bug(semantic): Incorrect ScopeId for IIFEs

[DonIsaac]

See this test case for a reproduction

function foo()

In this following scope ids are equal:

• the root scope id
• AstNode::scope_id(), when called on foo's declaration node
• SymbolTable::get_scope_id(), when called with foo's SymbolId

Now, consider this snippet:

(function foo() {})()

Here, the following scope ids are equal

• the root scope id
• AstNode::scope_id(), when called on foo's declaration node
  However, the symbol id obtained from SymbolTable::get_scope_id() is not the same as the other two.

I'm not certain what expected behavior here is, whether foo should be considered in the root scope or not, but the scope id for its AstNode and for the Symbol should definitely be the same.

[Dunqing]

This is expected behavior. The main difference here is between FunctionDeclaration and FunctionExpression. I explained in #4240 (comment)

[DonIsaac]

Oh boy that's confusing haha

Any chance we could write that down somewhere?

[Dunqing]

Oh boy that's confusing haha

Any chance we could write that down somewhere?

Maybe we should write down what the ScopeFlags are for each of the different SymbolFlags?

[overlookmotel]

As mentioned in #4240 (comment), I feel we should have separate types FunctionDeclaration and FunctionExpression. I think this would make it less confusing.