On Thu, May 28, 2020 at 9:49 AM Hubert Tong via SG16 <sg16@lists.isocpp.org> wrote:

On Thu, May 28, 2020 at 4:04 AM Corentin via SG16 <sg16@lists.isocpp.org> wrote:

Source character set is redefined as being the Unicode character set
It seems like we're encouraging homoglyph issues. Do we expect open source projects to maintain coding guidelines that restrict characters outside the ASCII range?

That ship has sailed:

// In many fonts the Greek letter 'Α', the Cyrillic letter 'А' and the Latin letter 'A' are visually identical, 
// as are the Latin letter 'a' and the Cyrillic letter 'а'
int Α = 0;
int А = 1;
int A = 2;
int a = 3;

int а = 4;
compiles out of the box with current gcc, and clang since version 3.x. This is new behavior for gcc. MSVC does not permit it with no flags, but does with /utf-8.

https://godbolt.org/z/WAMqrq

It's one of the reasons addressing identifiers is somewhat urgent, even though we're only addressing UAX 31, not 36 http://unicode.org/reports/tr36/ UNICODE SECURITY CONSIDERATIONS. The implementation costs for 36 are very high, though.