[XML-SIG] PyXML 0.6.6 is released

[Martin v. Loewis]

> Which bugs does it fix? 

That I had problems to establish, that's why I decided not to write a
complete list. It essentially contains corrections to 4DOM and the
boolean type. You'll find that a diff relative to 0.6.5 is quite
short. Here are a few details, if you find that a specific change is
not covered, please post it to the list.

- do not compile xmlwf into pyexpat.
- support floating point conversions in boolean.c better
- remove dangling semicolons in boolean.c
- expose BooleanType from boolean.c
- properly process PI in sgmlop.c
- fix isSameNode signature in FtNode
- support __iter__ in NamedNodeMap
- check for localName, not tagName, in minidom.getElementsByTagNameNS;
  actually return a result there
- add a minidom.CDATASection class
- fix typo in minitraversal
- record namespace mappings using the DOM namespaces namespace in pulldom
- fix encoding processing in dom.ext.Printer
- in ext.PrettyPrint and ext.StripHTML, do not assume that the Document
  has an ownerDocument
- in SeekNSS, deal with default namespaces
- in HtmlLib, do not empty the document first
- fix signature of ext.reader.Sax.unparsedEntityDecl,
- Sax2: delay unparsed entitiy processing if there is no _ownerDoc. 
- Sax2Lib: add self parameter to startNamespaceDeclScope,
  endNamespaceDeclScope
- ext.reader.Sgmlop: Do not create doctypes or PIs for HTML DOMs.
- correct spelling of the pound entity.
- marshal.generic: try to give more reproducable results across
  Python versions

Hope this helps,
Martin