Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 50 Next »

BrightSign Node.js Overview

Our customers often need to communicate between applications on a BrightSign player, or between a BrightSign player and the outside world. This can be done using various protocols: HTTP, serial, and UDP are the most popular. You can write applications in either BrightScript or JavaScript to create this functionality. If you choose to write your application in JavaScript, you will need Node.js® to implement features such as hosting a web server in JavaScript.

BrightSign makes a Node.js endpoint available so that our partners can use Node.js modules and features. 

BrightSign Node.js Implementation

BrightSign players support the Node.js runtime environment, which runs on the same V8 JavaScript engine used by Chromium. The Node.js® and Chromium instances share a single JavaScript execution context, so JavaScript applications can access both Node.js modules and DOM objects at the same time. BrightSign firmware pushes Node.js events to the Chromium event loop, ensuring that JavaScript applications receive Node.js and DOM events seamlessly.

The BrightSign Node.js implementation is based on the NW.js and Electron projects and shares many characteristics with them. Whereas Electron uses a JavaScript file as the entry point (i.e. the JavaScript file creates a webview, which also has access to the Node.js runtime), NW.js and BrightSign use an HTML file as the entry point: On a BrightSign player, BrightScript creates a Node.js-enabled roHtmlWidget instance; the initial URL, which is passed when roHtmlWidget is initialized, acts as the entry point for Node.js applications.

Like NW.js and Electron, BrightSign does not use sandboxing. Instead, it launches the render process with a Node.js user and storage group, which has write permissions for local storage and read permissions for the entire file system. It also can access networking interfaces and use privileged ports.

ON THIS PAGE


The integrated Node.js implementation in OS 8.2.x and OS 8.3.x is based on Node v10. For further documentation and usage information, consult the Node.js 10.0.0 API documentation

Note

Node.js is not enabled for iframes or Web Workers.

Enabling Node.js

Node.js object functionality is available in BrightAuthor:connected when users specify Node.js items in Presentation SettingsNode.js objects are not available in BrightAuthor.

Node.js is enabled for individual roHtmlWidget instances by including the nodejs_enabled:true entry in the initialization parameters of the roHtmlWidget object.

Example
r=CreateObject("roRectangle", 0,0,1920,1080)
is = {
	port: 3000
}
config = {
	nodejs_enabled: true
	inspector_server: is
	brightsign_js_objects_enabled: true
	url: "file:///sd:/nodehello.html"
}
h=CreateObject("roHtmlWidget", r, config)
h.Show()

Important

We do not recommend loading arbitrary websites with Node.js enabled. Some JavaScript libraries assume that Node.js is running on an instance with server-side capabilities, and they will attempt to load dependencies, causing playback to fail.

If you are using a BrightAuthor plugin to enable Node.js, you will need to set other desired parameters in the plugin, rather than with an HTML5 state. For example, if you want to enable the mouse cursor, you will need to set mouse_enabled:true during the roHtmlWidget initialization, rather than checking the box in the HTML5 state.


Cross-Domain Security

Chromium has default security measures for preventing cross-site scripting attacks: If the URL for the roHtmlWidget instance is a remote domain, JavaScript applications from that domain cannot make HTTP requests to other domains; on the other hand, if the URL points to local storage, requests to other, remote domains are acceptable.

If you want to reference other domains in remote applications, set the websecurity parameter to false when initalizing the roHtmlWidget, as shown below:

Example
r=CreateObject("roRectangle", 0,0,1920,1080)
is = {
	port: 3000
}
config = {
	nodejs_enabled: true
	inspector_server: is
	brightsign_js_objects_enabled: true
	url: "http://www.mysitehere.com"
	security_params: {websecurity: false} 
}
h=CreateObject("roHtmlWidget", r, config)
h.Show()

Web Storage

If you want to use JavaScript storage applications, you will need to specify a storage_path and storage_quota when initializaing the roHtmlWidget:

Example
r=CreateObject("roRectangle", 0,0,1920,1080)
is = {
	port: 3000
}
config = {
	nodejs_enabled: true
	inspector_server: is
	brightsign_js_objects_enabled: true
	url: "file:///sd:/nodehello.html"
	storage_path: "SD:"
	storage_quota: 1073741824
}
h=CreateObject("roHtmlWidget", r, config)
h.Show()

JQuery

JQuery® requires a workaround to operate correctly with Node.js. See this page for an example.


Packaging and Delivering Node.js Applications

To deploy your Node.js application to a BrightSign player, run "npm install" on your computer. This will create the node_modules directory. Copy this directory to the SD card along with the rest of the application.

When initialized, the BrightSign Node.js implementation seeks to the node_modules directory relative to the loaded HTML file. Like a standard Node.js application, it then loads all modules contained in the node_modules directory.

Note

The majority of Node.js modules contain JavaScript code only. However, some modules contain binary code. When a module containing binary parts is installed using "npm install", the binary parts are compiled for the local platform (usually Intel x64), and this code will not run on a BrightSign player. Currently, the BrightSign Node.js implementation is limited to JavaScript code only.

WebPack

The node_modules directory associated with a Node.js application may contain hundreds or thousands of unnecessary files. The webpack® bundler allows you to reduce the node_modules directory to a manageable size.

To use webpack®, you will need to place your Node.js entry-point function in a separate file (e.g. "index.js"), rather than as inline HTML. This file will contain the references to other JavaScript files and Node.js modules:

Example index.js
var myutils = require('./utilities'); // My utilities.js
var moment = require('moment'); // A node module
 
function main() {
 
  myutils.foo();
  moment.now();
  //...
}

The JavaScript file containing the Node.js entry point can then be referenced in the HTML file:

Example HTML
<script src='./index.js'></script>

To use webpack®, attach main() to the window object so that it can be found from the HTML file:

var myutils = require('./utilities'); // My utilities.js
var moment = require('moment'); // A node module
 
function main() {
 
  myutils.foo();
  moment.now();
  //...
}
window.main = main;

Change the HTML file so that it points to the bundled JavaScript and to window.main():

<script src='./bundle.js'></script>
 
<body onload="window.main()">

To build your bundle, run the following npm steps on your computer:

npm i -D webpack webpack-cli
npx webpack --mode production

Now you can publish the index.html and bundle.js files; there's no need to publish the node_modules directory. See the sample webpack® configuration below for more information.

Device Storage Paths

To load Node.js modules and read/write files, you must first define the root directory of the device storage. The following are common root directories:

  • microSD: "storage/sd/"
  • SSD: "storage/ssd/"
  • USB: "storage/usb1/"

We recommend using the process.chdir() call at the beginning of the script to change the process path:

Example
var process = require("process");
process.chdir("/storage/sd");


Alternatively, if you have modules located on multiple storage drives, you can append multiple search paths to a module:

Example
module.paths.push("/storage/sd/")
module.paths.push("/storage/ssd/")
module.paths.push("/storage/usb1/")

Debugging Applications

When Node.js modules are enabled, they become visible from the Chromium remote inspector, allowing you to debug applications. The console.log works like a normal web application: Output is redirected to both stderr and the remote inspector.

Downloading Large Files 

If your application uses the XMLHttpRequest object to download a large file (100-200MB, depending on the player model), the player will run out of memory and the download operation will fail. The XMLHttpRequest object first downloads the whole file into memory, then creates a blob object of equal size, so memory requirements for a download are effectively double that of the file size.

For large-file downloads, we recommend using the Fetch API, which has callbacks that are fired when fragments are downloaded. Data can be appended to the disk as it arrives, so the player won't run out of memory, no matter the file size. The "download-test" HTML/JavaScript example below uses the Fetch API and the Node.js File Service module to download a file in fragments.

Note

This example is specific to BrightSign players and requires node.js runtime, enabled by use of the roHtmlWidget nodejs_enabled flag, to write files to disk.

Node.js Examples

GitHub contains a Node.js-starter-project. This project was originally created for BrightAuthor but can also be used with BrightAuthor:connected.

Another example is the script below, which initializes an HTTP server on the BrightSign player at port 8000. When a client (for example, a desktop browser) connects to the server, it will send the model number and boot version of the player to the client. The script also displays the IP address of the connected client on the screen attached to the player. 

Example
<html>
<script>
function displayMessage()
{
  // Load the http module to create an http server.
  var http = require('http');

  // Configure our HTTP server to respond with Hello World to all requests.
  var server = http.createServer(function (request, response) {
    var device_info = new BSDeviceInfo();
    response.writeHead(200, {"Content-Type": "text/plain"});
    response.end("Device Information:\n" + device_info.model + "\n" + device_info.bootVersion + "\n");
    var ip = request.connection.remoteAddress;
    document.getElementById("Ip").innerHTML+="Server responded to: "+ ip + "<br>";
    console.log("Server responded to request from " + ip);
  });

  // Listen on port 8000, IP defaults to 127.0.0.1
  server.listen(8000);

  // Display it on brightsign browser
  var os = require('os');
  var interfaces = os.networkInterfaces();
  var addresses = [];
  for (var k in interfaces) {
      for (var k2 in interfaces[k]) {
          var address = interfaces[k][k2];
          if (address.family === 'IPv4' && !address.internal) {
              addresses.push(address.address);
          }
      }
  }
  var message = "Server running at: " + addresses[0] + ":8000<br>";
  document.getElementById("Ip").innerHTML+= message;

  // Print message on console
  console.log(message);

}
</script>
<body style="background-color:red" onload="displayMessage()">
  <div id = "Ip" style="font-size:60px; text-align:center;">
  </div>
</body>
</html>


Built-in modules, such as "os" and "http", can be initialized using the require() method. If the nodejs_enabled:true entry is not included when initializing the roHtmlWidget object (as shown above), the require() method will not be available.


  • No labels