Main Question
event System.Delegate d;
Code above will raise error:
Compiler Error CS0066
'event': event must be of a delegate type
Why System.Delegate can not be used as event?
If I want an event to accept any kind of delegate, what should I use?
Afterwards
(Added one week later. Below is not very related to the question. It is more about my own problem.)
Thanks to #John.
Thanks about his answer. And thanks about asking me "Why would you want such a thing?".
I said "If I need an event, but I'm not sure what kind of method I may use in the event. Then I would want to have an event for any kind."
The situation in my mind is something like below:
There is a complicated class like Enviroment, changing from different part of it may raise different kinds of event.
class Environment
{
float temperature;
float humidity;
float factor3;
float factor4;
float factor5;
//more factors may be added later.
}
/*
Need an event for Environment.
The event should invoke:
TemperatureInfluence.OnTemperatureChange(...)
HumidityInfluence.OnHumidityChange(...)
Factor3And4Influence.OnFactor3And4Change(...)
Factor4And5Influence.OnFactor4And5Change(...)
and it should be albe to handle mehtod of new kinds.
*/
class TemperatureInfluence
{
public void OnTemperatureChange(float temperature)
{
Console.WriteLine("Temperature changed, put on more or less clothes");
}
}
class HumidityInfluence
{
public void OnHumidityChange(float humidity)
{
Console.WriteLine("Humidity changed, drink more or less water");
}
}
class Factor3And4Influence
{
public void OnFactor3And4Change(float factor3, float factor4)
{
Console.WriteLine("Factor3 and Factor4 changed, do blablabla");
}
}
class Factor4And5Influence
{
public void OnFactor4And5Change(float factor4, float factor5)
{
Console.WriteLine("Factor4 and Factor5 changed, do hohohuhu");
}
}
I think my problem is that I didn't understand what event keyword is in C#.
The event keyword adds some features to delegate and that's it.(So it should be the same as delegate, with precise paramters.)
It is different from real word event, which can be of any kind.
If I really need an event of any kind, I should not count on event keyword. I think I may need to make a structure for this or just write new code for new event.
System.Delegate is a class, not a delegate.
An event is a member of a type. All members must have their own data type specified. You can use Object for fields and the like if you want to be able to refer to any type, or you can use a base type to refer to any derived type. Delegate types don't inherit each other so there's no base delegate type that you can specify to refer to derived delegate types.
It doesn't make sense anyway. Why would you want to accept delegates to methods with parameters that you have no way to pass arguments for? Just declare your event as type EventHandler and then you can pass the usual arguments, i.e. object sender, EventArgs e but methods with no parameters can still be used as handlers. Any other parameters make no sense as you can't provide arguments anyway.
I read a lot of materials but I still did not understand, can someone explain me, please?
What do I know:
Declare an event:
public event MyEvent myEvent;
Vs.
declare EventHandler:
public EventHandler MyEventHandler;
While EventHandler is a keyword in .NET:
public delegate void EventHandler (Send object, EventArgs e);
So, EventHandler is a delegate and not an event, since it is not used in the keyword 'event'?
And when should I use 'Event' and when 'EventHandler'?
Ah, the event vs. delegate question. I remember having that question as well...
So, suppose you're making a class called "Cat" and you want to let people know when your cat is hungry. You could do that in one of two ways: by exposing a delegate on the Cat class or by exposing an event.
You can think of a delegate as a pointer to a function (or method). So let's say there's a class called "Person" that has a method called FeedCat(Cat cat).
The delegate way
If your cat is exposing a delegate called HungryDelegate, the person can point the delegate to their FeedCat method, so that when the cat is hungry, it has a way to call the FeedCat method on the person.
The problem here is, only one person can feed the cat. Suppose you want multiple people to be able to feed the cat. People can be busy doing other things, so it's important for the cat to be able to tell multiple people that it's hungry. That way, people get notified of this, and they can check up on the cat when they have a chance, see if someone already fed the cat, and feed it if not.
Events to the rescue:
An event is basically a list of delegates (of a certain type). If you expose a Hungry event on your "Cat" class, multiple people can ADD pointers (delegates) to their FeedCat methods.
They can also remove the pointers (delegates) to their FeedCat methods if they like. Let's say a person moved out, and their sister is now taking care of the cat. The person can remove their delegate for their own FeedCat function from the cat, so that they no longer get notified that the darn cat is hungry.
Events vs. multicast delegates?
Technically, in order to be able to provide multiple delegates, you could use the so called MultiCastDelegates instead of events. They are composite delegates (a linked list or delegates). The problem there is, everyone can mess with them from the outside. An evil person could remove everyone else's "FeedCat" delegates, and the poor cat would starve (or would have to learn to hunt).
The important thing when using events is that the person cannot see other people's delegates that are added to the event, and it can't remove them or interact with them (in principle).
Exposing a (public) field
public EventHandler MyEventHandler;
is a bad practice: one can easily ruin the code with a small typo:
MyClass demo = new MyClass();
demo.MyEventHandler += MyMethod;
...
// Can you see the error? = instead of correct += ?
// MyMethod will not be called since this assignment
demo.MyEventHandler = MyReaction;
That's why you should use event which has been specially desinged for this
public event EventHandler MyEventHandler;
if we try the previous demo we'll get compile time error:
MyClass demo = new MyClass();
...
demo.MyEventHandler = MyReaction; // <- doesn't compile, insist on +=
In very rare cases you may want explicit delegate-based fields, but these field should be concealed, say, be private, not public:
// We don't expose the field
private EventHandler m_MyEventHandler;
// ... But event:
public event EventHandler MyEventHandler {
add {
//TODO: extra logic on += operation
m_MyEventHandler += value;
}
remove {
//TODO: extra logic on -= operation
m_MyEventHandler -= value;
}
}
DELEGATES, EVENTS(EVENT HANDLERS/EVENT LISTENERS), CONCEPTS(MULTICASTING/BROADCASTING), ACTION & FUNC
This will be a long one but its the simplest explanation, the problem this is such a nuisance of a topic is because people are just using different words to explain the same thing
First of all, you should know a few things
DELEGATES: It's nothing but a list of methods, why create a list? because when your code is being executed, that list is taken and every method there is executed one by one, just don't listen to textbook definitions take this and you will be all right
also called :
a pointer to a function
a wrapper for a method that can send and receive methods just like a variable
to create a delegate you go
[[access modifier] delegate [return type] [delegate name]([parameters])]
example: public delegate int demo(int a);
now to execute all these methods stored in a list called delegate, you go
1. demo.invoke(a);
2. demo(a); ..... both are valid
using the dot and explicitly saying invoke shines in async programming where you use beginInvoke, but that is out of the scope of this topic
there is one more thing called "Creating an object of the delegate/instantiate Delegate" which is pretty much as it sounds but just to avoid confusion it goes like (for the above example )
example : demo del = new demo(); (or) Public demo del = null;
to add any method to the list called delegate you go += and you also need to remove it once the "requirements of the methods are met" you go -=
(requirements of the methods are met mean you no longer need the method to be active or aka "listening") if you don't remove it, it could cause a "memory leak" meaning your computers ram will be eaten alive, technically allocated memory will not be released
example: say there is a method
public int calculate (int c)
to add this method to delegate you go
1. del = calculate;
2. del += calculate; .... all are valid
to remove
del -= calculate
first of all notice the similarities between the delegate and the method, the return type(output) and the input/parameters are the same, and that is a rule you just cannot add any random or a bunch of methods in a delegate it needs to follow the input-output rule
now why are there 2 different ways to do one thing, the only thing different is the assignment operators (+, =), this introduces a new topic called
EVENTS
which is nothing but a constrained version of a Delegate, It's still a List of methods don't confuse when people explain these terminologies, they change the name, so stick with this to understand
what is the constraint? you cannot do this del = calculate;
what's the harm in it, say a bunch of methods are added to the Delegate(List), you do that 👆 all are wiped out and only a single method "calculate" remains, so to prevent that Events are used,
Event Syntax
Public Event demo del = null;
One more thing you cannot do with events is invoke the delegate directly like demo.invoke since its public it can be accessed and invoked but with events, it can't
now you just add the methods to the event (a special type of delegate)
when to use an event vs a delegate, depends on your situation but pragmatically events are popular
few more keywords
MULTICASTING: nothing but adding more than one method to a delegate
BROADCASTING: adding more than one method to an event
PUBLISHER: the one that executes the method (term used in broadcasting), only a single entity
SUBSCRIBER: The methods that are being executed, can be multiple
LISTENER: the same thing as a subscriber but the term is used in multicasting
EVENT HANDLER: same thing as a subscriber/event listener so what the difference? it's basically the same thing, some say an eventlistener detect for the event to occur and the event handler "handles" or execute the code, ITS THE SAME THING PRACTICALLY!
action and func are just delegates that have been created and instantiated so 2 lines of code in a word, the difference is just in return types
ACTION: does not return anything while taking 0 or more than 1 input
FUNC: returns one thing and takes in parameters
if you don't do good with reading here is the best video on this topic
https://www.youtube.com/playlist?list=PLFt_AvWsXl0dliMtpZC8Qd_ru26785Ih_
How in C# do you create a true singlecast delegate. i.e a delegate instance that can reference one (and only one) method in its invocation list and so be used as a callback (for a single subscriber) rather than an event (which may have many subscribers).
The Framework has the classes System.Delegate and System.MulticastDelegate, which gives the mistaken impression that System.Delegate is singlecast and System.Multicast delegate adds multicast capability. But the MSDN documentation for System.Delegate http://msdn.microsoft.com/en-us/library/system.delegate.aspx indicates that System.Delegate is in fact multicast...
"The invocation list of a delegate is an ordered set of delegates in which each element of the list invokes exactly one of the methods represented by the delegate."
...while the documentation for System.MulticastDelegate does not really explain what extra behaviour it provides.
The official documentation in this area is pretty confusing, but one thing which is clear is that end-users cannot derive from either System.Delegate or System.MulticastDelegate. So is there any way supported by the framework to create a true singlecast delegate that can be used as a variable to store a reference to a single callback?
#dtb. If I can use a singlecast delegate then the need for the runtime check is eliminated. Of course it's true that the application logic could still fail in other ways, such as the wrong handler being assigned, but at least if I use a singlecast delegate then the problem of there being multiple handlers where I only expect one is one problem that simply cannot exist, hence there is one less thing to check, simpler unit tests, more elegant design. Also if a delegate for a method with a return value has multiple handlers in its invocation list, then it is the value returned by the last handler in the list that is returned to the caller, not the first.
If you really need a "singlecast" delegate with an event why don't you simply implement your own add/remove methods for that event?
The "problem" of Delegate is that MulticastDelegate is a derived class so if anyone assigns a MulticastDelegate object to a Delegate variable then you'll always have a MulticastDelegate.
For example we can simplify the default event implementation to this:
private ChangedEventHandler _changed;
public event ChangedEventHandler Changed
{
add
{
_changed += value;
}
remove
{
_changed -= value;
}
}
Now let's change the event implementation to:
private ChangedEventHandler _changed;
public event ChangedEventHandler Changed
{
add
{
_changed = value; // Do NOT combine delegates
}
remove
{
_changed -= value;
}
}
Now what you have is (almost) a singlecast delegate, because of event syntax users can't assign directly a multicast delegate and only the last assigned delegate is stored. If your users are really malicious they can create a MulticastDelegate and then add that delegate to your event handler. If you really need to prevent this you may add this check to the add method:
if (value.GetInvocationList().Length > 1)
throw new ArgumentException("MulticastDelegates are not allowed here.");
It's a bit more awkward than using a delegate, but one approach is to pass a callback interface with the method you want. Since there can only be a single implementation of the method, only one callback can be registered.
public interface IDoSomething
{
void DoSomething();
}
public sealed class MyClass
{
private IDoSomething _doer;
//We use a Set method rather than a property to prevent other classes from accessing the callback
//Another common (and generally better) pattern is to pass the instance into the constructor
public void SetSomethingDoer(IDoSomething doer)
{
_doer = doer;
}
//Other code can now access _doer to call back the method
}
This has the side benefit of allowing you to group multiple callback methods together, which often makes sense in these cases where you're trying to guarantee a single callback handler.
Possible Duplicate:
What is the difference between a delegate and events?
Possible Duplicate:
Difference between events and delegates and its respective applications
(copied from this duplicate)
When i have to raise an event i do this
public delegate void LogUserActivity(Guid orderGUID);
public event LogUserActivity ActivityLog;
even this works
public delegate void LogUserActivity(Guid orderGUID);
public LogUserActivity ActivityLog;
What is the difference between two of them
There are three things here:
Declaring a delegate type
Creating a public variable of a delegate type
Creating a public event of a delegate type
The variable is just a normal variable - anyone can read from it, assign to it etc. An event only exposes subscribe/unsubscribe abilities to the outside world. A field-like event as you've shown here effectively has a "default" subscribe/unsubscribe behaviour, stored in a field with the same name. Within the declaring class, you access the field; outside you access the event.
I have an article about events and delegates which explains in more detail.
EDIT: To answer the comment, you can easily initialize a field-like event with a "no-op" handler:
public event LogUserActivity ActivityLog = delegate{};
Events and public delegates differ in one big way that keeps me from using public delegates in most cases:
Event:
obj.ActivityLog = null; // invalid
Public Delegate:
obj.ActivityLog = null; // valid
This matters, because I only want the subscriber to add/remove themselves from the list in most cases. I don't want other objects unhooking events from other subscribers.
In cases where the delegate is less of an event and more of a callback, I tend to use public methods to do this and keep the public delegate from being exposed directly:
obj.RegisterActivityCallback(...)
An event is an abstraction over a delegate, just as a property is an abstraction over a field. And - just like a property - an event allows you to gain fine control over what happens when a handler is added/removed:
public event LogUserActivity ActivityLog
{
add { ... }
remove { ... }
}
Indeed, your event may not have a delegate backing it at all, just like a property may not necessarily have a field backing it.
In the first case you can assign multiple listeners to your ActivityLog event as in:
logger.ActivityLog += new LogUserActivity(Listener1);
logger.ActivityLog += new LogUserActivity(Listener2);
Both methods (Listener1 and Listener2) will be called when the event is fired. In the second case you are not creating an event but a delegate. In this case you can only assign one listener:
logger.ActivityLog = new LogUserActivity(Listener1);
In this case you do not fire an event but instead you call the delegate. An event is really just a chain of delegates that are called in turn.
to add to all the previous answers, An event IS a delegate, of a specific type. A delegate is a type that is designed to act as a "smart" function pointer, or, putting it another way, as a smart "wrapper" for a function pointer, that allows the compiler to intelligently decide at compile time whether the function(s) you are sticking into the delegate are consistent with the uses you are making of the delegate (consistency is based on the function signature).
An event is a delegate with a specific signature, specifically, one that does not return anything (returns void), and takes two parameters, a System.object, to hold a reference to whatever object triggered the event, and an instance of some type derived from System.EventArgs, to hold whatever other data the event needs to carry with it from the initialtor to the handler.
I understand the purpose of events, especially within the context of creating user interfaces. I think this is the prototype for creating an event:
public void EventName(object sender, EventArgs e);
What do event handlers do, why are they needed, and how do I to create one?
To understand event handlers, you need to understand delegates. In C#, you can think of a delegate as a pointer (or a reference) to a method. This is useful because the pointer can be passed around as a value.
The central concept of a delegate is its signature, or shape. That is (1) the return type and (2) the input arguments. For example, if we create a delegate void MyDelegate(object sender, EventArgs e), it can only point to methods which return void, and take an object and EventArgs. Kind of like a square hole and a square peg. So we say these methods have the same signature, or shape, as the delegate.
So knowing how to create a reference to a method, let's think about the purpose of events: we want to cause some code to be executed when something happens elsewhere in the system - or "handle the event". To do this, we create specific methods for the code we want to be executed. The glue between the event and the methods to be executed are the delegates. The event must internally store a "list" of pointers to the methods to call when the event is raised.* Of course, to be able to call a method, we need to know what arguments to pass to it! We use the delegate as the "contract" between the event and all the specific methods that will be called.
So the default EventHandler (and many like it) represents a specific shape of method (again, void/object-EventArgs). When you declare an event, you are saying which shape of method (EventHandler) that event will invoke, by specifying a delegate:
//This delegate can be used to point to methods
//which return void and take a string.
public delegate void MyEventHandler(string foo);
//This event can cause any method which conforms
//to MyEventHandler to be called.
public event MyEventHandler SomethingHappened;
//Here is some code I want to be executed
//when SomethingHappened fires.
void HandleSomethingHappened(string foo)
{
//Do some stuff
}
//I am creating a delegate (pointer) to HandleSomethingHappened
//and adding it to SomethingHappened's list of "Event Handlers".
myObj.SomethingHappened += new MyEventHandler(HandleSomethingHappened);
//To raise the event within a method.
SomethingHappened("bar");
(*This is the key to events in .NET and peels away the "magic" - an event is really, under the covers, just a list of methods of the same "shape". The list is stored where the event lives. When the event is "raised", it's really just "go through this list of methods and call each one, using these values as the parameters". Assigning an event handler is just a prettier, easier way of adding your method to this list of methods to be called).
C# knows two terms, delegate and event. Let's start with the first one.
Delegate
A delegate is a reference to a method. Just like you can create a reference to an instance:
MyClass instance = myFactory.GetInstance();
You can use a delegate to create an reference to a method:
Action myMethod = myFactory.GetInstance;
Now that you have this reference to a method, you can call the method via the reference:
MyClass instance = myMethod();
But why would you? You can also just call myFactory.GetInstance() directly. In this case you can. However, there are many cases to think about where you don't want the rest of the application to have knowledge of myFactory or to call myFactory.GetInstance() directly.
An obvious one is if you want to be able to replace myFactory.GetInstance() into myOfflineFakeFactory.GetInstance() from one central place (aka factory method pattern).
Factory method pattern
So, if you have a TheOtherClass class and it needs to use the myFactory.GetInstance(), this is how the code will look like without delegates (you'll need to let TheOtherClass know about the type of your myFactory):
TheOtherClass toc;
//...
toc.SetFactory(myFactory);
class TheOtherClass
{
public void SetFactory(MyFactory factory)
{
// set here
}
}
If you'd use delegates, you don't have to expose the type of my factory:
TheOtherClass toc;
//...
Action factoryMethod = myFactory.GetInstance;
toc.SetFactoryMethod(factoryMethod);
class TheOtherClass
{
public void SetFactoryMethod(Action factoryMethod)
{
// set here
}
}
Thus, you can give a delegate to some other class to use, without exposing your type to them. The only thing you're exposing is the signature of your method (how many parameters you have and such).
"Signature of my method", where did I hear that before? O yes, interfaces!!! interfaces describe the signature of a whole class. Think of delegates as describing the signature of only one method!
Another large difference between an interface and a delegate is that when you're writing your class, you don't have to say to C# "this method implements that type of delegate". With interfaces, you do need to say "this class implements that type of an interface".
Further, a delegate reference can (with some restrictions, see below) reference multiple methods (called MulticastDelegate). This means that when you call the delegate, multiple explicitly-attached methods will be executed. An object reference can always only reference to one object.
The restrictions for a MulticastDelegate are that the (method/delegate) signature should not have any return value (void) and the keywords out and ref is not used in the signature. Obviously, you can't call two methods that return a number and expect them to return the same number. Once the signature complies, the delegate is automatically a MulticastDelegate.
Event
Events are just properties (like the get;set; properties to instance fields) which expose subscription to the delegate from other objects. These properties, however, don't support get;set;. Instead, they support add; remove;
So you can have:
Action myField;
public event Action MyProperty
{
add { myField += value; }
remove { myField -= value; }
}
Usage in UI (WinForms,WPF,UWP So on)
So, now we know that a delegate is a reference to a method and that we can have an event to let the world know that they can give us their methods to be referenced from our delegate, and we are a UI button, then: we can ask anyone who is interested in whether I was clicked, to register their method with us (via the event we exposed). We can use all those methods that were given to us and reference them by our delegate. And then, we'll wait and wait.... until a user comes and clicks on that button, then we'll have enough reason to invoke the delegate. And because the delegate references all those methods given to us, all those methods will be invoked. We don't know what those methods do, nor we know which class implements those methods. All we do care about is that someone was interested in us being clicked, and gave us a reference to a method that complied with our desired signature.
Java
Languages like Java don't have delegates. They use interfaces instead. The way they do that is to ask anyone who is interested in 'us being clicked', to implement a certain interface (with a certain method we can call), then give us the whole instance that implements the interface. We keep a list of all objects implementing this interface and can call their 'certain method we can call' whenever we get clicked.
That is actually the declaration for an event handler - a method that will get called when an event is fired. To create an event, you'd write something like this:
public class Foo
{
public event EventHandler MyEvent;
}
And then you can subscribe to the event like this:
Foo foo = new Foo();
foo.MyEvent += new EventHandler(this.OnMyEvent);
With OnMyEvent() defined like this:
private void OnMyEvent(object sender, EventArgs e)
{
MessageBox.Show("MyEvent fired!");
}
Whenever Foo fires off MyEvent, then your OnMyEvent handler will be called.
You don't always have to use an instance of EventArgs as the second parameter. If you want to include additional information, you can use a class derived from EventArgs (EventArgs is the base by convention). For example, if you look at some of the events defined on Control in WinForms, or FrameworkElement in WPF, you can see examples of events that pass additional information to the event handlers.
Here is a code example which may help:
using System;
using System.Collections.Generic;
using System.Text;
namespace Event_Example
{
// First we have to define a delegate that acts as a signature for the
// function that is ultimately called when the event is triggered.
// You will notice that the second parameter is of MyEventArgs type.
// This object will contain information about the triggered event.
public delegate void MyEventHandler(object source, MyEventArgs e);
// This is a class which describes the event to the class that receives it.
// An EventArgs class must always derive from System.EventArgs.
public class MyEventArgs : EventArgs
{
private string EventInfo;
public MyEventArgs(string Text) {
EventInfo = Text;
}
public string GetInfo() {
return EventInfo;
}
}
// This next class is the one which contains an event and triggers it
// once an action is performed. For example, lets trigger this event
// once a variable is incremented over a particular value. Notice the
// event uses the MyEventHandler delegate to create a signature
// for the called function.
public class MyClass
{
public event MyEventHandler OnMaximum;
private int i;
private int Maximum = 10;
public int MyValue
{
get { return i; }
set
{
if(value <= Maximum) {
i = value;
}
else
{
// To make sure we only trigger the event if a handler is present
// we check the event to make sure it's not null.
if(OnMaximum != null) {
OnMaximum(this, new MyEventArgs("You've entered " +
value.ToString() +
", but the maximum is " +
Maximum.ToString()));
}
}
}
}
}
class Program
{
// This is the actual method that will be assigned to the event handler
// within the above class. This is where we perform an action once the
// event has been triggered.
static void MaximumReached(object source, MyEventArgs e) {
Console.WriteLine(e.GetInfo());
}
static void Main(string[] args) {
// Now lets test the event contained in the above class.
MyClass MyObject = new MyClass();
MyObject.OnMaximum += new MyEventHandler(MaximumReached);
for(int x = 0; x <= 15; x++) {
MyObject.MyValue = x;
}
Console.ReadLine();
}
}
}
Just to add to the existing great answers here - building on the code in the accepted one, which uses a delegate void MyEventHandler(string foo)...
Because the compiler knows the delegate type of the SomethingHappened event, this:
myObj.SomethingHappened += HandleSomethingHappened;
Is totally equivalent to:
myObj.SomethingHappened += new MyEventHandler(HandleSomethingHappened);
And handlers can also be unregistered with -= like this:
// -= removes the handler from the event's list of "listeners":
myObj.SomethingHappened -= HandleSomethingHappened;
For completeness' sake, raising the event can be done like this, only in the class that owns the event:
//Firing the event is done by simply providing the arguments to the event:
var handler = SomethingHappened; // thread-local copy of the event
if (handler != null) // the event is null if there are no listeners!
{
handler("Hi there!");
}
The thread-local copy of the handler is needed to make sure the invocation is thread-safe - otherwise a thread could go and unregister the last handler for the event immediately after we checked if it was null, and we would have a "fun" NullReferenceException there.
C# 6 introduced a nice short hand for this pattern. It uses the null propagation operator.
SomethingHappened?.Invoke("Hi there!");
My understanding of the events is;
Delegate:
A variable to hold reference to method / methods to be executed. This makes it possible to pass around methods like a variable.
Steps for creating and calling the event:
The event is an instance of a delegate
Since an event is an instance of a delegate, then we have to first define the delegate.
Assign the method / methods to be executed when the event is fired (Calling the delegate)
Fire the event (Call the delegate)
Example:
using System;
namespace test{
class MyTestApp{
//The Event Handler declaration
public delegate void EventHandler();
//The Event declaration
public event EventHandler MyHandler;
//The method to call
public void Hello(){
Console.WriteLine("Hello World of events!");
}
public static void Main(){
MyTestApp TestApp = new MyTestApp();
//Assign the method to be called when the event is fired
TestApp.MyHandler = new EventHandler(TestApp.Hello);
//Firing the event
if (TestApp.MyHandler != null){
TestApp.MyHandler();
}
}
}
}
//This delegate can be used to point to methods
//which return void and take a string.
public delegate void MyDelegate(string foo);
//This event can cause any method which conforms
//to MyEventHandler to be called.
public event MyDelegate MyEvent;
//Here is some code I want to be executed
//when SomethingHappened fires.
void MyEventHandler(string foo)
{
//Do some stuff
}
//I am creating a delegate (pointer) to HandleSomethingHappened
//and adding it to SomethingHappened's list of "Event Handlers".
myObj.MyEvent += new MyDelegate (MyEventHandler);
publisher: where the events happen. Publisher should specify which delegate the class is using and generate necessary arguments, pass those arguments and itself to the delegate.
subscriber: where the response happen. Subscriber should specify methods to respond to events. These methods should take the same type of arguments as the delegate. Subscriber then add this method to publisher's delegate.
Therefore, when the event happen in publisher, delegate will receive some event arguments (data, etc), but publisher has no idea what will happen with all these data. Subscribers can create methods in their own class to respond to events in publisher's class, so that subscribers can respond to publisher's events.
I recently made an example of how to use events in c#, and posted it on my blog. I tried to make it as clear as possible, with a very simple example. In case it might help anyone, here it is: http://www.konsfik.com/using-events-in-csharp/
It includes description and source code (with lots of comments), and it mainly focuses on a proper (template - like) usage of events and event handlers.
Some key points are:
Events are like "sub - types of delegates", only more constrained (in a good way). In fact an event's declaration always includes a delegate (EventHandlers are a type of delegate).
Event Handlers are specific types of delegates (you may think of them as a template), which force the user to create events which have a specific "signature". The signature is of the format: (object sender, EventArgs eventarguments).
You may create your own sub-class of EventArgs, in order to include any type of information the event needs to convey. It is not necessary to use EventHandlers when using events. You may completely skip them and use your own kind of delegate in their place.
One key difference between using events and delegates, is that events can only be invoked from within the class that they were declared in, even though they may be declared as public. This is a very important distinction, because it allows your events to be exposed so that they are "connected" to external methods, while at the same time they are protected from "external misuse".
Another thing to know about, in some cases, you have to use the Delegates/Events when you need a low level of coupling !
If you want to use a component in several place in application, you need to make a component with low level of coupling and the specific unconcerned LOGIC must be delegated OUTSIDE of your component ! This ensures that you have a decoupled system and a cleaner code.
In SOLID principle this is the "D", (Dependency inversion principle).
Also known as "IoC", Inversion of control.
You can make "IoC" with Events, Delegates and DI (Dependency Injection).
It's easy to access a method in a child class. But more difficult to access a method in a parent class from child. You have to pass the parent reference to the child ! (or use DI with Interface)
Delegates/Events allows us to communicate from the child to the parent without reference !
In this diagram above, I do not use Delegate/Event and the parent component B has to have a reference of the parent component A to execute the unconcerned business logic in method of A. (high level of coupling)
With this approach, I would have to put all the references of all components that use component B ! :(
In this diagram above, I use Delegate/Event and the component B doesn't have to known A. (low level of coupling)
And you can use your component B anywhere in your application !
I agree with KE50 except that I view the 'event' keyword as an alias for 'ActionCollection' since the event holds a collection of actions to be performed (ie. the delegate).
using System;
namespace test{
class MyTestApp{
//The Event Handler declaration
public delegate void EventAction();
//The Event Action Collection
//Equivalent to
// public List<EventAction> EventActions=new List<EventAction>();
//
public event EventAction EventActions;
//An Action
public void Hello(){
Console.WriteLine("Hello World of events!");
}
//Another Action
public void Goodbye(){
Console.WriteLine("Goodbye Cruel World of events!");
}
public static void Main(){
MyTestApp TestApp = new MyTestApp();
//Add actions to the collection
TestApp.EventActions += TestApp.Hello;
TestApp.EventActions += TestApp.Goodbye;
//Invoke all event actions
if (TestApp.EventActions!= null){
//this peculiar syntax hides the invoke
TestApp.EventActions();
//using the 'ActionCollection' idea:
// foreach(EventAction action in TestApp.EventActions)
// action.Invoke();
}
}
}
}
Great technical answers in the post! I have nothing technically to add to that.
One of the main reasons why new features appear in languages and software in general is marketing or company politics! :-) This must not be under estimated!
I think this applies to certain extend to delegates and events too! i find them useful and add value to the C# language, but on the other hand the Java language decided not to use them! they decided that whatever you are solving with delegates you can already solve with existing features of the language i.e. interfaces e.g.
Now around 2001 Microsoft released the .NET framework and the C# language as a competitor solution to Java, so it was good to have NEW FEATURES that Java doesn't have.
DELEGATES, EVENTS(EVENT HANDLERS/EVENT LISTENERS), CONCEPTS(MULTICASTING/BROADCASTING), ACTION & FUNC
This will be a long one but its the simplest explanation, the problem this is such a nuisance of a topic is because people are just using different words to explain the same thing
First of all, you should know a few things
DELEGATES: It's nothing but a list of methods, why create a list? because when your code is being executed, that list is taken and every method there is executed one by one, just don't listen to textbook definitions take this and you will be all right
also called :
a pointer to a function
a wrapper for a method that can send and receive methods just like a variable
to create a delegate you go
[[access modifier] delegate [return type] [delegate name]([parameters])]
example: public delegate int demo(int a);
now to execute all these methods stored in a list called delegate, you go
1. demo.invoke(a);
2. demo(a); ..... both are valid
using the dot and explicitly saying invoke shines in async programming where you use beginInvoke, but that is out of the scope of this topic
there is one more thing called "Creating an object of the delegate/instantiate Delegate" which is pretty much as it sounds but just to avoid confusion it goes like (for the above example )
example : demo del = new demo(); (or) Public demo del = null;
to add any method to the list called delegate you go += and you also need to remove it once the "requirements of the methods are met" you go -=
(requirements of the methods are met mean you no longer need the method to be active or aka "listening") if you don't remove it, it could cause a "memory leak" meaning your computers ram will be eaten alive, technically allocated memory will not be released
example: say there is a method
public int calculate (int c)
to add this method to delegate you go
1. del = calculate;
2. del += calculate; .... all are valid
to remove
del -= calculate
first of all notice the similarities between the delegate and the method, the return type(output) and the input/parameters are the same, and that is a rule you just cannot add any random or a bunch of methods in a delegate it needs to follow the input-output rule
now why are there 2 different ways to do one thing, the only thing different is the assignment operators (+, =), this introduces a new topic called
EVENTS
which is nothing but a constrained version of a Delegate, It's still a List of methods don't confuse when people explain these terminologies, they change the name, so stick with this to understand
what is the constraint? you cannot do this del = calculate;
what's the harm in it, say a bunch of methods are added to the Delegate(List), you do that 👆 all are wiped out and only a single method "calculate" remains, so to prevent that Events are used,
Event Syntax
Public Event demo del = null;
One more thing you cannot do with events is invoke the delegate directly like demo.invoke since its public it can be accessed and invoked but with events, it can't
now you just add the methods to the event (a special type of delegate)
when to use an event vs a delegate, depends on your situation but pragmatically events are popular
few more keywords
MULTICASTING: nothing but adding more than one method to a delegate
BROADCASTING: adding more than one method to an event
PUBLISHER: the one that executes the method (term used in broadcasting), only a single entity
SUBSCRIBER: The methods that are being executed, can be multiple
LISTENER: the same thing as a subscriber but the term is used in multicasting
EVENT HANDLER: same thing as a subscriber/event listener so what the difference? it's basically the same thing, some say an eventlistener detect for the event to occur and the event handler "handles" or execute the code, ITS THE SAME THING PRACTICALLY!
action and func are just delegates that have been created and instantiated so 2 lines of code in a word, the difference is just in return types
ACTION: does not return anything while taking 0 or more than 1 input
FUNC: returns one thing and takes in parameters
if you don't do good with reading here is the best video on this topic
https://www.youtube.com/playlist?list=PLFt_AvWsXl0dliMtpZC8Qd_ru26785Ih_