Robert Haas <robertmhaas@gmail.com> writes:
> On Wed, Aug 25, 2010 at 11:46 AM, Simon Riggs <simon@2ndquadrant.com> wrote:
>> I have a horrible nagging feeling this breaks something.
> Well, the only place it's used at the moment is in the drop-object
> code. Not sure if that might be enough to jog your memory?
Also, it's pretty hard to see how this could break anything that isn't
already broken, because AcceptInvalidationMessages() already happens
in the other cases in that code path.
regards, tom lane