websocket-sharp supports:
- RFC 6455
- WebSocket Client and Server
- Per-message Compression extension
- Secure Connection
- HTTP Authentication
- Query String, Origin header and Cookies
- Connecting through the HTTP Proxy server
- .NET 4.5 or later (includes compatible)
websocket-sharp is built as a single assembly, websocket-sharp.dll.
websocket-sharp is developed with MonoDevelop. So the simple way to build is to open websocket-sharp.sln and run build for websocket-sharp project with any of the build configurations (e.g. Debug
) in MonoDevelop.
You should add your websocket-sharp.dll (e.g. /path/to/websocket-sharp/bin/Debug/websocket-sharp.dll
) to the library references of your project.
websocket-sharp is available on the NuGet Gallery.
You can add websocket-sharp to your project using the NuGet Package Manager, the following command in the Package Manager Console.
PM> Install-Package WebSocketSharp.clone
using System;
using WebSocketSharp;
namespace Example
{
public class Program
{
public static void Main (string[] args)
{
using (var ws = new WebSocket ("ws://dragonsnest.far/Laputa")) {
ws.OnMessage = e => {
Console.WriteLine ("Laputa says: " + e.Data);
}
ws.Connect ();
ws.Send ("BALUS");
Console.ReadKey (true);
}
}
}
}
Required namespace.
using WebSocketSharp;
The WebSocket
class exists in the WebSocketSharp
namespace.
Creating a new instance of the WebSocket
class with the WebSocket URL to connect.
using (var ws = new WebSocket ("ws://example.com")) {
...
}
The WebSocket
class inherits the System.IDisposable
interface, so you can use the using
statement. And the WebSocket connection will be closed with close status 1001
(going away) when the control leaves the using
block.
Setting the WebSocket
events.
A WebSocket.OnOpen
event occurs when the WebSocket connection has been established.
ws.OnOpen = (sender, e) => {
...
};
e
has passed as the System.EventArgs.Empty
, so you don't use e
.
A WebSocket.OnMessage
event occurs when the WebSocket
receives a message.
ws.OnMessage += (sender, e) => {
...
};
e
has passed as a WebSocketSharp.MessageEventArgs
.
e.Type
property returns either WebSocketSharp.Opcode.Text
or WebSocketSharp.Opcode.Binary
that represents the type of the message. So by checking it, you can determine which item you should use.
If it returns Opcode.Text
, you should use e.Data
property that returns a string
(represents the Text message).
Or if it returns Opcode.Binary
, you should use e.RawData
property that returns a byte[]
(represents the Binary message).
if (e.Type == Opcode.Text) {
// Do something with e.Data.
...
return;
}
if (e.Type == Opcode.Binary) {
// Do something with e.RawData.
...
return;
}
A WebSocket.OnError
event occurs when the WebSocket
gets an error.
ws.OnError += (sender, e) => {
...
};
e
has passed as a WebSocketSharp.ErrorEventArgs
.
e.Message
property returns a string
that represents the error message. So you should use it to get the error message.
And if the error is due to an exception, you can get the System.Exception
instance that caused the error, by using e.Exception
property.
A WebSocket.OnClose
event occurs when the WebSocket connection has been closed.
ws.OnClose += (sender, e) => {
...
};
e
has passed as a WebSocketSharp.CloseEventArgs
.
e.Code
property returns a ushort
that represents the status code indicating the reason for the close, and e.Reason
property returns a string
that represents the reason for the close. So you should use them to get the reason for the close.
Connecting to the WebSocket server.
ws.Connect ();
If you would like to connect to the server asynchronously, you should use the WebSocket.ConnectAsync ()
method.
Sending a data to the WebSocket server.
ws.Send (data);
The WebSocket.Send
method is overloaded.
You can use the WebSocket.Send (string)
, WebSocket.Send (byte[])
, or WebSocket.Send (System.IO.FileInfo)
method to send a data.
If you would like to send a data asynchronously, you should use the WebSocket.SendAsync
method.
ws.SendAsync (data, completed);
And also if you would like to do something when the send is complete, you should set completed
to any Action<bool>
delegate.
Closing the WebSocket connection.
ws.Close (code, reason);
If you would like to close the connection explicitly, you should use the WebSocket.Close
method.
The WebSocket.Close
method is overloaded.
You can use the WebSocket.Close ()
, WebSocket.Close (ushort)
, WebSocket.Close (WebSocketSharp.CloseStatusCode)
, WebSocket.Close (ushort, string)
, or WebSocket.Close (WebSocketSharp.CloseStatusCode, string)
method to close the connection.
If you would like to close the connection asynchronously, you should use the WebSocket.CloseAsync
method.
using System;
using WebSocketSharp;
using WebSocketSharp.Server;
namespace Example
{
public class Laputa : WebSocketBehavior
{
protected override void OnMessage (MessageEventArgs e)
{
var msg = e.Data == "BALUS"
? "I've been balused already..."
: "I'm not available now.";
Send (msg);
}
}
public class Program
{
public static void Main (string[] args)
{
var wssv = new WebSocketServer ("ws://dragonsnest.far");
wssv.AddWebSocketService<Laputa> ("/Laputa");
wssv.Start ();
Console.ReadKey (true);
wssv.Stop ();
}
}
}
Required namespace.
using WebSocketSharp.Server;
The WebSocketBehavior
and WebSocketServer
classes exist in the WebSocketSharp.Server
namespace.
Creating the class that inherits the WebSocketBehavior
class.
For example, if you would like to provide an echo service,
using System;
using WebSocketSharp;
using WebSocketSharp.Server;
public class Echo : WebSocketBehavior
{
protected override void OnMessage (MessageEventArgs e)
{
Send (e.Data);
}
}
And if you would like to provide a chat service,
using System;
using WebSocketSharp;
using WebSocketSharp.Server;
public class Chat : WebSocketBehavior
{
private string _suffix;
public Chat ()
: this (null)
{
}
public Chat (string suffix)
{
_suffix = suffix ?? String.Empty;
}
protected override void OnMessage (MessageEventArgs e)
{
Sessions.Broadcast (e.Data + _suffix);
}
}
You can define the behavior of any WebSocket service by creating the class that inherits the WebSocketBehavior
class.
If you override the WebSocketBehavior.OnMessage (MessageEventArgs)
method, it's called when the WebSocket
used in the current session in the service receives a message.
And if you override the WebSocketBehavior.OnOpen ()
, WebSocketBehavior.OnError (ErrorEventArgs)
, and WebSocketBehavior.OnClose (CloseEventArgs)
methods, each of them is called when each event of the WebSocket
(the OnOpen
, OnError
, and OnClose
events) occurs.
The WebSocketBehavior.Send
method sends a data to the client on the current session in the service.
If you would like to access the sessions in the service, you should use the WebSocketBehavior.Sessions
property (returns a WebSocketSharp.Server.WebSocketSessionManager
).
The WebSocketBehavior.Sessions.Broadcast
method broadcasts a data to every client in the service.
Creating an instance of the WebSocketServer
class.
var wssv = new WebSocketServer (4649);
wssv.AddWebSocketService<Echo> ("/Echo");
wssv.AddWebSocketService<Chat> ("/Chat");
wssv.AddWebSocketService<Chat> ("/ChatWithNyan", () => new Chat (" Nyan!"));
You can add any WebSocket service to your WebSocketServer
with the specified behavior and path to the service, using the WebSocketServer.AddWebSocketService<TBehaviorWithNew> (string)
or WebSocketServer.AddWebSocketService<TBehavior> (string, Func<TBehavior>)
method.
The type of TBehaviorWithNew
must inherit the WebSocketBehavior
class, and must have a public parameterless constructor.
And also the type of TBehavior
must inherit the WebSocketBehavior
class.
So you can use the classes created in Step 2 to add the service.
If you create an instance of the WebSocketServer
class without a port number, the WebSocketServer
set the port number to 80 automatically. So it's necessary to run with root permission.
$ sudo mono example2.exe
Starting the WebSocket server.
wssv.Start ();
Stopping the WebSocket server.
wssv.Stop (code, reason);
The WebSocketServer.Stop
method is overloaded.
You can use the WebSocketServer.Stop ()
, WebSocketServer.Stop (ushort, string)
, or WebSocketServer.Stop (WebSocketSharp.CloseStatusCode, string)
method to stop the server.
websocket-sharp supports the Per-message Compression extension. (But it doesn't support with the extension parameters.)
If you would like to enable this extension as a WebSocket client, you should set such as the following.
ws.Compression = CompressionMethod.Deflate;
And then your client sends the following header with the connection request to the server.
Sec-WebSocket-Extensions: permessage-deflate
If the server supports this extension, it returns the same header. And when your client receives that header, it enables this extension.
websocket-sharp supports the Secure Connection with SSL/TLS.
As a WebSocket Client, you should create an instance of the WebSocket
class with the wss scheme WebSocket URL.
using (var ws = new WebSocket ("wss://example.com")) {
...
}
And if you would like to use the custom validation for the server certificate, you should set the WebSocket.SslConfiguration.ServerCertificateValidationCallback
property.
ws.SslConfiguration.ServerCertificateValidationCallback =
(sender, certificate, chain, sslPolicyErrors) => {
// Do something to validate the server certificate.
...
return true; // If the server certificate is valid.
};
If you set this property to nothing, the validation does nothing with the server certificate, and returns true
.
As a WebSocket Server, you should create an instance of the WebSocketServer
or HttpServer
class with some settings for secure connection, such as the following.
var wssv = new WebSocketServer (4649, true);
wssv.SslConfiguration.ServerCertificate =
new X509Certificate2 ("/path/to/cert.pfx", "password for cert.pfx");
websocket-sharp supports the HTTP Authentication (Basic/Digest).
As a WebSocket Client, you should set a pair of user name and password for the HTTP authentication, using the WebSocket.SetCredentials (string, string, bool)
method before connecting.
ws.SetCredentials ("nobita", "password", preAuth);
If preAuth
is true
, the WebSocket
sends the Basic authentication credentials with the first connection request to the server.
Or if preAuth
is false
, the WebSocket
sends either the Basic or Digest (determined by the unauthorized response to the first connection request) authentication credentials with the second connection request to the server.
As a WebSocket Server, you should set an HTTP authentication scheme, a realm, and any function to find the user credentials before starting, such as the following.
wssv.AuthenticationSchemes = AuthenticationSchemes.Basic;
wssv.Realm = "WebSocket Test";
wssv.UserCredentialsFinder = id => {
var name = id.Name;
// Return user name, password, and roles.
return name == "nobita"
? new NetworkCredential (name, "password", "gunfighter")
: null; // If the user credentials aren't found.
};
If you would like to provide the Digest authentication, you should set such as the following.
wssv.AuthenticationSchemes = AuthenticationSchemes.Digest;
As a WebSocket Client, if you would like to send the Query String with the WebSocket connection request to the server, you should create an instance of the WebSocket
class with the WebSocket URL that includes the Query string parameters.
using (var ws = new WebSocket ("ws://example.com/?name=nobita")) {
...
}
And if you would like to send the Origin header with the WebSocket connection request to the server, you should set the WebSocket.Origin
property to an allowable value as the Origin header before connecting, such as the following.
ws.Origin = "http://example.com";
And if you would like to send the Cookies with the WebSocket connection request to the server, you should set any cookie using the WebSocket.SetCookie (WebSocketSharp.Net.Cookie)
method before connecting, such as the following.
ws.SetCookie (new Cookie ("name", "nobita"));
As a WebSocket Server, if you would like to get the Query String included in each WebSocket connection request, you should access the WebSocketBehavior.Context.QueryString
property, such as the following.
public class Chat : WebSocketBehavior
{
private string _name;
...
protected override void OnOpen ()
{
_name = Context.QueryString["name"];
}
...
}
And if you would like to validate the Origin header, Cookies, or both included in each WebSocket connection request, you should set each validation with your WebSocketBehavior
, for example, using the AddWebSocketService<TBehavior> (string, Func<TBehavior>)
method with initializing, such as the following.
wssv.AddWebSocketService<Chat> (
"/Chat",
() => new Chat () {
OriginValidator = val => {
// Check the value of the Origin header, and return true if valid.
Uri origin;
return !val.IsNullOrEmpty () &&
Uri.TryCreate (val, UriKind.Absolute, out origin) &&
origin.Host == "example.com";
},
CookiesValidator = (req, res) => {
// Check the Cookies in 'req', and set the Cookies to send to the client with 'res'
// if necessary.
foreach (Cookie cookie in req) {
cookie.Expired = true;
res.Add (cookie);
}
return true; // If valid.
}
});
Also, if you would like to get each value of the Origin header and cookies, you should access each of the WebSocketBehavior.Context.Origin
and WebSocketBehavior.Context.CookieCollection
properties.
websocket-sharp supports to connect through the HTTP Proxy server.
If you would like to connect to a WebSocket server through the HTTP Proxy server, you should set the proxy server URL, and if necessary, a pair of user name and password for the proxy server authentication (Basic/Digest), using the WebSocket.SetProxy (string, string, string)
method before connecting.
var ws = new WebSocket ("ws://example.com");
ws.SetProxy ("http://localhost:3128", "nobita", "password");
I tested this with the Squid. And it's necessary to disable the following configuration option in squid.conf (e.g. /etc/squid/squid.conf
).
# Deny CONNECT to other than SSL ports
#http_access deny CONNECT !SSL_ports
websocket-sharp supports RFC 6455, and it's based on the following WebSocket references:
websocket-sharp is provided under The MIT License.