silverwind

csstree merged it, should be in next release.

Still waiting on next `css-tree` package release.

If stylelint had a proper parsers like proposed in https://github.com/stylelint/css-parser/issues/2, this case wouldn't even need a rule, as the invalid syntax would trip that parser.

Spec links are: https://drafts.csswg.org/css-syntax-3/ https://drafts.csswg.org/css-variables/ As for `@` rules, it seems it does not affect all of them. `@page` accepts `var`, but `@font-face` and `@counter-style` does not. So it seems...

Agree, syntax errors are topics for a parser, not a linter.

Agree, renamed title to `at-rule-no-invalid-descriptor`.

Right, it's only refering to the right side, so `at-rule-no-invalid-descriptor-value`.

Ah, misunderstood you. If `var()` is valid in some constallations, it'll make the rule more complicated. I wonder if there is data around somewhere which CSS functions are considered invalid...

`@custom-media` seems to not be part of any CSS standard, I wonder why stylelint chose to even support it.

Yeah, I guess they are more likely to just make variables work inside media queries than to implement another at-rule, I suppose 😆