Implement convenient API for reading contents, attrs out of Saxon nodes.

[pobocks]

Currently, the only part of Nokogiri's actual "interact with the processed XML" that is supported is XPath. Everything else is still raw Saxon. I was thinking that it would be nice to support, say:

node = my_doc.xpath('//something/that/pulls/at/least/one/node').first
node['my_attr'] # attr value as string
node.content    # text content of el and subels
node.xml          # string rep of fragment consisting of nodes.

I'm not sure of the right way to go about it - easiest way (at least for attrs and content on individual nodes) would be monkey-patching XdmNode, and maybe the Saxon sequence type. Since monkey-patches on Java objects don't propagate back into Java-land, and there's fairly little overlap between the Saxon API and the Noko-lite style stuff I'd want to implement, I think this might actually be reasonable?