That looks a lot like an endian issue, unfortunately at the time it was written I didn't have a lot of access to big-endian machines so it didn't get a whole lot of testing in big-endian environments. There are a whole lot of things on my Geo::ShapeFile todo list, and this is at the top, but sadly I don't have much time to work through that list as my current job doesn't involve mapping and I've recently been busy with other things.

If you can send me a sample of your shape data, it may help in tracking down some of those endian issues (or at the very least I could verify whether that is indeed the issue.)