A much faster parser

[andreas-zeller]

Make use of the Python built-in parser instead of using our own:

Split the .fan file into three parts:

1. Grammar rules, identified by ::=
2. Constraints, identified by ; or a new leading keyword (Introduce constraints by a where keyword #180)
3. Python code

Then, use the Python parser to parse (2) and (3); use a lightweight parser to parse (1) (or convert (1) into something that can be parsed by Python).

Note:

• This would also reduce the need for caching parse results.
• This may require to implement our own import (Allow import of .fan files #178)
• This could allow us to implement more grammar syntax features (FEAT: Charsets and regexes in grammars #65)