "scroll into view" logic. Does this fix that in general, or only if your page isn't scrolled down" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/157876 (owner: 10Catrope) [16:49:24] (03PS3) 10Jforrester: Update VE core submodule to master (9512223) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157986 [16:50:37] (03CR) 10jenkins-bot: [V: 04-1] Update VE core submodule to master (9512223) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157986 (owner: 10Jforrester) [16:52:15] (03PS4) 10Jforrester: Update VE core submodule to master (9512223) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157986 [17:32:04] 3VisualEditor: VisualEditor: Cannot type in Korean using Mozilla on Windows 7 with dual-script keyboard - 10https://bugzilla.wikimedia.org/70353 (10WhatamIdoing) 3NEW p:3Unprio s:3normal a:3None Kdammers is running the current version of Mozilla on Windows 7 with a dual-script keyboard. Typing in Han... [17:32:12] 3VisualEditor / 3Language: VisualEditor: CE eats up syllables except a last syllable of a word in Korean IME - 10https://bugzilla.wikimedia.org/50631 (10WhatamIdoing) [17:32:12] 3VisualEditor: VisualEditor: Cannot type in Korean using Mozilla on Windows 7 with dual-script keyboard - 10https://bugzilla.wikimedia.org/70353 (10WhatamIdoing) [17:36:10] (03PS5) 10Jforrester: Update VE core submodule to master (7004717) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157986 [17:36:18] edsanders: https://gerrit.wikimedia.org/r/157986 merge please. [17:40:42] 3VisualEditor / 3Language: VisualEditor: Cannot type in Korean using Mozilla on Windows 7 with dual-script keyboard - 10https://bugzilla.wikimedia.org/70353#c1 (10James Forrester) 5NEW>3ASSI p:5Unprio>3Normal s:5normal>3major This is probably a dupe of bug 50631, but leaving it open. [17:47:20] (03PS1) 10Trevor Parscal: Fix documentation typo [oojs/ui] - 10https://gerrit.wikimedia.org/r/158139 [17:47:41] (03CR) 10Jforrester: [C: 032] Fix documentation typo [oojs/ui] - 10https://gerrit.wikimedia.org/r/158139 (owner: 10Trevor Parscal) [17:48:58] (03Merged) 10jenkins-bot: Fix documentation typo [oojs/ui] - 10https://gerrit.wikimedia.org/r/158139 (owner: 10Trevor Parscal) [17:50:30] 3VisualEditor / 3Editing Tools: VisualEditor: Link target names are not appearing pressing command+K just applying the link text as link target straight off - 10https://bugzilla.wikimedia.org/70356 (10Rummana Yasmeen) 3NEW p:3Unprio s:3normal a:3None Steps to reproduce: 1.Open a page with VE 2.Selec... [17:50:36] (03CR) 10Jforrester: "I5d894f8a has merged for 1.24wmf20 for the wfUseMW call." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/156559 (https://bugzilla.wikimedia.org/50747) (owner: 10Alex Monk) [17:50:43] 3VisualEditor / 3Editing Tools: VisualEditor: Link target names are not appearing pressing command+K just applying the link text as link target straight off - 10https://bugzilla.wikimedia.org/70356 (10Rummana Yasmeen) p:5Unprio>3High s:5normal>3major a:3Roan Kattouw [18:04:05] RoanKattouw: this is the fresh and new brokenness in Betalabs: https://bugzilla.wikimedia.org/show_bug.cgi?id=70356 :) [18:06:11] 3VisualEditor / 3MediaWiki integration: VisualEditor: Toolbar utilities menu items should not be cut off against browser edge - 10https://bugzilla.wikimedia.org/57366#c9 (10James Forrester) *** Bug 70316 has been marked as a duplicate of this bug. *** [18:06:13] 3VisualEditor / 3Editing Tools: Page Options Dialog width is not wide enough - 10https://bugzilla.wikimedia.org/70316#c2 (10James Forrester) 5UNCO>3RESO/DUP Yeah, this is a painful issue; merging with bug 57366. *** This bug has been marked as a duplicate of bug 57366 *** [18:06:33] (03CR) 10Trevor Parscal: [BREAKING CHANGE] Element mixin refactor (031 comment) [oojs/ui] - 10https://gerrit.wikimedia.org/r/157274 (owner: 10Trevor Parscal) [18:07:39] (03PS5) 10Trevor Parscal: [BREAKING CHANGE] Element mixin refactor [oojs/ui] - 10https://gerrit.wikimedia.org/r/157274 [18:10:56] (03PS2) 10Trevor Parscal: Update VE core submodule to 5858265dcf05b7511bd951608d14d62da6f112ee [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157285 [18:11:08] (03CR) 10jenkins-bot: [V: 04-1] Update VE core submodule to 5858265dcf05b7511bd951608d14d62da6f112ee [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157285 (owner: 10Trevor Parscal) [18:18:08] (03PS2) 10Trevor Parscal: Update OOjs UI to v0.1.0-pre (5d6fc43de8) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/157286 [18:21:23] (03PS3) 10Trevor Parscal: Update VE core submodule to 369d3e2923e82bb206dbb4cf38b7b44c95709249 [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157285 [18:22:06] (03CR) 10jenkins-bot: [V: 04-1] Update VE core submodule to 369d3e2923e82bb206dbb4cf38b7b44c95709249 [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157285 (owner: 10Trevor Parscal) [18:22:24] (03PS5) 10Trevor Parscal: Automatically align popup toolgroups [oojs/ui] - 10https://gerrit.wikimedia.org/r/157295 [18:22:40] (03PS2) 10Trevor Parscal: [WIP] MediaWiki theme [oojs/ui] - 10https://gerrit.wikimedia.org/r/157866 [18:23:16] (03CR) 10jenkins-bot: [V: 04-1] Automatically align popup toolgroups [oojs/ui] - 10https://gerrit.wikimedia.org/r/157295 (owner: 10Trevor Parscal) [18:26:13] (03PS6) 10Trevor Parscal: Automatically align popup toolgroups [oojs/ui] - 10https://gerrit.wikimedia.org/r/157295 [18:26:54] (03CR) 10jenkins-bot: [V: 04-1] Automatically align popup toolgroups [oojs/ui] - 10https://gerrit.wikimedia.org/r/157295 (owner: 10Trevor Parscal) [18:27:15] (03PS3) 10Trevor Parscal: [WIP] MediaWiki theme [oojs/ui] - 10https://gerrit.wikimedia.org/r/157866 [18:34:24] (03PS6) 10Trevor Parscal: [BREAKING CHANGE] Element mixin refactor [oojs/ui] - 10https://gerrit.wikimedia.org/r/157274 [18:37:49] (03PS3) 10Trevor Parscal: Update OOjs UI to v0.1.0-pre (060cf0ccbc) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/157286 [18:42:34] (03PS4) 10Trevor Parscal: Update VE core submodule to 369d3e2923e82bb206dbb4cf38b7b44c95709249 [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157285 [18:43:06] (03CR) 10jenkins-bot: [V: 04-1] Update VE core submodule to 369d3e2923e82bb206dbb4cf38b7b44c95709249 [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157285 (owner: 10Trevor Parscal) [18:43:47] (03PS7) 10Trevor Parscal: Automatically align popup toolgroups [oojs/ui] - 10https://gerrit.wikimedia.org/r/157295 [18:43:57] (03PS4) 10Trevor Parscal: [WIP] MediaWiki theme [oojs/ui] - 10https://gerrit.wikimedia.org/r/157866 [18:44:31] (03CR) 10jenkins-bot: [V: 04-1] Automatically align popup toolgroups [oojs/ui] - 10https://gerrit.wikimedia.org/r/157295 (owner: 10Trevor Parscal) [18:53:13] 3VisualEditor / 3MediaWiki integration: VisualEditor should respect [[MediaWiki:Noarticletext]] - 10https://bugzilla.wikimedia.org/70322 (10James Forrester) 5NEW>3ASSI p:5Unprio>3Normal a:3Alex Monk [18:56:05] ryasmeen: I can't reproduce bug 70356 on beta labs in Chrome [18:56:17] Pressing Ctrl+K opens the link inspector and it says open [18:56:42] did you select a text first ? [18:57:19] Oh sorry, I didn't [18:57:25] Holy crap! [18:57:35] yes it is ! :P [18:58:43] We suspect we know which change caused this [18:58:59] okay thats great! [19:01:33] I though it is a side effect of the fix in https://bugzilla.wikimedia.org/show_bug.cgi?id=69122 [19:01:47] But that did not get merge yet right? [19:01:57] No it didn't [19:02:04] Also the link inspector is displayed in totally the wrong place [19:02:23] What is going ON [19:03:13] OK it's not the change we thought it was, but it's still Ed's fault [19:03:19] :D [19:03:52] (03CR) 10Catrope: "FIXME: caused https://bugzilla.wikimedia.org/show_bug.cgi?id=70356" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/157199 (owner: 10Esanders) [19:04:20] Ed: "Oh, balls" xD [19:05:13] 3VisualEditor / 3Editing Tools: VisualEditor: Link target names are not appearing pressing command+K just applying the link text as link target straight off - 10https://bugzilla.wikimedia.org/70356#c1 (10Roan Kattouw) Regression from https://gerrit.wikimedia.org/r/#/c/157199/ [19:05:24] RoanKattouw: :-) [19:06:34] "I'll fix it today. It's only code, how hard can it be." [19:11:13] RoanKattouw: Is edsanders going to "fix" DM to support generic types (mw:Error) today too? It's only code… :-) [19:13:47] TrevorParscal, so the patch I was talking about earlier is https://gerrit.wikimedia.org/r/#/c/156559/ + these changes on top: https://gist.github.com/Krenair/ff4b3121e05a7c93b4fc [19:14:56] obviously I need to deal with i18n in the dialog file, but that'll be easy. the less obvious part is getting this to show on top of the template/citation dialog [19:15:21] yeah, that nesting of promises sucks huh? [19:15:28] we gotta come up with a better way to handle that [19:16:01] I'm thinking maybe openWindow() should return the closing promise [19:16:17] but, I'm not sure yet [19:16:29] brb, gotta go to the bank [19:19:41] 3VisualEditor / 3Editing Tools: VisualEditor: Stop users from deleting parameters in the template dialog if they're marked as "required" in TemplateData - 10https://bugzilla.wikimedia.org/60358#c9 (10James Forrester) (In reply to Helder from comment #8) > (In reply to James Forrester from comment #6) > > (In... [19:38:53] TrevorParscal, could it give you access to both? [20:21:11] 3VisualEditor / 3MediaWiki integration: VisualEditor should respect [[MediaWiki:Noarticletext]] - 10https://bugzilla.wikimedia.org/70322#c1 (10Alex Monk) https://en.wikiversity.org/w/index.php?title=Lalala&action=edit&uselang=qqx is showing 'newarticletext' (not 'noarticletext'), which VE already does... 'n... [20:22:26] what are the plans for the future? when tables will be included? [20:22:51] Juandev: For VisualEditor? The main plans are in https://www.mediawiki.org/wiki/VisualEditor/Roadmap [20:22:51] Hi Juandev [20:22:59] oh, and hi James_F :p [20:23:34] Krenair: hui [20:23:56] (I was going to point Juandev to that page but James is too fast :p) [20:24:10] because, I think it is allready prety cool and fixed right [20:25:30] Juandev: Table editing is planned for next quarter (so, before Christmas). [20:25:41] ah good [20:26:08] well, I think Foundation would deploy you than to Commons [20:26:27] are there any plans to make Wikimedia Commons more user friendly? [20:26:29] Juandev: VisualEditor is available on Commons right now, as a Beta Feature. [20:26:43] maybe users would love to upload images via VE:-) [20:26:51] Yeah, uploading is something we want to do. [20:27:00] That's waiting for work by the Multimedia team. [20:27:12] I mean to upload from Wikipedia [20:27:16] But right now they're quite busy with structured data and MediaViewer, so we've not made much progress. [20:27:24] From Wikipedia to Commons? [20:27:36] oh yes, MediaViewer and protests:-) [20:27:46] James_F, Timo and Gabriel decided not to do the 'References' part of that page, see https://gerrit.wikimedia.org/r/#/c/120962/ [20:28:00] Krenair: It's not their call, however. [20:28:04] yes, from wp to commons, that should be somehow possible [20:28:22] the problem with commons is, it does not have multilingual category names [20:28:26] James_F, well if you want to post a comment... [20:28:29] Krenair: If they object they have a duty to provide the alternative. [20:28:38] It's certainly not been 'in progress' since April [20:28:39] Krenair: I told them at the time. Still waiting for them to come up with ideas. [20:28:40] James_F: read the bug ;) [20:28:47] readily documented alternative [20:28:54] No there isn't. [20:28:59] You want to remove user-facing functionality. [20:29:08] But rather than carry the can for it, you want me to sell it. [20:29:18] And you just wave your hands and say "three styles is enough". [20:29:24] are you talking about the list-style part of it? [20:29:27] Yes. [20:29:41] can you describe the issue? [20:29:53] We want to replace {{reflist}} with [20:30:27] which has my full support [20:30:43] Whether that's (ick) or references type="foo"> (like I suggested) or… [20:30:58] we discussed this on the bug IIRC [20:31:15] so your point is that each references tag needs different styling? [20:31:29] Currently {{reflist}} is parameterised. [20:31:45] sure, as there's no automatic way to do it [20:31:45] It's used (badly) to produce multiple columns, and multiple list styling types. [20:32:11] I was suggesting limiting this down to column widths (never column count) as it's more appropriate. [20:32:48] But you want this done by the skin instead? [20:32:51] my instinct would be that this should be device-specific etc [20:33:13] column-width is the W3C-provided tool for device-agnostic list column styling. [20:33:15] my main concern is with the numbering styles though [20:33:23] What's wrong with those? [20:33:42] is your point that those vary wildly per wiki? [20:33:54] The styles? [20:33:58] list styles [20:34:07] They're used semantically for different cases. [20:34:08] I mean, between references section within a single wiki [20:34:15] *sections [20:34:16] Yeah. [20:34:52] References in 1,2,3…; notes in i,ii,iii…; further reading in a,b,c…. [20:35:00] could you add some example use cases to the bug? [20:35:27] I think I did when I created the bug. [20:36:09] I don't see it in the first comment [20:36:17] https://bugzilla.wikimedia.org/show_bug.cgi?id=51260 [20:36:19] gwicke: https://en.wikipedia.org/wiki/Special:PrefixIndex/MediaWiki:Cite_link_label_group from MatmaRex [20:36:50] gwicke: (Auto-styled based on group name.) [20:37:04] those look like possible styles to me [20:37:07] not use cases [20:37:16] Those are all created manually to be used. [20:37:25] gwicke: we have some "semantic" ones on pl.wp [20:37:28] https://pl.wikipedia.org/wiki/Specjalna:Strony_według_prefiksu/MediaWiki:Cite_link_label_group [20:37:28] And I believe are all used. [20:38:17] the first is for references and reflists inside infoboxes, used for e.g. update dates - example: https://pl.wikipedia.org/wiki/Lionel_Messi [20:38:25] no idea what the second one is… [20:38:36] well, I proposed to offer a drop-down of styles per wiki, which would translate to css classes [20:38:39] and the third is for notes (in contract to citations) [20:38:51] contrast* [20:39:00] https://bugzilla.wikimedia.org/show_bug.cgi?id=51260#c24 [20:39:11] I proposed to offer a drop-down to styles, which would translate to CSS attributes. [20:39:53] And we could devise standard names for the semantics (so reference -> decimal; note -> arabic; etc.) which could be over-ridden per-wiki if needed. [20:40:06] James_F: Hi! This page https://www.mediawiki.org/wiki/VisualEditor/Design/Realtime_collaboration conveys that there's already some planning done around realtime collab editing. Is there a PoC or some code towards that goal already? [20:40:20] James_F: please describe the virtues of inline styles somewhere ;) [20:40:52] gwicke: I'm not advocating inline styles… [20:40:53] semantic class names sound much better to me [20:41:12] gwicke: I'm advocating inline wikitext attributes. [20:41:32] Adityab: Hey. There's no real PoC that's current, sadly. [20:41:39] well, ... [20:42:03] gwicke: I'm not a huge fan of using HTML "class" when we mean "type". [20:42:14] gwicke: We want to limit this to known values, right? [20:42:19] Hmm ok [20:42:36] Adityab: There's https://git.wikimedia.org/log/mediawiki%2Fextensions%2FVisualEditor.git/refs%2Fheads%2Frealtimeve from Ashish, but it's very out of date. [20:42:41] James_F: if we want to really lock it down we could also use a 'type' attribute [20:42:56] but having a whitelist of classes isn't that hard either [20:42:58] Adityab: Also, cscott did some awesome work using Mozilla's TowTruck (now "Together.js") to provide this. [20:43:13] gwicke: Whitelist of classes == a bug to fix. [20:43:13] and classes are a generic way to mark up semantics in HTML [20:43:18] * cscott pricks his ears [20:43:25] gwicke: Formal types == an actual discussion about need. [20:43:35] James_F: Ah! by 'this' do you mean the link you just supplied, or? [20:43:40] same thing, only different syntax [20:43:54] both need the whitelist in any case [20:43:59] Adityab: No, cscott's system is a real-time collaboration system. [20:44:12] gwicke: Different syntax on HTML level, but also a different expectation. [20:44:14] which reminds me that i should put my slides up [20:44:31] gwicke: https://www.mediawiki.org/wiki/Help:Images#Mode_parameter [20:44:38] damn, 15 minutes left to get lunch [20:44:41] For instance. "Mode" works. [20:44:49] * James_F stops talking so gwicke can eat. :-) [20:45:05] James_F: as long as a class pops out in the end I'll be happy ;) [20:45:19] * James_F grins. [20:45:24] anyone know what the story is with wikimania and videos of the sessions? [20:45:31] James_F: nice. where can I find that? [20:45:35] Krenair: ^^^ Want to summarise on the bug and pick this back up? [20:45:39] cscott: ^ :) [20:45:40] cscott: Not sure, sorry. [20:45:41] 3VisualEditor / 3Language: VisualEditor: Cannot type in Korean using Mozilla on Windows 7 with dual-script keyboard - 10https://bugzilla.wikimedia.org/70353#c2 (10Andre Klapper) What is "the current version of Mozilla"? Firefox? Seamonkey? [20:45:48] Krenair: (Sorry. :-)) [20:46:11] ah this https://github.com/cscott/mediawiki-extensions-togetherjs [20:48:08] [[Extension:TogetherJS]] [20:48:39] the visual editor piece still requires some unmerged patches to VE. i've been working on them as time permits. [20:48:56] 3VisualEditor / 3Editing Tools: VisualEditor: Stop users from deleting parameters in the template dialog if they're marked as "required" in TemplateData - 10https://bugzilla.wikimedia.org/60358#c10 (10Helder) This kind of relationship between parameters was requested on bug 50407. [20:49:06] Adityab: but you can use https://en.wikipedia.org/wiki/User:Cscott/TogetherJS.js to try it out on enwiki [20:53:07] ah very nice [20:56:10] James_F: as a active contributor I definitely prefer wikitext editor at Commons. there are so many particles I need to change that I dont think Ill be using there VE [20:56:41] Juandev: Oh well. :-( [20:56:57] James_F: but newbies will like it [20:57:23] Hopefully not just them [20:58:32] cscott: That said; I'm a bit lost. How do I try it out? [21:00:56] well see [21:01:10] I cannot image VE on wv either:-) [21:01:39] Adityab: https://docs.google.com/presentation/d/1O-mxmygpSIJE_gjTlEYh4pl1cxW6CYL_0LM4vhEkVKc/edit?usp=sharing slide 42 [21:01:55] thanks [21:03:10] Juandev: That's coming too. :-) [21:03:12] 3MediaWiki extensions / 3TemplateData: TemplateData: Add a way to express dependencies between parameters - 10https://bugzilla.wikimedia.org/50407#c7 (10James Forrester) [Moved from bug 60358 discussion.] (In reply to Helder from bug 60358 comment #10) > (In reply to James Forrester from bug 60358 comment #... [21:21:34] James_F: what about real time collaboration, will this be enabled on wp? [21:23:04] yah "Upload – Editors can upload a file to Commons and use it in the page without leaving the edit, including starting it with "drag-and-drop". " thats what I said [21:23:17] Juandev: When it's eventually done, yes, but that it probably a long way away. [21:23:36] Juandev: Yeah, that (file uploading) will take some time, sadly. :-( [21:23:43] cscott: I could enable TogetherJS with that but edits are still not relayed. I do get the toolbar on the right [21:24:02] so I can see, you dont have to be afraid of loosing job [21:24:06] lotsa goals [21:24:16] yes, like i said, visual editor requires some patches which have not yet been merged yet [21:24:24] ack [21:24:27] but the standard wikitext box ought to work. [21:24:42] (oh, hm, maybe i need to update the hub, i should double-check) [21:24:43] yup mediasearch is still kind of strange [21:24:54] I add a string and it offers media I dont want [21:27:32] (03PS1) 10Bartosz Dziewoński: readme: Fix "ECMAScript" typo [oojs/core] - 10https://gerrit.wikimedia.org/r/158247 [21:27:42] this is probably the last one. :> [21:28:31] hmm, why do oojs and oojs ui readmes use different syntax for code blocks? [21:30:53] MatmaRex: Because I wrote them at different times and haven't synchronised? [21:31:23] haha, okay [21:39:51] MatmaRex: If you want to fix go ahead. ISTR I fixed one to work with gitblit's Markdown support [21:40:28] nah :>, was just wondering [21:40:40] which is the good one? [21:40:44] or i guess i can check myself [21:41:14] (03CR) 10Esanders: [C: 032] Update VE core submodule to master (7004717) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157986 (owner: 10Jforrester) [21:43:11] (03Merged) 10jenkins-bot: Update VE core submodule to master (7004717) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/157986 (owner: 10Jforrester) [21:43:26] MatmaRex: I don't remember, sorry. :-) [21:43:32] MatmaRex: Probably OOjs. [21:43:49] (03PS1) 10Bartosz Dziewoński: readme: Use 
 instead of ``` [oojs/ui] - 10https://gerrit.wikimedia.org/r/158252 
[21:43:52] 	 James_F: found it already :) ^
[21:44:40] 	 (03CR) 10Krinkle: [C: 032] readme: Fix "ECMAScript" typo [oojs/core] - 10https://gerrit.wikimedia.org/r/158247 (owner: 10Bartosz Dziewoński)
[21:45:03] 	 MatmaRex: 
 is good
[21:45:09] 	 ``` is github specific
[21:45:13] 	 James_F: by the way, i can't +2 in oojs/ and VisualEditor/. just sayin'
[21:45:33] 	 I suppose we should fix that
[21:45:34] 	 That's because it's not part of mediawiki/
[21:46:07] 	 yeah :)
[21:46:26] <_ob>	 hey, ve-team.. greetingz from Linz, Austria...
[21:46:41] 	 Hey _ob 
[21:46:49] 	 (03CR) 10Krinkle: "I think the line break between block-level is needed by some parsers. If not GitBlit, maybe GitHub." [oojs/ui] - 10https://gerrit.wikimedia.org/r/158252 (owner: 10Bartosz Dziewoński)
[21:46:57] <_ob>	 RoanKattouw: hey
[21:47:24] 	 Krinkle: there isn't one in core oojs readme
[21:47:28] 	 (03CR) 10Krinkle: [C: 032] readme: Use 
 instead of ``` [oojs/ui] - 10https://gerrit.wikimedia.org/r/158252 (owner: 10Bartosz Dziewoński)
[21:47:31] 	 Hey _ob.
[21:47:33] 	 Yeah, Just checkd
[21:47:45] 	 MatmaRex: There is actually, but not in all.
[21:47:45] <_ob>	 +James_F: hey
[21:48:10] 	 (03Merged) 10jenkins-bot: readme: Use 
 instead of ``` [oojs/ui] - 10https://gerrit.wikimedia.org/r/158252 (owner: 10Bartosz Dziewoński)
[21:48:44] 	 3VisualEditor / 3Editing Tools: VisualEditor: Reference groups created from Reference list dialog does not populate the Reference group input combo box with those values - 10https://bugzilla.wikimedia.org/70365 (10Rummana Yasmeen) 3NEW p:3Unprio s:3normal a:3None Steps to reproduce:  1.Open a page wi...
[21:49:27] <_ob>	 +RoanKattouw: listening on model changes and updating the html element, I would do that in ce.TableCellNode, or elsewhere?
[21:49:42] 	 Yes, ce.TableCellNode is responsible for rerendering
[21:49:56] <_ob>	 Ok. I try that
[21:49:58] 	 3VisualEditor / 3Editing Tools: VisualEditor: Reference group input combo box does not populate the Reference group names created from Reference list dialog - 10https://bugzilla.wikimedia.org/70365 (10Rummana Yasmeen) a:3Alex Monk
[21:50:13] 	 Not many CE nodes rerender themselves on attribute change, but you can look at ve.ce.InlineImageNode for an example
[21:50:43] 	 Basically you bind onAttributeChange to this.model's attributeChange event, and in that function you call this.$element.attr() to change attributes or do whatever else you need to do
[21:50:43] <_ob>	 Ok... thanks for the hint
[21:50:58] <_ob>	 yep... got it
[21:52:04] 	 (03CR) 10jenkins-bot: [V: 04-1] readme: Fix "ECMAScript" typo [oojs/core] - 10https://gerrit.wikimedia.org/r/158247 (owner: 10Bartosz Dziewoński)
[21:52:35] <_ob>	 +RoanKattouw: ah... should I black-list the html attributes i manage... or is it fine with storeHTMLAttributes=true
[21:52:41] 	 3VisualEditor / 3MediaWiki integration: VisualEditor should respect [[MediaWiki:Noarticletext]] - 10https://bugzilla.wikimedia.org/70322#c2 (10James Forrester) 5ASSI>3RESO/INV Indeed.
[21:52:47] 	 (03CR) 10Bartosz Dziewoński: "Eh… I see this repo isn't free of stupid test failures, either." [oojs/core] - 10https://gerrit.wikimedia.org/r/158247 (owner: 10Bartosz Dziewoński)
[21:53:36] 	 _ob: No it'll be fine
[21:53:50] 	 Well, hmm, let me think
[21:54:01] 	 You might be in a situation where you need to remove one of those attributes, right?
[21:54:06] 	 In that case I think you do need to blacklist them
[21:54:24] <_ob>	 yep
[22:01:12] 	 (03CR) 10Krinkle: [C: 032] readme: Fix "ECMAScript" typo [oojs/core] - 10https://gerrit.wikimedia.org/r/158247 (owner: 10Bartosz Dziewoński)
[22:01:36] 	 MatmaRex: Yeah, concurrency in saucelabs.
[22:01:55] 	 works API wise, but there's limited quota of parallel workers.
[22:02:19] 	 Works for a small repo, but doesn't scale well.
[22:02:35] 	 James_F, sorry, I had to go and plan university visit stuff. Back now
[22:02:43] 	 Krenair: No worries at all. :-)
[22:02:44] 	 James_F, the result was " James_F: as long as a class pops out in the end I'll be happy ;)" ?
[22:03:03] 	 Krenair: Pretty much. I like "mode=""'
[22:03:06] 	 Bah.
[22:03:21] 	 Krenair: Pretty much. I like 'mode="…"' for consistency with  but beyond that, what do you think?
[22:03:50] 	 James_F, rather than "list-style"?
[22:04:46] 	 Krenair: Yeah; maybe we should invent some semantic names to map to list styles?
[22:05:00] 	 Krenair: That way they can be over-ridden.
[22:05:17] 	 Well as we're gonna do this with classes
[22:06:10] 	 We can just output class="cite-reference-"
[22:06:18] 	 set that to default to what we would expect, and then people can override?
[22:06:36] 	 Sure, the mapping can be done in referenceslists.css or whatever. :-)
[22:06:54] 	 But Gabriel suggested a whitelist of accepted 'modes'.
[22:06:59] 	 Which I think is reasonable.
[22:07:41] 	 Well we can just accept all of the values we know HTML defines for list-style-type
[22:08:12] 	 (03CR) 10jenkins-bot: [V: 04-1] readme: Fix "ECMAScript" typo [oojs/core] - 10https://gerrit.wikimedia.org/r/158247 (owner: 10Bartosz Dziewoński)
[22:08:36] 	 Oy.
[22:08:44] 	 heh.
[22:08:44] 	 ?
[22:09:04] 	 third time's a charm?
[22:09:09] 	 no
[22:09:32] 	 or are the tests actually proper broken?
[22:10:07] 	 James_F, which is what my patch was already doing: https://gerrit.wikimedia.org/r/#/c/120962/6/Cite_body.php
[22:10:28] 	 Krenair: Yeah.
[22:11:41] * gwicke would much prefer semantics over presentation
[22:11:48] 	 Krenair: Restore and fix up?
[22:12:42] 	 James_F, the parsoid part will still need to be done by someone else
[22:13:30] 	 (03CR) 10Catrope: [C: 032] [BREAKING CHANGE] Element mixin refactor [oojs/ui] - 10https://gerrit.wikimedia.org/r/157274 (owner: 10Trevor Parscal)
[22:13:55] 	 (03Restored) 10Alex Monk: WIP Support new column-width & list-style attributes to Cite's  [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/123093 (https://bugzilla.wikimedia.org/51260) (owner: 10Alex Monk)
[22:13:56] 	 Krenair: It's a start.
[22:14:25] 	 Rebasing is going to be such fun I'm sure
[22:14:33] 	 :-)
[22:14:42] 	 Krenair: Cite doesn't actually change that much…
[22:14:46] 	 James_F, has Terry been around the office in the last week by the way?
[22:14:46] 	 yeah
[22:14:54] 	 Krenair: Not seen him, no.
[22:14:59] 	 Okay.
[22:17:07] <_ob>	 +RoanKattouw: that works :)
[22:18:26] 	 3VisualEditor / 3Editing Tools: VisualEditor: Reference group input combo box does not populate the Reference group names if created from Reference list dialog - 10https://bugzilla.wikimedia.org/70365 (10James Forrester) 5NEW>3ASSI p:5Unprio>3Normal
[22:19:37] 	 Krinkle: Did you get added to https://github.com/Rillke/oojs-ui-distro too?
[22:23:46] 	 hmm. `npm install` in oojs/core on windows failed with cryptic errors, but then i ran it again and it succeeded. okay…
[22:24:00] 	 MatmaRex: Lovely.
[22:24:22] 	 can't complain, that's better than most of development-oriented software
[22:24:46] 	 Still, not good.
[22:25:10] 	 James_F: Yes
[22:25:11] 	 want the log file? it's 1.6 MB of text :P
[22:25:29] 	 MatmaRex: I'm sure Krinkle doesn't. ;-)
[22:26:43] 	 Windows isn't officially supported by CI. It works when it does.
[22:30:36] 	 I'm checking oojs locally
[22:30:38] 	 seems to run fine
[22:30:50] 	 checking karma-ci1/2 to be sure as well
[22:30:54] 	 might be a saucelabs issue
[22:32:15] 	 James_F: http://tools.wmflabs.org/oojs-ui/
[22:34:59] 	 James_F,  https://gerrit.wikimedia.org/r/#/c/120962/7..8/Cite_body.php
[22:36:14] 	 Krinkle: Oh, interesting.
[22:36:43] 	 Krinkle: That's really helpful.
[22:37:10] 	 I don't know. I think it should be obsolete, but nice effort.
[22:37:41] 	 Krinkle: Sure, but "should be obsolete" for a change we were working on 9 months ago without effect.
[22:38:01] 	 I was referring to the download, not the demo.
[22:38:45] 	 Krenair: I thought we said we were going with semantic rather than stylistic names for the attribute (and that it'd be called 'mode' not 'list_style')?
[22:39:01] 	 Krinkle: The two being side by side is the value.
[22:39:07] 	 Krinkle: And the history, presumably?
[22:40:52] 	 James_F, yeah I've been fiddling with the parser tests rather than doing that bit
[22:41:00] 	 Krenair: Ah, OK. :-)
[22:41:12] 	 I just made a class pop out at the end to see what breaks
[22:46:33] 	 heh, we use a npm package called 'hooker'.
[22:47:38] 	 James_F: yeah karma-ci2 is consistently failing for me locally
[22:47:39] 	 not a quota thing
[22:47:55] 	 Krinkle: What changed?
[22:48:00] 	 the world
[22:48:09] 	 Krinkle: Is Safari-5 no longer Safari-5?
[22:50:21] 	 James_F: Nah, Safari 5 isnt available
[22:50:27] 	 I get the same via their GUI on saucelabs.com
[22:50:32] 	 Wasn't that what failed before?
[22:50:56] 	 before today? No, never.
[22:52:29] 	 No, in the failures.
[22:52:43] 	 James_F, okay see PS10
[22:53:02] 	 Krinkle: https://integration.wikimedia.org/ci/job/oojs-core-npm/280/console and https://integration.wikimedia.org/ci/job/oojs-core-npm/281/console
[22:53:20] 	 Krinkle: Or is "safari 5 (OSX 10.6)" not Safari 5?
[22:53:28] 	 That was both today
[22:53:41] 	 I just retriggered it.. ?
[22:53:53] 	 Yes, it's deterministic
[22:53:56] 	 I get the same locally
[22:53:57] 	 their VM is fucked
[22:54:06] 	 well gn
[22:54:54] 	 Krinkle: Ah, SauceLabs broke?
[22:55:23] 	 Krenair: Looks good, except for naming of allowed values. Also, should those be i18n able?
[22:56:58] 	 (03PS1) 10Esanders: Only focus the paste target if focusedNode has changed [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/158269 (https://bugzilla.wikimedia.org/70356) 
[22:57:44] 	 (03CR) 10Catrope: [C: 032] [BREAKING CHANGE] Element mixin refactor [oojs/ui] - 10https://gerrit.wikimedia.org/r/157274 (owner: 10Trevor Parscal)
[22:58:08] 	 (03CR) 10jenkins-bot: [V: 04-1] Only focus the paste target if focusedNode has changed [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/158269 (https://bugzilla.wikimedia.org/70356) (owner: 10Esanders)
[22:58:20] 	 James_F: will debug once then going to bed
[22:58:28] 	 Krinkle: Sure.
[22:58:45] 	 (03Merged) 10jenkins-bot: [BREAKING CHANGE] Element mixin refactor [oojs/ui] - 10https://gerrit.wikimedia.org/r/157274 (owner: 10Trevor Parscal)
[22:58:51] 	 Krinkle: ISTR we have issues with Safari 5.1.x but not 5.0.x – but I thought that was OOUI not OOjs.
[22:59:12] 	 Krinkle: If they upgraded their VM from 5.0.x to 5.1.x that could have broken things for us.
[22:59:21] 	 (03PS8) 10Catrope: Automatically align popup toolgroups [oojs/ui] - 10https://gerrit.wikimedia.org/r/157295 (owner: 10Trevor Parscal)
[23:01:52] 	 James_F: Nah, both 5.0 and 5.1 shuld be fine
[23:01:54] 	 no known issues
[23:02:01] 	 remember we support IE6
[23:02:33] 	 Krinkle: We certainly have 5.0 vs. 5.1 issues in VE, though.
[23:02:36] 	 Krinkle: Not sure why.
[23:02:47] 	 Krinkle: (Not been a priority to support. :-))
[23:03:54] 	 Hm.. it's working for me now
[23:04:18] <_ob>	 +RoanKattouw: i need your advice....
[23:04:41] 	 _ob: OK, what's up?
[23:04:42] <_ob>	 +RoanKattouw: i need to manage a redundant data structure (matrix) in ve.ce.TableNode...
[23:04:50] <_ob>	 +RoanKattouw: might there be a good way to detect structural changes on the level of ve.ce.TableNode (row inserts, cell inserts, span changes) to invalidate that cached data structure?
[23:05:00] 	 Maybe
[23:05:03] 	 First question: why do you need it?
[23:05:04] <_ob>	 ;)
[23:05:12] 	 (03CR) 10Krinkle: [C: 032] "I can reproduce from my local command line in isolation (w/ just safari 5) against a SauceLabs account not used by other people. It's not " [oojs/core] - 10https://gerrit.wikimedia.org/r/158247 (owner: 10Bartosz Dziewoński)
[23:05:19] 	 (Not that I don't believe you, but I want to check if there's not some other approach that might be better)
[23:05:30] <_ob>	 the problem are spanning cells...
[23:05:59] <_ob>	 it is really hard to work with them without a global notion of spacial extents
[23:06:06] <_ob>	 locally (cell) this is not possible
[23:06:11] <_ob>	 not even row wise
[23:06:41] <_ob>	 with html cells it is actually known only on table scale
[23:07:00] 	 Right, because you can have cells span vertically as well as horizontally
[23:07:13] <_ob>	 yep...
[23:07:26] 	 What do you use this data structure for?
[23:07:36] 	 (Trying to decide if it makes sense to be in CE, or whether it should be in DM instead)
[23:07:38] <_ob>	 e.g., when you insert a row....
[23:07:57] <_ob>	 you need either insert cells or adapt rowspans of some other row
[23:08:02] <_ob>	 same with cols
[23:08:11] 	 Right!
[23:08:16] 	 You might insert a row in the middle of a span
[23:08:19] 	 That's clever
[23:08:20] 	 (03Merged) 10jenkins-bot: readme: Fix "ECMAScript" typo [oojs/core] - 10https://gerrit.wikimedia.org/r/158247 (owner: 10Bartosz Dziewoński)
[23:08:25] <_ob>	 yep....
[23:08:35] 	 OK so you need this information to determine what should happen when the user says "insert a row here"
[23:08:43] <_ob>	 what i have is a matrix containing cell nodes plus placeholders linking to the owning cells
[23:08:54] <_ob>	 yep...
[23:09:04] 	 "placeholders linking to the owning cells" --> what are those?
[23:09:15] 	 Oh, hold on I think I understand
[23:09:22] <_ob>	 imagine: colspan = 2 -> [cell, placeholder]
[23:09:30] 	 Riiight
[23:09:47] 	 You have a full NxN matrix but there are fewer than N^2 cells, so some cells are referred to more than once
[23:09:54] <_ob>	 the placeholders allow me to go back to the origin of the problem easily
[23:10:01] <_ob>	 yep... sparse
[23:10:11] 	 And you have primary and secondary references
[23:10:21] <_ob>	 primary, secondary?
[23:10:48] <_ob>	 what u mean with that?
[23:10:51] 	 What you call a placeholder is what I call a secondary reference I think
[23:11:03] 	 You have a "cell" in your NxN matrix (let's call that a "position")
[23:11:04] <_ob>	 ok... it is basically a reference
[23:11:15] 	 And each position refers to a cell, but multiple positions can refer to the same cell
[23:11:19] <_ob>	 placeholder is not the correct term i think
[23:11:25] <_ob>	 yep
[23:11:47] 	 OK
[23:11:49] 	 So, let's see
[23:11:50] <_ob>	 colspan=2, rowspan=2 -> [cell, ref; ref, ref]
[23:12:01] 	 (03PS1) 10Jforrester: Revert "Build with clean CSS" [oojs/ui] - 10https://gerrit.wikimedia.org/r/158271 
[23:12:16] 	 It's probably sensible to maintain this information in dm.TableNode
[23:12:34] 	 We don't have much of a precedent for maintaining that much auxiliary information yet
[23:12:41] <_ob>	 hmmm... is there a nice way to get from dm.Node to ce.Node?
[23:12:53] 	 Not really. Why do you need ot?
[23:13:12] 	 (It's possible but there isn't a direct reference, on purpose)
[23:13:16] <_ob>	 good question ;)
[23:13:27] <_ob>	 yes that should be that way
[23:13:49] 	 CE has a reference to DM but DM has no reference to CE. That way we theoretically allow for multiple views (of multiple different types) to exist on the same model
[23:13:53] <_ob>	 I think in certain moments I would love to attach a class to certain cells...
[23:14:01] <_ob>	 then the $elements are nice...
[23:14:04] 	 Yeah
[23:14:24] 	 OK well it is doable, as long as you have a reference to the ce.Document
[23:14:30] <_ob>	 that's why I am working in ce domain --- but felt bad from the first moment...
[23:14:33] 	 But it sounds like maybe we can solve that later?
[23:14:57] <_ob>	 Let me explain the complexity estimation currently
[23:15:10] 	 (03PS9) 10Catrope: Automatically align popup toolgroups [oojs/ui] - 10https://gerrit.wikimedia.org/r/157295 (owner: 10Trevor Parscal)
[23:15:27] <_ob>	 I currently update that on every model change received in TableNode
[23:15:33] 	 (03CR) 10Catrope: [C: 032] Revert "Build with clean CSS" [oojs/ui] - 10https://gerrit.wikimedia.org/r/158271 (owner: 10Jforrester)
[23:15:39] <_ob>	 update is O(n) cells
[23:16:02] 	 Yeah that was my first lazy suggestion, listen to the update event on dm.TableNode and rebuild from scratch every time. Obviously not the most efficient way
[23:16:23] <_ob>	 But it is actually not too expensive too...
[23:16:38] 	 Well the table might be large relative to the change
[23:16:39] 	 (03Merged) 10jenkins-bot: Revert "Build with clean CSS" [oojs/ui] - 10https://gerrit.wikimedia.org/r/158271 (owner: 10Jforrester)
[23:17:04] <_ob>	 Actually, now on each input (changes outer range)
[23:17:08] 	 It would be nice if the event gave you a way of knowing what changed
[23:17:16] 	 Ouch!
[23:17:16] <_ob>	 exactly...
[23:17:22] 	 Wait, you store ranges?
[23:17:27] <_ob>	 no...
[23:17:57] <_ob>	 but i tried to listen to onModelUpdate
[23:18:10] 	 Oh, and that gives you selection changes too?
[23:18:24] 	 You mean onModelUpdate on the surface?
[23:18:40] <_ob>	 not selection... but impacts on node.getOuterRange()... (not too exact sorry)
[23:18:58] <_ob>	 I could look at debug stack.... and be more precise ;)
[23:21:03] <_ob>	 dm.Node.setLength emits 'update'
[23:21:27] <_ob>	 and setLength is called on every child's range change
[23:21:42] <_ob>	 maybe there is a better event to listen too?
[23:21:52] 	 Oh, size changes
[23:21:53] 	 Yeah that makes sense
[23:21:57] <_ob>	 yep
[23:22:14] 	 When you type in a cell, its length will change, and an update event will bubble up to the TableNode
[23:22:22] <_ob>	 correct
[23:22:42] 	 So yeah it would be nice to 1) filter events that don't alter the structure of the table, and/or 2) for each event, narrow down the scope of the change
[23:22:49] 	 If we have #2 we don't really need #1 as much
[23:22:50] <_ob>	 unfortunately to Tansaction is provided or anything to skip those primitive ones
[23:22:58] 	 3VisualEditor / 3Editing Tools: VisualEditor: Category suggestion lists no longer existing categories - 10https://bugzilla.wikimedia.org/70025 (10WhatamIdoing)
[23:22:58] 	 3VisualEditor: Category suggestions include non-existent and implausible suggestions - 10https://bugzilla.wikimedia.org/69927 (10WhatamIdoing)
[23:23:05] 	 Oh I think Transaction objects might have an affected range getter now!
[23:23:37] <_ob>	 sorry, typo ... _no_ transaction delivered in that emit
[23:23:37] 	 Yeah tx.getModifiedRange()
[23:23:44] 	 Ugh
[23:23:53] 	 Well, you could also listen for 'transact' on the dm.Document
[23:23:59] <_ob>	 hmpf ;)
[23:24:01] 	 And ignore events whose transactions are not in your range
[23:24:05] 	 Which sucks, I know
[23:24:07] <_ob>	 yep
[23:24:13] <_ob>	 but that is an option
[23:24:16] <_ob>	 :)
[23:24:47] 	 Alternatively we could change the update event so that it passes the affected child/descendant as a parameter
[23:26:13] <_ob>	 maybe a more pragmatic approach would be to have 'update' events provide something more than undefined ;)
[23:26:40] 	 Yeah that's what I meant, maybe do something like parentNode.emit( 'update', changedNode );
[23:26:48] <_ob>	 ok.
[23:27:10] 	 (In practice it would probably have to be an array of nodes I suppose)
[23:27:11] <_ob>	 I think it would be an achievment not to invalidate on primitive length changes
[23:27:18] <_ob>	 ok
[23:27:24] 	 Yeah that would be ince
[23:27:30] 	 Or if we do, just invalidate the one cell
[23:28:26] <_ob>	 for sake of code simplicity i would regenerate anyways...
[23:28:54] <_ob>	 to hard to transfer local changes incrementally to global matrix
[23:29:15] <_ob>	 or maybe not ;)
[23:29:54] 	 (03CR) 10Catrope: [C: 032] Automatically align popup toolgroups [oojs/ui] - 10https://gerrit.wikimedia.org/r/157295 (owner: 10Trevor Parscal)
[23:30:42] 	 (03Merged) 10jenkins-bot: Automatically align popup toolgroups [oojs/ui] - 10https://gerrit.wikimedia.org/r/157295 (owner: 10Trevor Parscal)
[23:31:58] <_ob>	 +RoanKattouw: i should rename 'Placeholder' to 'CellReference'?
[23:32:23] 	 Maybe you could explain what Placeholder is exactly?
[23:32:38] 	 It's only used in places where there would be a cell but there isn't one due to colspan/rowspan, is that right?
[23:32:39] <_ob>	 hmmm ... stupid choice ;)
[23:32:45] 	 And then it points to the cell that "owns" it?
[23:32:49] <_ob>	 right
[23:32:55] <_ob>	 it's a ref 
[23:33:01] 	 OK
[23:33:08] 	 But at the cell itself, what kind of data structure do you have?
[23:33:14] 	 3VisualEditor: VisualEditor:  Do not permit empty section headings (wikitext like == == or similar) - 10https://bugzilla.wikimedia.org/70368 (10WhatamIdoing) 3NEW p:3Unprio s:3normal a:3None Sometimes users set the format of a line to be a == Section heading ==, but never put anything in it.  In that c...
[23:33:17] 	 Is there more data there than just a pointer to the TableCellNode?
[23:33:22] <_ob>	 currently... ce.TableCellNode
[23:33:27] <_ob>	 nop
[23:33:33] 	 Right. So is there a reason to have indirection?
[23:33:38] <_ob>	 cells and {owner: cell}
[23:33:50] 	 Is it important to know which one is the "real" cell and which one is a placeholder?
[23:33:51] <_ob>	 yep... for look up the owner
[23:34:03] <_ob>	 only ref -> cell
[23:34:17] <_ob>	 distinction is important
[23:34:43] <_ob>	 reasoning on cells is done by looking at colspan rowspan
[23:34:57] <_ob>	 no reasoning on those refs
[23:35:05] 	 Oh OK
[23:36:06] 	 Then I don't really care if it's called a placeholder or a reference
[23:36:22] 	 If I start caring in the future, we can worry about it in the future :)
[23:36:25] <_ob>	 yes.. but... names are important ;)
[23:36:59] <_ob>	 what about getting that from ce to dm....
[23:37:11] 	 So, one thing I didn't get: when you say we reason on cells but not on refs/placeholders, are the "cells" you reason on the entries in the matrix (but only those that are "real"), or are they ce.TableCellNodes?
[23:37:42] <_ob>	 they are the real nodes
[23:37:45] <_ob>	 no copies
[23:37:47] 	 OK
[23:38:13] 	 So then in a way, all entries in the matrix are references, right? And there isn't really a difference between the "real" reference to a cell and the (possibly many) "placeholder" references to it?
[23:38:24] <_ob>	 and when I need to know if a cell needs to span more then i look into the actual span attribute
[23:38:35] <_ob>	 as an example...
[23:39:20] <_ob>	 it would be great if i could get this down to dm.TableNode...
[23:40:04] <_ob>	 only, that I need to be able to get $elements at certain times 'easily'
[23:40:33] 	 Regarding partial rebuilds: I think it should be reasonably doable to do a row-scoped rebuild, because row-scoped is how things are organized in the DOM and the DM tree already
[23:40:42] 	 Regarding getting $element: you can get it, as follows
[23:41:17] 	 If you are in CE land, but you're reaching into the DM to get a node reference, then you can translate that node reference back to CE
[23:41:46] 	 Say you have var dmCellNode = ceTableNode.getModel().getCellAt( 5, 2 );
[23:42:26] 	 Then you can do ceCellNode = ceTableNode.getDocument().getNodeFromOffset( dmCellNode.getOffset() );
[23:42:59] <_ob>	 yep... but wasn't there a comment telling that this is rather 'inefficient'?
[23:43:01] 	 Basically this uses the fact that nodes in the DM and CE trees have the same offsets, and does a traversal twice (once to compute the DM node's offset, then another to find the CE node at that offset)
[23:43:08] 	 Yeah it's not the most efficient way to do things ever
[23:43:42] 	 So if you need $element often, storing this matrix in CE might be better
[23:43:42] <_ob>	 ok.... I will think about that... :)
[23:43:47] 	 What do you need $element for, though?
[23:44:10] <_ob>	 highlighing selected cells, displaying overlays....
[23:44:33] <_ob>	 that changes in scale of selectionChanges
[23:45:51] 	 Krinkle|detached, James_F: do you know a good person to ask for fix-ups in enwiki's MediaWiki:Vector.css ?
[23:45:53] 	 But to highlight selected cells, don't you already need to translate from offsets to nodes?
[23:46:00] 	 Because the selection is a ve.Range
[23:46:10] <_ob>	 yep... that I do already
[23:46:32] <_ob>	 but only for two cells per selection change
[23:46:43] <_ob>	 cells defining the selected region
[23:47:13] <_ob>	 let me think about it a bit more...
[23:47:32] <_ob>	 probably you are right... and i can derive what i need form those two...
[23:47:59] 	 gwicke: i have global editinterface if it's urgent
[23:48:24] 	 gwicke: otherwise i recommend {{editprotected}} on the talk page
[23:48:41] 	 MatmaRex: it's not super urgent, more a clean-up
[23:48:42] <_ob>	 I should keep traversal code general (for dm+ce)... then I maybe could solve that with same code.
[23:48:50] 	 the people who respond to these on JS and CSS pages are very reasonable people (unlike those who like to respond for interface and template changes)
[23:49:05] 	 okay, thanks!
[23:49:11] 	 will ask on the talk page then
[23:49:41] 	 (03PS4) 10Catrope: Update OOjs UI to v0.1.0-pre (99f02229d7) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/157286 (owner: 10Trevor Parscal)
[23:50:39] <_ob>	 +RoanKattouw: I will meditate on that... thank you for now :)
[23:51:28] 	 _ob: Don't worry too much about moving that from CE to DM for now, do what's easiest for now
[23:51:33] 	 We can clean things up alter
[23:51:34] 	 *later
[23:51:36] 	 (03CR) 10jenkins-bot: [V: 04-1] Update OOjs UI to v0.1.0-pre (99f02229d7) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/157286 (owner: 10Trevor Parscal)
[23:53:13] 	 wtf
[23:53:20] 	 Krinkle|detached: Why the hell does jsduck run on *.css files?
[23:53:26] * RoanKattouw replaces /** with /* in CSS files
[23:53:47] 	 RoanKattouw: haha, /*! was the convention last time i had to deal with that
[23:54:07] 	 jsduck can generate docs for css, too… not sure what for
[23:56:31] 	 what the hell was I thinking when I wrote this code 5 months ago?
[23:58:15] 	 (03PS1) 10Catrope: Don't use /** in CSS, use /* instead [oojs/ui] - 10https://gerrit.wikimedia.org/r/158292 
[23:58:32] 	 James_F: ---^^ +2 please
[23:58:41] 	 3VisualEditor: VisualEditor: Duplicating categories (and possibly also something else) - 10https://bugzilla.wikimedia.org/54169#c3 (10WhatamIdoing) https://fr.wikipedia.org/w/index.php?title=Bello_de_Carcassonne&diff=107097528&oldid=107091155 duplicated two categories and added others, but not within a ref tag...
[23:59:19] 	 (03PS2) 10Catrope: Only focus the paste target if focusedNode has changed [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/158269 (https://bugzilla.wikimedia.org/70356) (owner: 10Esanders)
[23:59:28] 	 (03PS2) 10Jforrester: Don't use /** in CSS or LESS, use /* instead [oojs/ui] - 10https://gerrit.wikimedia.org/r/158292 (owner: 10Catrope)
[23:59:34] 	 (03CR) 10Jforrester: [C: 032] Don't use /** in CSS or LESS, use /* instead [oojs/ui] - 10https://gerrit.wikimedia.org/r/158292 (owner: 10Catrope)
[23:59:56] 	 (03CR) 10Catrope: [C: 032] Only focus the paste target if focusedNode has changed [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/158269 (https://bugzilla.wikimedia.org/70356) (owner: 10Esanders)