Re: eliminate xl_heap_visible to reduce WAL (and eventually set VM on-access) - Mailing list pgsql-hackers

From Andrey Borodin
Subject Re: eliminate xl_heap_visible to reduce WAL (and eventually set VM on-access)
Date
Msg-id BB281DFF-1260-433C-89A3-14953297D800@yandex-team.ru
Whole thread Raw
In response to Re: eliminate xl_heap_visible to reduce WAL (and eventually set VM on-access)  (Kirill Reshke <reshkekirill@gmail.com>)
List pgsql-hackers
Well, converting to TAP seems feasible to me. This will make test more stable.

> On 30 Jan 2026, at 14:59, Kirill Reshke <reshkekirill@gmail.com> wrote:
>
> I use 'autovacuum-worker-start' injection point to 'disable'
> autovacuum until test is done

If you just set autovacuum off - the test will be executed on bf animals that has no injection points.

Also
x4mmm@x4mmm-osx postgres % git am ~/Downloads/v1-0001-Reimplement-regression-tests-from-21796c267-as-TA.patch
Applying: Reimplement regression tests from 21796c267 as TAP-test.
.git/rebase-apply/patch:148: trailing whitespace.
# Check that vacuum phase I does not need to modify the heap buffer.  warning: 1 line adds whitespace errors.

There are some typos: "flakky", "introduces", "actuvity", "actaully", "hilding".

Test descriptions are off: you used 'page_header returned as expected' when it's actually about
pg_visibility_map_summary.

And, of course, total lack of comments is not good.

However, in principle approach seems good to me.


(FWIW I'm looking into patches 0003-0005 of v33, I'll post when I'll find some nits)


Best regards, Andrey Borodin.


pgsql-hackers by date:

Previous
From: "Andy Fan"
Date:
Subject: Re:Re: A out of date comment of WaitForWALToBecomeAvailable
Next
From: David Rowley
Date:
Subject: Re: More speedups for tuple deformation