fix(WebVTT): Tags in the WebVTT subtitle are not parsed (#4960)

Fixes #4956
shaka-project · Feb 9, 2023 · 15afb4b · 15afb4b
1 parent 762b1e7
commit 15afb4b
Show file tree

Hide file tree

Showing 2 changed files with 30 additions and 7 deletions.
diff --git a/lib/text/vtt_text_parser.js b/lib/text/vtt_text_parser.js
@@ -369,8 +369,7 @@ shaka.text.VttTextParser = class {
     end += timeOffset;
 
     // Get the payload.
-    const payload = VttTextParser.htmlUnescape_(
-        text.slice(1).join('\n').trim());
+    const payload = text.slice(1).join('\n').trim();
 
     let cue = null;
     if (styles.has('global')) {
@@ -427,7 +426,7 @@ shaka.text.VttTextParser = class {
         const childNode = childNodes[0];
         if (childNode.nodeType == Node.TEXT_NODE ||
             childNode.nodeType == Node.CDATA_SECTION_NODE) {
-          rootCue.payload = payload;
+          rootCue.payload = VttTextParser.htmlUnescape_(payload);
           return;
         }
       }
@@ -438,7 +437,7 @@ shaka.text.VttTextParser = class {
       rootCue.nestedCues = cues;
     } else {
       shaka.log.warning('The cue\'s markup could not be parsed: ', payload);
-      rootCue.payload = payload;
+      rootCue.payload = VttTextParser.htmlUnescape_(payload);
     }
   }
 
@@ -682,7 +681,7 @@ shaka.text.VttTextParser = class {
         }
         if (text.length > 0) {
           const textCue = nestedCue.clone();
-          textCue.payload = text;
+          textCue.payload = VttTextParser.htmlUnescape_(text);
           cues.push(textCue);
         }
         isFirst = false;

diff --git a/test/text/vtt_text_parser_unit.js b/test/text/vtt_text_parser_unit.js
@@ -776,11 +776,35 @@ describe('VttTextParser', () => {
   it('support escaped html payload', () => {
     verifyHelper(
         [
-          {startTime: 20.1, endTime: 40.505, payload: '"Test & 1"\u{a0}'},
+          {
+            startTime: 20.1,
+            endTime: 40.505,
+            payload: '"Test & 1"\u{a0}',
+          },
+          {
+            startTime: 41,
+            endTime: 42,
+            payload: '',
+            nestedCues: [
+              {
+                startTime: 41,
+                endTime: 42,
+                payload: 'Test',
+                fontStyle: Cue.fontStyle.ITALIC,
+              },
+              {
+                startTime: 41,
+                endTime: 42,
+                payload: '&',
+              },
+            ],
+          },
         ],
         'WEBVTT\n\n' +
         '00:00:20.100 --> 00:00:40.505\n' +
-        '&quot;Test &amp; 1&quot;&nbsp;',
+        '&quot;Test &amp; 1&quot;&nbsp;\n\n' +
+        '00:00:41.000 --> 00:00:42.000\n' +
+        '<i>Test</i>&amp;',
         {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
   });