Editorial: Replace the single-row time zone code points table

[gibson042]

The single-row Time Zone Offset String Code Points table takes up a large amount of visual space for little practical benefit. This PR replaces it with direct use of <U+2212 MINUS SIGN> in the only grammar production that references the abbreviation defined by the table.

[michaelficarra]

Personally, I'd prefer to do this for tables 34 and 35 as well.

[michaelficarra]

This is fine, but we typically parenthesise the code point name.

Suggested change

	<U+2212 MINUS SIGN>
	<U+2212 (MINUS SIGN)>

[gibson042]

We're not consistent about the parentheses, although it's true that they are much more common than not (exceptions being concentrated in Canonicalize, plus one outlier in Appendix F).

But this particular suggestion also has the drawback that it will be rejected by the just-released version of ESMeta (which AFAICT was only extended to allow <U[+]([1-9A-F]|10)?[0-9A-F]{4}( \w+)*>), and beyond that I'd like to get to a point where ECMA-262 and 402 hew more closely to Unicode conventions (i.e., "U+…" followed by an unpunctuated uppercase name [in small caps where possible]). Given that, is appearance inside a grammar as a nonterminal sufficient distinction for omitting the parentheses here?

[michaelficarra]

I think either way is fine, I was just trying to match our current conventions. I'll bring it up in editor call.

[gibson042]

Personally, I'd prefer to do this for tables 34 and 35 as well.

Agreed! Should I extend this PR to do so?

[michaelficarra]

@gibson042 We'll probably want to do it in a separate PR. Let's wait until after we resolve this one.

[michaelficarra]

From editor call today:

• let's also do tables 34, 35, and 36 as well
• no parens is fine
• it can all be done in this PR

[bakkot]

Doing it for everything ends up being kind of noisy:

But I guess I'm ok with it.