Anyhow, I rewrote it from scratch, and doing that is never as tedious as I think. Indeed, as many people will tell you, tossing out your first try at something and rewriting, is usually a pretty good idea.
So now I have the hardest parts done of the new Pod parser module. It's downhill from here, just a bunch of busywork. The one remaining notable hurdle that I see, is the processing of L<...> codes. That's a mess and then some.