You can overload the 2 unary operators:

Overloading is the same for both types (++ and --). Scroll down for explanation

Overloading outside of class/struct:

//Prefix operator ++foo
T& operator++(T& lhs)
{
    //Perform addition
    return lhs;
}

//Postfix operator foo++ (int argument is used to separate pre- and postfix) 
//Should be implemented in terms of ++foo (prefix operator)
T operator++(T& lhs, int)
{
    T t(lhs);
    ++lhs;
    return t;
}

Overloading inside of class/struct:

//Prefix operator ++foo
T& operator++()
{
    //Perform addition
    return *this;
}

//Postfix operator foo++ (int argument is used to separate pre- and postfix) 
//Should be implemented in terms of ++foo (prefix operator)
T operator++(int)
{
    T t(*this);
    ++(*this);
    return t;
}

Note: The prefix operator returns a reference to itself, so that you can continue operations on it. The first argument is a reference, as the prefix operator changes the object, that’s also the reason why it isn’t const (you wouldn’t be able to modify it otherwise).


The postfix operator returns by value a temporary (the previous value), and so it cannot be a reference, as it would be a reference to a temporary, which would be garbage value at the end of the function, because the temporary variable goes out of scope). It also cannot be const, because you should be able to modify it directly.

The first argument is a non-const reference to the “calling” object, because if it were const, you wouldn’t be able to modify it, and if it weren’t a reference, you wouldn’t change the original value.

It is because of the copying needed in postfix operator overloads that it’s better to make it a habit to use prefix ++ instead of postfix ++ in for loops. From the for loop perspective, they’re usually functionally equivalent, but there might be a slight performance advantage to using prefix ++, especially with “fat” classes with a lot of members to copy. Example of using prefix ++ in a for loop:

for (list<string>::const_iterator it = tokens.begin();
     it != tokens.end();
     ++it) { // Don't use it++
    ...
}