bk_text and bk_info both need to know the on-screen width of
characters in order to wrap and align them properly. Therefore, they
should be using wcwidth(). So here are a couple of wrappers on
wcwidth(), one which filters out the Unicode characters not
representable in the target charset, and one which converts _from_ a
charset to Unicode before calling wcwidth(). bk_text and bk_info
should now align correctly even in the face of unsupported
characters and Japanese.
git-svn-id: svn://svn.tartarus.org/sgt/halibut@4116
cda61777-01e9-0310-a592-
d414129be87e