Re: BUG #5611: SQL Function STABLE promoting to VOLATILE

Brian Ceccarelli <bceccarelli(at)net32(dot)com> writes:
> STABLE no longer means STABLE. This behavior is killing my performance. I am getting 500% to 30000% increase in latency.

You seem to be under the illusion that "stable" is a control knob for a
function result cache. It is not, and never has been. Postgres doesn't
do function result caching.

If you've constructed your app in such a way that it depends on not
inlining SQL set-returning functions, it's fairly easy to defeat that.
>From memory, marking a SRF as either strict or volatile will do it
(although volatile might cause you problems elsewhere --- I suspect
your design is pretty brittle in this area).

regards, tom lane