[utils] js_to_json: various improvements

now JS object literals like { /* " */ 0: ",]\xaa<\/p>", } will be correctly converted to JSON.

test/test_utils.py

@@ -640,6 +640,18 @@ class TestUtil(unittest.TestCase):
         on = js_to_json('{"abc": "def",}')
         self.assertEqual(json.loads(on), {'abc': 'def'})
 
+        on = js_to_json('{ 0: /* " \n */ ",]" , }')
+        self.assertEqual(json.loads(on), {'0': ',]'})
+
+        on = js_to_json(r'["<p>x<\/p>"]')
+        self.assertEqual(json.loads(on), ['<p>x</p>'])
+
+        on = js_to_json(r'["\xaa"]')
+        self.assertEqual(json.loads(on), ['\u00aa'])
+
+        on = js_to_json("['a\\\nb']")
+        self.assertEqual(json.loads(on), ['ab'])
+
     def test_extract_attributes(self):
         self.assertEqual(extract_attributes('<e x="y">'), {'x': 'y'})
         self.assertEqual(extract_attributes("<e x='y'>"), {'x': 'y'})

youtube_dl/utils.py

@@ -1914,24 +1914,26 @@ def js_to_json(code):
         v = m.group(0)
         if v in ('true', 'false', 'null'):
             return v
-        if v.startswith('"'):
+        elif v.startswith('/*') or v == ',':
-            v = re.sub(r"\\'", "'", v[1:-1])
+            return ""
-        elif v.startswith("'"):
+
-            v = v[1:-1]
+        if v[0] in ("'", '"'):
-            v = re.sub(r"\\\\|\\'|\"", lambda m: {
+            v = re.sub(r'(?s)\\.|"', lambda m: {
-                '\\\\': '\\\\',
-                "\\'": "'",
                 '"': '\\"',
-            }[m.group(0)], v)
+                "\\'": "'",
+                '\\\n': '',
+                '\\x': '\\u00',
+            }.get(m.group(0), m.group(0)), v[1:-1])
+
         return '"%s"' % v
 
-    res = re.sub(r'''(?x)
+    return re.sub(r'''(?sx)
-        "(?:[^"\\]*(?:\\\\|\\['"nu]))*[^"\\]*"|
+        "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
-        '(?:[^'\\]*(?:\\\\|\\['"nu]))*[^'\\]*'|
+        '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
-        [a-zA-Z_][.a-zA-Z_0-9]*
+        /\*.*?\*/|,(?=\s*[\]}])|
+        [a-zA-Z_][.a-zA-Z_0-9]*|
+        [0-9]+(?=\s*:)
         ''', fix_kv, code)
-    res = re.sub(r',(\s*[\]}])', lambda m: m.group(1), res)
-    return res
 
 
 def qualities(quality_ids):