Unicode simple case conversion of the 32-bit code point
chr, replacing towupper(), towlower() and towctrans().
Code points that have no mapping are returned unchanged.
Simple means one code point in, one code point out, so conversion is
length preserving: U+00DF LATIN SMALL LETTER SHARP S and the ligatures
are returned unchanged rather than expanded to SS,
FI, ... as SpecialCasing.txt would have it.
As with the classes, the mapping comes from the Unicode Character
Database and is therefore the same in every locale --- notably it does
not apply the Turkish and Azeri dotted/dotless i rules ---
and on every platform, and it works for non-BMP code points also on
Windows.
