Function overloads

[jchitel]

Overloads

New syntactic features:

• None

New semantic features:

• Functions can have the same name, so long as the return types are the same
• Functions will now be stored with a unique id, and a lookup table will point names to lists of ids
• Anytime a function is used, its name is used to look up the list of overloads, and the arguments are used to determine the correct id, if applicable
• The most specific argument types are attempted first, in argument order
• If no assignable function with the right number of arguments is found, it's an error
• Function references are stored by id for the next step
• Exported functions will store the list of exported ids
• Imported functions will import that list

Transformation:

• The typechecking step stores the correct function id in the module for lookup
• Nothing else in the process needs to change

[jchitel]

Tasks:

• Refactor Module registry tables to use ids instead of names. Names will be in a separate table that correspond to a list of ids.
• Refactor function name clash type checking to check for signature instead of name.
• Refactor identifier type resolution to return the list of possible ids, and function application type checking should narrow that list down to one function.