Sam Atkins

This has conflicts.

> I think you have forgotten to give semicolon(;) after red.So I guess that's the issue. That doesn't affect the outcome. CSS doesn't require a semicolon if there isn't another...

IIRC we do the correct thing for borders here, so it should just be a case of copying that logic to the outlines painting pass.

And I should have read the test case properly. 😅 That's weird, I'm sure I made this work a while ago.

If and when it becomes an official web specification, we'll consider it. Currently, as far as I can tell, it's still an experimental API only available in Chrome-derived browsers.

> This should only be reviewed and merged once that one has been. Please mark PRs as drafts if they shouldn't be merged yet. There's a link on the right...

MacOS "Actually ran to completion fine" timeouts are my favourite.

This is a consequence of us walking up the tree until we find an element with a custom property of the name we're after, instead of cascading them properly. A...

Hmm. It's probably a good candidate for optimising because it does one thing repeatedly. One option otherwise would be to mark some tests (slower, more exhaustive ones like this) as...

https://github.com/w3c/csswg-drafts/pull/12261 now solves the issue of two places saying what to do to absolutize a URL, by making the computed-value step explicit. Going to give that a review now, but...