.NET中委托的目的是什么？

Question

So I'm a little bit confused about delegates in C#.... what do they do and how are they useful? I've read a few tutorials, and I don't really get exactly what they're supposed to do (everyone relates them to function pointers in C, and I've never programmed in C).

So... what do delegates do? What's a scenario in which I should use them? How would I then use them?

Answer 1

The other answers are good, but here's another way to think about delegates that might help. Imagine that a delegate is nothing more than an interface. When you see:

delegate void Action();

think:

interface IAction
{
    void Invoke();
}

And when you see:

Action myAction = foo.Bar;

think:

class FooBarAction : IAction
{
    public Foo Receiver { get; private set; }
    public FooBarAction(Foo foo)
    {
        this.Receiver = foo;
    }
    public void Invoke()
    {
        this.Receiver.Bar();
    }
}
...
IAction myAction = new FooBarAction(foo);

And when you see

myAction();

think

myAction.Invoke();

The actual details of what types get constructed are a bit different, but fundamentally that's what's happening. A delegate is simply an object with a method called Invoke, and when you call that method, it calls some other method on some other object on your behalf . That's why it's called a "delegate" -- because it delegates the call to another method of another object.

Answer 2

Delegates are sort of like objects that represent a method call. One useful way they can be used are as callbacks. For example, imagine you have a method that does something asynchronous, and you want the caller to be able to specify what they want to happen once it completes ( Action is a type of delegate):

public void DoSomething(Action whatToDoWhenDone)
{
    // Your code

    // See how the delegate is called like a method
    whatToDoWhenDone();
}

A user of DoSomething can now specify the callback as a parameter:

public void AnotherMethod()
{
    DoSomething(ShowSuccess); // ShowSuccess will be called when done
}

public void ShowSuccess()
{
    Console.WriteLine("Success!");
}

You can also use lamba expressions as a shorter way of writing your delegate:

public void AnotherMethod()
{
    DoSomething(() => Console.WriteLine("Success!"));
    // Also DoSomething(delegate() { Console.WriteLine("Success!"); });
}

Callbacks are far from the only use cases for delegates. Hopefully this shows you some of their power: the ability to have code to be executed as a variable.

Answer 3

Delegates allow you to treat functions as if they were any other variable. A delegate type defines the signature of the function, that is, what the function returns, and the number and type of arguments that it takes:

// This is the delegate for a function that takes a string and returns a string.
// It can also be written using the framework-provided Generic delegate Func, as
// Func<String, String>
delegate String StringToStringDelegate(String input);

You can define a variable of this type, and assign it to an existing method. I use the generic as an example, because that is the more common usage in .net since 2.0:

String Reverse(String input) {
    return input.Reverse();
}

Func<String, String> someStringMethod = new Func<String, String>(Reverse);

// Prints "cba":
Console.WriteLine(someStringMethod("abc"));

You can also pass functions around this way:

String Reverse(String input) {
    return input.Reverse();
}

String UpperCase(String input) {
    return input.ToUpper();
}

String DoSomethingToABC(Func<String, String> inputFunction) {
    return inputFunction("abc");
}

var someStringMethod = new Func<String, String>(Reverse);

// Prints "cba":
Console.WriteLine(DoSomethingToABC(someStringMethod));

var someOtherStringMethod = new Func<String, String>(UpperCase);

// Prints "ABC":
Console.WriteLine(DoSomethingToABC(someOtherStringMethod));

Answer 4

In a big application it is often required to other parts of the application based on some condition or something else. The delegate specifies the address of the method to be called. In simple manner a normal event handler implements the delegates in the inner layers.

Answer 5

The oversimplified answer is that a delegate is basically a "pointer" to a block of code, and the benefit is that you can pass this block of code into other functions by assigning your block of code to a variable.

The reason people relate Delegates to C function pointers is because this is in essence what delegation is all about, Ie: Pointers to methods.

As an example:

public void DoSomething(Action yourCodeBlock)
{
    yourCodeBlock();
}

public void CallingMethod()
{
    this.DoSomething(
    {
        ... statements
    });

    this.DoSomething(
    {
        ... other statements
    });
}

There are naturally lots of ways to invoke delegates as all of the tutorials will show you. The point is though that it allows you to "delegate" functionality in such a way that you can call into methods without necessarily knowing how they work, but simply trusting that they will be taken care of. In other words, I might create a class that implements a "DoSomething()" function, but I can leave it up to someone else to decide what DoSomething() will do later on.

I hope that helps. :-)

Answer 6

Delegates are a way to call back into your code when a long running operation completes or when an event occurs. For example, you pass a delegate to a method that asynchronously downloads a file in the background. When the download is complete, your delegate method would be invoked and it could then take some action such as processing the file's contents.

An event handler is a special type of delegate. For example, an event handler delegate can respond to an event like a mouse click or key press. Events are by far the most common type of delegate. In fact, you will typically see the event keyword used far more often in C# code than the delegate keyword.

Answer 7

You can think of it as a type in which you may store references to functions. That way you can in effect, store a function in a variable so you may call it later like any other function.

eg,

 public delegate void AnEmptyVoidFunction();

This creates a delegate type called AnEmptyVoidFunction and it may be used to store references to functions that return void and has no arguments.

You could then store a reference to a function with that signature.

public static void SomeMethod() { }
public static int ADifferentMethod(int someArg) { return someArg; }

AnEmptyVoidFunction func1 = new AnEmptyVoidFunction(SomeMethod);
// or leave out the constructor call to let the compiler figure it out
AnEmptyVoidFunction func2 = SomeMethod;
// note that the above only works if it is a function defined
// within a class, it doesn't work with other delegates

//AnEmptyVoidFunction func3 = new AnEmptyVoidFunction(ADifferentMethod);
                              // error wrong function type

Not only can it store declared functions but also anonymous functions (ie, lambdas or anonymous delegates)

// storing a lambda function (C#3 and up)
AnEmptyVoidFunction func4 = () => { };

// storing an anonymous delegate (C#2)
AnEmptyVoidFunction func5 = delegate() { };

To call these delegates, you can just invoke them like any other function call. Though since it is a variable, you may want to check if it is null beforehand.

AnEmptyVoidFunction func1 = () =>
{
    Console.WriteLine("Hello World");
};
func1(); // "Hello World"

AnEmptyVoidFunction func2 = null;
func2(); // NullReferenceException

public static void CallIt(AnEmptyDelegate func)
{
    // check first if it is not null
    if (func != null)
    {
        func();
    }
}

You would use them any time you needed to pass around a method that you wish to invoke. Almost in the same way that you may pass instances of objects so you may do what you wish with them. The typical use case for delegates is when declaring events. I have written another answer describing the pattern so you can look at that for more information on how to write those.