SignalR

SignalR is an open-source library that simplifies adding real-time web functionality to apps. Real-time web functionality enables server-side code to push content to clients instantly.

Historically it was tied to ASP.NET Core but the protocol is open and implementable in any language.

This repository contains an implementation of a SignalR server and a SignalR client in go. The implementation is based on the work of David Fowler at https://github.com/davidfowl/signalr-ports. Client and server support transport over WebSockets, Server Sent Events and raw TCP. Protocol encoding in JSON and MessagePack is fully supported.

Install
Getting Started
Debugging

Install

With a correctly configured Go toolchain:

go get -u github.com/philippseith/signalr

Getting Started

SignalR uses a signalr.HubInterface instance to anchor the connection on the server and a javascript HubConnection object to anchor the connection on the client.

Server side

Implement the HubInterface

The easiest way to implement the signalr.HubInterface in your project is to declare your own type and embed signalr.Hub which implements that interface and will take care of all the signalr plumbing. You can call your custom type anything you want so long as it implements the signalr.HubInterface interface.

package main

import "github.com/philippseith/signalr"

type AppHub struct {
    signalr.Hub
}

Add functions with your custom hub type as a receiver.

func (h *AppHub) SendChatMessage(message string) {
    h.Clients().All().Send("chatMessageReceived", message)
}

These functions must be public so that they can be seen by the signalr server package but can be invoked client-side as lowercase message names. We'll explain setting up the client side in a moment, but as a preview, here's an example of calling our AppHub.SendChatMessage(...) method from the client:

    // javascript snippet invoking that AppHub.Send method from the client
    connection.invoke('sendChatMessage', val);

The signalr.HubInterface contains a pair of methods you can implement to handle connection and disconnection events. signalr.Hub contains empty implementations of them to satisfy the interface, but you can "override" those defaults by implementing your own functions with your custom hub type as a receiver:

func (c *chat) OnConnected(connectionID string) {
    fmt.Printf("%s connected\n", connectionID)
}

func (c *chat) OnDisconnected(connectionID string) {
   fmt.Printf("%s disconnected\n", connectionID)
}

Serve with http.ServeMux

import (
    "net/http"
	
    "github.com/philippseith/signalr"
)

func runHTTPServer() {
    address := 'localhost:8080'
    
    // create an instance of your hub
    hub := AppHub{}
	
    // build a signalr.Server using your hub
    // and any server options you may need
    server, _ := signalr.NewServer(context.TODO(),
        signalr.SimpleHubFactory(hub)
        signalr.KeepAliveInterval(2*time.Second),
        signalr.Logger(kitlog.NewLogfmtLogger(os.Stderr), true))
    )
    
    // create a new http.ServerMux to handle your app's http requests
    router := http.NewServeMux()
    
    // ask the signalr server to map it's server
    // api routes to your custom baseurl
    server.MapHTTP(signalr.WithHTTPServeMux(router), "/chat")

    // in addition to mapping the signalr routes
    // your mux will need to serve the static files
    // which make up your client-side app, including
    // the signalr javascript files. here is an example
    // of doing that using a local `public` package
    // which was created with the go:embed directive
    // 
    // fmt.Printf("Serving static content from the embedded filesystem\n")
    // router.Handle("/", http.FileServer(http.FS(public.FS)))
    
    // bind your mux to a given address and start handling requests
    fmt.Printf("Listening for websocket connections on http://%s\n", address)
    if err := http.ListenAndServe(address, router); err != nil {
        log.Fatal("ListenAndServe:", err)
    }
}

Client side: JavaScript/TypeScript

Grab copies of the signalr scripts

Microsoft has published the client-side libraries as a node package with embedded typescript annotations: @microsoft/signalr.

You can install @microsoft/signalr through any node package manager:

package manager	command
npm	`npm install @microsoft/signalr@latest`
yarn	`yarn add @microsoft/signalr@latest`
LibMan	`libman install @microsoft/signalr@latest -p unpkg -d wwwroot/js/signalr --files dist/browser/signalr.js --files dist/browser/signalr.min.js --files dist/browser/signalr.map.js`
none	you can download the version we are using in our `chatsample` from here (the minified version is here)

Use a HubConnection to connect to your server Hub

How you format your client UI is going to depend on your application use case but here is a simple example. It illustrates the basic steps of connecting to your server hub:

import the signalr.js library (or signalr.min.js);
create a connection object using the HubConnectionBuilder;
bind events
- UI event handlers can use connection.invoke(targetMethod, payload) to send invoke functions on the server hub;
- connection event handlers can react to the messages sent from the server hub;
start your connection

<html>
<body>
    <!-- you may want the content you send to be dynamic -->
    <input type="text" id="message" />
    
    <!-- you may need a trigger to initiate the send -->
    <input type="button" value="Send" id="send" />
    
    <!-- you may want some container to display received messages -->
    <ul id="messages">
    </ul>

    <!-- 1. you need to import the signalr script which provides
            the HubConnectionBuilder and handles the connection
            plumbing.
    -->
    <script src="js/signalr.js"></script>
    <script>
    (async function () {
        // 2. use the signalr.HubConnectionBuilder to build a hub connection
        //    and point it at the baseurl which you configured in your mux
        const connection = new signalR.HubConnectionBuilder()
                .withUrl('/chat')
                .build();

        // 3. bind events:
        //    - UI events can invoke (i.e. dispatch to) functions on the server hub
        document.getElementById('send').addEventListener('click', sendClicked);
        //    - connection events can handle messages received from the server hub
        connection.on('chatMessageReceived', onChatMessageReceived);

        // 4. call start to initiate the connection and start streaming events
        //    between your server hub and your client connection
        connection.start();
        
        // that's it! your server and client should be able to communicate
        // through the signalr.Hub <--> connection pipeline managed by the
        // signalr package and client-side library.
        
        // --------------------------------------------------------------------
       
        // example UI event handler
        function sendClicked() {
            // prepare your target payload
            const msg = document.getElementById('message').value;
            if (msg) {
                // call invoke on your connection object to dispatch
                // messages to the server hub with two arguments:
                // -  target: name of the hub func to invoke
                // - payload: the message body
                // 
                const target = 'sendChatMessage';
                connection.invoke(target, msg);
            }
        }

        // example server event handler
        function onChatMessageReceived(payload) {
            // the payload is whatever was passed to the inner
            // clients' `Send(...)` method in your server-side
            // hub function.
           
            const li = document.createElement('li');
            li.innerText = payload;
            document.getElementById('messages').appendChild(li);
        }
    })();
    </script>
</body>
</html>

Client side: go

To handle callbacks from the server, create a receiver class which gets the server callbacks mapped to its methods:

type receiver struct {
	signalr.Hub
}

func (c *receiver) Receive(msg string) {
	fmt.Println(msg)
}

Receive gets called when the server does something like this:

hub.Clients().Caller().Send("receive", message)

The client itself might be used like that:

// Create a Connection (with timeout for the negotiation process)
creationCtx, _ := context.WithTimeout(ctx, 2 * time.Second)
conn, err := signalr.NewHTTPConnection(creationCtx, address)
if err != nil {
    return err
}
// Create the client and set a receiver for callbacks from the server
client, err := signalr.NewClient(ctx,
	signalr.WithConnection(conn),
	signalr.WithReceiver(receiver))
if err != nil {
    return err
}
// Start the client loop
c.Start()
// Do some client work
ch := <-c.Invoke("update", data)
// ch gets the result of the update operation

Debugging

Server, Client and the protocol implementations are able to log most of their operations. The logging option is disabled by default in all tests. To configure logging, edit the testLogConf.json file:

{
  "Enabled": false,
  "Debug": false
}

If Enabled is set to true, the logging will be enabled. The tests will log to os.Stderr.
If Debug ist set to true, the logging will be more detailed.

wesraph / signalr