Andres Freund <andres@anarazel.de> writes:
> Wouldn't it be better to reset an appropriate context after each
> invocation? Yes, that'd require some care to manage the lifetime of
> tuples returned by triggers, but that seems OK?
I'm not sure that we can change much here without effectively breaking
the trigger-function API. I doubt that there's enough win to be had
to justify forcing external PLs etc to change.
Having said that, I recall that that API was kind of a pain in the
rear when I was redoing plpgsql's handling of composite variables
last year. Don't recall details right now (ENOCAFFEINE). Maybe
a wholesale rethink would be justified? But I'm not excited about
just twiddling things at the margin.
regards, tom lane