Comment by tapia

Nice work. Maybe you could do some preprocessing of the XML data, so that you actually have a diff of the content and not the whole XML block.