Joe Conway <mail@joeconway.com> writes:
> Should I just accept that the cursor advances on a row type mismatch
> error, fix using the attached patch and adding SCROLL to dblink_open(),
> or something else? Any opinions out there?
I would go with accepting (and documenting) the cursor advance. Trying
to undo it seems too risky, and it's not like this is a situation that
would arise in a properly-debugged application anyway. I can't see
expending a great amount of effort on it --- protecting against a crash
seems sufficient.
regards, tom lane