fix regexs to deal with hard delimiters

[maxkorp]

A few key changes here

• Added the concept of a "hard" delimiter (something that cant be contained in another word, ie }} or <)
• MakeEntry now generates 2 tokens for "soft" delimiters, one checking for it being broken by a hard delimiter, and one checking for it hitting the end of the template
• The Tokenizer checks for captures[1] instead of captures[0]. Since the tokens can now mach something like each%}, it became necessary to land each in a group, and I was able to unify them all on capture[1], rather than flipflopping between 0 and 1 conditionally.
• moved whitespace to check \s (which is \\s in js sometimes because js is retarded, as much as i love it). Reads cleaner, and way, way easier to debug.

[johnhaley81]

I added a test inside of test/tests/lib/tokenParsing.js that this pull request needs to pass before it can be merged.

[maxkorp]

Tweaked it to look for literals first, which are presumed to be able to contain anything else. Nothing will break a literal value, not spaces or hard delimiters or anything else. Works for both double and single quoted strings

[johnhaley81]

Some more test cases were added that need to be passed. The example case is a property that's inside of a string.

"{{ hello }} world" will always ignore the hello property

[maxkorp]

This needs a bigger refactor of the grammar and the tokenizer to deal with some of the "quirkier" scenarios (and some expected behaviors need to be defined). Creating an issue for this instead.