fix(WebVTT): Add support to  , ‎ and ‏ (shaka-project#4920)

lib/text/vtt_text_parser.js

@@ -940,8 +940,8 @@ shaka.text.VttTextParser = class {
   }
 
   /**
-   * This method converts the HTML entities &, <, >, ", and
-   * ' in string to their corresponding characters.
+   * This method converts the HTML entities &, <, >, ", ',
+   *  , ‎ and ‏ in string to their corresponding characters.
    *
    * @param {!string} input
    * @return {string}
@@ -955,10 +955,13 @@ shaka.text.VttTextParser = class {
       '>': '>',
       '"': '"',
       ''': '\'',
+      ' ': '\u{a0}',
+      '‎': '\u{200e}',
+      '‏': '\u{200f}',
     };
 
     // Used to match HTML entities and HTML characters.
-    const reEscapedHtml = /&(?:amp|lt|gt|quot|#(0+)?39);/g;
+    const reEscapedHtml = /&(?:amp|lt|gt|quot|#(0+)?39|nbsp|lrm|rlm);/g;
     const reHasEscapedHtml = RegExp(reEscapedHtml.source);
     // This check is an optimization, since replace always makes a copy
     if (input && reHasEscapedHtml.test(input)) {

test/text/vtt_text_parser_unit.js

@@ -852,11 +852,11 @@ describe('VttTextParser', () => {
   it('support escaped html payload', () => {
     verifyHelper(
         [
-          {startTime: 20.1, endTime: 40.505, payload: '"Test & 1"'},
+          {startTime: 20.1, endTime: 40.505, payload: '"Test & 1"\u{a0}'},
         ],
         'WEBVTT\n\n' +
         '00:00:20.100 --> 00:00:40.505\n' +
-        '"Test & 1"',
+        '"Test & 1" ',
         {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
   });