Node.js
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.
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.
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 Settings. Node.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()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). This workaround requires modifying the content, so if you don't have control over the webpage enabling node in your HTML widget, this can cause an intractable problem. You should only enable Node.js if you are planning to use it (for example with our JavaScript APIs).
See also HTML Best Practices.
Node SerialPort
The BrightSign player provides JavaScript serial port bindings for the Node SerialPort package. These bindings can be used using with the @brightsign/serialport API:
Also see our reference implementation on GitHub.
Reloading Node.js Enabled Web Applications
Any customer application that uses @brightsign objects, and some of the "BS" objects, will throw a JavaScript exception when reloaded in BrightSignOS 8.0 to 8.4 (even though it may have worked in OS 7.x and earlier) due to limitations with native modules in Node 10. This problem will be fixed in OS 8.5 and above, which ships with Node 14.x. You can also disable Node.js (and @brightsign objects) if you don’t need it..
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.
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:
The JavaScript file containing the Node.js entry point can then be referenced in the HTML file:
Example HTML:
To use webpack®, attach main() to the window object so that it can be found from the HTML file:
Change the HTML file so that it points to the bundled JavaScript and to window.main():
To build your bundle, run the following npm steps on your computer:
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:
Alternatively, if you have modules located on multiple storage drives, you can append multiple search paths to a module:
Example:
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.
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:
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.