[02:20:46] 10VisualEditor: Visual Editor save dialog hangs, causing data loss - https://phabricator.wikimedia.org/T91949#1099444 (10LuisVilla) 3NEW [02:30:14] 10VisualEditor: Visual Editor save dialog hangs, causing data loss - https://phabricator.wikimedia.org/T91949#1099453 (10Krenair) What was the error shown in the console? [06:18:20] (03CR) 10Mattflaschen: [C: 04-1] "I reviewed it in detail again, so please refrain from changing things that are not mentioned in the comments." (038 comments) [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/189333 (owner: 10Paladox) [06:45:54] 10VisualEditor, 10VisualEditor-MediaWiki-Links, 7Design, 3VisualEditor 2014/15 Q3 blockers: Search interface for link dialogue - https://phabricator.wikimedia.org/T76397#1099535 (10santhosh) Content Translation also need this component to enhance our link adaptation features, also for providing a better so... [07:43:28] 10OOjs-UI, 10UI-Standardization, 7Design: Document tipsy - https://phabricator.wikimedia.org/T91856#1099575 (10Pginer-WMF) @violetto yes, tooltips are used in different pats of Flow such as the indication that you are anonymous at the reply form or the description of the "edit header" icon. [07:59:05] 10OOjs-UI, 10UI-Standardization, 7Design: PopupButtonWidget should be shown at hover not on-click. - https://phabricator.wikimedia.org/T88630#1099576 (10Pginer-WMF) Another aspect to take into account is we support hover to show/hide those is to be careful not to accidentally hide them while the user is movi... [09:49:29] edsanders: Actually my commit asserts that it should be B), while right now we make a half-assed attempt to do A) [09:49:49] edsanders: Full rant + fix: https://gerrit.wikimedia.org/r/194980 [11:20:26] Is there a way to insert code from the toolbar? [11:20:59] Ve parses existing syntaxhighlight codes perfectly, but I can't seem to insert a new one from the toolbar [11:37:03] RagBal: Yeah unfortunately you can't insert a new one right now :( [11:41:40] Last year a Google Summer of Code student worked on a tool to insert those and edit them more nicely, but that code wasn't quite finished and never ended up getting more love [11:45:00] RagBal, you can copy and paste them though [11:45:29] RoanKattouw, should we throw an exception if someone tries to apply such a transaction? [11:49:36] edsanders: I'm not too convinced we should [11:49:52] I mean what ends up happening may or may not be what you expect [11:50:23] But I think taking the replacement literally (without applying annotations to the inserted content) is more consistent than the alternative [11:50:41] (because otherwise you get unclean reverts and stuff) [11:51:10] I guess this also means that transaction composition isn't completely trivial [11:51:30] If you have a replace transaction followed by an annotate transaction, you could compose them into one transaction trivially [11:51:49] But if they're ordered the other way around, the composition logic would have to be a bit cleverer [11:52:45] However, 1) if you choose the other behavior, you just swap which one is weird, and 2) this is in a hypothetical case where we even care about transaction composition [11:53:16] RoanKattouw, edsanders, thanks for the clarification [11:55:00] Also, it seems like sometimes the tags won't get picked up by ve [11:55:22] RagBal: :O when does that happen? [11:55:37] After some refreshes they magickly appear [11:55:55] Perhaps cache, let me test it in incognito [11:56:50] Nope, not cache =( [11:58:06] RoanKattouw, I can't seem to find a logical routine in which they do and don't appear [11:59:56] I'm using MW 1.24.1 with VE REL1_24, ULS REL1_24 and syntaxhighlight REL1_24 [12:02:58] Odd [12:03:10] What does it look like when it doesn't work? Are they just not there at all? [12:11:12] It displays the text like: code blablabla [12:11:59] hah [12:12:03] That's... probably a Parsoid issue? [12:12:05] Not one syntaxhighlight is parsed on the page, all show the open and closing tag with the code between them [12:12:14] But still very werid [12:12:48] I'm using the master branch of parsoid [12:13:31] v0.10.36 node [12:15:15] WTF, why are my tests passing [12:19:52] RagBal: I have never seen that issue before and I don't know what would cause it [12:20:01] RagBal: Is your wiki publicly accessible? [12:20:42] RoanKattouw, nope private [12:21:53] Then I can't help you figure out what's going on, sorry :( [12:23:42] Is there any way to enable debug logs or something? [12:24:28] (03PS1) 10Catrope: When replacing e.g. heading1 with heading2, use attribute changes [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195242 [12:25:13] RagBal: This is so far outside of the realm of issues I've heard of that I don't even know where to start [12:25:21] I'd have to poke at the actual wiki directly [12:27:01] RoanKattouw, let me check if the problems also occur in my private wiki, no secret stuff there that you aren't allowed to see =) [12:38:24] (03CR) 10Esanders: [C: 032] Queue linear model modifications in TransactionProcessor [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194981 (owner: 10Catrope) [12:38:33] (03CR) 10Esanders: [C: 032] Remove unnecessary support for nesting annotation and replace operations [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194980 (owner: 10Catrope) [12:39:40] (03CR) 10Catrope: "Duplicate of https://gerrit.wikimedia.org/r/195236 and related stack. This is what happens when the IRC bot breaks :(" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195228 (https://phabricator.wikimedia.org/T91831) (owner: 10Esanders) [12:39:56] (03CR) 10Esanders: [C: 032] Factor attribute setting code into ElementLinearData [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194977 (owner: 10Catrope) [12:42:23] (03Merged) 10jenkins-bot: Factor attribute setting code into ElementLinearData [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194977 (owner: 10Catrope) [12:42:45] (03Merged) 10jenkins-bot: Move ve.dm.Document#spliceMetadata to dm.MetaLinearData [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194978 (owner: 10Catrope) [12:42:47] (03Merged) 10jenkins-bot: Remove unused variables in dm.TransactionProcessor#applyAnnotations [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194979 (owner: 10Catrope) [12:42:51] (03Merged) 10jenkins-bot: Remove unnecessary support for nesting annotation and replace operations [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194980 (owner: 10Catrope) [12:42:53] (03Merged) 10jenkins-bot: Queue linear model modifications in TransactionProcessor [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194981 (owner: 10Catrope) [12:44:51] (03CR) 10Catrope: [C: 04-1] Create CE view initialization event for when $element is created (034 comments) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195228 (https://phabricator.wikimedia.org/T91831) (owner: 10Esanders) [12:46:16] edsanders: Looks like we both came up with our own solution for the initialize thing [12:46:19] And grrrit-wm was broken so it didn't notify us :| [12:46:57] hm [12:50:18] RoanKattouw, your fix only works for original dom elements? [12:50:33] My fix for what? [12:51:11] Headings? [12:51:11] initialize [12:51:12] Ahm I don't understand the question [12:51:12] https://gerrit.wikimedia.org/r/#/c/195238/1/src/ce/ve.ce.View.js [12:51:39] if I change the colspan, then convert td to th [12:51:52] where is the correct colspan re-applied? [12:52:05] ce.TableCellNode#initialize [12:52:35] oh, I see the parent commit [12:52:42] I deliberately submitted this change separately from the rest of the initialize stuff because I felt like there was probably some pitfall somewhere that I was missing [12:52:44] so I did an initialize method first [12:52:49] but how does that work with mixins? [12:52:55] Yeah it doesn't :S [12:53:00] Hence my comment on your change [12:53:22] OTOH, a method can just be called in the ce.View constructor whereas for an event that doesn't work because there are no listeners yet [12:53:32] Presumably that's why you put the emit call in NodeFactory [12:55:14] I guess we could have both the FocusableNode constructor and onFocusableSetup call initializeFocusable() ? [12:55:14] (We can't define ve.ce.FocusableNode.prototype.initialize because there's already an initialize in the class it's being mixed into) [12:56:47] (03CR) 10Jdlrobson: [C: 032] Remove half-baked touch event handling [oojs/ui] - 10https://gerrit.wikimedia.org/r/195067 (https://phabricator.wikimedia.org/T87818) (owner: 10Bartosz Dziewoński) [13:05:46] (03CR) 10Catrope: "recheck" [extensions/VisualEditor] - 10https://gerrit.wikimedia.org/r/195223 (owner: 10Catrope) [13:06:06] RoanKattouw, yes, the old problem of not being able to do something after construction [13:06:36] I think we do setup and teardown as events for this reason, so we should probably stick with it? [13:07:35] But they happen significantly later, after attachment [13:07:44] Not immediately upon construction [13:09:57] (03CR) 10Esanders: Create CE view initialization event for when $element is created (032 comments) [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195228 (https://phabricator.wikimedia.org/T91831) (owner: 10Esanders) [13:10:16] (03PS2) 10Esanders: Create CE view initialization event for when $element is created [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195228 (https://phabricator.wikimedia.org/T91831) [13:12:33] (03CR) 10Esanders: "FIXME: This breaks the positioning of the table context." [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/194420 (owner: 10Jforrester) [13:28:24] (03PS1) 10Esanders: Fix margins of indicators in table context [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195252 [13:34:41] Krinkle|detached: https://phabricator.wikimedia.org/T91968 [13:35:21] Also our merge pipeline is broken, we can't merge anything in VE-MW right now. It's because of https://phabricator.wikimedia.org/T91967 in Flow, I'll bug the relevant people as soon as they wake up [13:53:22] Morning all [13:53:50] gi11es rightly told me to not create an interface from scratch because it's in OOUI, but I'm wondering now what I should try to fix myself and what I should try to get into the library [13:54:44] Which interface? [13:55:07] I have a PopupButtonWidget that 1. is restyling my $button element unexpectedly, 2. has neither background colour nor border, making it impossible to distinguish from the rest of the page, and 3. the arrow seems like it's formatted wrong, but that may be the lack of border and background. [13:56:19] Interesting [13:56:47] marktraceur: Is this in an MMV commit in Gerrit somewhere that I can download and poke at? [13:56:59] RoanKattouw: For the button styling, I really didn't want it to *be* a button, but I did want the link to be the origin, I saw no other way to do that [13:57:07] RoanKattouw: Not MMV, core, plus pdfhandler if you want a demo [13:57:18] Let me upload the OOUI version that is broken. [13:57:18] Oh core OK [13:57:24] Cool [13:57:44] I'll probably just try to jank out a minimal breaking example from your code into an OOUI sandbox [13:57:48] Speaking of, we need a real OOUI sandbox [13:57:55] Sure. [13:58:01] RoanKattouw: https://gerrit.wikimedia.org/r/194565 [13:58:36] The hovering thing isn't working yet because I haven't gotten 'round to it - maybe the answer here is a new "PopupHoverWidget" class that doesn't restyle the element you tell it to aim at [13:59:24] RoanKattouw: FYI, pginer made us a mockup... [13:59:25] https://phab.wmfusercontent.org/file/data/mvdjj6ioz54w4f7mep2o/PHID-FILE-nhpfn2ujb6ohslyfgsxq/hecj5ynbh5lynwtp/pdf-warning.png [14:00:05] marktraceur: That link is dead and broken [14:00:22] Oh wait no [14:00:31] Oh, yeah, it is. [14:00:32] I'm supposed to be looking at a diagram of someone's house :D [14:00:51] Well. Something like that, yes. [14:01:08] Yeah for a popup that opens on hover rather than click, talk to MatmaRex and/or TrevorP|Away [14:01:41] * marktraceur will wait patiently for an opportunity to do so. [14:01:56] I imagine .hover( function () { dialog.open(); } ) won't be too hard. [14:02:01] Hmm [14:02:07] How do I reproduce this with PdfHandler? [14:02:18] I was gonna use a sandbox but there's too much MW-specific stuff going on [14:02:20] There's a linked patchset in the core commit summary... [14:02:37] https://gerrit.wikimedia.org/r/194564 [14:02:47] Oh, oops, that's the stupid wmf15 mistake I made. [14:02:48] Which is abandoned :O [14:03:00] Oh I found the other one [14:04:50] OK, so I have that PdfHandler patchset [14:04:55] Now do I have to upload a PDF to my wiki? [14:05:04] Ah, yes. [14:05:17] Then the file page should have the warning on it. [14:07:00] OK I see [14:07:11] The button looks very button-y, and the popup's styling is totally broken [14:07:15] Yup. [14:07:24] Hence my stumbling in here. [14:07:54] I'm 80% sure that my very few style rules are not interfering with the popup. [14:08:11] And the button thing, well, I can fix that on my own but I wasn't sure if I should do it as an ooui patch. [14:08:24] Oh I see [14:08:33] You're not meant to use content with popups, probably [14:09:20] Also where the hell is config.content even defined :O [14:09:36] Oh it's in Element [14:09:39] I thought that was $content ! [14:10:06] Oh brilliant [14:10:11] There's both config.content and config.$content [14:10:24] PopupWidget correctly mangles config.$content to compensate for the strange things it does [14:10:37] But config.content was introduced later, and PopupWidget wasn't updated for that [14:11:55] Ah, so I should use $content perhaps? [14:13:24] Yes [14:13:28] Amendment inbound [14:13:39] (Because $content only takes a jQuery object, not an array of jQuery objects :S ) [14:13:56] OK, updated [14:13:57] Right. [14:14:00] That fixes the background issue [14:14:13] Could you file a bug about the content vs $content thing? [14:14:16] Sure! [14:14:40] Hm. [14:14:53] There are two problems there: PopupWidget doesn't mangle content, only $content, and 2) the way it does its mangling is hideous [14:15:10] RoanKattouw: Also, $info and $footer aren't showing up now [14:15:12] Like, it reimplements $content but differently (requires config.$content instanceof jQuery while the base class is more liberal) [14:15:28] Oh, hmm [14:15:30] They aren't in the div. [14:15:50] .add takes an array, or only one element, maybe? [14:15:56] Is that because I'm incompetent and don't know how .add() works? [14:16:13] Yeah that's it [14:16:15] Thanks jQuery [14:16:17] Could be, I won't tell anyone if you don't :) [14:17:05] haha [14:17:23] marktraceur: Yeah you can't do $foo.add( $bar, $baz, $quux ) you have to do $foo.add( $bar ).add( $baz ).add( $quux ) [14:17:34] My fault for assuming jQuery was helpful [14:17:41] So $content: $header.add( $main, $info, $footer ) doesn't work [14:17:58] Yup, I can fix that [14:18:08] Thanks [14:18:24] (Realistically what might be better is to create a wrapper, stick all your stuff into that wrapper, and pass that wrapper into $content ) [14:18:58] Might, yes [14:19:34] Also the border goes away if you pass framed: false (in the top-level config object, i.e. the button's not the popup's) [14:19:56] Weird. [14:20:00] That does result in the text getting misaligned unfortunately, but I'm guessing that's probably a bug in OOUI [14:21:09] OK, so next up: button-y link [14:21:24] Yeah that's what that is [14:21:32] framed: false gives you a frameless button which is basically a button-y link [14:21:37] Or a link-like button or whatever you wanna call it [14:21:38] Oh! [14:21:45] I thought you meant the border on the dialog [14:21:51] Sorry, reading comprehension [14:21:53] Oh, sorry [14:21:55] My bad [14:22:08] It's a popup BTW, a dialog is something different in OOUI parlance (dialogs are modal) [14:22:12] Right [14:22:19] * marktraceur will attempt to alter behaviour [14:22:32] Oh man, that's great [14:22:41] I'd guess the alignment is somehow related to the icon I added [14:22:45] Also you should not have to insert the icon yourself [14:22:53] Yeah I'm guessing your CSS is screwing that up somehow [14:23:02] But if you inspect the button you'll see icon and indicator spans [14:23:15] I guess I can just stick it in there, yeah [14:23:21] This was from before I was using ooui [14:24:54] RoanKattouw: Hm, pginer seemed to want it to be orange though, can I somehow use the custom icon? [14:25:02] * marktraceur reads through icon docs [14:27:45] marktraceur: Sorry, computer rebooted [14:27:47] So I tried removing your CSS and setting indicator: 'alert' [14:27:50] And exposed a wonderful bug in OOUI where the indicator overlaps with the label, oops [14:27:56] Heh [14:28:08] I set "icon: 'alert'" and it seems fine [14:28:10] But not orange [14:28:20] Wow this shit is seriously broken [14:28:26] Somewhere, a single tear is rolling down pginer's cheek. [14:28:29] Icons are supposed to be to the left of the label, not to the right [14:28:33] Oh. [14:28:39] Well, in this case it's what I wanted [14:28:39] But somehow in frameless PopupButtonWidgets they're to the right?! [14:29:04] * marktraceur gives RoanKattouw leave to call it a feature [14:29:06] Oh no I see why [14:29:13] You're not using the label feature correctly [14:29:29] ...oh? [14:29:44] Oh, right, I dump an existing element in [14:29:50] Hmm and when I try to do it correctly it breaks :D [14:29:59] Doing things correctly is overrated. [14:31:17] Hmm [14:31:47] OK here we go [14:31:54] Apparently what you needed was label: $mimetype [14:32:00] As opposed to $button: $mimetype [14:32:02] Ah. [14:32:11] Semantics. [14:32:17] Yeah [14:32:23] I tried $label: $mimetype first and that exploded even worse [14:32:46] But then I realized that passing in a jQuery object for $label vs label means something very different [14:32:52] Ah. [14:33:07] Shall I amend? [14:33:11] Yeah go for it [14:33:32] You probably also want framed: false, and either one of icon: 'alert' or indicator: 'alert' [14:33:39] Yup [14:33:56] As for making it yellow, that should be a simple matter of coming up with a new flag ('warning' perhaps) [14:34:21] Define "simple" [14:34:22] The only thing that sucks is that currently I don't believe we are able to define custom flags on core icons, so it'd have to be added to OOUI itself [14:34:29] Ah, fun. [14:34:45] Simple as in a simple change to the JSON file that describes the icons and indicators [14:34:50] Oh, sure. [14:35:08] Except that we'd like to be able to do that in MW but for now have to do that in OOUI [14:35:17] Because we don't yet have full-matrix extensibility [14:36:01] Hello lovely people. [14:36:26] RoanKattouw: So shall I add this to ooui or wait for core? [14:37:00] marktraceur: Add it to OOUI [14:37:14] Don't bother waiting for us to get our extensibility act together, that'll take a while [14:37:15] 'kay [14:37:32] oojs/ui/src/themes/mediawiki/images.json [14:37:44] Should be fairly self-explanatory, use 'destructive' as an example [14:39:17] (03PS1) 10MarkTraceur: Add warning icon to MediaWiki set [oojs/ui] - 10https://gerrit.wikimedia.org/r/195257 [14:39:27] RoanKattouw: Sorry, were you still talking? :P [14:39:40] Oh, ahm, sorry [14:39:45] I might not be able to find what I want in ooui documentation, but I can damn sure grep [14:40:31] I can probably add it to the mediawiki/core.git copy temporarily? [14:40:31] marktraceur: Is that the one we have in VE or a new one? [14:40:39] James_F: Uh [14:40:41] marktraceur: Locally? [14:40:44] I think new. It's orangey. [14:40:45] The thing is [14:40:56] It's basically the same as alert.svg except with a different color scheme [14:41:02] marktraceur: That colour isn't in theme. [14:41:07] (And also different dimensions than everything else) [14:41:14] I tend to listen to our designer on design questions [14:41:15] marktraceur: You're going to have to get Pau to get Design to alter the theme. [14:41:22] * marktraceur sighs, will do [14:41:26] marktraceur: Sure. And I listen to WMF's, who've said no new colours. [14:41:28] marktraceur: We have programmatic ways to apply colors to icons/indicators, all the base versions are black [14:41:49] Oh. [14:41:50] That's what I meant when I said "use 'destructive' as an example" [14:41:51] OOjs UI is designed to make it hard to randomly add an off-theme icon. [14:42:00] I was not being remotely clear enough, sorry [14:42:21] RoanKattouw: Use the excuse that it's early. ;-) marktraceur doesn't need to know you're in Europe! [14:43:02] What I wanted you to do was add a new flag 'warning' which applies a yellow color, and have alert be able to take that flag [14:43:07] Then apply that flag to your button [14:43:26] Yup, sounds much better anyway [14:43:35] And as James said, on the way that commit will probably be embroiled in some sort of scandal about adding a new color [14:44:24] Of course. [14:44:50] But if you apply { icon: 'alert', flags: ['warning'] } (or indicator, depending on which of the two images you want, whether you want it to be on the left or right, and what size you want it to be), then it'll start out being black, and if and when 'warning' starts being mapped to a yellow color in OOUI, it'll automatically change from black to yellow [14:44:57] (03PS2) 10MarkTraceur: Add warning variant to MediaWiki set [oojs/ui] - 10https://gerrit.wikimedia.org/r/195257 [14:45:04] Variant, right [14:45:08] That's the word I was looking for [14:45:22] :) [14:45:29] We figured it out eventually! [14:45:32] (03CR) 10Jforrester: [C: 04-1] "Need to add to demos too." [oojs/ui] - 10https://gerrit.wikimedia.org/r/195257 (owner: 10MarkTraceur) [14:45:42] I'll warn Pau about the possible embroilment [14:45:57] marktraceur: It's May and Jared who "own" the design, FWIW. [14:46:04] marktraceur: Ugh and because our code sucks you also need to update getElementClasses in MediaWikiTheme.js [14:48:29] AOK! [14:48:41] Sigh, this is what I get for committing early and committing often [14:48:53] I have to change the modification queueing system I wrote on Friday that Ed merged a few hours ago :S [14:48:56] marktraceur: Also, build it to proove that it works. [14:48:59] Oh well [14:49:10] RoanKattouw: Whoops. [14:49:22] Turns out I need my modifications to be reversible [14:49:35] So maybe they shouldn't directly map to ElementLinearData/MetaLinearData methods [14:50:03] James_F: Build what now? [14:50:10] marktraceur: OOjs UI. [14:50:22] marktraceur: `npm install && npm build` [14:50:37] marktraceur: Then inspect /dist to check that the icon you want really does get created in orange. [14:51:42] Oh, sure. [14:51:59] James_F: Also, add it to demos how? There's no variant demo that I can see so far. [14:52:10] Yeah there's no variant demo :( [14:52:14] * RoanKattouw blames TrevorP|Away [14:52:18] marktraceur: Good point. [14:52:26] There are some stray varianted icons in the widgets demo [14:52:28] That's actually meant to be fixed by "my" icon patch. [14:52:31] You could add one there [14:52:34] RoanKattouw: Eww. [14:52:55] James_F: That's what TrevorP|Away did when I forced him to prove the variants stuff worked in the demo [14:53:01] Least effort required [14:53:11] RoanKattouw: Clearly. :-) [14:55:42] Hm. [14:56:17] Oh, it worked. Sweet. [14:56:28] (03PS3) 10MarkTraceur: Add warning variant to MediaWiki set [oojs/ui] - 10https://gerrit.wikimedia.org/r/195257 [14:56:34] eog scared me for a sec [14:56:42] RoanKattouw: Can https://gerrit.wikimedia.org/r/#/c/195222 ("also restore rowspan and colspan on TableCellNodes") likely be trivially cherry-picked? It's a wmf20 regression. [14:56:51] marktraceur: eog? [14:57:14] marktraceur: You probably need to put the variant stuff into the PHP code as well as JS. [14:57:15] Eye Of Gnome, the default image viewer [14:57:20] marktraceur: Just to be super-unhelpful. [14:57:30] Oh, does it suck? [14:57:37] For SVGs, apparently yes. [14:57:46] It showed me black and misformed. [14:57:59] Fun. [14:58:09] marktraceur: MatmaRex is the export. [14:58:12] Err. Expert. [14:58:19] hm? [14:58:51] James_F: Are you sure it's wmf20? [14:59:00] (03PS4) 10MarkTraceur: Add warning variant to MediaWiki set [oojs/ui] - 10https://gerrit.wikimedia.org/r/195257 [14:59:10] James_F: https://gerrit.wikimedia.org/r/#/c/193813/ says Included In: master [14:59:14] Oh, meh, it's lib/ve [14:59:18] RoanKattouw: The bug says wmf20. [14:59:18] what is the question? [14:59:20] that means that's inaccurate, nm [14:59:21] MatmaRex: I was about to ask you about PopupButtonWidgets appearing on hover. [14:59:29] RoanKattouw seems to think you have the magic bullet. [14:59:40] MatmaRex: marktraceur wants to add a new colour to the MediaWiki theme for icon variants. [14:59:54] Also that, but less urgent from my standpoint that we deal with that [14:59:58] James_F: You're right, it's wmf20, and yes, it can be trivially cherry-picked, that's why I split it out [15:00:02] eww. [15:00:07] MatmaRex: Inorite. [15:00:07] why? [15:00:09] RoanKattouw: Cool. [15:00:17] It's *not my idea* [15:00:17] MatmaRex: Because the designers don't talk to each other. [15:00:25] dude. [15:00:30] Talking is overrated [15:00:34] :-) [15:00:38] edsanders: Could you merge https://gerrit.wikimedia.org/r/#/c/195222 which is the minimally invasive fix for the colspan regression so that it can be backported? [15:00:59] Apparently MatmaRex didn't even. [15:01:28] James_F: grrrit-wm broke for an hour or so this afternoon, and apparently Ed and I are so reliant on it that we both independently wrote basically the same code but with different approaches and didn't notice until much later [15:01:46] RoanKattouw: I saw. :-) [15:02:28] (sorry, got disconnected) [15:03:23] RoanKattouw: Where's the script for creating wmf branches? Adding VE core would help. [15:04:22] RoanKattouw: We'd have to check it was pulled through already, but… [15:04:47] Yeah maybe [15:04:54] Actually no, we can do that cleverly [15:05:07] At the time ext/VE is branched, check what its submodule points to, and tag/branch that [15:05:20] Well, sure. [15:05:26] But that's much more work. :-) [15:06:47] Not really [15:07:15] Find me the script that creates the existing extension wmf branches and I can do it in like 10 mins [15:08:38] RoanKattouw: Searching for "wmf branch .sh" didn't find me anything. [15:08:57] RoanKattouw: Didn't you write this script, or was it a Reedy invention post-you? [15:09:17] Well after my time [15:09:20] It was a Reedy invention [15:09:34] Ah. [15:11:07] Aha. [15:11:16] RoanKattouw: https://github.com/wikimedia/mediawiki-tools-release/blob/HEAD/make-wmf-branch/make-wmf-branch [15:11:37] RoanKattouw: It's in PHP, because obviously that makes sense. [15:13:29] Right [15:15:47] (03CR) 10Jforrester: [C: 04-1] "Fixes the horizontal one (for columns) but not the vertical one (for rows): https://imgur.com/MgTE86c" [VisualEditor/VisualEditor] - 10https://gerrit.wikimedia.org/r/195252 (owner: 10Esanders) [15:20:39] hello editor-builders [15:20:47] if I futz with parsoid on betalabs over the next few days [15:20:50] am I going to be shot at? [15:21:22] What are you gonna be futzing with? [15:21:38] I mean if you break Parsoid, people will be upset, especially on Tuesdays and Wednesdays [15:21:45] YuviPanda: possibly. if it's going to be down, let us know in #wikimedia-releng [15:22:04] (03PS27) 10Paladox: Convert .css to .less [extensions/WikiEditor] - 10https://gerrit.wikimedia.org/r/189333 [15:22:14] YuviPanda: and yeah, Thu and Fri are better futzing days [15:22:20] RoanKattouw: https://gerrit.wikimedia.org/r/#/c/193082/ basically. [15:22:33] RoanKattouw: because the prod / beta parsoid roles had become wildly different for strange purposes. [15:22:41] chrismcmahon: RoanKattouw alright. I shall not futz with it until Thursday. [15:23:28] RoanKattouw: it’s the last ::beta role :) [15:23:33] RoanKattouw: BTW, added a wmf20 branch for VE core. [15:23:47] YuviPanda: Oh dear God I'm so happy that's getting cleaned up [15:24:01] That should make Parsoid in beta break much less often [15:24:21] RoanKattouw: I’ve basically spent a month and a half getting rid of every ::beta or $::realm == ‘labs’ branch... [15:24:27] at least in our core stack [15:24:32] parsoid is the last ‘terrible’ one. [15:24:55] YuviPanda: you rawk. thanks. [15:25:29] RoanKattouw: do you know if jenkins based parsoid deploys work at all? [15:25:30] James_F: pginer linked me to https://trello.com/c/IRqbu8p4/15-color-swatches which has the colour I used. [15:25:43] RoanKattouw: oooh, adn do you know anything about how parsoid picsk which config file to use? (labs vs prod) [15:25:48] feel free to tell me to go away and come back later [15:25:56] As "med severity" which seems reasonable. [15:26:07] marktraceur: That looks like it's from MediaWiki UI, which is no longer current. [15:26:29] Well [15:26:39] marktraceur: I'm not remotely saying the colour isn't right, just that this is the very first time someone's suggested we need it. [15:26:41] Back to designers not talking, but he also said it was probably out of date [15:26:47] * James_F nods. [15:30:46] YuviPanda: I have no idea if or how Jenkins based deploys work [15:31:02] YuviPanda: You would also have to talk to the Parsoid people to ask if deploys from master at merge time are even desirable [15:31:11] RoanKattouw: ah, right. I shall do that. [15:31:32] RoanKattouw: either way, I’ll avoid futzing it till thursday. I’ll give you, James_F and chrismcmahon a headsup before I start futzing [15:31:36] Cool [15:31:40] Thanks. [15:31:51]