node package manager


Volatile namespaces for Primus


Version npmBuild StatusDependenciesCoverage StatusIRC channel

SubStream is a simple stream multiplexer for Primus. It allows you to create simple message channels which only receive the information you send to it. These channels are in fact, streams, which is why this module is called SubStreams because it adds small streams on top of the main stream and intercepts them.


npm install --save substream

The module can only be used in conjunction with Primus so make sure that your application is using that as its real-time backend.

Getting started

In all the code examples, we assume that the following code is present:

'use strict';
var Primus = require('primus')
  , http = require('http');
var server = http.createServer()
  , primus = new Primus(server);
// Custom code here, just above the listen call. 

Which is the most minimal bootstrapping code required to create a Primus powered server. Once you've setup the server you need to add SubStream as a plugin in to Primus:

// The `primus.use` method adds the plugin to primus. It requires a unique 
// name in order to easily retrieve it again. For the 
// sake of clarity, we're going to use 'substream' as the name. 
primus.use('substream', require('substream'));

After you've added plugins, you might want to re-compile the client library that Primus serves, as it automatically adds the client-side plugin to the framework, as well as the custom substream.js library to create the actual name spaces. To save the client just run: +'/primus.js');

But this is only needed if you serve the file manually and not through the automatically generated /primus/primus.js path (which is regenerated on the fly). Now that we've set everything up correctly we can start creating some substreams.

The client

To create or access a substream in the Primus client start off with making a connection:

var primus = new Primus('http://<your url here:whateverportnumber>');
var foo = primus.substream('foo');

The substream method automatically creates a namespaced stream if you didn't create it before. Or, it will return your previously created stream when you call it again. So now we have a foo stream we can just write to:


Awesome, all works as intended. But this was just one single substream, we can add more:

var bar = primus.substream('bar')
  , baz = primus.substream('baz');

You can create an infinite number of substreams on top of one single base stream. The data is not leaked between streams. It's all "sandboxed".

As the returned substreams are streams or eventemitters we can just listen to data, end or close events. But it also proxies all the other events that Primus emits such as the reconnect, offline events etc. (The full list is in the Primus So for receiving and writing data you can just do:

bar.on('data', function () {
bar.write('hello from bar');
foo.on('data', function (data) {
  console.log('recieved data', data);
}).on('end', function () {
  console.log('connection has closed or substream was closed');

The server

The server portion of this module isn't that different than the client portion. It follows the same API stream/eventemitter API:

primus.on('connection', function (spark) {
  var foo = spark.substream('foo')
    , bar = spark.substream('bar')
    , baz = spark.substream('baz');
  foo.on('data', function (data) {
    console.log('foo received:', data);
  // You can even pipe data 
  fs.createReadSteam(__dirname + '/example.js').pipe(bar, {
    end: false
  // To stop receiving data, simply end the substream: 


  • When the spark/connection closes all SubStreams will automatically close, this only happens for the end event, the close event is proxied.
  • We add an extra streams object to the spark/connection which contains all active SubStreams for the given connection mapped by name->substream-instance