Lispy Syntax

[Vandesm14]

Stack is great, but having things reversed around isn't great. What if you could have both the concatenation syntax and the lispy syntax??

Functions:

• Added a dedicated function type to ExprKind (deprecates Add a dedicated function type #52)
• Rewrote list syntax to create functions instead of lists with fn in the beginning
• Updated intrinsics to support functions with most list ops
• Updated tests to use the new ExprKind

S-Expressions

• Added an SExpr (s-expression) type to ExprKind
• Changed lists to use [] instead of () (parenthesis are now reserved for functions and s-expressions)
• Added parser code to parse s-expressions (similar to the aforementioned function parsing code)
• Added underscore type to ExprKind for special s-expression syntax
• Added auto-evaluation and eager calling of args
• Updated existing tests
• Added new tests
• Updated docs

[leonskidev]

Given the additions in this PR, I think it would make sense for lists [] to be more akin to plain old data, whilst still allowing functions and symbols to be stored within. This would mean that [1 2 +] would be stored verbatim on stack rather than eagerly evaluating to 3. If you need to call the list, you would cast it into either a function or s-expression and then call it instead.

It should also have the added benefit of only intrinsics and functions being callable, which should simplify they number of pathways for items to be evaluated internally, helping with consistency.