[scroll-animations-1] Should `none` be part of the list in scroll/view-timeline-name?

[andruud]

The scroll-timeline-name and view-timeline-name properties have this grammar: none | <custom-ident>#. That is fine in isolation, but it's inconsistent with animation-name, which has [ none | <custom-ident> ]#, and that made me wonder if it's intentional.

If it is intentional, then feel free to close this.

@fantasai

[ydaniv]

I think it's intentional since these are timeline declarations that can be matched by definitions of animation-timeline's and are matched by name, and not by position in the list. So it's either a list of names or none.

[tabatkins]

Yeah, animation-name's use of none as any item in the list (which just "turns off" that entry in the list) is pretty odd actually.

[andruud]

OK, in that case, how am I supposed to interpret the grammar of scroll-timeline:

[ <'scroll-timeline-name'> [ <'scroll-timeline-axis'> || <'scroll-timeline-attachment'> ]? ]#

I'm aware of css-values-4:

These types do not include CSS-wide keywords such as inherit, and also do not include any top-level comma-separated-list multiplier (i.e. if a property named pairing is defined as [ <custom-ident> <integer>? ]#, then <'pairing'> is equivalent to [ <custom-ident> <integer>? ], not [ <custom-ident> <integer>? ]#).

But scroll-timeline-name doesn't have top-level comma-separated-list multipliers.

Even if you ignore that, the grammar would suggest that scroll-timeline: none, none, none is allowed, even though it's not allowed by the longhand?

Yeah, animation-name's use of none as any item in the list (which just "turns off" that entry in the list) is pretty odd actually.

Yet if scroll-timeline worked like that it seems to answer all my questions, so maybe not that odd. :-)

[fantasai]

Hm, yes, it does seem we have a problem. :) I think the two options are:

• Allow none as an alternative to <custom-ident> in *-timeline-name, essentially disables that entry in the coordinated properly list.
• Rework the grammars explicitly so that both *-timeline-name and the *-timeline shorthand have none | <whatever>.

[ydaniv]

ah, for coordinating between longhand lists, I see.
So the options mean either keep allowing none for -name for disabling a single entry, or only allowing it as the entire value for disabling all timelines of that type, right?
Although it won't be necessary in most cases, I guess the first option allows more flexibility, and it may seem strange to developers that they have to remove an entry from all properties to disable it.

[tabatkins]

Yeah, I suppose following the animation-name model works. It's weird and I don't particularly like it, but it works slightly better with a few of our abstractions.

[css-meeting-bot]

The CSS Working Group just discussed [scroll-animations-1] Should `none` be part of the list in scroll/view-timeline-name?, and agreed to the following:

• RESOLVED: Allow a list of nones

The full IRC log of that discussion

<dael> fantasai: This issue was brought b/c right now the way values are spec, scroll-timeline name is none or a list of comma spec idents. But background-image and the like is a list of nones or whatever so can have a list of nones.
<dael> fantasai: Question here is shorthand drafted as inconsitent. Do we want pattern to have a list of nones or make that explicitly not possible
<dael> fantasai: Seem to be leaning toward a list of nones for consistency and it allows an author to turn off a timeline without taking out definition from all properties you can flip it to none
<dael> flackr: I support allowing a list of nones
<dael> fantasai: Prop: Allow list of nones
<dael> Rossen_: Any opinions?
<dael> florian: Seems reasonable. No strong opinion.
<dael> Rossen_: Objections:
<dael> RESOLVED: Allow a list of nones