   The methods to get a character's combining class, bi-directional category,
and numeric value are not yet implemented.  The method to get the decomposition
mapping simply returns that field as a string, but this is sufficient to tell
whether that character has a decomposed equivalent, at least.
   The GNUstep Base Library currently defines type 'unichar' as a 16-bit value.
It does not yet support surrogate pairs.  Hence, the GNUstep Unicode Character
Data Library ignores characters with values above 65,536.
