Recognize cases where rascal is greedy instead of non-greedy

[DavyLandman]

I see a pattern where rascal's greedy matches (a+ !>> a) are mapped to non greedy regexes with a follow guard. a+? (?!a).

We could optimize this as: a+ in regex.

Some heuristics on when A+ or A* is greedy:

• it's followed by a non optional literal that does not match A. example: [0-9]+ "e"
• it has a a follow restriction with same chars as itself: example: [0-9]+ !>> [0-9]