| From: | "Jelte Fennema-Nio" <me(at)jeltef(dot)nl> |
|---|---|
| To: | "Nikolay Samokhvalov" <nik(at)postgres(dot)ai> |
| Cc: | "pgsql-hackers" <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: Python/pytest test framework take two |
| Date: | 2026-11-04 21:07:14 |
| Message-ID: | DMMQEFKMPQ5S.144FVVXW8SPSI@jeltef.nl |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Thu, 29 Oct 2026 at 20:04, Nikolay Samokhvalov <nik(at)postgres(dot)ai> wrote:
> 1. result(timeout=...) and exception(timeout=...) mark a future consumed
> before waiting. A caught timeout can therefore allow connection reuse
> while the query still runs. The first patch fixes this: eight baseline
> failures become passes (20/20 total).
Hmm, yeah that's a bug (a small one though). I solved it differently
than your suggestion. (see attached v3)
> 2. The bulk pg_visibility conversion (8a8f008f7596) misses
> close_portal("") after begin / select txid_current(), retaining a snapshot
> that psql releases. The second patch adds two passing comparison tests
> for this documented behavior.
This pg_visibility conversion wasn't part of the patchset I submitted on
the list. That referenced commit is part of a branch on my own postgres
fork that I didn't submit for review because it was not ready. So I'm
not adding this one.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Kirk Wolak | 2026-11-04 21:07:39 | Re: PG19 regression in ecd9288624a: JSON_ARRAY() indexes that go stale |
| Previous Message | Lukas Fittl | 2026-11-04 20:50:33 | Re: Fix cumulative WAL statistics double-counting from parallel workers |