Fixing strange parses

[jaredponn]

In light of issue #116, this fixes some of the strange somewhat nonintuitive parses.

High level overview of changes

• This proposes a more structured way of handling whitespace with parsec. The key idea is to maintain the invariant that all parsers begin at a non whitespace token.
• We maintain the invariant by ensuring that the initial parser skips whitespace / comments, and every "token" eats up whitespace/comments after it finishes parsing the token.
• The structure of parsing lead to changes in the test suite where we had lots of test cases which had "invalid" whitespace at the beginning. This structure makes these tests no longer relevant.
• I also included a specification of the syntactic forms of a lambda buffer file.

[bladyjoker]

@jaredponn thank you so much! this is incredible work and I learned a lot.

This is pure gold.

The only thing I want you to do is to extract that big comment about the syntax and put it in a chapter.

[jaredponn]

Thank you @bladyjoker for the feedback, it's very appreciated.

As requested, I wrote a quick chapter over here.

[bladyjoker]

Very lovely documentation! Left some comments for you to attend to, but feel free to merge this at your leisure. We will definitely iterate on this chapter some more so no need to get it perfect the first time around.

[jaredponn]

Thanks again for the feedback -- it was really helpful! #119 #120 will track the remaining issues.

I'll rebase this and merge it now.