[CM] Use ICU4C to support Unicode character case-mapping.

[jingkaimori]

Hi all,
    I have forked s7 and works on unicode string. S7 hard-coded case
mapping for ASCII, and load them during startup. It's impossible to
use this method for unicode characters, because unicode alphabets is
too large to be coded in source file or loaded during startup. So use
3rd library such as ICU4C to convert case and handle utf encode may be
necessary.
    Are these modification containing 3rd library port welcomed to snd
project? Or a fork is necessary?