feat: Allow VTT files with erroneous linebreaks (#2394)

Bad linebreaks will now cause cues to be skipped (with a warning), rather than throwing an error. Closes #2358 Co-authored-by: Álvaro Velad Galván <[email protected]>
shaka-project · Feb 9, 2023 · 9b1c614 · 9b1c614
1 parent dc9bd27
commit 9b1c614
Show file tree

Hide file tree

Showing 3 changed files with 48 additions and 31 deletions.
diff --git a/lib/text/vtt_text_parser.js b/lib/text/vtt_text_parser.js
@@ -380,11 +380,9 @@ shaka.text.VttTextParser = class {
     let end = VttTextParser.parseTime_(parser);
 
     if (start == null || expect == null || end == null) {
-      throw new shaka.util.Error(
-          shaka.util.Error.Severity.CRITICAL,
-          shaka.util.Error.Category.TEXT,
-          shaka.util.Error.Code.INVALID_TEXT_CUE,
-          'Could not parse cue time range in WebVTT');
+      shaka.log.alwaysWarn(
+          'Failed to parse VTT time code. Cue skipped:', id, text);
+      return null;
     }
 
     start += timeOffset;

diff --git a/lib/util/text_parser.js b/lib/util/text_parser.js
@@ -21,7 +21,7 @@ shaka.util.TextParser = class {
      * @const
      * @private {string}
      */
-    this.data_ = data;
+    this.data_ = data || '';
 
     /** @private {number} */
     this.position_ = 0;

diff --git a/test/text/vtt_text_parser_unit.js b/test/text/vtt_text_parser_unit.js
@@ -8,7 +8,6 @@ describe('VttTextParser', () => {
   const Cue = shaka.text.Cue;
   const CueRegion = shaka.text.CueRegion;
   const originalLogWarning = shaka.log.warning;
-  const anyString = jasmine.any(String);
 
   /** @type {!jasmine.Spy} */
   let logWarningSpy;
@@ -161,38 +160,30 @@ describe('VttTextParser', () => {
   });
 
   it('rejects invalid time values', () => {
-    errorHelper(shaka.util.Error.Code.INVALID_TEXT_CUE,
+    verifyHelper([],
         'WEBVTT\n\n00.020    --> 0:00.040\nTest',
-        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0},
-        anyString);
-    errorHelper(shaka.util.Error.Code.INVALID_TEXT_CUE,
+        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
+    verifyHelper([],
         'WEBVTT\n\n0:00.020  --> 0:00.040\nTest',
-        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0},
-        anyString);
-    errorHelper(shaka.util.Error.Code.INVALID_TEXT_CUE,
+        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
+    verifyHelper([],
         'WEBVTT\n\n00:00.20  --> 0:00.040\nTest',
-        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0},
-        anyString);
-    errorHelper(shaka.util.Error.Code.INVALID_TEXT_CUE,
+        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
+    verifyHelper([],
         'WEBVTT\n\n00:100.20 --> 0:00.040\nTest',
-        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0},
-        anyString);
-    errorHelper(shaka.util.Error.Code.INVALID_TEXT_CUE,
+        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
+    verifyHelper([],
         'WEBVTT\n\n00:00.020 --> 0:00.040\nTest',
-        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0},
-        anyString);
-    errorHelper(shaka.util.Error.Code.INVALID_TEXT_CUE,
+        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
+    verifyHelper([],
         'WEBVTT\n\n00:00:00:00.020 --> 0:00.040\nTest',
-        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0},
-        anyString);
-    errorHelper(shaka.util.Error.Code.INVALID_TEXT_CUE,
+        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
+    verifyHelper([],
         'WEBVTT\n\n00:61.020 --> 0:00.040\nTest',
-        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0},
-        anyString);
-    errorHelper(shaka.util.Error.Code.INVALID_TEXT_CUE,
+        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
+    verifyHelper([],
         'WEBVTT\n\n61:00.020 --> 0:00.040\nTest',
-        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0},
-        anyString);
+        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
   });
 
   it('supports vertical setting', () => {
@@ -1223,6 +1214,34 @@ describe('VttTextParser', () => {
         {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
   });
 
+  it('supports an extra newline inside the cue body', () => {
+    verifyHelper(
+        [
+          {startTime: 20, endTime: 40, payload: 'Test'},
+          {startTime: 40, endTime: 50, payload: 'Test2'},
+        ],
+        'WEBVTT\n\n' +
+        '00:00:20.000 --> 00:00:40.000\n' +
+        'Test\n\nExtra line\n\n' +
+        '00:00:40.000 --> 00:00:50.000\n' +
+        'Test2',
+        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
+  });
+
+  it('supports an extra newline before the cue body', () => {
+    verifyHelper(
+        [
+          {startTime: 20, endTime: 40, payload: ''},
+          {startTime: 40, endTime: 50, payload: 'Test2'},
+        ],
+        'WEBVTT\n\n' +
+        '00:00:20.000 --> 00:00:40.000\n' +
+        '\nTest\n\n' +
+        '00:00:40.000 --> 00:00:50.000\n' +
+        'Test2',
+        {periodStart: 0, segmentStart: 0, segmentEnd: 0, vttOffset: 0});
+  });
+
   /**
    * @param {!Array} cues
    * @param {string} text