In this section we will see how we can use the operators of the C# language on instances of our own classes, or on values of our own structs.

21.1 Why operator overloading? 21.3 An example of overloaded operators: Interval 21.2 Overloadable operators in C# 21.4 Some details of operator overloading

 

21.1.  Why operator overloading?
Contents   Up Previous Next   Slide Annotated slide Aggregated slides    Subject index Program index Exercise index 

In this section we will describe how to program operations that can be called in expressions that make use of the conventional operators (such as +, &, >>, and !) of C#. Thus, in a client of a class C, we will provide for notation such as aC1 + aC2 * aC3 instead of aC1.Plus(aC2.Mult(aC3)) or (with use of class methods) C.Plus(aC1, C.Mult(aC2,aC3)).

When operator notation is natural for objects or values of type C, clients of C can often be programmed with a more dense and readable notation. The example in Program 21.1 (only on web) provides additional motivation. In Program 21.1 (only on web) we use the class MyInt, which is shown in Program 21.2 (only on web). MyInt just wraps the operators +, -, *, /, and % in static methods.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38

using System;

public class OperatorsOrNot {

 public static int With (int a, int b, int c, int m){
   return (a % m + b % m + c % m) / 3;
 }

 public static int Without (int a, int b, int c, int m){
   return 
     MyInt.Quotient(
       MyInt.Plus(
         MyInt.Plus(MyInt.Remainder(a,m), MyInt.Remainder(b,m)),
         MyInt.Remainder(c,m)),
       3);
 }

// Without class qualification:
//   Quotient(
//     Plus(
//       Plus(Remainder(a,m), Remainder(b,m)),
//       Remainder(c,m)),
//     3);


// In some languages, such as Lisp.
// Alternative in languages with more liberal rules for identifiers:
//   (/ (+ (% a m) (% b m) (% c m)) 3) 


 public static void Main(){

   Console.WriteLine(With(18,19,25, 7));
   Console.WriteLine(Without(18,19,25, 7));
  
 }

}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

using System;

public class MyInt{
  
  public static int Plus(int a, int b){
    return a + b;
  }

  public static int Minus(int a, int b){
    return a - b;
  }

  public static int Multiply(int a, int b){
    return a * b;
  }

  public static int Quotient(int a, int b){
    return a / b;
  }

  public static int Remainder(int a, int b){
    return a % b;
  }

}

21.2.  Overloadable operators in C#
Contents   Up Previous Next   Slide Annotated slide Aggregated slides    Subject index Program index Exercise index 

We have already once studied the operator table of C#, see Section 6.7. In Table 21.1 below we show a version of operator table of C# (with operator priority and associativity) in which we have emphasized all the operators that can be overloaded in C#.

All the gray operators in Table 21.1 cannot be overloaded directly. Many of them can, however, be overloaded indirectly, or defined by other means. We will now discuss how this can be done.

A notation similar to a[x] (array indexing) can be obtained by use of indexers, see Chapter 19.

The conditional (short circuited) operators && and || can be overloaded indirectly by overloading the operators & and |. In addition, the operators called true and false must also be provided. true(x) tells if x counts as boolean true. false(x) tells if x counts as boolean false. (Notice that x belongs to the type - class or struct - in which we have defined the operators). The operators && and || are defined by the following equivalences:

• x && y    is equivalent to    false(x) ? x : (x & y)
• x || y    is equivalent to    true(x) ? x : (x | y)

Thus, when x && y is encountered we first evaluate the expression false(x). If the value is true, x is returned. If it is false, y is also evaluated, and the value of x && y becomes x & y. A similar explanation applies for x || y.

You can define the unary true and false operators in your own classes, and hereby control if the object is considered to be true or false in some boolean contexts. If you define one of them, you will also have to define the other. Recall that an expression of the form    a ? b : c    uses the conditional operator ?: with the meaning    if a then b else c.

All the assignment operators, apart from the basic assignment operator =, are implicitly overloadable. As an example, the assignment operator *= is implicitly overloaded when we explicitly overload the multiplication operator *.

The type cast operator (T)x can in reality also be overloaded. In a given class C you can define explicit and/or implicit conversion operators that converts to and from C. We will see an example of an explicit type conversion in Program 21.3.

21.3.  An example of overloaded operators: Interval
Contents   Up Previous Next   Slide Annotated slide Aggregated slides    Subject index Program index Exercise index 

We will now study the type Interval. This type allows us to represent and operate on intervals of integers. The Interval type makes a good case for illustration of overloaded operators. We program all interval operations in a functional style. We want intervals to be non-mutable, and the type is therefore programmed as a struct.

An interval is characterized by two integer end points from and to. The interval [from - to] denotes the interval that starts in from and goes to to. Notice that from is not necessarily less than to. The following are concrete examples: [1 - 5] represents the sequence 1, 2, 3, 4, and 5. [5 - 1] represents the sequence 5, 4, 3, 2, and 1. These two sequences are different.

In Program 21.3 we see the struct Interval. The private instance variables from and to represent the interval in a simple and straightforward way, and the constructor is also simple. Just after the constructor there are two properties, From and To, that access the end points of the interval.

After the two properties we have highlighted a number of overloaded operators. These are our main interest in this section. Notice the syntax for definition of the operators. There are two definitions of the + operator. One of the form anInterval + i and one of the form i + anInterval. Both have the same meaning, namely addition of i to both end-points. Thus [1 - 5] + 3 and 3 + [1 - 5] are both equal to the interval [4 - 8].

In similar ways we define multiplication of intervals and integers. We also define subtraction of an integer from an interval (but not the other way around). The shift operators << and >> provide nice notations for moving one of the end-points of an interval. Thus, [1 - 5] >> 3 is equal to the interval [1 - 8].

Finally, the unary prefix operator ! reverses an interval (internally, by making an interval with swapped end-points). Thus, ![1 - 5] is equal to the interval [5 - 1].

The private class IntervalEnumerator (shown only in the web version) and the method GetEnumerator make it possible to traverse an interval in a convenient way with use of foreach. Interval traversal is what makes intervals useful. This is illustrated in Program 21.4. We will, in great details, discuss IntervalEnumerator later in this material, see Section 31.6 - in particular Program 31.9.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95

using System;
using System.Collections;

public struct Interval{

  private readonly int from, to;  

  public Interval(int from, int to){   
    this.from = from;
    this.to = to;
  }

  public int From{   
    get {return from;}
  }

  public int To{
    get {return to;}
  }

  public int Length{  
    get {return Math.Abs(to - from) + 1;}                
  }

  public static Interval operator +(Interval i, int j){  
    return new Interval(i.From + j, i.To + j);
  }

  public static Interval operator +(int j, Interval i){  
    return new Interval(i.From + j, i.To + j);
  }

  public static Interval operator >>(Interval i, int j){ 
    return new Interval(i.From, i.To + j);
  }

  public static Interval operator <<(Interval i, int j){ 
    return new Interval(i.From + j, i.To);
  }

  public static Interval operator *(Interval i, int j){  
    return new Interval(i.From * j, i.To * j);
  }

  public static Interval operator *(int j, Interval i){  
    return new Interval(i.From * j, i.To * j);
  }

  public static Interval operator -(Interval i, int j){  
    return new Interval(i.From - j, i.To - j);
  }

  public static Interval operator !(Interval i){   
    return new Interval(i.To, i.From);
  }    

  public static explicit operator int[] (Interval i){  
    int[] res = new int[i.Length];                     
    for (int j = 0; j < i.Length; j++) res[j] = i[j];
    return res; 
  }

  private class IntervalEnumerator: IEnumerator{    
                                                    
    private readonly Interval interval;             
    private int idx;

    public IntervalEnumerator (Interval i){
      this.interval = i;
      idx = -1;   // position enumerator outside range
    }
 
    public Object Current{ 
         get {return (interval.From < interval.To) ? 
                       interval.From + idx :
                       interval.From - idx;}
    }

    public bool MoveNext (){
      if ( idx < Math.Abs(interval.To - interval.From))
         {idx++; return true;}
      else
         {return false;}
    }

    public void Reset(){
      idx = -1;         
    }
  }    
    
  public IEnumerator GetEnumerator (){
    return new IntervalEnumerator(this);
  }

}

Take a look at Program 21.4 in which we use intervals. Based on the constructed intervals iv1 and iv2 we write expressions that involve intervals. These are all highlighted in Program 21.4. Let me explain the expression !(3 + !iv2 * 2). When we evaluate this expression we adhere to normal precedence rules and normal association rules of the operators. We cannot change these rules. Therefore, we first evaluate !iv2, which is [5 - 2]. Next we evaluate !iv2 * 2, which is [10 - 4]. To this interval we add 3. This gives the interval [13 - 7]. Finally we reverse this interval. The final value is [7 - 13].

Also emphasized in Program 21.3 we show iv3[0] and iv3[iv3.Length-1]. These expressions use interval indexers. In Exercise 5.6 it is an exercise to program this indexer.

Emphasized with blue in Program 21.3 and Program 21.4 we show how to program and use an explicit type cast from Interval to int[].

You should follow the evaluations of all highlighted expressions in Program 21.4 and compare your results with the program output in Listing 21.5.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

using System;

public class app {

  public static void Main(){

    Interval iv1 = new Interval(17,14),
             iv2 = new Interval(2,5),
             iv3;

    foreach(int k in !(3 + iv1 - 2)){  
      Console.Write("{0,4}", k);       
    }
    Console.WriteLine();

    foreach(int k in !(3 + !iv2 * 2)){  
      Console.Write("{0,4}", k);
    }
    Console.WriteLine();

    iv3 = !(3 + !iv2 * 3) >> 2 ;        
    Console.WriteLine("First and last in iv3: {0}, {1}", 
                       iv3[0], iv3[iv3.Length-1]);  
                                                    
    int[] arr = (int[])iv3;   
    foreach(int j in arr){    
      Console.Write("{0,4}", j);
    }

  }

}

1
2
3
4

  15  16  17  18
   7   8   9  10  11  12  13
First and last in iv3: 9, 20
   9  10  11  12  13  14  15  16  17  18  19  20

In Program 21.6 we show yet another example of programming overloaded operators. We overload ==, !=, <, and >. This example brings us back to the playing card class which we have discussed already in Program 12.7 of Section 12.6 and Program 14.3 of Section 14.3.

Emphasized with colors in Program 21.6 we show operators that compare two cards. Notice, as above, that the operator definitions always are static. Also notice that if we define == we also have to define != . In addition, we must make sure Equals is redefined to be in accordance with the == operator. If we redefine Equals we must also redefine GetHashCode. All together, a lot of work! Similarly, if we define <= we have also have to define >= .

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83

using System;

public enum CardSuite { Spades, Hearts, Clubs, Diamonds };
public enum CardValue { Ace = 1, Two = 2, Three = 3, Four = 4, Five = 5, 
                        Six = 6, Seven = 7, Eight = 8, Nine = 9, Ten = 10, 
                        Jack = 11, Queen = 12, King = 13};


public class Card{

   private CardSuite suite;
   private CardValue value;

   public Card(CardSuite suite, CardValue value){
     this.suite = suite;
     this.value = value;
   }

   public Card(CardSuite suite, int value){
     this.suite = suite;
     this.value = (CardValue)value;
   }

   public Card(int suite, int value){
     this.suite = (CardSuite)suite;
     this.value = (CardValue)value;
   }

   public CardSuite Suite{
     get { return this.suite; }
   }

   public CardValue Value{
     get { return this.value; }
   }

   public System.Drawing.Color Color{
     get{
       System.Drawing.Color result;
       if (suite == CardSuite.Spades || suite == CardSuite.Clubs)
          result = System.Drawing.Color.Black;
       else
          result = System.Drawing.Color.Red;
          return result;
       }
   }

   public override String ToString(){
     return String.Format("Suite:{0}, Value:{1}, Color:{2}", 
                           suite, value, Color.ToString());
   }

   public override bool Equals(Object other){
      return (this.suite == ((Card)other).suite) && 
             (this.value == ((Card)other).value);
   }

   public override int GetHashCode(){
      return (int)suite ^ (int)value;
   }

   public static bool operator ==(Card c1, Card c2){
     return c1.Equals(c2);
   }

   public static bool operator !=(Card c1, Card c2){
     return !(c1.Equals(c2));
   }

   public static bool operator <(Card c1, Card c2){
     bool res;
     if (c1.suite < c2.suite)
        res = true;
     else if (c1.suite == c2.suite)
        res = (c1.value < c2.value);
     else res = false;
        return res;
   }

   public static bool operator >(Card c1, Card c2){
     return !(c1 < c2) && !(c1 == c2);
   }
}

21.4.  Some details of operator overloading
Contents   Up Previous Next   Slide Annotated slide Aggregated slides    Subject index Program index Exercise index 

Below we summarize the syntax of operator definition, which overloads a predefined operators symbol.

public static return-type operator symbol(formal-par-list){
  body-of-operator
}

There are many detailed rules that must be observed when we overload the predefined operator symbols. Some of them were mentioned in Section 21.3. Others are brought up below.

Operators must be public and static One or two formal parameters must occur, corresponding to unary and binary operators At least one of the parameters must be of the type to which the operator belongs Only value parameters apply Some operators must be defined in pairs (either none or both): ==   and   !=            <   and   >            <=   and   >= The special unary boolean operators true and false define when an object is playing the role as true or false in relation to the conditional logical operators Overloading the binary operator op causes automatic overloading of the assignment operator op=

This concludes our coverage of operator overloading. Notice that we have not discussed all details of this subject. You should consult a C# reference manual for full coverage.