> The 'codePoint' typedef may be problematic:
>
> // Unicode code points (4-byte int on most systems)
> typedef wchar_t codePoint;
How often is UTF-32 actually in use? I would have thought
that UTF-16 and UTF-8 are the dominant encodings?
On Expat it works like this:
Depending on compile options we define XML_Char to be
char (for UTF-8), or to be either ushort or wchar_t for UTF-16.
In the latter case the use of wchar_t depends on a given compiler's
ability to define it as a two byte type (which is an option on gcc,
and the default on Windows, covering a big portion of the market).
Karl
