
On 30/06/2020 13:16, Richard Damon wrote:
On 6/30/20 7:53 AM, M.-A. Lemburg wrote:
Since the C world has adopted wchar_t for this purpose, it's the natural choice.
I would disagree with this comment. Microsoft Windows has chosen to use 'wchar_t' for Unicode, because they adopted UCS-2 before it morphed into UTF-16 due to the expansion of Unicode above 16 bits. The *nix side of the world has chosen to use UTF-8 as the preferred way to store Unicode characters.
Also, in Windows, wchar_t doesn't really meet the requirements for what C defines wchar_t to mean, as wchar_t is supposed to represent every character as a single unit, and thus would need to be at least a 21 bit type (typically, it would be a 32 bit type), but Windows makes it a 16 bit type due to ABIs being locked before the Unicode expansion.
Seconded. I've had to do cross-platform (Linux and Windows)* unicode work in C. Using wchar_t was eventually rejected as infeasible. * Sorry, I had a Blues Brothers moment. -- Rhodri James *-* Kynesim Ltd