more steps toward UTF-8 support: port all the parts of DB's UTF-8 patch

more steps toward UTF-8 support: port all the parts of DB's UTF-8 patch
that I currently understand to current CVS, with modifications of mine
to autodetect UTF-8 support and to display multibyte strings instead of
wide strings

git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@2193 35c25a1d-7b9e-4130-9fde-d3aeb78583b8