Reimplement the parser

[IovoslavIovchev]

In its current implementation, the parser works recursively, which is fine if you are doing simple expression parsing. But once you start parsing a long series of tokens, a stack overflow may occur (such is the case in #162).

What I propose is to start reimplementing the parser in small logical chunks. This includes:

• eliminating recursive calls and making the parsing iterative
• trying to keep the expressions as linear as possible (e.g. binary operators being represented in reverse polish notation, instead of a tree)

---

As I said, the work can be split into small chunks, such as:

• var
• let
• const
• return
• functions
• blocks
• new
• if, else if, else & ternary
• loops
• objects
• arrays
• switch
• typeof
• binary operators
• unary operators
• regular expressions

Note: this list might not be exhaustive

[jasonwilliams]

Good to match up with https://tc39.es/ecma262/#sec-ecmascript-language-lexical-grammar

Good to support concurrent tokens coming through channel.

good to be broken out into functions for specific expressions.

[jasonwilliams]

@IovoslavIovchev whats the state of this issue now that the parser has been rebuilt?
There is conversation on how we break the parser up happening jasonwilliams#281 (comment)