operator+() and operator+=()

Question

My C++ lecturer claims that whenever you have to implement the operator+() you must provide operator+=() as well?

I could not understand this logic, can someone explain me the logic behind this decision?

It makes sense to do both if you are going to do one, but there is no "must" for this. — crashmstr, Nov 13 '14 at 20:22
No, nothing forces you to do that it is just recommended because it is logical. — clcto, Nov 13 '14 at 20:22
It is like whenever you use a = b + c; it is possible that you might use a = a + b; that is a += b; — Venkatesh, Nov 13 '14 at 20:22
What your lecturer said is not true. It's _recommended_ to implement thte `+=` operator as well because users might assume a class that can use a `+` operator might also assume they can use a `+=` operator. — Coda17, Nov 13 '14 at 20:23
For an immutable value type, you definitely should _not_ implement `+=`. For a mutable value-holding type, you usually should implement it—and, more often than not, you should implement `+` in terms of copy and `+=`. But just flat out saying "always" is misleadingly incomplete at best. The reason for that "usually" is just that most mutable types for which `+` makes sense from a user point of view, `+=` does too. After all, you can do `i += 1` for an `int`, so why not your type? — abarnert, Nov 13 '14 at 20:24

AnT stands with Russia · Accepted Answer · 2014-11-13T20:29:47.500

Your lecturer is wrong.

The fact that you can sum two objects forming a new third object (which is what binary + does) does not necessarily imply that your objects are supposed to be modifiable "in place" (which is what += does).

In other words, if and only if your class supports the concept of being modifiable in place, then it is a fairly good programming practice to provide += whenever you provide both + and =. But in any case there's no "must" here.

operator+() and operator+=()

1 Answers1