Tom Lane writes:
> Let's use int16 (2-byte integers) as format selector codes; this seems a
> reasonable compromise between bandwidth and flexibility. As of 7.4 the
> only supported values will be 0 = text and 1 = binary, but future versions
> can add more codes.
I think the "future versions" in this are going to be making this choice a
datatype-specific session state. How can we make this transition
smoother? Maybe 0 can be default, 1 text, 2 binary? Or maybe we should
implement a simple version of the session-state system right now with only
two predefined transform groups for each type? Then we don't need to send
a format to the server at all, and the messages to the client would
contain an OID field in the RowDescription message.
--
Peter Eisentraut peter_e@gmx.net