Tom Lane wrote:
> The fix actually uses the "expected tuple desc" that's now passed by
> ExecMakeTableFunctionResult as the target descriptor. This should mean
> that it'd be possible to support plpgsql functions returning RECORD, but
> I didn't have time to look into that. Anyone want to try?
I haven't really looked too closely at Neil plpgsql stuff, so I'm a bit
concerned with the spin-up time I'd need to figure this out. But if Neil
doesn't show up and volunteer between now and Saturday morning, I'll
take a look.
Joe