Explicitly disallow unnamed Unicode codepoints in http://eel.is/c++draft/lex.charset#2

[tahonermann]

http://eel.is/c++draft/lex.charset#2
Says:

The character designated by the universal-character-name \UNNNNNNNN is that character whose character short name in ISO/IEC 10646 is NNNNNNNN; the character designated by the universal-character-name \uNNNN is that character whose character short name in ISO/IEC 10646 is 0000NNNN. If the hexadecimal value for a universal-character-name corresponds to a surrogate code point (in the range 0xD800–0xDFFF, inclusive), the program is ill-formed.

Should it also say:

If the hexadecimal value for a universal-character-name is not named by ISO/IEC 10646 then the program is ill-formed

Or is that implied?

[rmartinho]

After some review, I have the feeling this stems from an outdated 10646 version. The current version doesn't have a "short name" concept (well, except for jamo short names, but that's clearly not the intention here). There's a "short identifier" concept (in 6.5) but that cannot be more than six digits (the familiar U+ syntax). I propose the following.

The character designated by the universal-character-name \U00NNNNNN is that character whose code point short identifier in ISO/IEC 10646 is U+NNNNNN; the character designated by the universal-character-name \uNNNN is that character whose code point short identifier in ISO/IEC 10646 is U+NNNN. If the hexadecimal value for a universal-character-name corresponds to a surrogate code point (in the range 0xD800–0xDFFF, inclusive), the program is ill-formed. If a universal-character-name has any other hexadecimal value, the program is ill-formed.

(Ab)use of the C-word notwithstanding.

[steve-downey]

Also, IIUC, all short identifiers are valid even if they refer to a reserved codepoint. So you may be indicating something meaningless, but well defined meaningless.

[rmartinho]

At Rapperswil I discussed this issue with people from Core and it was agreed that the current wording with "short name" will be fixed editorially. We didn't discuss whether we should explicitly make those without such a short identifier illformed.

[tahonermann]

That's good. I think the right process is to submit a PR to https://github.com/cplusplus/draft. Anyone want to volunteer to do so?

[rmartinho]

I can do that.

[rmartinho]

This PR cplusplus/draft#2201 will be merged fixing the "short name" issue editorially. As discussed with Jens, I will write a short paper to clean up the wording there and clarify the nonexistent code point case.

[tahonermann]

@rmartinho I think this issue is complete, yes?

[tahonermann]

@rmartinho I think this issue is complete, yes?

Ah, no, this isn't complete yet. The accepted editorial PR only addressed the "short name" vs "short identifier" terminology issue. Martinho's draft D1139 (which I think is yet to be submitted to a mailing) addresses the concern tracked by this issue.

[tahonermann]

Closing this issue as resolved following the adoption of P1139R2 in Kona.