Lexer is a mess

[michaelgwelch]

Perhaps I had a good reason 6 years ago, but it looks like the Lexer is trying to do half the job of the parser. All kinds of crazy state is being kept track of.

I recommend taking a new look at this and seeing if I can't identify X different types of tokens and simplify the Lexer to simply worrying about the tokens and not the context.

Issues:

• Worrying about startOfLine.
• Blank lines being skipped rather than just returning Token.EndOfLine (what did I think Token.EndOfLine couldn't be handled as an empty statement in Parser?)
• other weird stuff.