Amit Kapila <> writes:
> Attached patch tightens the error handling.  However, on debugging, I
> found that CreateFileMapping() always set error code to 0 on success.

Ah, interesting.

> Now, it is quite possible
> that error code is set to 0 on success in my windows environment
> (Win7) and doesn't work in some other environment.  In any case, if we
> want to go ahead and don't want to rely on CreateFileMapping(), then
> attached patch should suffice the need.

Yeah, seeing that this is not mentioned in MS' documentation I don't
think we should rely on it.  The patch looks fine to me, pushed.

                        regards, tom lane

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

Reply via email to