Robert Haas <robertmhaas@gmail.com> writes:
> On Wed, Aug 1, 2018 at 3:36 PM, Andres Freund <andres@anarazel.de> wrote:
>> What precisely are you proposing?
> Inserting AcceptInvalidationMessages() in some location that
> guarantees it will be executed at least once per SQL statement. I
> tentatively propose the beginning of parse_analyze(), but I am open to
> suggestions.
What will that accomplish that the existing call in transaction start
doesn't? It might make the window for concurrency issues a bit narrower,
but it certainly doesn't close it.
regards, tom lane