Did you read the article? That downloaded file was valid XML file. The problem was that it did not contain data they expected it will always contain. They did not check existence of nodes and blindly requested them. And, by the way, they ARE using libxml2 to parse those files.

As far I know russian. Google translated it correctly.