Fix to htmlentityreplace_errors when a character has no named entity. Also now takes into account the fact that entity names could end with a semicolon.

tbroyer · tbroyer · commit fc63d15ed2cd · 2007-07-05T23:30:30.000Z
--HG--
extra : convert_revision : svn%3Aacbfec75-9323-0410-a652-858a13e371e0/trunk%40867
diff --git a/src/html5lib/serializer/htmlserializer.py b/src/html5lib/serializer/htmlserializer.py
@@ -33,11 +33,12 @@ def htmlentityreplace_errors(exc):
         if isinstance(exc, (UnicodeEncodeError, UnicodeTranslateError)):
             res = []
             for c in exc.object[exc.start:exc.end]:
-                c = encode_entity_map.get(c)
-                if c:
+                e = encode_entity_map.get(c)
+                if e:
                     res.append("&")
-                    res.append(c)
-                    res.append(";")
+                    res.append(e)
+                    if not e.endswith(";"):
+                        res.append(";")
                 else:
                     res.append(c.encode(exc.encoding, "xmlcharrefreplace"))
             return (u"".join(res), exc.end)
