Second attempt at closing loophole

[quephird]

To clarify, the loophole was that although the user could properly create instances of LoxList by using bracket notation for a list literal in the REPL, they could also explicitly create an instance of List. This would result in an object in a hybrid state, without an elements property and no way to modify it to behave like an actual list. In this PR, we introduced the instanceType property for LoxClass so that when instances of it are created via LoxClass.call(), we create a LoxList instance if the class' name is List, and LoxInstance instance otherwise. This is what closes the loophole. (It should be noted that the name gets set to List when the interpreter is first instantiated and processes the class declaration for List in the standard library.)

We also included the following changes:

• Introduced makeList() in Interpreter to get rid of code duplication but also made it public so that it could be used in unit tests
• Moved equality conformance to LoxValue, which is used only for unit testing, for clarity