The plugin we use for this only uses classes found in the current codebase:

Note also that with an agreed prefixMap it could be set upstream in jsdoc-wmf-theme so extension never had to re-define it.

Data shows clicks to permalink timestamps on both non-talk (or subject) pages and talk pages. Is this possible/expected?

We use mix-blend-mode: darken so the text stays black under the highlight. There isn't a blend mode that works well in dark and light mode, so you'll need to use lighten in dark mode. mix-blend-mode is used in a few place so perhaps you will want to make this a re-usable token.

Deleted

Since T345281 all non-inverting images are marked with mw-no-invert, or cdx-no-invert.

In T362939#9844095, @JScherer-WMF wrote:

In T362939#9840767, @Jack_who_built_the_house wrote:

Then, boxes should be evenly spaced from top and bottom (p:last-child { margin-bottom: 0.5em }):

Yes, boxes should have equal top and bottom margins like the message component in codex.

Looks like https://gerrit.wikimedia.org/r/c/mediawiki/skins/Vector/+/1036715 which introduces negative margins.

Possibly related to T362939

You can prepare the upgrade patches before we do the release by installing from gerrit (npm i --save-dev git+https://gerrit.wikimedia.org/r/jsdoc/wmf-theme#refs/changes/47/1041147/2)

(executeAction is called directly quite rarely, so I doubt we would have a problem with the above breaking change: https://codesearch.wmcloud.org/search/?q=executeAction%5C%28&files=&excludeFiles=&repos=#Extension:VisualEditor)

2. We could fix this at the OOUI level, checking if there are visible actions, and if those actions are disabled. It would technically be a breaking change, although I don't know how likely it is that someone has deliberately programmatically triggered an action that is disabled in the UI. One could add an override to allow it to go through, or have it skip the UI check by default.

but the functionality is known to other people

I did not open all rMW2af1c3c901a6117fe062e1fd88c0146cffa1481d but I could imagine that the same bug has been injected in many many codes.

Looking at the code I can't see where this might be going wrong either. In T368029, there was an issue with the first argument being undefined, but in this case the first argument is a literal object, and only the second argument can be undefined, which isn't a problem.

> Object.assign( {a: 1}, undefined )
< {a: 1}

When on source text editing something is marked and string overwriting is requested, undefined is additionally inserted into wikitext after selection end.

Users expect to be able to follow the links in the preview - to check on the source while editing. Link contexts work the same way. If you feel this functionality needs to change we should file a task explicitly for that and discuss the pros and cons there.

I've never really understood the hidden file standard (leading .) so happy to go without it. Other tools are inconsistent, so I don't think it matters too much which we pick:

• .eslintrc.json
• .stylelintrc.json
• .bananaconfig.json
• tsconf.json
• package.json
• bundlesize.config.json

@Novem_Linguae : https://gerrit.wikimedia.org/r/c/jsdoc/wmf-theme/+/1041147

Should be fixed by https://www.mediawiki.org/w/index.php?title=MediaWiki%3AGadget-tabbedwindow.js&diff=6588257&oldid=6583748

Should be fixed by https://www.mediawiki.org/w/index.php?title=MediaWiki%3AGadget-tabbedwindow.js&diff=6583748&oldid=6583747

MediaWiki:Gadget-tabbedwindow.js sets a hash fragment for the current tab, but doesn't put an ID element in the actual document. It should pass an ID attribute to either the tabPanel or panelLayout.

Filed remaining work as T367739

See T367735: Drop the user-agent-based browser support list from VE (VisualEditorBrowserUnsupportedList)

The extension in question doesn't exist anymore. I still don't think we need to support any future extension that kills essential web APIs.

See T367735: Drop the user-agent-based browser support list from VE (VisualEditorBrowserUnsupportedList)

As a chromium based browser this should work. If not it is unlikely to ever be prioritised.

See T367735: Drop the user-agent-based browser support list from VE (VisualEditorBrowserUnsupportedList)

Unrelated to night mode issues this seems like a sensible fix, and is quite small change.

@akosiaris This is just a technical exploration at the moment, there is no timeline or prioritisation yet. Knowing that this would a first within our infrastructure is useful information, thanks.

In T366095#9841427, @kostajh wrote:

Is there instrumentation to understand the usage of this feature? I guess we could look at API usage of discussiontoolsthank and/or data associated with Echo notifications that comment thanking will generate?

Update eslint-config-wikimedia to default to commonjs mode

Will need to be backported to wmf.9

Bisect blames my refactor at https://gerrit.wikimedia.org/r/c/mediawiki/extensions/Echo/+/1038337

PatchDemo: https://patchdemo.wmflabs.org/wikis/560a719650/w/index.php?title=Arsenal_F.C.&mobileaction=toggle_view_mobile

Quoting relevant parts of previous task:

Filed as T367413

We should make the references drawer white per T367413.

There is still an open patch to make the reference drawer white, which I think should be merged. If we don't want to repurpose this task for it, can a new one be created.

I would consider this done now. Feel free to close.

We're waiting for a few weeks to switch over to the version of the message with a separate header and body. Until then, it will look like this:

While the extra button is the responsibility of FlaggedRevs, it's VE who is responsible for refreshing all relevant parts of the UI after an edit happened.

Or "your edit session expired ..."

In T138401#9850102, @Ladsgroup wrote:

Flow will be undeployed soon. I'm not sure it's worth working on. Just my 2c.

In T366220#9847941, @Jdlrobson wrote:

When we switch to MathML (T271001) this will not be a problem and we may want to consider undo-ing this rule .

These tests were written by WMDE as part of their focus on the Template dialog. Not sure what project to tag this with now. CC @thiemowmde @Lina_Farid_WMDE

You'll want to use the full colour-rotating inversion rule, as math can contain colours.

This doesn't appear to be a regression, and has been broken since we merged the linear data and metadata.

Minimal test case in VE standalone:
<p>Foo</p><meta><h2>Bar</h2>

The 'mwAlienMeta' in question is the auto-generated TOC marker that gets inserted after the first paragraph:
<meta property="mw:PageProp/toc" data-mw="{&quot;autoGenerated&quot;:true}" id="mwDw">

For fr.wiki you can achieve what you need with :has selectors, which are supported by modern browsers:

dd:has( + dd ):not( :has( dl ) ) > span[data-mw-comment-end] {
	display: block;
	margin: 0.5em 0 0.5em -0.5em;
	border-top: 1px solid #a7d7f9;
}