Hi all, since 5.8.0 Perl strings are handled internally either as octet strings or as UTF-8 strings.

I need to link a C++ library that uses wide-characters (wchar_t) with perl to extend an existing CPAN package.

Is it true to assume that the stanard XS typemap that comes with perl is not capable to do the correct UTF-8 to wchar_t conversion or vice-versa ?

The standard typemap contains the following:

wchar_t *               T_PV
...
INPUT
T_PV
        $var = ($type)SvPV_nolen($arg)
...
OUTPUT
T_PV
        sv_setpv((SV*)$arg, $var);
[download]

What would be a good replacement for this that would work with both forms of Perl internal strings ?