Alvaro Herrera <> writes:
> Robert Haas wrote:
>> In this case,
>> I'll blame the fact that we allow a role to be dropped while there are
>> users connected using that role.

> Actually, my first comment when Pavan mentioned this on IM was that we
> should look into fixing that problem sometime.  It's not terribly urgent
> since it doesn't seem to hurt anything too badly, but it's still a bug.

My feeling is that it's going to be unreasonably expensive.  Are we
going to take a lock every time we call a SECURITY DEFINER function?

                        regards, tom lane

Sent via pgsql-hackers mailing list (
To make changes to your subscription:

Reply via email to