web-platform-tests/wpt#58814 introduces a test that passes in chrome by minting the inheritance tree, see web-platform-tests/wpt#58814 (comment)
It basically asserts that non-existing view-transition pseudo-elements should inherit from non-existing ancestor pseudo-elements rather than the originating element.
While this can be implemented for this specific view transitions case, it's a really odd special-case (which I think was wrongly implemented in chrome, see the comment above), and is inconsistent with all other pseudo-elements.
I think it's more consistent to say that when CSSOM needs to resolve a style of a non-existent pseudo-element, you just match the rules and inherit from the originating element. That's what every other pseudo-element does.
I think there's no need to change the spec necessarily (but might be worth clarifying / being explicit).
cc @lilles @kevers-google
web-platform-tests/wpt#58814 introduces a test that passes in chrome by minting the inheritance tree, see web-platform-tests/wpt#58814 (comment)
It basically asserts that non-existing view-transition pseudo-elements should inherit from non-existing ancestor pseudo-elements rather than the originating element.
While this can be implemented for this specific view transitions case, it's a really odd special-case (which I think was wrongly implemented in chrome, see the comment above), and is inconsistent with all other pseudo-elements.
I think it's more consistent to say that when CSSOM needs to resolve a style of a non-existent pseudo-element, you just match the rules and inherit from the originating element. That's what every other pseudo-element does.
I think there's no need to change the spec necessarily (but might be worth clarifying / being explicit).
cc @lilles @kevers-google