Change sequence parser API.
Use [a, b, c, d].r instead of a < b < c < d
Benefits: can specify token name and inner skip.
[a, b, c, d].r(token: some_spec, skip: /[\ \t]/)
Unify helper methods of sequence and repeat parsers.
[a, b, c, d].r[3]
(a ** 0)[-1]
(a ** 0).operator_table(infix('a'), postfix('b'), rinfix('c'))
Note: don't add values_at and many other Array operations, it would be confusing.
Remove right associative things.
Because they are not very useful.
Rewrite >> and <<
Make them the same as rparsec.
