On May 2, 2005, at 09:34 , Tom Lane wrote:
> I think you could demonstrate that if the spec is "make it look
> like the
> original query was retyped as source each time", then *every* DDL
> change
> in the database potentially requires invalidating every cached plan.
> I don't find that a desirable spec.
I agree. It seems to me that if someone is doing that sort of
chicanery, then one should not use prepared statements. IOW, I would
view it as an edge case.
Regards,
David