Peter Eisentraut wrote:
>
> I know we left this issue open, but I came to the conclusion that it would
> be wiser to keep it like it used to be in that operator comments should be
> indexed on the underlying functions. The reason is simply that there is a
> a one-to-one relationship between operators and their function, so we'd
> end up writing everything double with little purpose. That would mean
> you'd have to tweak your code a little.
>
If that's the way you want it, so it shall be. I have to write up
a diff for pg_dump this weekend anyways to generate the
appropriate COMMENT ON statements for version 7.0.
Mike Mascari