I found an example where in the a class definition, there were 2 member functions defined, but with the same name and same number of arguments:
const MyClass& operator [] (int index) const;
MyClass& operator [] (int index);
My question is how will the compiler know which operator definition it needs to use?
They are different methods, primarily due to the const suffix.
The return value (alone) cannot be used to resolve overloaded methods or operators.
Edit 1:
You understanding is not correct.
The first function returns a reference, which cannot be modified, to an object. It does not return a copy. The method is not allowed to modify class members.
The second function returns a reference to an object, the object can be modified. The method is allowed to modify class members.
If you attempt to use the operator in a constant function, the constant operator will be called, otherwise, the non-constant one will be called. This is a subtlety that often trips people up as if you only define the non-constant operator but attempt to use it in a constant function it will result in errors.
When two methods differ only in the declaration of const, the const version will be selected if it is called from a pointer or reference that is also declared const. Otherwise you will get the non-const version.
This is most commonly seen in the implementation of vector. If you have a const vector, you don't want to be able to change any of the elements, so a const reference to an element is returned.
