Re: Parallelized polymorphic aggs, and aggtype vs aggoutputtype

David Rowley <david.rowley@2ndquadrant.com> writes:
> I've gone and implemented the dummy argument approach for
> deserialization functions.

How do you feel about the further idea of locking down the signatures
to be exactly "serialize(internal) returns bytea" and "deserialize(bytea,
internal) returns internal", and removing pg_aggregate.aggserialtype?
I don't see very much value in allowing any other nominal transmission
type besides bytea; and the less flexibility in these function signatures,
the less chance of confusion/misuse of other internal-related functions.
        regards, tom lane