Re: [PATCH] Revised: utf-16 and utf-32 support in C and C++

[Jason Merrill <jason at redhat dot com>]

Kris Van Hees wrote:
> The complications with wchar_t not being at least a 32-bit value is of
> course that U"a" L"b" as a concatenation could have two interpretations:
>
> 	U"a" L"b"	-> U"ab" (if sizeof(char32_t> > sizeof(wchar_t))
> 	U"a" L"b"	-> L"ab" (if sizeof(char32_t> <= sizeof(wchar_t))
>
> Would it be preferable to just not support additional concatenations like
> this (to avoid the complication altogether), or is target specific behaviour
> acceptable in cases like this?

I think my preference would be just not to support such concatenations, 
as the encoding for wchar_t will be different from the encoding for 
char16/32_t in most locales; I doubt that people will want to freely mix 
the two.

Jason