UTF-8 support - char or wchar_t?

[Martin Sebor]

Paolo Carlini wrote:
...
>> fstream must use the codecvt facet that's installed in the locale 
>> imbued in it. If
>> the fstream object is imbued with locale("fi_FI.UTF-8") it must
>> use codecvt_byname<char, char>.
> 
> 
> Could you please cite the specific section of the standard requiring this?
> Definitely, *not* 27.8.1.1, p5.

No, but I'd say 22.1.1.1.2, p4 strongly suggests that it should
be the case. There might be a more explicit requirement somewhere
else in clause 22, but I think it just makes sense. Otherwise
what would be the point of any of the byname facets at all if
an implementation was allowed to construct a named locale with
just the base facets in it?

> 
>> There isn't any codecvt<char, char> in any named locale other than C 
>> or POSIX.
> 
> 
> For that matter, I don't think any instantiation of codecvt_byname is 
> required...

You must be looking at the old standard. Table 52 of C++ '03
lists the required instantiations of all facets, including
codecvt_byname<char, char>. See DR 21.

Martin