On Thu, Jul 3, 2025 at 11:54 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> I hadn't noticed (or maybe I forgot) this thread, so when the
> same problem was reported at [1] I just went ahead and pushed the
> submitted patch, which is only cosmetically different from your 0001.
> Apologies for treading on your toes.
No worries, as long as it's fixed I'm happy!
(And many thanks to Greg for the review; sorry for not getting to it
fast enough.)
> As for the question of how to test this sort of thing, I'm not
> too excited about the narrow-gauge test case your 0002 proposes.
> What I did for manual testing in [1] was to hack the postgres_fdw
> tests to connect using hostaddr instead of the default. I think
> formalizing that sort of approach would yield much better coverage.
I agree that overriding connection defaults probably gets us better
overall coverage -- I just think I got pushback in the past for adding
"multipliers" in that way. But I won't argue against test coverage as
long as we get it in the end. :D
That said, I am planning to get noisier about the lack of "TCP suite".
The number of tests we've discarded just because we don't have a
current place to put them keeps slowly growing, and my long-term
intent with 0002 was to actually add a new place for them. Whatever
formalization we choose, let's please keep a TCP-only cluster
somewhere instead of forcing people to try to find a least-bad suite
to slot new tests into.
Thanks!
--Jacob
