[css3-conditional] Should CSS.supports() accept properties which are only valid in certain rules?

[AlaskanEmily]

The exact case I'm looking at is page-size ( https://drafts.csswg.org/css-page-3/#page-size-prop ), which I'm currently implementing in Firefox, and which is only specified to be valid in @page rules.

It appears that size is also valid in style declarations in WebKit in Chrome with the same syntax, which is hit in some web-platform tests, for instance https://github.com/web-platform-tests/wpt/blob/master/css/css-conditional/js/CSS-supports-CSSStyleDeclaration.html

If a property like size is supported only in a specific at-rule, should it accepted by CSS.supports?

[tabatkins]

I personally believe the answer should be no, because it's possible to have a property and an at-rule descriptor of the same name, and that would be ambiguous here.

[SebastianZ]

What @tabatkins says applies to the current syntax of property checks, though what about introducing a new syntax that includes the at-rule in the condition? I.e. something like this for example:

@supports (@page { size: 10cm; }) {
...
}

That would work like the existing parsing checks for properties and selectors, so no new logic here. It only requires to extend the syntax of @supports to take at-rules.

Sebastian

[Loirooriol]

This is related to #2463, but discussion there is focused on a @supports variant for descriptors, while this seems more about the JS counterpart of that.

[fantasai]

As @tabatkins explained, there are cases where a property and descriptor have the same name, but these are nonetheless separate features: this is most evident for @font-face. For @page rules they behave almost exactly like properties, but technically this distinction still exists. I've slightly clarified the spec to say it's talking about property declarations within style rules (not descriptor declarations within at-rules). @AlaskanEmily Let me know if you think any additional clarifications are needed.

[gsnedders]

It appears that size is also valid in style declarations in WebKit in Chrome with the same syntax, which is hit in some web-platform tests, for instance https://github.com/web-platform-tests/wpt/blob/master/css/css-conditional/js/CSS-supports-CSSStyleDeclaration.html

Note that this test only checks that CSSStyleDeclaration and CSS.supports agree, which per the current spec is true (both only contain properties, because we don't actually define e.g. CSSPageRule.style as having attributes for like page-size, c.f. #5649).