http://dryice.name/blog Tue, 31 Aug 2010 12:01:11 +0000 http://backend.userland.com/rss092 en I need to change this line from true_encoding = http_encoding or 'us-ascii' to true_encoding = http_encoding or xml_encoding or 'us-ascii' for those buggy sites that don't obey the standard. And set content type to text/* but don't offer a charset, set their encoding in the xml file. Share This http://dryice.name/blog/python/feedparsertext-content-type/ According to the XML spec http://www.w3.org/TR/REC-xml/#NT-EncodingDecl whitespace is allowed around the quotes of encoding Here is a simple patch: --- /usr/ports/textproc/py-feedparser/work/feedparser/feedparser.py.old Sat Jul 2 16:17:11 2005 +++ /usr/ports/textproc/py-feedparser/work/feedparser/feedparser.py Sat Jul 2 16:18:25 2005 @@ -2101,7 +2101,7 @@ else: # ASCII-compatible pass - xml_encoding_match = re.compile('^<\?.*encoding=[\'"](.*?)[\'"].*\?>').match(xml_data) + ... http://dryice.name/blog/python/feedparserwhitespace/ Looks Feedparser was written with Python 2.3. With python 2.4, the CJKcodecs is included in the official release. So the line import cjkcodecs.aliases should be changed to import encodings.aliases Share This http://dryice.name/blog/python/feedparserencoding/