A number of tests attempt to dereference URLs in the www.example.com domain, and expect this to fail. An example of such a test is K2-SeqDocFunc-5, which fetches the URL 

http://www.example.com/example.com/example.org/does/not/exist/doesNotExist/works-mod.xml

and expects error FODC0002.

In fact the domain name www.example.com is owned by IANA, and the request for this URL actually delivers a document, whose true URL is http://www.iana.org/domains/example/. This document (at the time of writing) claims to be XHTML, but is invalid (it contains an attribute value that is not in quotes). So the test does in fact throw error FODC0002, but not for the expected reasons, and it could stop working as expected if ever IANA fix their XHTML validation problems.

I suggest we try and find a more reliable way to fetch non-existent documents.