66.6. Database Page Layout

On 8 Jul 2026, at 12:43, PG Doc comments form <noreply@postgresql.org> wrote:

The following documentation comment has been logged on the website:

Page: https://www.postgresql.org/docs/18/storage-page-layout.html
Description:

The second field contains the page checksum if -k are enabled.

-k -- This is enabled by default; use --no-data-checksums to disable
checksums.

Why does this need to be clarified?

I'm not entirely sure which part you are referring to, but I agree that the
current wording isn't very good.

"The second field contains the page checksum if -k are enabled."

I propose the attached patch which changes the sentence to "The second field
contains the page checksum if data checksums are enabled." with a link to the
Checksums chapter in the documentation instead.

--
Daniel Gustafsson

On Thu, 2026-07-09 at 23:58 +0200, Daniel Gustafsson wrote:

I propose the attached patch which changes the sentence to "The second field
contains the page checksum if data checksums are enabled." with a link to the
Checksums chapter in the documentation instead.

That patch looks good to me.

Yours,
Laurenz Albe

On 10 Jul 2026, at 08:33, Laurenz Albe <laurenz.albe@cybertec.at> wrote:

On Thu, 2026-07-09 at 23:58 +0200, Daniel Gustafsson wrote:

I propose the attached patch which changes the sentence to "The second field
contains the page checksum if data checksums are enabled." with a link to the
Checksums chapter in the documentation instead.

That patch looks good to me.

I took another look at this today to close the thread and found that there were
one more occurrence of this pattern stemming from 67846550dc6d which removed
the initdb xreflabels. The attached patches fix the other one as well as
changes the link in the amcheck docs to point to the main checksums page which
is more relevant than the (now default) initdb option. Patches for master down
to v18 attached.

--
Daniel Gustafsson

On Wed, 2026-07-15 at 15:17 +0200, Daniel Gustafsson wrote:

I took another look at this today to close the thread and found that there were
one more occurrence of this pattern stemming from 67846550dc6d which removed
the initdb xreflabels.  The attached patches fix the other one as well as
changes the link in the amcheck docs to point to the main checksums page which
is more relevant than the (now default) initdb option.  Patches for master down
to v18 attached.

Referencing "checksums" rather than "app-initdb-data-checksums" is an improvement
in all these places, and the only other reference to "app-initdb-data-checksums"
in the documentation is good the way it is and should stay.

So the patches look good to me.

Yours,
Laurenz Albe

On 15 Jul 2026, at 16:50, Laurenz Albe <laurenz.albe@cybertec.at> wrote:

On Wed, 2026-07-15 at 15:17 +0200, Daniel Gustafsson wrote:

I took another look at this today to close the thread and found that there were
one more occurrence of this pattern stemming from 67846550dc6d which removed
the initdb xreflabels. The attached patches fix the other one as well as
changes the link in the amcheck docs to point to the main checksums page which
is more relevant than the (now default) initdb option. Patches for master down
to v18 attached.

Referencing "checksums" rather than "app-initdb-data-checksums" is an improvement
in all these places, and the only other reference to "app-initdb-data-checksums"
in the documentation is good the way it is and should stay.

So the patches look good to me.

Thanks for review!

--
Daniel Gustafsson