Add option to preserve/track whitespace during parsing

[bitwiseman]

There is an option to track the comments before a token in the parser. There are scenarios for beautification that are only possible by keeping the whitespace found before a token during a parsing.

For example, given how tokens were indented relative to each other during input, we could make some guesses about how they'd like to beautifier to format them.

It looks like this would involve making skip_whitespace() return the whitespace and they tracking that through the token and the resulting AST.