[css-syntax-3] Rename <an+b> to <an-b>

[tidoust]

See context in #9473. Rename is aimed at avoiding interpretation of + as a component value multiplier by parsing tools. The ID of the definition does not change.

The rename is also done in the definition of the serialize an <an+b> value algorithm, preserving the previous ID since the algorithm was exported (and referenced by other specs).

Once construct has been renamed, the following specs will need to be slightly updated as well (they either reference <an+b> or serialize an <an+b> value):

• cssom-1
• css-gcpm-4
• selectors-4

I'll prepare a follow-up PR for these updates once the cross-references database gets the new definitions.

[cdoublev]

I do not see any mention in CSS GCPM 4: are you are thinking to :nth-of-page(n)? I am not sure if the spec authors expect <integer> or <an-b>. There are mentions in CSS GCPM 3 and CSS Overflow 4 though.

This PR might also be an opportunity to replace "the An+B notation" with "the <an-b> type", and An+B with <an-b> in value definitions, to fix #9494.

[tidoust]

I do not see any mention in CSS GCPM 4: are you are thinking to :nth-of-page(n)? I am not sure if the spec authors expect <integer> or <an-b>. There are mentions in CSS GCPM 3 and CSS Overflow 4 though.

Sorry, css-gcpm-4 is a typo, it should have been css-gcpm-3.

I didn't include CSS Overflow 4 because, while it mentions an+b, it does not reference the <an+b> type, and the mentions are only there to alert about open issues in any case.

This PR might also be an opportunity to replace "the An+B notation" with "the <an-b> type", and An+B with <an-b> in value definitions, to fix #9494.

I leave that up to spec editors, but how would that change help with #9494, which is more about updating Selectors 4 to use more formal definitions, and reference <an-b>, for pseudo-class selectors?

[cdoublev]

I didn't include CSS Overflow 4 because, while it mentions an+b, it does not reference the <an+b> type [...]

Yeah, I realized it afterwards.

I am also fine with making changes for css-gcpm-4 and selectors-4 in another PR.

Selectors 4

• :nth-child(An+B [of S]?) should be :nth-child(<an-b> [of <complex-real-selector-list>]?)
• :nth-last-child(An+B [of S]?) should be :nth-last-child(<an-b> [of <complex-real-selector-list>]?)
• :nth-of-type(An+B) should be :nth-of-type(<an-b>)
• :nth-last-of-type(An+B) should be :nth-of-type(<an-b>)
• :nth-col(An+B) should be :nth-col(<an-b>)
• :nth-last-col(An+B) should be :nth-last-col(<an-b>)

CSS GCPM 4

• :nth-of-page(n) should be :nth-of-page(<an-b>) (to be confirmed)

[tabatkins]

Yes, go ahead and change the selectors references.

Ignore GCPM 4, I'm not sure what's intended (and the spec isn't maintained anyway).