Centralize definition for name conflict handling of naming at-rules

[SebastianZ]

As @cdoublev pointed out in #12973, most of the at-rules that define a name have a definition that indicates that the last one wins. Though there are currently a few at-rules like @color-profile and @font-palette-values, for which handling of duplicate names is undefined.

Therefore, I think we should establish a general term for these name-defining at-rules and centralize the definition for name conflict handling, so all those rules can refer to it.

I am not sure where that definition could live. We currently have one for at-rules in general in CSS Syntax 3, but that might not be the best place for a definition like that.

Sebastian

[cdoublev]

"The last one wins" may have been a simplification. For @function and @mixin, it is the rule in the stronger cascade layer, for example.

Similarly, many specs defining at-rules have to repeat how to handle multiple declarations of the same descriptor, noting that descriptor declarations can have a priority in some contexts.

[SebastianZ]

"The last one wins" may have been a simplification. For @function and @mixin, it is the rule in the stronger cascade layer, for example.

Yeah, that definition deviates them from the other specs. So CSS Mixins might still refer to the central definition for the term but keep its definition for how to resolve name conflicts.

Similarly, many specs defining at-rules have to repeat how to handle multiple declarations of the same descriptor, noting that descriptor declarations can have a priority in some contexts.

Right. We might cover this here, as well. Though to me this seems separate enough to be discussed in its own issue. @cdoublev Maybe you can create one for that?

Sebastian