Forkparser

namespaces, as opposed to 2.6 when I said I did but didn't really;
sanitize HTML markup within some elements; added mxTidy support (if
installed) to tidy HTML markup within some elements; fixed indentation
bug in _parse_date (FazalM); use socket.setdefaulttimeout if available
(FazalM); universal date parsing and normalization (FazalM): 'created', modified',
'issued' are parsed into 9-tuple date format and stored in 'created_parsed',
'modified_parsed', and 'issued_parsed'; 'date' is duplicated in 'modified'
and vice-versa; 'date_parsed' is duplicated in 'modified_parsed' and vice-versa

nested divs within content (JohnD); fixed missing sys import (JohanS);
fixed regular expression to capture XML character encoding (Andrei);
added support for Atom 0.3-style links; fixed bug with textInput tracking;
added support for cloud (MartijnP); added support for multiple
category/dc:subject (MartijnP); normalize content model: 'description' gets
description (which can come from description, summary, or full content if no
description), 'content' gets dict of base/language/type/value (which can come
from content:encoded, xhtml:body, content, or fullitem);
fixed bug matching arbitrary Userland namespaces; added xml:base and xml:lang
tracking; fixed bug tracking unknown tags; fixed bug tracking content when
<content> element is not in default namespace (like Pocketsoap feed);
resolve relative URLs in link, guid, docs, url, comments, wfw:comment,
wfw:commentRSS; resolve relative URLs within embedded HTML markup in
description, xhtml:body, content, content:encoded, title, subtitle,
summary, info, tagline, and copyright; added support for pingback and
trackback namespaces

textInput, and also to return the character encoding (if specified)

inline <xhtml:body> and <xhtml:div> as used in some RSS 2.0 feeds

User-Agent (otherwise urllib2 sends two, which confuses some servers)

snapshot of July 1 <http://www.intertwingly.net/blog/1506.html>; changed
project name

2.5 - 7/25/2003 - MAP - changed to Python license (all contributors agree);
removed unnecessary urllib code -- urllib2 should always be available anyway;
return actual url, status, and full HTTP headers (as result['url'],
result['status'], and result['headers']) if parsing a remote feed over HTTP --
this should pass all the HTTP tests at <http://diveintomark.org/tests/client/http/>;
added the latest namespace-of-the-week for RSS 2.0