[css-cascade] Does 'revert-layer' roll back to preshints?

[Loirooriol]

https://drafts.csswg.org/css-cascade-5/#preshint

All document language-based styling must be translated to corresponding CSS rules and enter the cascade as rules in either the UA-origin or a special-purpose author presentational hint origin between the regular user origin and the author origin. For the purpose of cascading this author presentational hint origin is treated as an independent origin, but for the purpose of the revert keyword it is considered part of the author origin.

What about the revert-layer keyword?

[fantasai]

It's not the revert keyword, so it's treated as an independent origin, and isn't reverted by revert-layer in an author style sheet.

IIRC this was intentional.

[Loirooriol]

That section was entirely copied from L4 where there is no revert-layer, so it's not obvious that this was intentional. Maybe add a note? Also, is it possible to test this? Is there any normative preshint? HTML rendering only says "expected", not "must".

[Alohci]

Using revert-layer to revert height and width of an image back to the HTML attribute values seems distinctly useful. Chrome and Firefox already do this nicely.

[Loirooriol]

OK, filed https://bugs.webkit.org/show_bug.cgi?id=237532

[Loirooriol]

Actually, Chrome was not doing this well. Filed https://bugs.chromium.org/p/chromium/issues/detail?id=1306511

[tabatkins]

@Loirooriol Added a clarification. Let us know if the wording works for you, or if you think something different would be better.

[Loirooriol]

LGTM