charset for a locale

[andrew at pimlott dot ne dot mediaone dot net (Andrew Pimlott)]

I hope this is an ok place to ask this question.

I have read the glibc 2.2 documentation on charset handling and
locales, and a very basic concept still escapes me.  How can I
figure out the default multibyte encoding for the current locale
(ie, the encoding used to interpret the char type)?  It seems that I
can use the mbtowc functions to convert any string to Unicode (well,
this assumes I know the wchar encoding, but in practice this is
easier), but can't actually query the multibyte encoding used by
mbtowc.  This would seem further to imply that mbtowc has
functionality that cannot be obtained with the iconv functions,
which I find strange.

Did I miss something?  Is there a reason that most people don't need
this functionality?

Thanks,
Andrew