I'm reading Scott Meyers' book and come across the following example:
class Rational { ... };
const Rational operator*(const Rational& lhs, const Rational& rhs);
Rational a, b, c;
...
(a * b) = c; // invoke operator= on the
// result of a*b!
He said that it was really wierd thing, but I didn't see why. What's wrong with the invokation of the operator= on the result of a*b?
The result of a*b is a temporary value, which will disappear at the end of the statement. Assigning to it would be weird, since you wouldn't be able to do anything with it after the assignment.
However, the fact that it's weird doesn't necessarily justify adding more weirdness to prevent it. In modern C++, it's a bad idea to return a const object like this since it inhibits move semantics.
operator=() is a member function, and you can call member functions on rvalues.
= applies to the built-in = operations (i.e. not operator= calls in user-defined types), not to literally any utterance of the symbol.const member function operator=() on a const object - doing so would discard the const qualifier, which isn't allowed. The const prevents the assignment, which is why Meyers suggested it (long before move semantics made it a bad idea).What is wrong or weird is that the operator= is invoked on a object which will be destructed at the end of the statement (ie at ;).
