Re: making pg_regress less noisy by removing boilerplate

Hi,

On 2022-02-22 08:55:23 -0500, Andrew Dunstan wrote:
> I'm pretty sure all my Windows machines with buildfarm animals are
> sufficiently modern except the XP machine, which only builds release 10
> nowadays.

Cool.


> What's involved in moving to require Unix socket support?

It works today (the CI scripts use it on windows for example).

But it's awkward because make_temp_sockdir() defaults to /tmp/ if TMPDIR isn't
set. Which it is not by default on windows. There's PG_REGRESS_SOCK_DIR, which
kind of works for the msvc build, because pg_regress tests aren't run
concurrently (whereas tap tests can be run concurrently with
PROVE_FLAGS-j).

I think we just make make_temp_sockdir() a tad smarter. Perhaps by lifting the
code in src/bin/psql/command.c:do_edit() to src/port/path.c or such?

Greetings,

Andres Freund