NL029 05.10 [tab:lex.name.allowed] Disallow zero-width and control characters

[wg21bot]

Allowed characters include those from U+200b until U+206x; these are zero-width and control characters that lead to impossible to type names, indistinguishable names and unusable code & compile errors (such as those accidentally including RTL modifiers).

Proposed change:
Disallow invisible characters in this range

[jfbastien]

See sg16-unicode/sg16#48

[tahonermann]

SG16 plans to discuss this comment Tuesday morning:

• http://wiki.edg.com/bin/view/Wg21belfast/SG16

[tahonermann]

SG16 in Belfast: Recommend to accept with modification to deprecate identifiers that do not conform to Unicode TR#31 with exception for use of underscore.

Rationale: Future direction via D1949R0 intends to make these identifiers ill-formed in C++Future, so we want to put programmers on notice now to give them time to react. Additionally, we don't want to break compatibility with C and we lack implementation experience with D1949R0.

We had strong consensus for addressing this concern in C++20. The recommendation above had slightly better consensus than a poll to, rather than deprecating identifiers that do not conform to Unicode TR#31, adding a non-normative note containing guidance that programmers should use conforming identifiers.

Removing the SG16 label. This now needs to be seen by EWG.

SG16 minutes are available at:

• http://wiki.edg.com/bin/view/Wg21belfast/SG16NBNL029

[davidstone]

JF believes he may have created a CWG issue as the resolution for this NB comment, if not, one will need to be created. EWG recommended to adopt Unicode TR#31 with no particular ship vehicle. We did not have consensus to adopt NL029 for C++20.

[jfbastien]

I can't find it. I think I ended up just sending the email to SG16, not CWG chair.

[tahonermann]

I can't find it. I think I ended up just sending the email to SG16, not CWG chair.

I can't find one either and the above sounds right to me. sg16-unicode/sg16#48 contains a link to the following message that JF sent to the SG16 mailing list.

• http://www.open-std.org/pipermail/unicode/2019-May/000367.html

I'll request a CWG issue.

[jensmaurer]

CWG in Belfast: Given the EWG recommendation: Rejected. There was no consensus to adopt this change.

[tahonermann]

I'll request a CWG issue.

SG16 has continued to discuss changes to allowed identifiers in line with UAX #31; most recently in our January 8th, 2020 telecon. We have not yet identified a precise solution to propose. I am therefore not going to pursue filing a CWG issue in the near term. We'll continue pursuing a solution for C++23 and, based on those results, I may consider filing a CWG issue at a later time.