On Fri, Jul 28, 2023 at 07:00:01AM +0300, Alexander Lakhin wrote:
> 28.07.2023 05:17, Noah Misch wrote:
> >On Tue, Jun 20, 2023 at 07:49:52AM -0400, Russell Foster wrote:
> >>/*
> >>* We don't use Unix-domain sockets on Windows by default, even if the
> >>* build supports them. (See comment at remove_temp() for a reason.)
> >>* Override at your own risk.
> >>*/
> >>
> >>Is there some sort of race condition in the SSPI code that sometimes
> >>doesn't gracefully finish/close the connection when the backend
> >>decides to exit due to error?
> >No. remove_temp() is part of test driver "pg_regress". Non-test usage is
> >unaffected. Even for test usage, folks have reported no failures from the
> >cause mentioned in the remove_temp() comment.
>
> It seems to me that it's just another manifestation of bug #16678 ([1]).
> See also commits 6051857fc and 29992a6a5.
>
> [1] https://www.postgresql.org/message-id/flat/16678-253e48d34dc0c376%40postgresql.org
That was about a bug that appears when using TCP sockets. The remove_temp()
comment involves code that doesn't run when using TCP sockets. I don't think
they can be manifestations of the same phenomenon.