[EXTERNAL] Re: Basic Unicode character/string support absent even in modern C++

[Tom Honermann]

On 4/23/20 4:24 PM, Murray Sargent wrote:
>
> I write a lot of C++ code to process math text and use the Visual 
> Studio C++ and clang compilers to compile for Win32, Mac, iOS, and 
> Android platforms. These compilers support UTF-8 strings, literals and 
> comments. They even allow math italic characters like 𝑎 (U+1D44E) to 
> be used as variables! Using real math symbols instead of notation like 
> 0x222B makes the code so much more readable. It’s easy to have 
> Devanagari string constants, braille constants (see Unicode 2800 
> block), etc.
>
There is a proposal in action to place new restrictions on what 
characters can be used in identifiers following the guidelines in UAX 
#31 <https://unicode.org/reports/tr31/>.  Follow the progress of P1949 
<https://wg21.link/p1949> if you are interested (monitoring the SG16 
mailing list and meeting notes (see 
https://github.com/sg16-unicode/sg16) is one way to do so).  It is 
*possible* that some of the identifiers you've been using may become 
invalid if that proposal is adopted.
>
> C++ has grown up considerably past ASCII. But you still have to use 
> ASCII operators such as != instead of ≠. It would be so fine if C++ 
> would allow standard math operators to be used as aliases for the 
> ASCII operator pairs like !=, <=, >=
>
I'm not aware of any serious proposals to do so.  But we are being 
careful in our evaluation of P1949 to ensure that doing so will remain a 
possibility in the future.

Note that P1949 will *not* allow doing something like the following.  
This is to ensure the availability of ≠ for use as a future operator.

#define ≠ !=
int f(int a, int b) {
   return a ≠ b;
}

Tom.

> Murray
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://corp.unicode.org/mailman/private/unicode/attachments/20200423/22d0c910/attachment.htm>