My server downloaded diggnation three times over the weekend. This is not good. Turns out that the guys at Revision3 have tested one of the greatest assumptions of podcasting -- that only 1 URL will be put in an RSS feed for a given item. They are now dynamically generating the URLs every time you request the RSS feed. This means that if your podcast client uses the URL as a unique identifier (such as podracer/bashpodder in my case) it will continually download the file.
I had to edit podracer to fix this, not sure what the best solution is but I filtered the dynamic parts out of the URLs leaving the domain name and the file name as the unique identifier -- anyway heads up.