add u8 variants

[rurban]

C18 thought about adding the char8_t type, to match the new u8 string prefix: http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2231.htm
They only wanted to add some converters mbrtoc8, c8rtomb, no other functions for char8_t, instead of the ASCII str and widechar wcs variants.

80% use UTF-8, not widechar. wcs only makes sense on windows, everywhere else u8 won.

Think of adding u8 variants for all functions (disabled by default).
look at julia's utf8lib and utf8.h (which is not as good).

[rurban]

A minor proof that nobody uses wchar besides Windows is that gcc waits for __attribute__(format(wprintf,,) support since 2008. Patches exist since 2015.

Just kill wchar and replace it with char8 or utf8 or u8 as default. POSIX is a joke.
They dont support case-folding (wcsfc), nor wcscmp nor wcsnorm. Which means you cannot search for strings, just byte buffers. There is a difference which didn't make it to the committee yet. (neither the various kernel's, which treat identifiers as byte-buffers. they are not. file paths, names, ...) http://perl11.org/blog/foldcase.html