This pull request breaks a convention: every other method on Char (like alpha? and upcase) deals only with ASCII text.

Crystal needs extensive support for Unicode and its categories, but the language creators have expressed the wish to keep that in a Unicode module.

Note that I'm not against generally supporting Unicode in String and Char, it's just that I think it's super hard to do right now, so I consider it low priority (for myself, if anyone wants to implement it please go ahead)

#blank? Should be aligned with #whitespace?. I kind of prefer defaulting to unicode operations. Even on case related functions. I think we should settle down what we want before making only some methods unicode friendly.

Unicode sounds good as standard for everything on Char. Perhaps there would be room for an AsciiChar (or ASCIIChar according to idiomatic naming?) for when needing max-juice to process text-records and transactions specified to being ASCII. Or not.

extract to here https://github.com/kostya/blank, because use in some projects.