String normalization: canonical forms and identifier case fold

[egao1980]

This should allow implementation of the Unicode efficient canonical identifier caseless match. Something like that:

(defun indentifiers-match-p (a b)
  (equal 
    (cl-unicode:identifier-case-fold-mapping (cl-unicode:normalization-form-d a))
    (cl-unicode:identifier-case-fold-mapping (cl-unicode:normalization-form-d b))))