As far as I can tell, right now the only way to get a char normalized is to convert it into a String first, then "iterate" over it. This is a bit awkward and inefficient. It would be nice to have UnicodeNormalization implemented for char.

The UnicodeNormalization trait looks fine for this: you still need to produce an iterator because a single char might normalize to multiple chars.

The project looks big enough that I'm not going to try a pull request myself without some feedback. It looks like the normalize module provides the necessary functions, so I guess it's mostly just typing it in?