normalizeUnicode

This optimisation can convert unicode-range descriptors to use the shorter wildcard ranges when a particular value meets the wildcard criteria. Values will be converted when the code matches 0 & f in the same place on both sides of the range. So, u+2000-2fff can be converted to u+2???, but u+2100-2fff will be left as it is.