[css-cascade-6] fix @import definition - #14419

To parse a <media-query-list> production, parse a comma-separated list of component values, then parse each entry in the returned list as a <media-query>. Its value is the list of <media-query>s so produced.

Note: This definition of parsing intentionally accepts an empty list.

https://drafts.csswg.org/mediaqueries-5/#mq-syntax

But that isn't reflected in the syntax definition.
Maybe it should be?

A conditional <media-import-condition> in @import makes it possible to use that definition in linters. Otherwise such linters flag @import 'foo.css' as invalid.

I have no opinion on whether this should be reflected in the syntax. But I prefer that a superfluous ? not be added if an empty input is also defined as valid in prose.

<forgiving-selector-list> and <media-query-list> are the only two productions that accept an empty input, if I am not mistaken. To accomodate linter's needs while considering my opinion, an alternative could therefore be to redefine how they are parsed, and add ? wherever it is needed. E.g. @media {} is valid and @media is currently defined with @media <media-query-list> {...}.

Note that :is() is generally defined to match : <function-token> <any-value> ) before being validated/parsed against :is(<forgiving-selector-list>). <any-value> does not accept an empty input while <forgiving-selector-list> does, and all browsers currently accept an empty :is().

I think that in this case author needs go before technical purity :)
I also think that if something can be trivially expressed in a syntax definition that we should always do that.

Other than that I don't have any preference as to how or where we fix this.

I tend to agree with you because spec authors can make mistake such as @custom-media <extension-name> [ <media-query-list> | true | false ] ;: I do not think @custom-media --name; is intended to be valid, so it should be @custom-media <extension-name> [<media-query-list> | true | false]! ;.

I will work on an alternative PR later, if that works for you.

That would be great, thank you so much 🙇

I preferred to open #14431 since it seems more complicated than I thought, for the reasons explained in this issue.

Thank you @cdoublev 🙇