Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
> It strikes me that we really need to try reconnecting to the shared
> memory here several times, and maybe the backoff need to increase each
> time.
Adding a backoff would make the code significantly more complex, with
no gain that I can see. Just loop a few times around the
one-second-sleep-and-retry logic.
I concur with Greg's opinion that the need for a sleep here at all
is pretty fishy, but I doubt anyone really cares enough to find out
exactly what's happening (and it being Windows, there may be no better
solution anyway ...)
regards, tom lane