On Sun, Jan 12, 2020 at 7:06 PM Steve Downey via SG16 <sg16@lists.isocpp.org> wrote:

Module names are, I believe, a superset of identifiers, where '.' is allowed between the parts that correspond to the identifier fragment.

Module names are actually a dotted sequence of identifiers (they can have whitespace between the dots and the identifiers, and don't allow consecutive dots).

- Michael Spencer

 

I will add details. That header file names are not identifiers is an interesting compatibility issue. 

As currently proposed, I am restricting syntax elements to the basic source character set. That might make sense for headers (not header files), but expressing that constraint might be complicated. I doubt anyone will insist on a header with characters outside the basic set soon. 

On Sun, Jan 12, 2020, 21:39 Alisdair Meredith <alisdairm@me.com> wrote:

Minor suggestion: call out where identifiers are used (class/variable/function names etc.) and a couple of examples where they are not (header names, raw string prefix/suffix, etc.). In particular - would this apply to module names?

AlisdairM

Sent from my iPhone

On Jan 13, 2020, at 1:36 AM, Steve Downey via SG16 <sg16@lists.isocpp.org> wrote:



Current draft of "C++ Identifier Syntax using Unicode Standard Annex 31" https://github.com/steve-downey/papers/blob/master/D1949R1.md
Final format will by html, produced by MPark's wg21 tool. Expanded discussion of everything, added explicit character lists, added authors, per last telecon.

--
SG16 mailing list
SG16@lists.isocpp.org
https://lists.isocpp.org/mailman/listinfo.cgi/sg16

--
SG16 mailing list
SG16@lists.isocpp.org
https://lists.isocpp.org/mailman/listinfo.cgi/sg16