Overload operators to work with class objects?

问题

How can I modify the following code in such way I don't need to repeat f2=11; f3=12; in the main function. The code is for overloading the most common operators.

class FLOAT{
    private:
        float x;
    public:
        FLOAT(){    x=0.0;  }
        void setFloat(float f)      {   x=f;    }
        float getFloat()            {   return x;};
        FLOAT operator+(FLOAT obj)  {x=x+obj.x; return *this;};
        FLOAT operator-(FLOAT obj)  {x=x-obj.x; return *this;};
        FLOAT operator*(FLOAT obj)  {x=x*obj.x; return *this;};
        FLOAT operator/(FLOAT obj)  {x=x/obj.x; return *this;};
        FLOAT& operator=(const FLOAT& obj)  {this->x=obj.x; return *this;   };
        FLOAT& operator=(const float& y)    {this->x=y; return *this;   };
};

int main() {
    FLOAT f,f2,f3;
    f2=11;
    f3=12;

    f=f3-f2;
    cout<<"f3-f2 ="<<f.getFloat()<<endl;


    f2=11;
    f3=12;
    f=f3+f2;
    cout<<"f3+f2 ="<<f.getFloat()<<endl;

    f2=11;
    f3=12;
    f=f3*f2;
    cout<<"f3*f2 ="<<f.getFloat()<<endl;

    f2=11;
    f3=12;
    f=f3/f2;
    cout<<"f3/f2 ="<<f.getFloat()<<endl;

    system("pause"); // to pause console screen
    return 0;
}

回答1:

@Oli's answer pretty much says you what minimal thing you need to do in order to make your code work. However, I see (and I know even @Oli sees) that your implementation of the class has many flaws.

Since you've implemented FLOAT, I'm explaining you the implementation of Double (the implementation of FLOAT would be similar).

class Double {
  double data;
public:
  Double (double p=0.0) : data(p){}
  double value() { return data; }
  Double & operator+=(Double const & other) 
  {
         data += other.data; 
         return *this;
  }
  Double & operator-=(Double const & other) 
  {
         data -= other.data; 
         return *this;
  }
  //...
};

Note that you don't need to implement operator=(Double const&) and Double(Double const&). The compiler generated ones would be enough. Since the constructor takes one argument of type double, you don't need to implement operator=(double const &) also. The compiler generated copy-semantics, along with the constructor, would take care of that.

Now see this,

//implement operator+ and operator- as non-member functions
Double operator+(Double a,  Double const & b)
{
    a += b; //a is local copy, so we can change it
    return a;
}
Double operator-(Double a,  Double const & b)
{
    a -= b; //a is local copy, so we can change it
    return a;
}

Note that I've implemented operator+ and operator- in terms of operator+= and operator-= respectively.

Similarly, you can implement operator/= and operator*= as member functions, and then implement operator/ and operator* in terms of the them!

回答2:

Your operators should create a new instance; they shouldn't be modifying themselves (in fact, they should be declared const to prevent this).

e.g.:

FLOAT operator+(FLOAT obj) const
{
    FLOAT tmp;
    tmp.setFloat(x + obj.x);
    return tmp;
}

Note there are much more idiomatic ways of defining operator overloads (e.g. defining operator+ in terms of operator+=, and defining a constructor that takes a float). But the above should suffice.

来源：https://stackoverflow.com/questions/8464172/overload-operators-to-work-with-class-objects

标签

c++

class

overloading