base32, base64, and ascii85 all need to check for particular characters:
CONVERTER logic in the decoding functions was misinterpreted for this
case.  CONVERTER is really for turning elements of the input sequence
into integers and was inappropriately being used to try to convert the
special characters into integers, which didn't work when decoding octet
vectors.

For those checks, explicitly check the type of sequence we're decoding
from and take the appropriate transformation on the special character.

This is more consistent with what Common Lisp already
does (e.g. READ-SEQUENCE), as well as being slightly more efficient.

Use dummy constructors for the decoding bits, since we're not using the
new scheme there yet.  base85 tests do not yet pass.

That's what I get for attempting to be clever.