Log Message:
Update to a prerelease 4.2 snapshot, as 4.1 no longer copes
particularly well with many feeds and there's no indication that
a release is imminent. From the changelog:

* Support for parsing microformats, including rel=enclosure, rel=tag,
  XFN, and hCard.
* Updated the whitelist of acceptable HTML elements and attributes based
  on the latest draft of the HTML 5 specification.
* Support for CSS Sanitization. (Previous versions of Universal Feed
  Parser simply stripped all inline styles.) Many thanks to Sam Ruby for
  implementing this, despite my insistence that it was impossible.
* Support for SVG Sanitization.
* Support for MathML Sanitization. Many thanks to Jacques Distler for
  patiently debugging this feature.
* IRI support for every element that can contain a URI.
* Ability to disable relative URI resolution.
* Command-line arguments and alternate serializers, for manipulating
  Universal Feed Parser from shell scripts or other non-Python sources.
* More robust parsing of author email addresses, misencoded win-1252
  content, rel=self links, and better detection of HTML content in
  elements with ambiguous content types.