Migrating from 3.x to 4.0

The 4.0.0 release adds quite a lot of new features, which are detailed below, but it also contains a few API breaking changes (hence the major bump).

Please note that these breaking changes only impact the API on the server side. The Socket.IO protocol itself was not updated, so a v3 client will be able to reach a v4 server and vice-versa. Besides, the compatibility mode (allowEIO3: true) is still available between a Socket.IO v2 client and a Socket.IO v4 server.

Here is the complete list of changes:

• Breaking changes
  • io.to() is now immutable
  • wsEngine option
• Configuration
  • Ensure compatibility with Swift v15 clients
  • The default value of pingTimeout was increased
• New features
  • Allow excluding specific rooms when broadcasting
  • Allow to pass an array to io.to()
  • Additional utility methods
  • Typed events
  • autoUnref option
• Known migration issues

Breaking changes

io.to() is now immutable

Previously, broadcasting to a given room (by calling io.to()) would mutate the io instance, which could lead to surprising behaviors, like:

io.to("room1");
io.to("room2").emit(/* ... */); // also sent to room1

// or with async/await
io.to("room3").emit("details", await fetchDetails()); // random behavior: maybe in room3, maybe to all clients

Calling io.to() (or any other broadcast modifier) will now return an immutable instance.

Examples:

const operator1 = io.to("room1");
const operator2 = operator1.to("room2");
const operator3 = socket.broadcast;
const operator4 = socket.to("room3").to("room4");

operator1.emit(/* ... */); // only to clients in "room1"
operator2.emit(/* ... */); // to clients in "room1" or in "room2"
operator3.emit(/* ... */); // to all clients but the sender
operator4.emit(/* ... */); // to clients in "room3" or in "room4" but the sender

wsEngine option

The format of the wsEngine option was updated in order to get rid of the following error:

Critical dependency: the request of a dependency is an expression

when bundling the server with webpack.

Before:

const io = require("socket.io")(httpServer, {
  wsEngine: "eiows"
});

After:

const io = require("socket.io")(httpServer, {
  wsEngine: require("eiows").Server
});

Configuration

Ensure compatibility with Swift v15 clients

Before version 16.0.0, the Swift client would not include the EIO query parameter in the HTTP requests, and the Socket.IO v3 server would infer EIO=4 by default.

That's why a Swift client v15 was not able to connect to the server, even when the compatibility mode was enabled (allowEIO3: true), unless you explicitly specified the query param:

let manager = SocketManager(socketURL: URL(string: "http://localhost:8080")!, config: [
  .log(true),
  .connectParams(["EIO": "3"])
])
let socket = manager.defaultSocket

The Socket.IO v4 server will now infer EIO=3 if the EIO query param is not included.

The default value of pingTimeout was increased

The default value of pingTimeout (used in the heartbeat mechanism) value was updated from 60000 to 5000 in socket.io@2.1.0 (March 2018).

The reasoning back then:

Some users experienced long delays between disconnection on the server-side and on the client-side. The "disconnect" event would take a long time to fire in the browser, probably due to a timer being delayed. Hence the change.

That being said, the current value (5s) caused unexpected disconnections when a big payload was sent over a slow network, because it prevents the ping-pong packets from being exchanged between the client and the server. This can also happen when a synchronous task blocks the server for more than 5 seconds.

The new value (20s) thus seems like a good balance between quick disconnection detection and tolerance to various delays.

New features

Allow excluding specific rooms when broadcasting

Thanks to the awesome work of Sebastiaan Marynissen, you can now exclude a specific room when broadcasting:

io.except("room1").emit(/* ... */); // to all clients except the ones in "room1"
io.to("room2").except("room3").emit(/* ... */); // to all clients in "room2" except the ones in "room3"

socket.broadcast.except("room1").emit(/* ... */); // to all clients except the ones in "room1" and the sender
socket.except("room1").emit(/* ... */); // same as above
socket.to("room4").except("room5").emit(/* ... */); // to all clients in "room4" except the ones in "room5" and the sender

Allow to pass an array to io.to()

The to() method now accepts an array of rooms.

Before:

const rooms = ["room1", "room2", "room3"];
for (const room of rooms) {
  io.to(room);
}
// broadcast to clients in "room1", "room2" or "room3"
// WARNING !!! this does not work anymore in v4, see the breaking change above
io.emit(/* ... */);

After:

io.to(["room1", "room2", "room3"]).emit(/* ... */);

socket.to(["room1", "room2", "room3"]).emit(/* ... */);

Additional utility methods

Some (long-awaited) methods were added:

• socketsJoin: makes the matching socket instances join the specified rooms

// make all Socket instances join the "room1" room
io.socketsJoin("room1");

// make all Socket instances of the "admin" namespace in the "room1" room join the "room2" room
io.of("/admin").in("room1").socketsJoin("room2");

• socketsLeave: makes the matching socket instances leave the specified rooms

// make all Socket instances leave the "room1" room
io.socketsLeave("room1");

// make all Socket instances of the "admin" namespace in the "room1" room leave the "room2" room
io.of("/admin").in("room1").socketsLeave("room2");

• disconnectSockets: makes the matching socket instances disconnect

// make all Socket instances disconnect
io.disconnectSockets();

// make all Socket instances of the "admin" namespace in the "room1" room disconnect
io.of("/admin").in("room1").disconnectSockets();

// this also works with a single socket ID
io.of("/admin").in(theSocketId).disconnectSockets();

• fetchSockets: returns the matching socket instances

// return all Socket instances of the main namespace
const sockets = await io.fetchSockets();

// return all Socket instances of the "admin" namespace in the "room1" room
const sockets = await io.of("/admin").in("room1").fetchSockets();

// this also works with a single socket ID
const sockets = await io.in(theSocketId).fetchSockets();

The sockets variable in the example above is an array of objects exposing a subset of the usual Socket class:

for (const socket of sockets) {
  console.log(socket.id);
  console.log(socket.handshake);
  console.log(socket.rooms);
  socket.emit(/* ... */);
  socket.join(/* ... */);
  socket.leave(/* ... */);
  socket.disconnect(/* ... */);
}

Those methods share the same semantics as broadcasting, and the same filters apply:

io.of("/admin").in("room1").except("room2").local.disconnectSockets();

Which makes all Socket instances of the "admin" namespace

• in the "room1" room (in("room1") or to("room1"))
• except the ones in "room2" (except("room2"))
• and only on the current Socket.IO server (local)

disconnect.

Typed events

Thanks to the awesome work of Maxime Kjaer, TypeScript users can now type the events sent between the client and the server.

First, you declare the signature of each event:

interface ClientToServerEvents {
  noArg: () => void;
  basicEmit: (a: number, b: string, c: number[]) => void;
}

interface ServerToClientEvents {
  withAck: (d: string, cb: (e: number) => void) => void;
}

And you can now use them on the client side:

import { io, Socket } from "socket.io-client";

const socket: Socket<ServerToClientEvents, ClientToServerEvents> = io();

socket.emit("noArg");

socket.emit("basicEmit", 1, "2", [3]);

socket.on("withAck", (d, cb) => {
    cb(4);
});

Your IDE should now properly infer the type of each argument:

Similarly on the server side (the ServerToClientEvents and ClientToServerEvents are reversed):

import { Server } from "socket.io";

const io = new Server<ClientToServerEvents, ServerToClientEvents>(3000);

io.on("connect", (socket) => {
    socket.on("noArg", () => {
      // ...
    });

    socket.on("basicEmit", (a, b, c) => {
      // ...
    });

    socket.emit("withAck", "42", (e) => {
        console.log(e);
    });
});

By default, the events are untyped and the arguments will be inferred as any.

autoUnref option

And finally, thanks to the awesome work of KC Erb, the autoUnref option was added.

With autoUnref set to true (default: false), the Socket.IO client will allow the program to exit if there is no other active timer/TCP socket in the event system (even if the client is connected):

const socket = io({
  autoUnref: true
});

Note: this option only applies to Node.js clients.

Known migration issues

• cannot get emit of undefined

The following expression:

socket.to("room1").broadcast.emit(/* ... */);

was working in Socket.IO v3 but is now considered invalid, as the broadcast flag is useless because the to("room1") method already puts the Socket instance in broadcasting mode.

// VALID
socket.broadcast.emit(/* ... */); // to all clients but the sender
socket.to("room1").emit(/* ... */); // to clients in "room1" but the sender

// VALID (but useless 'broadcast' flag)
socket.broadcast.to("room1").emit(/* ... */);

// INVALID
socket.to("room1").broadcast.emit(/* ... */);