[xml] more document fragment issues

[Aaron Patterson <aaron patterson gmail com>]

When parsing HTML document fragments (via xmlParseInNodeContext),
libxml2 seems to add surrounding p tags to dangling text fields.

For example, parsing a fragment like this: "sir<div>phil</div>collins"
will result in 3 text nodes and 3 element nodes when using an HTML
document.  However, if you're using an XML document, the fragment will
result in 3 text nodes and only one element node.  This doesn't seem
like correct behavior.

I've put together a sample program that uses the latest git builds:

  http://gist.github.com/292934

Are there any bits I can twiddle to deal with this?  Thanks!

-- 
Aaron Patterson
http://tenderlovemaking.com/