Incrementally parse documents

[euclidianAce]

At the very least we can incrementally update the token list without re-parsing the entire doc.
Whether this speedup is worth it, idk. We could bring in my ltreesitter module to do this, but then theres the issue of translating the cst that tree-sitter generates to an ast that tl understands, which might not be worth it.

We could get a real speedup if we could figure out a way to incrementally update the type checking environment instead of regenerating it each time

[euclidianAce]

With neovim master being able to get incremental updates I will try to implement this soon