> On 24 Mar 2023, at 00:33, Michael Paquier <michael@paquier.xyz> wrote:
>
> On Thu, Mar 23, 2023 at 10:46:56PM +0100, Daniel Gustafsson wrote:
>> I'm fairly convinced it's a timeout in the interactive psql session. Given how
>> ugly the use of that is I'm sort of waiting for Andres' refactoring patch [0] to
>> commit this such that I can rewrite the test in a saner and more robust way.
>
> FWIW, I'd be OK here even if you don't have a test for libpq in the
> first change as what you have sent is already testing for the core
> machinery in scram-common.c. You could always add one later.
Yeah, that's my fallback in case we are unable to get the TAP refactoring done
in time for the end of the CF/feature freeze.
I've actually ripped out the test in question in the attached v9 to have it
ready and building green in CFbot.
--
Daniel Gustafsson