fix(iframe_api): toggle audio/video race condition

If toggle audio or video is executed too early and the local
tracks don't exist we fail to execute the operation. Now we store
the mute state and we are executing it after the tracks are
created

conference.js

 
 const eventEmitter = new EventEmitter();
 
-let room, connection, localAudio, localVideo;
+let room, connection, localAudio, localVideo,
+    initialAudioMutedState = false, initialVideoMutedState = false;
 
 /**
  * Indicates whether extension external installation is in progress or not.
                 analytics.init();
                 return createInitialLocalTracksAndConnect(options.roomName);
             }).then(([tracks, con]) => {
+                tracks.forEach(track => {
+                    if((track.isAudioTrack() && initialAudioMutedState)
+                        || (track.isVideoTrack() && initialVideoMutedState)) {
+                        track.mute();
+                    }
+                });
                 logger.log('initialized with %s local tracks', tracks.length);
                 con.addEventListener(
                     ConnectionEvents.CONNECTION_FAILED,
         return this.audioMuted;
     },
     /**
-     * Simulates toolbar button click for audio mute. Used by shortcuts and API.
+     * Simulates toolbar button click for audio mute. Used by shortcuts
+     * and API.
+     * @param {boolean} force - If the track is not created, the operation
+     * will be executed after the track is created. Otherwise the operation
+     * will be ignored.
      */
-    toggleAudioMuted () {
+    toggleAudioMuted (force = false) {
+        if(!localAudio && force) {
+            initialAudioMutedState = !initialAudioMutedState;
+            return;
+        }
         this.muteAudio(!this.audioMuted);
     },
     /**
     },
     /**
      * Simulates toolbar button click for video mute. Used by shortcuts and API.
+     * @param {boolean} force - If the track is not created, the operation
+     * will be executed after the track is created. Otherwise the operation
+     * will be ignored.
      */
-    toggleVideoMuted () {
+    toggleVideoMuted (force = false) {
+        if(!localVideo && force) {
+            initialVideoMutedState = !initialVideoMutedState;
+            return;
+        }
         this.muteVideo(!this.videoMuted);
     },
     /**

modules/API/API.js

     commands = {
         "display-name":
             APP.conference.changeLocalDisplayName.bind(APP.conference),
-        "toggle-audio": APP.conference.toggleAudioMuted.bind(APP.conference),
-        "toggle-video": APP.conference.toggleVideoMuted.bind(APP.conference),
-        "toggle-film-strip": APP.UI.toggleFilmstrip,
+        "toggle-audio": () => APP.conference.toggleAudioMuted(true),
+        "toggle-video": () => APP.conference.toggleVideoMuted(true),
+        "toggle-film-strip": APP.UI.toggleFilmStrip,
         "toggle-chat": APP.UI.toggleChat,
         "toggle-contact-list": APP.UI.toggleContactList,
         "toggle-share-screen":