Just in case anyone was curious, I ended up using a regex to store the original doctype tag. After this, I did my LibXML processing. Finally, I restored the original doctype, using a substition regex, after LibXML processing was complete. This is not the ideal solution, but it is the best that I have for now.