[utils] Decode HTML5 entities

[youtube-dl] / youtube_dl / utils.py

diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index 229de4b39a0ee6408277fcf1f999b53c1d78ee78..f77ab865093604f217ab6b2da824f340327df652 100644 (file)
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -39,6 +39,7 @@ from .compat import (
     compat_chr,
     compat_etree_fromstring,
     compat_html_entities,
+    compat_html_entities_html5,
     compat_http_client,
     compat_kwargs,
     compat_parse_qs,
@@ -456,12 +457,19 @@ def orderedSet(iterable):
     return res
 
 
-def _htmlentity_transform(entity):
+def _htmlentity_transform(entity_with_semicolon):
     """Transforms an HTML entity to a character."""
+    entity = entity_with_semicolon[:-1]
+
     # Known non-numeric HTML entity
     if entity in compat_html_entities.name2codepoint:
         return compat_chr(compat_html_entities.name2codepoint[entity])
 
+    # TODO: HTML5 allows entities without a semicolon. For example,
+    # '&Eacuteric' should be decoded as 'Éric'.
+    if entity_with_semicolon in compat_html_entities_html5:
+        return compat_html_entities_html5[entity_with_semicolon]
+
     mobj = re.match(r'#(x[0-9a-fA-F]+|[0-9]+)', entity)
     if mobj is not None:
         numstr = mobj.group(1)
@@ -486,7 +494,7 @@ def unescapeHTML(s):
     assert type(s) == compat_str
 
     return re.sub(
-        r'&([^;]+);', lambda m: _htmlentity_transform(m.group(1)), s)
+        r'&([^;]+;)', lambda m: _htmlentity_transform(m.group(1)), s)
 
 
 def get_subprocess_encoding():