[00:06:50] (03PS17) 10Catrope: Make category widget items draggable [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/173360 (https://bugzilla.wikimedia.org/50809) (owner: 10Alex Monk) [00:07:44] 3VisualEditor-EditingTools, VisualEditor: Permit sorting of categories via drag-and-drop - https://phabricator.wikimedia.org/T52809#547141 (10Jdforrester-WMF) [00:07:51] (03CR) 10Catrope: [C: 032] Make category widget items draggable [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/173360 (https://bugzilla.wikimedia.org/50809) (owner: 10Alex Monk) [00:08:48] RoanKattouw: https://phabricator.wikimedia.org/T88667 [00:11:36] (03Merged) 10jenkins-bot: Make category widget items draggable [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/173360 (https://bugzilla.wikimedia.org/50809) (owner: 10Alex Monk) [00:20:18] RoanKattouw, James_F so yeah I took another look and it seems there are no hooks for checking the result of a save attempt :( [00:20:37] RoanKattouw: i can merge, but do you want to do a quick scan of https://gerrit.wikimedia.org/r/#/c/188949/ for sanity? [00:23:13] doesn't appear to be a good place in EditPage to place a new hook for it [00:24:51] (03PS1) 10Trevor Parscal: [BREAKING CHANGE] Remove window isolation [oojs/ui] - 10https://gerrit.wikimedia.org/r/188951 [00:26:17] (03PS2) 10Trevor Parscal: [BREAKING CHANGE] Remove window isolation [oojs/ui] - 10https://gerrit.wikimedia.org/r/188951 [00:27:51] (03PS3) 10Catrope: [BREAKING CHANGE] Remove window isolation [oojs/ui] - 10https://gerrit.wikimedia.org/r/188951 (owner: 10Trevor Parscal) [00:30:22] 3VisualEditor-EditingTools, VisualEditor: Permit sorting of categories via drag-and-drop - https://phabricator.wikimedia.org/T52809#1019840 (10Jdforrester-WMF) 5Open>3Resolved [00:33:35] 3OOjs-UI, UI-Standardization: Shadow at the bottom of buttons, dropdown menus is intentionally vertically mis-aligned by a pixel or two in the MediaWiki theme - https://phabricator.wikimedia.org/T78083#1019850 (10Etonkovidova) [00:41:22] 3VisualEditor: Media Settings - Infinite loop when an image is inserted - https://phabricator.wikimedia.org/T88764#1019864 (10Etonkovidova) 3NEW a:3Mooeypoo [00:42:18] (03PS6) 10Mooeypoo: Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) [00:44:11] (03CR) 10jenkins-bot: [V: 04-1] Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) (owner: 10Mooeypoo) [00:49:04] 3VisualEditor: Media Settings: images displayed with uneven heights - https://phabricator.wikimedia.org/T88765#1019894 (10Etonkovidova) 3NEW [00:50:59] (03PS7) 10Catrope: Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) (owner: 10Mooeypoo) [00:52:11] 3OOjs-UI: Textfield widget uses role=textbox for no reason - https://phabricator.wikimedia.org/T87830#1019903 (10Etonkovidova) [00:52:20] (03PS8) 10Catrope: Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) (owner: 10Mooeypoo) [00:53:59] (03CR) 10jenkins-bot: [V: 04-1] Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) (owner: 10Mooeypoo) [00:58:51] (03PS9) 10Mooeypoo: Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) [01:01:02] (03PS1) 10Trevor Parscal: Liberation from the oppressive this.$ regime [oojs/ui] - 10https://gerrit.wikimedia.org/r/188957 [01:01:31] (03PS2) 10Trevor Parscal: Liberation from the oppressive this.$ regime [oojs/ui] - 10https://gerrit.wikimedia.org/r/188957 [01:02:41] (03PS10) 10Mooeypoo: Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) [01:03:18] (03CR) 10Krinkle: [C: 04-1] [BREAKING CHANGE] Remove window isolation (032 comments) [oojs/ui] - 10https://gerrit.wikimedia.org/r/188951 (owner: 10Trevor Parscal) [01:05:12] (03CR) 10Krinkle: "U+2020" [oojs/ui] - 10https://gerrit.wikimedia.org/r/188957 (owner: 10Trevor Parscal) [01:20:58] 3WikiEditor: Special Character inserter in wikitext editor is missing some IPA combinations - https://phabricator.wikimedia.org/T88767#1019946 (10Whatamidoing-WMF) 3NEW [01:45:01] (03PS11) 10Mooeypoo: Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) [03:11:05] (03PS1) 10Ryasmeen: [BrowserTest] [WIP]Fixing the test media_general_settings.feature [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188969 [03:11:09] (03CR) 10jenkins-bot: [V: 04-1] [BrowserTest] [WIP]Fixing the test media_general_settings.feature [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188969 (owner: 10Ryasmeen) [03:24:28] ori: So as I remarked earlier, the unit for up/down throughput throttling isn't documented. I found out that it's bytes/sec by tracking down the commit that introduced the UI for this feature (if you click the phone icon in the inspector, you can control the network as well as the screen size) and reading the code mapping UI strings like '2Mbps' to values like 2048 * 1024 / 8 [03:24:30] Oh, Crome [03:24:33] *Chrome [05:58:54] (03PS1) 10Krinkle: demo: Use properties instead of attributes for [oojs/ui] - 10https://gerrit.wikimedia.org/r/188974 [05:59:28] edsanders|away: James_F|Away: https://www.w3.org/Bugs/Public/show_bug.cgi?id=27677#c2 [09:16:20] 3MediaWiki-Page-editing, MediaWiki-Core-Team: Use parentRevId field for section change merging instead of timestamps - https://phabricator.wikimedia.org/T88734#1020348 (10Schnark) Sounds similar to what I suggested in T34037#366435. [12:05:15] 3MediaWiki-Page-editing, MediaWiki-extensions-AbuseFilter: Incorrect tagging of edits on some Wikipedias – AbuseFilter's diff and the real diff are not the same - https://phabricator.wikimedia.org/T73947#1020438 (10matmarex) [13:44:01] 3OOjs-UI: Compress OOjs UI PNGs with pngcrush -brute -reduce - https://phabricator.wikimedia.org/T76148#1020578 (10Nemo_bis) 5declined>3Open Can't really be declined until the development guidelines are changed for this to be in compliance with them. [13:50:39] 3OOjs-UI: Compress OOjs UI PNGs with pngcrush -brute -reduce - https://phabricator.wikimedia.org/T76148#1020583 (10Ricordisamoa) Why not to wait for T76473 and https://gerrit.wikimedia.org/r/188289/ instead? [13:58:53] 3OOjs-UI: Compress OOjs UI PNGs with pngcrush -brute -reduce - https://phabricator.wikimedia.org/T76148#1020592 (10matmarex) [14:04:22] Hello I would like to use visual editor on my Mediawiki but I php error, that tells me that I should update to the latest wmf branch, but I do not know how to switch to 1.25 ver. 16 [14:44:30] (03PS6) 10Esanders: Migrate to ve.config (except in init where it doesn't exist) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/187144 [14:46:31] (03CR) 10jenkins-bot: [V: 04-1] Migrate to ve.config (except in init where it doesn't exist) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/187144 (owner: 10Esanders) [14:58:30] (03CR) 10Esanders: [C: 04-1] "I also talked to Roan about the window transition/timings. He can fill you in and tell you if your CSS change is safe." (031 comment) [oojs/ui] - 10https://gerrit.wikimedia.org/r/188951 (owner: 10Trevor Parscal) [14:59:32] (03CR) 10Esanders: [C: 032] demo: Use properties instead of attributes for [oojs/ui] - 10https://gerrit.wikimedia.org/r/188974 (owner: 10Krinkle) [15:00:51] (03Merged) 10jenkins-bot: demo: Use properties instead of attributes for [oojs/ui] - 10https://gerrit.wikimedia.org/r/188974 (owner: 10Krinkle) [15:09:28] 3VisualEditor-EditingTools, VisualEditor: Various dialog windows gain and then lose the focus after opening - https://phabricator.wikimedia.org/T88411#1020641 (10Juandev) Actually still for a template dialog at cs.wp. [15:11:41] (03CR) 10Esanders: [C: 04-1] "The issue here is that we allow arbitrary input types. Clearly passing type='checkbox' would be insane for a TextInputWidget, so valid one" [oojs/ui] - 10https://gerrit.wikimedia.org/r/188425 (owner: 10Ricordisamoa) [15:13:03] (03CR) 10Esanders: [C: 032] Deprecate GridLayout [oojs/ui] - 10https://gerrit.wikimedia.org/r/188826 (owner: 10Bartosz Dziewoński) [15:14:21] (03Merged) 10jenkins-bot: Deprecate GridLayout [oojs/ui] - 10https://gerrit.wikimedia.org/r/188826 (owner: 10Bartosz Dziewoński) [15:44:30] (03CR) 10Bartosz Dziewoński: "This most definitely broke button spacing, check out the PHP demo." [oojs/ui] - 10https://gerrit.wikimedia.org/r/188874 (owner: 10Catrope) [15:52:47] 3VisualEditor-EditingTools, VisualEditor: Various dialog windows gain and then lose the focus after opening - https://phabricator.wikimedia.org/T88411#1020679 (10Aklapper) >>! In T88411#1020641, @Juandev wrote: > Actually still for a template dialog at cs.wp. Please provide specific links (or steps to reproduce). [16:21:05] 3VisualEditor-EditingTools, VisualEditor: Various dialog windows gain and then lose the focus after opening - https://phabricator.wikimedia.org/T88411#1020691 (10matmarex) The fix for T85484 was not deployed yet, I think – the bug is marked for "WMF-deploy-2015-02-04_(1.25wmf16)", cs.wp is on wmf15 according to... [16:42:09] 3OOjs-UI: Have a larger size of dialog than 'large' - https://phabricator.wikimedia.org/T86396#1020717 (10Etonkovidova) [16:51:36] 3VisualEditor-MediaWiki, VisualEditor: Link existence checker in LinkTargetInputWidget doesn't work for interwikis, language links - https://phabricator.wikimedia.org/T72504#1020734 (10Etonkovidova) [16:52:31] 3VisualEditor-MediaWiki-Media, VisualEditor: Make the media dialog bigger when searching for media to insert - https://phabricator.wikimedia.org/T86397#1020735 (10Etonkovidova) [16:54:39] 3OOjs-UI: In MediaWiki theme, some thin grey lines in dialogs do not appear in some configurations - https://phabricator.wikimedia.org/T85226#1020737 (10Etonkovidova) [16:58:39] 3VisualEditor-EditingTools, VisualEditor: Make it easy to identify the file type (e.g., .png versus .svg) when adding an image - https://phabricator.wikimedia.org/T74061#1020739 (10Etonkovidova) [17:02:17] 3OOjs-UI, VisualEditor: MediaWiki theme: Border radius used inconsistently - https://phabricator.wikimedia.org/T85973#1020756 (10Etonkovidova) [17:04:07] 3VisualEditor-MediaWiki-Templates, VisualEditor: Insert/Template inspector should put focus on input text_field when invoked but no longer does - https://phabricator.wikimedia.org/T85484#1020768 (10Etonkovidova) [17:06:53] 3VisualEditor-MediaWiki-Media, VisualEditor: In the VisualEditor media selector, show the name of each file so it can be identified - https://phabricator.wikimedia.org/T54459#1020788 (10Etonkovidova) [17:18:23] 3VisualEditor-EditingTools, VisualEditor: A number of VisualEditor dialogs do not cover full screen when 'full' size - https://phabricator.wikimedia.org/T76863#1020806 (10Etonkovidova) [17:30:35] Corruption alert: visualeditor-needcheck on frwiki: https://fr.wikipedia.org/?diff=111594269 [17:30:35] Corruption alert: visualeditor-needcheck on frwiki: https://fr.wikipedia.org/?diff=111594357 [17:30:35] Corruption alert: visualeditor-needcheck on frwiki: https://fr.wikipedia.org/?diff=111594572 [17:30:35] Corruption alert: visualeditor-needcheck on frwiki: https://fr.wikipedia.org/?diff=111594794 [17:30:35] Corruption alert: visualeditor-needcheck on frwiki: https://fr.wikipedia.org/?diff=111594968 [17:30:36] Corruption alert: visualeditor-needcheck on frwiki: https://fr.wikipedia.org/?diff=111603712 [17:30:36] Corruption alert: visualeditor-needcheck on zhwiki: https://zh.wikipedia.org/?diff=34228286 [17:30:37] Corruption alert: visualeditor-needcheck on zhwiki: https://zh.wikipedia.org/?diff=34228349 [17:30:37] Corruption alert: visualeditor-needcheck on zhwiki: https://zh.wikipedia.org/?diff=34228697 [17:30:38] Corruption alert: visualeditor-needcheck on zhwiki: https://zh.wikipedia.org/?diff=34228716 [17:30:38] Corruption alert: visualeditor-needcheck on hewiki: https://he.wikipedia.org/?diff=16518914 [17:30:39] Corruption alert: visualeditor-needcheck on hewiki: https://he.wikipedia.org/?diff=16522138 [17:30:39] Corruption alert: visualeditor-needcheck on ruwiki: https://ru.wikipedia.org/?diff=68412484 [17:30:40] Corruption alert: visualeditor-needcheck on ruwiki: https://ru.wikipedia.org/?diff=68412488 [17:42:47] mooepoo: Media Settings box - Cancel, Back, Use this image do not work in test2. Is it still work in progress? [17:43:30] (03PS12) 10Mooeypoo: Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) [17:43:43] etonkovidova, that's the patch ^^ [17:43:58] great! thx [17:44:05] etonkovidova, can you give me a test link, though? it worked for me before on test2 and it works on beta [17:44:08] and master [17:46:32] mooepooo: hmm... let me log out/log in again [17:49:32] mooeypoo: I think it's more specific case: try to search for After Cornelis Cornelisz. van Haarlem 001.jpg [17:50:05] mooeypoo: seems like the search results cannot be fully loaded [17:50:38] mooeypoo: oh, and Cancel etc works - generally [18:16:08] etonkovidova, yeah, that's the case of only 1 image [18:16:17] etonkovidova, that's going to be fixed with the patch above [18:16:45] ok [18:18:16] RoanKattouw_away, so ApiEditPage (and presumably friends such as the caller we're after) calls $editpage->internalAttemptSave which just returns the status object [18:18:26] We could run a hook in each of the callers [18:32:28] Ugh [18:36:11] RoanKattouw, two slightly urgent reviews for bugfixes -- this fixes a bug in category widget (it doesn't pop up) https://gerrit.wikimedia.org/r/#/c/187618/ and this fixes a rather major loop bug in media search https://gerrit.wikimedia.org/r/#/c/188725/ [18:37:03] (03PS13) 10Mooeypoo: Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) [18:39:49] Krenair: I was going to suggest running a hook from internalAttemptSave before returning, but there are like a dozen return points :( [18:39:59] yup. [18:41:49] Looks like WikiPage::doEditContent() (OMG the name) does fire the PageContentSave/ArticleSave hooks, but there are other failure paths in EditPage that you would miss [18:41:52] Urrrgh [18:42:31] Hmm I think there is one corner you can cut [18:42:37] You're instrumenting the wikitext editor, right? [18:42:42] So you don't really care about API edits? [18:43:07] In that case could you put something in EditPage::attemptSave()? [18:43:58] I don't think that would get approved. [18:44:02] At least it probably shouldn't. [18:45:33] There's not much sense doing it for EditAction but not ApiActionEdit... [18:47:15] RoanKattouw, could just call it in EditPage::attemptSave and ApiEditPage::execute of course, but ugh [18:47:31] Oh, I see what you're saying [18:47:44] If you're gonna put a hook in core, it should be generic [18:47:58] RoanKattouw: would you like me to move vbench to a separate, git-deployed repository so you could merge patches to it? (I can even move it to the VisualEditor repo if you like) [18:48:09] Yes, not run for EditAction but not ApiActionEdit. It should work for both. [18:48:12] Well I suppose you could rearrange some code in internalAttemptSave so it's easier to run a hook at the end, but that sounds a bit painful [18:48:15] ori: Yes please [18:48:28] ori: Also once you're here I need you to look at an issue where I can't get images to load at all any more [18:48:51] ok [18:48:55] Not sure if it's due to Chromium config or something else but I now can't reproduce the waves of image requests even with clean code [18:49:04] I got latency simulation working well though [18:49:36] handle status returned by attemptSave in EditPage::edit, use attemptSave to fire hook instead, convert ApiEditPage::execute to call attemptSave? [18:50:11] Unless something external is calling EditPage::attemptSave. That would break stuff. [18:50:21] Sigh. [18:51:23] hi. [18:52:15] Ooh I see [18:52:35] I was about to say "you can't call attemptSave from the API because it calls handleStatus" but you're proposing moving up handleStatus [18:52:41] That could potentially work [18:52:41] yes [18:52:46] except for breaking non-core callers [18:52:53] As long as you grep for any callers of attemptSave [18:53:12] Which given the "API" of that function (it may randomly decide to call $wgOut->redirect() for you) I would be surprised if there were any [18:53:44] (Confusingly, there is a hook called "EditPage::attemptSave" already - it runs at the start of internalAttemptSave) [18:54:04] So I guess we'd add EditPage::beforeAttemptSave or something [18:54:14] sorry, not before [18:54:25] after [18:54:26] "EditPage::afterSaveAttempt" [18:54:28] right [18:59:41] (03CR) 10Catrope: [C: 04-1] Use css class instead of jquery show/hide (031 comment) [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/187618 (https://phabricator.wikimedia.org/T87420) (owner: 10Mooeypoo) [19:01:13] RoanKattouw: was https://gerrit.wikimedia.org/r/#/c/188874/ supposed to fix some bug ir was it just code cleanup? [19:01:22] MatmaRex: The button margin thing? [19:01:55] The reason I did that is because ButtonWidgets are used as DraggableElements in MWCategoryWidget, and DraggableGroupElement doesn't like it when its DraggableElements have margins [19:02:23] I don't know why the PHP demo buttons don't have their own margin [19:02:38] But in the cases I could find, ButtonElements already had 0.75em margin providing plenty of spacing [19:02:47] I'll look at the PHP demo to check why that's not being applied there [19:02:57] RoanKattouw: https://phabricator.wikimedia.org/T76643 [19:03:25] Oooh [19:04:08] (03PS11) 10Mooeypoo: Use css class instead of jquery show/hide [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/187618 (https://phabricator.wikimedia.org/T87420) [19:04:18] I couldn't find a case where there was a problem for ButtonWidget but I must not have looked closely enough. I'll look [19:04:40] I might revert my change and instead fix the issue in DraggableGroupElement, which may not be as unfixable as I first thought [19:04:56] basically this is meant to behave nicely when you have several inline-block widget siblings [19:05:07] Yeah, which is exactly what CategoryWidget has [19:05:24] And there it worked fine without the inline spacing stuff because ButtonWidget already provided a 0.75em margin [19:05:34] (03CR) 10Catrope: [C: 032] Use css class instead of jquery show/hide [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/187618 (https://phabricator.wikimedia.org/T87420) (owner: 10Mooeypoo) [19:05:42] where is the 0.75em margin? i see no such thing [19:05:45] We also don't appear to have a test case for this in the JS demo? [19:06:11] yeah [19:06:12] .ve-ui-mwCategoryItemWidget > .oo-ui-buttonElement-button { margin: 0.75em 0.75em 0 0; } [19:06:15] Well that would explain why [19:06:17] I feel stupid now [19:06:24] Go ahead and revert my change [19:06:46] (03PS14) 10Mooeypoo: Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) [19:07:10] hah, yeah, so the combination of the two caused doubled margins, basically [19:07:25] Yeah [19:07:27] 3MediaWiki-ResourceLoader, Mobile-Web: Introduce MediaWiki:Mainpage.css - https://phabricator.wikimedia.org/T78418#1021016 (10Legoktm) [19:08:03] (03Merged) 10jenkins-bot: Use css class instead of jquery show/hide [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/187618 (https://phabricator.wikimedia.org/T87420) (owner: 10Mooeypoo) [19:08:29] RoanKattouw, another fun thing to filter out of the grep was a bunch of 'attemptSave' functions in Wikibase/Wikidata and all their callers [19:08:52] (03PS1) 10Bartosz Dziewoński: Revert "Remove inline spacing from ButtonWidget" [oojs/ui] - 10https://gerrit.wikimedia.org/r/189041 [19:09:06] RoanKattouw: ^ [19:09:15] but... otherwise seems there's no direct callers [19:10:02] (03CR) 10Catrope: [C: 032] "Sorry :S" [oojs/ui] - 10https://gerrit.wikimedia.org/r/189041 (owner: 10Bartosz Dziewoński) [19:12:02] (03Merged) 10jenkins-bot: Revert "Remove inline spacing from ButtonWidget" [oojs/ui] - 10https://gerrit.wikimedia.org/r/189041 (owner: 10Bartosz Dziewoński) [19:12:22] RoanKattouw: i'll fix up MWCategoryWidget [19:12:28] Thanks [19:12:42] I'll fix DraggableGroupElement [19:14:15] MatmaRex: One thing that would be nice is if the MWCategoryItem boxes were symmetrical [19:14:38] Right now DraggableGroupElement draws insertion lines between elements based on their box boundaries [19:15:04] But for things spaced with .oo-ui-inline-spacing() those lines will visually appear to be at the left edge of the rightmost element [19:15:08] Because the margin is asymmetrical [19:15:39] If instead of 0.5em on the right we had 0.25em on the left and 0.25em on the right, the line would be drawn correctly [19:16:12] (There's also the problem that DraggableGroupElement's current width measurements don't take margin into account, and so the right edge of element N isn't the same as the left edge of element N+1, but I can fix that) [19:16:27] hmph [19:17:22] wouldn't the margin collapse then, anyway? [19:17:33] Collapse how? [19:17:54] nope, nevermind [19:18:14] (I don't actually know as much about CSS as I might appear to, so you may have to talk to me like I'm a noob. I have no idea what margin collapsing is) [19:18:16] (03PS15) 10Mooeypoo: Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) [19:18:20] (03CR) 10jenkins-bot: [V: 04-1] Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) (owner: 10Mooeypoo) [19:18:26] RoanKattouw, justify this test case: [19:18:27] range: new ve.Range( 0 ), [19:18:28] mode: 'leaves', [19:18:28] expected: [ [19:18:28] // document [19:18:28] { [19:18:28] node: [], [19:18:30] range: new ve.Range( 0 ), [19:18:32] // no 'index' because documentNode has no parent [19:18:34] indexInNode: 0, [19:18:36] nodeRange: new ve.Range( 0, 63 ), [19:18:38] nodeOuterRange: new ve.Range( 0, 63 ) [19:18:40] } [19:18:42] ], [19:18:51] how can 'leaves' return the document? [19:19:09] what with it being a branch... [19:19:36] edsanders: 'leaves', despite the name, sometimes doesn't return a leaf node. This happens if your offset is between two children of a branch node [19:20:12] e.g.

...

|

...

returns the div as the node [19:20:21] In that case (and only in that case) indexInNode is set [19:20:40] ' * - `leaves`: Return all leaf nodes in the given range (descends all the way down)' [19:21:05] Sure, but if you are at an offset between two children, you can't descend any farther [19:21:13] Maybe that should be clarified [19:21:22] I'm not saying it's wrong, it's just lies [19:21:27] Also, note that it says "all leaf nodes in the given range". The given range is empty [19:21:42] since when is an empty range not a range? [19:21:53] It is a range. But it doesn't "contain" anything [19:22:25] selectNodes is weird and its documentation should probably be rewritten [19:22:35] (03PS16) 10Mooeypoo: Refactor MWMediaSearchWidget to use a queue and providers [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188725 (https://phabricator.wikimedia.org/T78161) [19:22:39] I think for

Fo|o

it behaves as described [19:22:41] But what you're seeing there is entirely intended [19:22:43] RoanKattouw: bah, we cannot make .oo-ui-inline-spacing symmetrical. you'd want something like this, i reckon: http://pastebin.com/9fUNHeHU – but if we do that, then inline-block elements that wrap to second line would be "indented" because of the margin-left [19:22:46] Yes, it does [19:22:50] it doesn't say anything about containing nodes [19:23:05] If you give it a zero-length selection, it descends as far as it can [19:23:13] It returns the deepest node that still contains the offset you gave it [19:23:19] Which for

Fo|o

is a text node [19:23:31] But for

Foo

|

Bar

is the div [19:23:44] Offsets between children of branch nodes are special-cased for that reason [19:24:10] And IIRC, all results that don't have .indexInNode set are in fact leaf nodes [19:24:16] RoanKattouw: if .oo-ui-inline-spacing is more trouble than it's worth, then we might want to revert that entirely, but until we decide that, let me know of bugs it causes (as phab reports, preferably :) ) and i'll look [19:25:04] If you give it a non-empty range of e.g.
|

Foo

Ba|r

, then IIRC you will get 1) the div with indexInNode:0, 2) the text node Foo, 3) the text node Bar with a range property indicating only the first two chars are covered [19:25:39] MatmaRex: Oh, right, crap. [19:25:50] MatmaRex: I could instead make Draggable much smarter [19:26:01] Have it measure the gap between the elements and compute the middle [19:31:43] (03PS1) 10Esanders: Fix definition of ve.Range#containsRange [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/189045 [19:31:45] (03PS1) 10Esanders: Fix rangeInsideOneLeafNode [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/189046 (https://phabricator.wikimedia.org/T88612) [19:31:51] RoanKattouw, ^^ [19:33:07] hah yeah ranges should probably contain themselves, good catch [19:33:12] (03CR) 10Catrope: [C: 032] Fix definition of ve.Range#containsRange [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/189045 (owner: 10Esanders) [19:34:11] Oh wow, you found the cause of Adam's bug [19:34:13] Well done [19:34:37] (03CR) 10Catrope: [C: 032] Fix rangeInsideOneLeafNode [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/189046 (https://phabricator.wikimedia.org/T88612) (owner: 10Esanders) [19:35:03] (03Merged) 10jenkins-bot: Fix definition of ve.Range#containsRange [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/189045 (owner: 10Esanders) [19:35:41] MatmaRex: Actually it turns out that Draggable can't be fixed [19:36:18] (03Merged) 10jenkins-bot: Fix rangeInsideOneLeafNode [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/189046 (https://phabricator.wikimedia.org/T88612) (owner: 10Esanders) [19:36:24] Because if you call getElementFromPoint() on a point that's in the margin of an element, it doesn't return the element, it returns the container [19:36:51] Which... I guess is kind of reasonable because the margin is outside the element's box [19:37:34] But then margins would be fundamentally incompatible with drag&drop, and we'd have to unset inline spacing for those cases [19:37:43] RoanKattouw: i'm afraid i don't know enough about draggable to have an intelligent conversation with you about this [19:37:44] And instead wrap the widgets in divs that we give padding or something [19:37:55] MatmaRex: OK, let me simplify the problem [19:38:29] MatmaRex: When the user is dragging, at any given point in time I need to know where they are dragging to be able to draw a line between the two items nearest their mouse pointer. Sound reasonable? [19:38:48] right [19:39:09] So, there's a native function for that called getElementFromPoint() with takes (x,y) coords and returns an element [19:39:32] The problem is, if the (x,y) coords you pass it are in the margin of an element, it doesn't return that element, but instead it returns its parent element [19:39:42] This is annoying but probably defensible [19:40:25] But it means that if you have a DraggableGroupElement containing a bunch of DraggableElements, then if those DraggableElements have non-zero margins, I will sometimes get useless data from getElementFromPoint() and I won't know where I am [19:40:47] rrrright. [19:40:58] The only way that this will work properly is if the DraggableElements have no margins at all and their boxes are fully adjacent [19:41:10] Because then I always know where the mouse pointer is [19:41:34] So this is what I mean when I say "draggable is fundamentally incompatible with margins" [19:42:25] We could work around this problem by wrapping each draggable thing in another div [19:42:39] Then that div itself wouldn't have margins, but the things inside it could safely have margins [19:42:41] yeah. we'd have to keep track of dimensions and positions ourselves. or probe an area of several pixels around the cursor, but that's probably going to be nasty [19:42:58] and yeah, too. that's what mwcategorywidget apparently does now [19:43:13] Yes, pretty much [19:43:14] (adds margins to inner buttonelements of buttonwidgets, rather than buttonwidgets themselves) [19:43:41] Except that the inline spacing thing then adds some more margins, which caused dragging issues, which is why I was so trigger-happy yesterday [19:44:09] We could unset inline spacing in categorywidget and move the margin somewhere else (or turn it into padding or whatever) [19:44:22] So option 1 is, we declare that draggable things can't have margins and it's their job to make that work [19:44:38] Option 2 is, DraggableGroupElement wraps every DraggableElement in a wrapper div [19:45:14] Actually now I think about it, option 2 might even eliminate the need for having a DraggableElement mixin at all [19:45:58] 3VisualEditor-Performance, Parsoid, VisualEditor: Parsoid should return the full srcset for high DPI devices - https://phabricator.wikimedia.org/T88827#1021211 (10Esanders) 3NEW [19:46:21] do we currently have anything else that is draggable? categories and parameters in templatedata dialog? [19:46:30] 3VisualEditor-Performance, Parsoid, VisualEditor: Add original dimensions information from the API to data-mw - https://phabricator.wikimedia.org/T64881#1021231 (10Esanders) [19:50:20] Only categories I think [19:50:51] Ok RoanKattouw, uploading a patch for MW core [19:50:52] MatmaRex: So I think option 2 is probably best [19:51:14] The only thing about that is that we would still need to manually ensure that draggable things are symmetrical [19:52:24] Right now wrapping buttonWidgets in individual divs would remove their spacing margin (because of the :last-child { margin-right: 0; } rule), but if you did have a margin-right: 0.5em; in there, the lines would be drawn in awkward places [19:52:27] Anyway, we can deal with that [19:52:44] eh, i'm not liking any of this [19:52:46] MatmaRex: Thanks for being my rubber duck :P [19:53:34] 3VisualEditor-Performance, Parsoid, VisualEditor: Parsoid should return the full srcset for high DPI devices - https://phabricator.wikimedia.org/T88827#1021270 (10Catrope) [20:04:31] RoanKattouw, I remember considering the #containsRange fencepost issue before but can't remember why I went with the containsOffset-ish definition [20:04:48] I must've been on something [20:08:46] 3VisualEditor-MediaWiki-Media, VisualEditor: Increase the spacing between Media Title and description as much as it is between the description and meta-data - https://phabricator.wikimedia.org/T87451#1021298 (10Etonkovidova) [20:13:53] 3VisualEditor: [Regression pre-wmf17] Save Page button is missing - https://phabricator.wikimedia.org/T88831#1021311 (10Ryasmeen) 3NEW [20:16:11] 3VisualEditor: [Regression pre-wmf17] Save Page button is missing - https://phabricator.wikimedia.org/T88831#1021329 (10Krenair) p:5High>3Unbreak! [20:16:16] 3VisualEditor: [Regression pre-wmf17] Cannot open any page with VE in Betalabs , "TypeError: Cannot read property 'setDisabled' of null" appears in the console - https://phabricator.wikimedia.org/T88831#1021331 (10Ryasmeen) p:5Unbreak!>3High [20:16:37] 3VisualEditor: [Regression pre-wmf17] Cannot open any page with VE in Betalabs , "TypeError: Cannot read property 'setDisabled' of null" appears in the console - https://phabricator.wikimedia.org/T88831#1021311 (10Ryasmeen) p:5High>3Unbreak! [20:39:10] (03PS2) 10Ryasmeen: [BrowserTest] [WIP]Fixing the test media_general_settings.feature [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188969 [20:39:44] 3VisualEditor: Media Settings "Uploaded by..." not consistently accurate - https://phabricator.wikimedia.org/T88835#1021380 (10Etonkovidova) 3NEW [20:49:42] 3VisualEditor: Save fails on beta labs with js error - https://phabricator.wikimedia.org/T88837#1021405 (10Cmcmahon) 3NEW [20:51:16] (03CR) 10Cmcmahon: [C: 031] "LGTM so fare, but I want to check again before merging when T88837 is fixed" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188969 (owner: 10Ryasmeen) [20:57:20] 3VisualEditor-Performance, Parsoid, VisualEditor: Parsoid should return the full srcset for high DPI devices - https://phabricator.wikimedia.org/T88827#1021443 (10brion) Just a warning note -- native browser support for srcset is still a little spotty, so if you're not engaging the jquery.hidpi() transformation... [20:59:55] Does anyone knows where $wgVisualEditorNamespaces is used by VisualEditor code? [20:59:56] RoanKattouw_away: our draggables would've been so much better if they worked like http://jqueryui.com/sortable/#display-grid (but with variable-sized thingies, not just a grid), with the placeholder being a gap the size of the element being dragged, rather than just a line [21:00:05] I only found $wgVisualEditorNamespaces being set, but not it being used [21:00:10] 3VisualEditor-Performance, Parsoid, VisualEditor: Parsoid should return the full srcset for high DPI devices - https://phabricator.wikimedia.org/T88827#1021445 (10brion) Also I kind of prefer things like srcset here to be implementation details that a syntax parser shouldn't know about - sizing the links require... [21:00:20] Helder: search without $wg [21:00:32] I did that [21:01:04] Helder: so did i and got this: http://pastebin.com/WkU9aUn8 [21:01:05] MatmaRex, http://dpaste.com/3P0WACQ [21:03:11] oh... [21:03:21] without wg too =/ [21:03:29] I removed only the $ [21:03:39] (forgot about the deprecation of globals) [21:20:51] 3VisualEditor: [Regression pre-wmf17] Cannot open any page with VE in Betalabs , "TypeError: Cannot read property 'setDisabled' of null" appears in the console - https://phabricator.wikimedia.org/T88831#1021485 (10Etonkovidova) "[Edit] Value null is the wrong type for property "action.abort.timing" (integer expe... [21:22:26] RoanKattouw_away: eurgh, draggable is really hopeless [21:24:24] we already have the problem with gaps. try something like this: http://i.imgur.com/aJvKA8z.png - you can't drag the category there, unless you drag it *over an existing one* [21:31:47] 3VisualEditor: Save fails on beta labs with js error - https://phabricator.wikimedia.org/T88837#1021519 (10matmarex) [21:31:48] 3VisualEditor: [Regression pre-wmf17] Cannot open any page with VE in Betalabs , "TypeError: Cannot read property 'setDisabled' of null" appears in the console - https://phabricator.wikimedia.org/T88831#1021520 (10matmarex) [21:34:05] 3VisualEditor: [Regression pre-wmf17] Cannot open any page with VE in Betalabs , "TypeError: Cannot read property 'setDisabled' of null" appears in the console - https://phabricator.wikimedia.org/T88831#1021538 (10matmarex) a:3matmarex [21:39:34] (03PS1) 10Bartosz Dziewoński: ve.ui.MWTocWidget: Rename .toggle property [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/189112 (https://phabricator.wikimedia.org/T88831) [21:40:05] MatmaRex: Yeah so we discussed that right before lunhc [21:40:25] MatmaRex: We decided that we should use proximity measurements if e.target is not available due to margin crap [21:40:50] (03CR) 10Bartosz Dziewoński: "Caused T88831, follow-up: I4dd955d015d2bda50e4146b0ae2ddff557dcd346." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/187618 (https://phabricator.wikimedia.org/T87420) (owner: 10Mooeypoo) [21:42:25] (03CR) 10Krinkle: [C: 031] ve.ui.MWTocWidget: Rename .toggle property [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/189112 (https://phabricator.wikimedia.org/T88831) (owner: 10Bartosz Dziewoński) [21:43:00] However, that'll be a lot of effort and we have better things to do this quarter [21:43:54] yeah [21:44:05] i'm just finishing fixing up the existing mwcategorywidget [21:44:07] So for now we'll just ensure margins aren't used in a way that breaks things, and file a tech debt bug about the proximity thing [21:44:13] Oh, awesome [21:44:17] Didn't realize you were on that already, thanks man [21:44:20] i got held up by beta being broken :P [21:44:30] the patch above fixes that, you might want to merge ^ [21:46:28] MatmaRex: WOOOOW [21:46:37] That's STUPID [21:46:50] this.toggle as a property? Really? COME ON [21:46:52] :D [21:47:43] all fault of the english language, for having one word for the verb and the noun [21:47:48] 3VisualEditor, VisualEditor-MediaWiki: [Regression wmf14] Response from parseFragment is blank content, meaning newly-inserted or -modified content is blank until saving - https://phabricator.wikimedia.org/T86046#1021580 (10Etonkovidova) [21:49:59] (03CR) 10Krinkle: [C: 032] "Confirmed exception and fix." [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/189112 (https://phabricator.wikimedia.org/T88831) (owner: 10Bartosz Dziewoński) [21:51:47] (03Merged) 10jenkins-bot: ve.ui.MWTocWidget: Rename .toggle property [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/189112 (https://phabricator.wikimedia.org/T88831) (owner: 10Bartosz Dziewoński) [22:01:18] 3OOjs-UI: Tech debt: Rewrite DraggableGroupElement/DraggableElement to be more robust and generalized - https://phabricator.wikimedia.org/T88854#1021674 (10Mooeypoo) 3NEW a:3Mooeypoo [22:05:10] (03PS1) 10Bartosz Dziewoński: MWCategoryItemWidget: Update CSS for OOUI changes [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/189117 [22:05:30] that took a lot longer than it should've, i kept trying to clean up things, but whenever i touched anything, it'd all blow up [22:05:57] 3OOjs-UI: Tech debt: Rewrite DraggableGroupElement/DraggableElement to be more robust and generalized - https://phabricator.wikimedia.org/T88854#1021704 (10Catrope) Proximity algorithm: - For each element (except the element being dragged), compute the distance to the element being dragged - Compute distance as... [22:12:40] (03PS1) 10Catrope: Track originalHtml in internalItems [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/189120 (https://phabricator.wikimedia.org/T88650) [22:13:18] 3OOjs-UI: Tech debt: Rewrite DraggableGroupElement/DraggableElement to be more robust and generalized - https://phabricator.wikimedia.org/T88854#1021718 (10Mooeypoo) If the items are separated by margins, then --- 1. The proximity algorithm should reveal not only closest item, but also closest "side" (left/righ... [22:14:50] (03PS2) 10Catrope: Support specifying reference contents with data-mw.body.id [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188741 (https://phabricator.wikimedia.org/T88650) [22:14:52] (03PS1) 10Catrope: Update tests for internalItem originalHtml change [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/189121 (https://phabricator.wikimedia.org/T88650) [22:14:59] (03Abandoned) 10Catrope: Allow access to DM document during model->DOM conversion [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/188736 (https://phabricator.wikimedia.org/T88650) (owner: 10Catrope) [22:15:04] (03Abandoned) 10Catrope: Pass original document in model->DOM tests [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/188737 (https://phabricator.wikimedia.org/T88650) (owner: 10Catrope) [22:16:13] 3VisualEditor: [Regression pre-wmf17] Cannot open any page with VE in Betalabs , "TypeError: Cannot read property 'setDisabled' of null" appears in the console - https://phabricator.wikimedia.org/T88831#1021732 (10Ryasmeen) 5Open>3Resolved [22:17:11] 3VisualEditor: [Regression pre-wmf17] Cannot open any page with VE in Betalabs , "TypeError: Cannot read property 'setDisabled' of null" appears in the console - https://phabricator.wikimedia.org/T88831#1021311 (10Ryasmeen) verified the fix in Betalabs [22:17:19] (03CR) 10jenkins-bot: [V: 04-1] Support specifying reference contents with data-mw.body.id [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188741 (https://phabricator.wikimedia.org/T88650) (owner: 10Catrope) [22:17:36] 3VisualEditor: [Regression pre-wmf17] Cannot open any page with VE in Betalabs , "TypeError: Cannot read property 'setDisabled' of null" appears in the console - https://phabricator.wikimedia.org/T88831#1021738 (10Ryasmeen) [22:18:03] 3VisualEditor: In Reference list(and 'Cite web' template) - help link for CS1 errors info cannot be clicked - https://phabricator.wikimedia.org/T76666#1021739 (10Etonkovidova) Also, Re-use citation Reference box will display the link to CS1 errors page - cannot be accessed from VE. Would be great to have help... [22:18:22] 3VisualEditor: In Reference list(and 'Cite web' template) - help link for CS1 errors info cannot be clicked - https://phabricator.wikimedia.org/T76666#1021740 (10Etonkovidova) p:5Triage>3Normal [22:21:10] (03CR) 10jenkins-bot: [V: 04-1] Update tests for internalItem originalHtml change [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/189121 (https://phabricator.wikimedia.org/T88650) (owner: 10Catrope) [22:27:22] MatmaRex, are you familiar with the css "+" usage? [22:27:40] probably? [22:28:52] MatmaRex, I think I found the cause for an ooui bug, but I am not sure if there's a better way of doing it. [22:29:46] hmm? [22:29:52] MatmaRex, https://phabricator.wikimedia.org/T88035 <-- this bug. So, if you look at the css, there's a rule that applies the bottom margin to the checkbox, but it specifies: .oo-ui-checkboxInputWidget input[type="checkbox"]:hover + span::before { [22:30:52] but as far as I've read from css specs, this behavior is supposedly applied to the second element. In any case, the issue seems to be that the hover is calculated on the span rather than on the checkbox, and the span is outside the checkbox and tin [22:30:54] y [22:31:39] 3OOjs-UI, UI-Standardization: Radio button hover state only appears when I mouse over the bottom part of it. - https://phabricator.wikimedia.org/T88035#1021780 (10matmarex) [22:32:10] (03PS3) 10Cmcmahon: [BrowserTest] [WIP]Fixing the test media_general_settings.feature [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/188969 (owner: 10Ryasmeen) [22:32:43] mooeypoo: hm, that might be intentional? the pretty checkbox is actually the span element [22:33:02] MatmaRex, if that's the case, we should make sure the span encompasses the checkbox [22:33:24] Otherwise, you have to be very talented in finding te "sweet spot" of the span to get the checkbox to light up [22:35:09] 3VisualEditor-MediaWiki-References, VisualEditor, VisualEditor-MediaWiki: [Regression pre-wmf14] While trying to open Re-use Reference dialog immediately after inserting a citation (not basic reference), throws the TypeError: Cannot read property 'length' of null and the newly added citation does not appear in the... [22:36:06] hmm [22:36:10] actually [22:36:19] the :hover should probably be on the span, not the input. [22:38:00] MatmaRex, the span is tiny, though [22:38:30] MatmaRex, that indeed fixes it, though [22:38:32] 3VisualEditor, VisualEditor-DataModel: "Uncaught TypeError: Cannot read property 'getRange' of undefined" when unmerging cells - https://phabricator.wikimedia.org/T75223#1021806 (10Etonkovidova) [22:39:27] mooeypoo: the span is tiny, but its ::before is the actual checkbox/radio display [22:39:35] * mooeypoo nods [22:39:48] MatmaRex, I just tested, if we move the :hover to the span::before it works [22:40:15] the input shouldn't even be hoverable at all [22:40:23] since the span::before is supposed to cover it entirely [22:40:29] …i think. [22:40:40] yeah I wasn't sure how that works. The ::before and the + confused me [22:40:49] I'm testingnow and will submit a patch for your review in a minute [22:45:26] (03PS1) 10Mooeypoo: CheckboxInputWidget: Apply hover event properly [oojs/ui] - 10https://gerrit.wikimedia.org/r/189125 (https://phabricator.wikimedia.org/T88035) [22:45:52] MatmaRex, ^^ though there is still a tiny weirdness there. IF you hover from a different angle it seems to not work (im testing the group checkboxes in the demo) [22:46:31] computer pixies [22:46:39] mooeypoo: radio input probably needs the same fix? [22:46:49] checking now [22:48:34] MatmaRex, i fixed both of them in that paatch, but I'm testing [22:49:45] MatmaRex, I have a problem in the RadioSelectWidget still. This works on the individual check/radio but the RadioSelectWidget is off. It makes no sense [22:50:52] mooeypoo: chrome's :hover gets reallllllly weird when there are s and