make isuppercase and islowercase agree with Unicode standard

[stevengj]

Currently, islowercase checks whether a character is in category Ll, Letter: Lowercase, and isuppercase checks for category Lu, Letter: Uppercase or Lt, Letter: Titlecase.

However, it was recently brought to my attention that there are actually official Unicode derived properties called Lowercase and Uppercase which differ from these definitions.

• Titlecase characters like Dž (U+01c5) are not considered uppercase. (Note that uppercase('Dž') yields a different character 'DŽ', so this makes a certain sense.)
• Some Lo, Letter: Other characters like ª are included as Lowercase (or Uppercase in other cases like Ⓐ).

The next version of utf8proc will provide islower and isupper functions compliant with these definitions (JuliaStrings/utf8proc#196), so we may want to switch to them.

(My guess is that it makes little difference in practice — I'm not clear how useful these functions are for general Unicode strings — but the standard here seems fairly sensible. Apparently this is what Python's isupper/islower functions do.)