Re: Volatile functions under Memoize node

On 20/9/2024 04:36, David Rowley wrote:
> On Fri, 20 Sept 2024 at 04:47, Andrei Lepikhov <lepihov(at)gmail(dot)com> wrote:
> I'm not sure if it's a good idea to penalise your case when we're not
> all that consistent to start with. Is this causing some sort of
> breakage?
I skimmed the code entries with such checks and found out that the
initial reason was to avoid index scans, with a reason that such a
routine should be applied to each tuple of the table.
The second reason - to postpone expression evaluation (9118d03) - is
also reasonable for me. It was introduced to be consistent with the
clause's syntactical level in the SQL.
It seems to follow the same idea as disabling subquery pull-ups: to
avoid multiple evaluations and change the syntactical level.
At the same time, Material doesn't care about volatility. So, what was
the idea behind the commit 990c365 you added?

--
regards, Andrei Lepikhov