>> The execution level
>> itself was almost trivial; it's getting the tuplestore reference
>> through the parse analysis and planning phases that is painful for
>> me.
> I am not sure why you think we would need to refer the tuplestore in
> the parse analysis and planner phases. It seems that we would need
> them only in execution phase. Or may be I didn't understand your
> point.
Ah I think I understand now. That might be because you are thinking of
having an infrastructure common to triggers and materialized views,
right ?