Multi-Dispatch in C#

Software Development‎ > ‎

As of C# version 4 you can do:

interface IA { }

interface IB { }

class CA1 : IA {}

class CA2 : IA {}

class CA11 : CA1 {}

class CB1 : IB {}

class CB2 : IB {}

class MD

{

        public enum X { X } ;
        public static void Foo(IA a, IB b, X dispatch = X.X) { Foo((dynamic)a, (dynamic)b); }
        static void Foo(IA a, IB b) { Console.WriteLine("IA IB"); }

static void Foo(CA1 a, CB1 b) { Console.WriteLine("CA1 CB1"); }

static void Foo(CA2 a, CB1 b) { Console.WriteLine("CA2 CB1"); }

static void Foo(CA1 a, CB2 b) { Console.WriteLine("CA1 CB2"); }

static void Foo(CA2 a, CB2 b) { Console.WriteLine("CA2 CB2"); }

static void Foo(CA11 a, CB2 b) { Console.WriteLine("CA11 CB2"); }

}

class Program

{

static void Main(string[] args)

{

var a1 = new CA1();

var a11 = new CA11();

var a2 = new CA2();

var b1 = new CB1();

var b2 = new CB2();

MD.Foo(a1, b1);

MD.Foo(a2, b1);

MD.Foo(a1, b2);

MD.Foo(a2, b2);

MD.Foo(a11, b1);

MD.Foo(a11, b2);

}

Yes, it uses the evil "dynamic" keyword. The usage is safely hidden behind type safe methods so it's really not that evil.

The enum X is used to differentiate the catch all IA, IB method from the main mutli-dispatch handler. The catch all method is used when no combination of overloads matches. Without this the code would infinitely loop if past in a concrete type for which no overload method existed.