As I've said, I managed to get the full expansion of the macros in question, and everything looked correct, so I was looking for the actual number that the compiler was evaluating the fully expanded macro to.
It turned out that I was misevaluating the order of operations when I looked at the...