UTF-8 patch -- library issues

[Michail Vidiassov <master iaas msu ru>]

Dear UTF-8 patch developers/mainteiners/users,

I would like to bring to your attention the following points:

1) While the patch relies on a custom patched version of slang,
   there is wide character support in newer versions of ncurses,
   and beta versions of slang 2 claim UTF-8 support.

2) New ncurses, if compiled with widec char support, provide
   addnwstr function, that seems to be a complete analog of SLsmg_write_nwchars.
   I have made mc with UTF patches compile and run when
   linked against ncrurses 5.4 by replacing
   SLsmg_write_nwchars by addnwstr.

   It may be a good idea to wrap wide character output routine
   in some sort of mc_ function with #ifdef HAVE_SLANG inside.

3) I did not give Slang 2 beta a try, but as far as I can see from
   the docs, there is no wide character API there (curses compatible
   or proprietary). Existing functions have been enhanced to work
   with UTF-8 data.

4) New version of ncurses requires some configure changes -
   the library may be called ncursesw, with header file
   <ncursesw/curses.h>. Plus some check for wide char support
    being compiled in.

                         Sincerely, Michail