AdapterJS Javascript Polyfill and Tools for WebRTC - Skylink WebRTC
JavaScript Other
Jacques-Olivier Haché
Jacques-Olivier Haché Merge branch 'development'
Latest commit 4d5e274 Sep 5, 2017

README.md

Temasys AdapterJS

Creating a common API for WebRTC in the browser

Find the most recent version hosted on our CDN.

  • Minified version: //cdn.temasys.io/adapterjs/0.15.x/adapter.min.js
  • Debug version //cdn.temasys.io/adapterjs/0.15.x/adapter.debug.js
  • Minified version (with screensharing changes): //cdn.temasys.io/adapterjs/0.15.x/adapter.screenshare.min.js
  • Debug version (with screensharing changes)//cdn.temasys.io/adapterjs/0.15.x/adapter.screenshare.js

Part of the Skylink WebRTC toolkit.

Compatibility

AdapterJS provides polyfills and cross-browser helpers for WebRTC. It wraps around the native APIs in Chrome, Opera and Firefox and provides support for WebRTC in Internet Explorer and Safari on Mac and Windows through the available Temasys Browser Plugins.

Supported Browsers & Platform functionality

Browsers Min. Version OS Platform Screensharing
Chrome \ Chromium 38 MacOS / Win / Ubuntu / Android Yes (w Extension)
Firefox 33 MacOS / Win / Ubuntu / Android Yes (w Extension for 51 and below)
Opera 26 MacOS / Win / Ubuntu / Android Yes (if configured with extension)
Edge 13.10547^ Win -
Bowser 0.6.1 iOS 9.x only** -
Safari (Native) 11 MacOS 10.13.x and iOS 11 No. Use Safari with the Temasys plugin***
Safari (Plugin) 7 MacOS Yes (custom build Plugin)
IE (Plugin) 9 Win Yes (custom build Plugin)

*Note that currently Edge doesn't support RTCDataChannel API.

**There seems to be issues for Bowser version 0.6.1 working with iOS 10.x version.

*** To use the Temasys plugin on Safari 11 and above, set the flag AdapterJS.options.forceSafariPlugin = true BEFORE including AdapterJS.

How it looks like if WebRTC is not supported by browser

Plugin Install Bar in IE and Safari In versions of IE and Safari that don't support WebRTC natively, AdapterJS will suggest to the user to install the Temasys WebRTC plugin for Mac or Windows when you try to access the getUserMedia or RTCPeerConnection API.

Using and setting up the AdapterJS dependency

Required initialization

In order to execute any WebRTC related code, you have to wait for the AdapterJS.webRTCReady function to trigger the callback parameter as it is triggered when the WebRTC interfaces is ready to be used by the browser.

Note that AdapterJS.onwebrtcready is deprecated.

AdapterJS.webRTCReady(function(isUsingPlugin) {
    // The WebRTC API is ready.
    //isUsingPlugin: true is the WebRTC plugin is being used, false otherwise
    getUserMedia(constraints, successCb, failCb);
});

To find more information about how to optimize your application for the Temasys WebRTC Plugin, read it more in the Temasys WebRTC Plugin Documentation.


Functionality

Polyfills

  • RTCPeerConnection
  • RTCIceCandidate
  • RTCSessionDescription
  • MediaStreamTrack
  • navigator.getUserMedia
  • navigator.mediaDevices.getUserMedia
  • navigator.mediaDevices.enumerateDevices

Helper functions

attachMediaStream(element, stream) : None
  getUserMedia({ audio: true, video: true }, function (stream) {
    attachMediaStream(videoElm, stream);
  }, ...);

This is taken over for compatibility with the original adapter.js from Google. Feeds a MediaStream object into video and audio tags. Calling attachMediaStream(element, null) will detach any existing stream from the element. The stream will still be running and available to be attached to any rendering element.

  • Parameters:

    • element: (DOM Element) The <video> or <audio> DOM element.
    • stream: (DOM Element) The MediaStream object.
  • Returns: None

reattachMediaStream(elementFrom, elementTo) : None
  reattachMediaStream(videoWithStreamElm, videoElm);

This is taken over for compatibility with adapter.js from Google. Feeds a MediaStream from one video or audio tag into another.

  • Parameters:

    • elementFrom: (DOM Element) The <video> or <audio> DOM element that has been attached with an existing MediaStream from attachMediaStream() API.
    • elementTo: (DOM Element) The <video> or <audio> DOM element to copy the MediaStream from the elementFrom DOM element parameter.
  • Returns: None

createIceServer(url, username, password) : JSON
  createIceServers('turn:ice.com', 'xxx', 'xxxx');

This creates a valid RTCIceServer object from one url, username and password.

  • Parameters:

    • url: (String) The ICE server url.
    • username: (String) The ICE server username only for TURN.
    • password: (String) The ICE server credential password only for TURN.
  • Returns: JSON

    • url: (String) The ICE server url.
    • username: (String) The ICE server username only for TURN.
    • credential: (String) The ICE server credential password only for TURN.
createIceServers(urls, username, password) : JSON
  createIceServers(['turn:ice.com', 'stun:ice.com'], 'xxx', 'xxxx');

This creates a valid an Array of RTCIceServer objects for browsers that supports the .url format only when constructing new RTCPeerConnection({ iceServers: [xx] }).

  • Parameters:

    • urls: (Array) The array of ICE server urls.
      • #index: (String) The ICE server url.
    • username: (String) The ICE server username only for TURN.
    • password: (String) The ICE server credential password only for TURN.
  • Returns: Array

    • #index: (JSON) The ICE server object.
      • url: (String) The ICE server url.
      • username: (String) The TURN ICE server username.
      • credential: (String) The TURN ICE server credential password.
checkIceConnectionState(peerId, iceConnectionState, callback) : None

Note that this function has been deprecated since expected triggering of ICE connection states should not be handled.

peerConnection.oniceconnectionstatechange = function () {
  checkICEConnectionState(peerId, peerConnection.iceConnectionState, function (updatedIceConnectionState) {
    // do Something every time there's a new state ['checking', 'connected', 'completed']
  });
};

This handles all the iceConnectionState return value differences cross-browsers when RTCPeerConnection.oniceconnectionstate is fired.

Expected outcome should be: checking > connected > completed. What was received in Chrome/Opera as offerer: checking > completed > completed. What was received in Chrome/Opera as answerer: checking > connected. What was received in Firefox as offerer: checking > connected. What was received in Firefox as answerer: checking > connected.

  • Parameters:

    • peerId: (String) The unique identifier for the peer to store all fired states tied specifically to this peer.
    • iceConnectionState: (String) The RTCPeerConnection.iceConnectionState received.
    • callback: (Function) The callback fired once the parsing is completed.
      • Returned callback parameters:
        • updatedIceConnectionState: (String) The updated RTCPeerConnection.iceConnectionState that Peer should trigger.
  • Returns: None

checkMediaDataChannelSettings(peerAgentBrowser, peerAgentVersion, callback, constraints) : None

Note that this function has been deprecated as this is based off older versions of browsers interoperability in which is lower than our minimum supported browser versions.

// Right now we are not yet doing the offer. We are just checking if we should be the offerer instead of the other peer
checkMediaDataChannelSettings(peerAgentBrowser, peerAgentVersion
  function (beOfferer, unifiedOfferConstraints) {
  if (beOfferer) {
    peerConnection.createOffer(function (offer) {
      // success
    }, function (error) {
      // failure
    }, unifiedOfferConstraints);
  } else {
    // let the other peer do the offer instead.
  }
}, inputConstraints);

This handles all MediaStream and DataChannel differences for interoperability cross-browsers. The method has to be called before creating the offer to check if peer should create the offer.

For some older (20+) versions of Firefox and Chrome MediaStream interoperability, MozDontOfferDataChannel has to be used, and hence Firefox cannot establish a DataChannel connection as an offerer, and results in no DataChannel connection. To achieve both MediaStream and DataChannel connection interoperability, Chrome or other browsers has to be the one creating the offer.

  • Parameters:

    • peerAgentBrowser: (String) The browser agent or name. E.g. Chrome.
    • peerAgentVersion: (Number) The browser agent version. E.g. 35.
    • callback: (Function) The callback fired after the check has been made.
      • Returned callback parameters:
        • beOfferrer: (Boolean) Returns a true or a false. If true, user should do the offer. If false, inform the other peer to do the offer.
        • unifiedConstraints: (JSON) The updated RTCOfferOptions offer constraints for interoperability.
    • constraints: (JSON) The RTCOfferOptions offer constraints.
  • Returns: None


Helper variables

webrtcDetectedBrowser : String

This displays the browser WebRTC browser agent name. Defined as null when it's an unknown WebRTC supported browser.

  • Expected values:

    • "chrome": Chrome or Chromium implemented browser.
    • "firefox": Firefox browser.
    • "safari": Safari browser.
    • "edge": Edge browser.
    • "opera": Opera browser.
    • "IE": IE browser.
    • "bowser": Bowser browser.
webrtcDetectedVersion : Number

This displays the browser WebRTC browser agent version. Defined as null when it's an unknown WebRTC supported browser.

webrtcMinimumVersion : Number

This displays the browser WebRTC browser agent version. Defined as null when it's an unknown WebRTC supported browser. The values follows the Supported Browsers minimum versions section.

webrtcDetectedType : String

This displays the browser WebRTC implementation type. Defined as null when it's an unknown WebRTC supported browser.

  • Expected values:

    • "webkit": Webkit implementation of webrtc.
    • "moz": Mozilla implementation of webrtc.
    • "plugin": Temasys plugin implementation of webrtc.
    • "ms": Edge implementation of webrtc (polyfilled from ORTC).
webrtcDetectedDCSupport : String

This displays the browser WebRTC Datachannel support type. Defined as null when it's an unknown WebRTC supported browser or when WebRTC supported browser does not support the Datachannel API.

  • Expected values:

    • "SCTP": SCTP enabled datachannel.
    • "RTP": RTP enabled datachannel.

Using Screensharing functionality

Note that the Firefox add-on not installed detection will not work with window.navigator.mediaDevices on Firefox browsers to prevent errors.

AdapterJS 0.12.0+ offers cross-browser screensharing in Chrome 34+, Firefox 33+ and with a licensed copy of our Temasys WebRTC Plugin in IE 9+ and Safari 7.1+. (For plugin licensing interest please contact sales (a) temasys.com.sg)

To use the screensharing functionality, reference publish/adapter.screenshare.js and add the mediaSource: 'window' setting to the video media constraints. This requires HTTPS!

Note that the the mediaSource property takes in String, or Array in which multiple sources is supported for Chrome / Opera sources. For Firefox, if an Array is provided, it takes the first item in the Array.

Example:

window.navigator.getUserMedia({
  video: {
    mediaSource: 'window' || 'screen' || ['tab', 'audio'] || ['window' || 'screen']
  }
}, function (stream) {
  console.log('Received stream', stream);
}, function (error) {
  console.log('Failed getting stream', error);
});

List of valid screensharing (mediaSource) sources:

Sources Description Browsers that supports it
"window" Fetches the list of application windows Chrome, Opera, Firefox, IE, Safari
"screen" Fetches the list of display screens Chrome, Opera, Firefox, IE, Safari
"tab" Fetches the list of browser tabs Chrome, Opera
"browser" Fetches the list of browser windows Firefox (Requires to configure about:config to enable media.getusermedia.browser.enabled)
"application" Fetches the list of applications Firefox
"camera" Fetches the list of cameras Firefox
["tab", "audio"] Fetches "tab" with its audio Chrome, Opera (Enable constraints.audio to retrieve tab audio)
["window", "screen"] Fetches "screen" and "window" Chrome, Opera, IE, Safari
["window", "screen", "tab"] Fetches "screen", "window" and "tab" Chrome, Opera
["window", "tab"] Fetches "window" and "tab" Chrome, Opera
["screen", "tab"] Fetches "screen" and "tab" Chrome, Opera
AdapterJS.WebRTCPlugin.plugin.screensharingKeys.screenOrWindow Same as ["window","screen"] IE, Safari
AdapterJS.WebRTCPlugin.plugin.screensharingKeys.screensharingKey Deprecated. Same as ["window","screen"]. IE, Safari
AdapterJS.WebRTCPlugin.plugin.screensharingKeys.screen Same as "screen" IE, Safari
AdapterJS.WebRTCPlugin.plugin.screensharingKeys.window Deprecated. Same as "window". IE, Safari

Configuring extension settings:

To configure your AdapterJS screensharing extensions, configure this before referencing the script for adapter.screenshare.js as an example:

var AdapterJS = {};

/**
 * Configure extension settings. For developers using "adapter.screenshare.js"
 */
AdapterJS.extensionInfo = {
  chrome: {
    // Configure the extension ID for Chrome
    extensionId: 'xxx',
    // Configure the extension webstore link for Chrome
    extensionLink: 'xxx',
    // Configure the extension iframe link (detectRTC) for Chrome - for older Chrome extension codebase
    iframeLink: 'xxx'
  },
  firefox: {
    // Configure the addon link (legacy if available and uploaded before restriction) for Firefox 51 and below
    extensionLink: 'xxx'
  },
  opera: {
    // Configure the extension ID for Opera
    extensionId: 'xxx',
    // Configure the extension webstore link for Opera
    extensionLink: 'xxx'
  }
};

Setup this project

  1. Install or update to at lest version 0.10.26 of node and version 1.4.6of npm.
  2. Install grunt-cli. (See: http://gruntjs.com/getting-started)
  3. Run npm install to install dev dependencies.
  4. Run npm install -g browserify and npm install -g testling (might require sudo) to install the necessary tools to test locally

Development

  • Run grunt jshint to run jshint on its own.
  • Run grunt publish to create production debug and minified copies of the source/ files with the webrtc/adapter dependency compiled in the publish/ folder.

Folders

publish/

Contains the debug version of the library and a minified copy of it.

  • adapter.debug.js: Contains the compiled AdapterJS with webrtc/adapter dependency.
  • adapter.min.js: Contains the minified and production ready version of adapter.debug.js.
  • adapter.screenshare.js: Contains the compiled adapter.debug.js with Screensharing polyfill changes.
  • adapter.screenshare.min.js: Contains the minified and production ready version of adapter.screenshare.js.
source/

Contains the AdapterJS library development files.

  • adapter.js: Contains the polyfills and APIs for Temasys Plugin WebRTC interface.
  • adapter.screenshare.js: Contains the polyfills for Screensharing changes.
  • pluginInfo.js: Contains the Temasys Plugin information. Modify this for your custom Temasys Plugin.
tests/

Run grunt test to generate the published versions of adapter.js (same as grunt publish) and run the automated test suite on it. You can configure the browser to test in Gruntfile.js (see the karma target). You can also run grunt karma to run the test and bypass the publish step.

(Mac only) If you are testing the Temasys WebRTC Plugin, you can run osascript tests/mac.watcher.scpt to automatically validate the permission popup.

License

APACHE 2.0 - http://www.apache.org/licenses/LICENSE-2.0.html