WIP: ASCIIHexDecode and ASCII85Decode filters

Work in progress for the time being, but I think it's ready to be looked at. I want to reduce code duplication in the semantic actions for partial groups of ascii85 characters. Additionally, it would probably be better if the a85string rule flattens the entire thing into a sequence of bytes. And the parsers allocated might need to be explicitly freed once the decoder function stops running.