[css-values-5] Clarify how to parse `<if-args-branch>`

[cdoublev]

https://drafts.csswg.org/css-values-5/#typedef-if-args-branch

<if-args-branch> = <declaration-value> : <declaration-value>?

<declaration-value> accepts : so : <declaration-value>? seems unreachable... unless a CSS parser is supposed to backtrack, drop the last token matched for <declaration-value>, then resume parsing? Or should it be <declaration-value> [ : <declaration-value>? ]??

I have a more specific question about the spread syntax. (edit) Never mind: since . . . var(--custom) is valid at parse time, I guess ...var() is too.

I could open a separate issue but the spread syntax seems only relevant for <if-args-branch> if I am not mistaken, since the argument grammars of other arbitrary substitutions take a comma-separated <declaration-value>, so they will always get a match.

This behavior can be worked around by immediately preceding an arbitrary substitution function with the spread syntax ..., indicating that it must be resolved "early", before division into arguments.

What is accepted (at parse time) after ...? Only an arbitrary substitution? Should it be valid itself? (eg. if(...if()))

[AtkinsSJ]

I'd like to add that this same issue happens for the argument-grammars of other ASFs. eg, attr():

<attr-args> = attr( <declaration-value>, <declaration-value>? )

A comma token is also valid in <declaration-value>. The way I handled this in Ladybird is to specifically disallow comma tokens when parsing the first <declaration-value>. I'm not familiar with if() but I think a similar thing could be done here with colons. It really needs mentioning in the prose of these though.

[cdoublev]

I think you forgot the requirements for comma-containing productions in a functional notation:

If it does not start with a "{" token, then it cannot contain commas or {} blocks

I do not know why <declaration-value> and <any-value> are defined in CSS Syntax instead of CSS V&U, collocated with these requirements.

[AtkinsSJ]

I think you forgot the requirements for comma-containing productions in a functional notation:

I was completely unaware of that whole section. 😅 Thanks for letting me know.

[AtkinsSJ]

And yes, putting them in the same spec would make this less confusing.