[css-values-5] Should an arbitrary substitution nested within another be accepted in context?

[cdoublev]

@page { size: env(safe-area-inset-top, attr(size length)) }

Is size declared with a valid value?

var(--custom, var(1)) must be invalid in any property value: the basic syntax of var(1) is checked, but should its validity in the context be checked?

[tabatkins]

Other than the most basic grammar check, we don't do any other checking at parse time. I'm inclined to be consistent there, and not check contextual validity either.

[cdoublev]

I am not sure I remember why I was unsure of the answer to my question. Perhaphs because of the current text for var()/attr()/env()/random-item() (emphasize added):

If a property contains one or more [arbitrary substitution] functions, and those functions are syntactically valid, the entire property’s grammar must be assumed to be valid at parse time.

But I would definitely say that @page { size: attr(size length) } should be invalid, ie. those functions [should be] syntactically valid and accepted in the context.

[cdoublev]

Related (to "we don't do any other checking at parse time"): #8771.

[cdoublev]

I changed my mind and agree that contextual validity is not needed: color: var(--custom, 1px) is valid so size: env(..., attr(...)) should be valid too.

It is a inconsistent with var(--custom, var(--)), which is currently invalid in all browsers (#8771) but... ok.