Re: [PATCH] Teach pg_waldump to extract FPIs from the WAL

On Tue, Nov 8, 2022 at 4:45 PM Justin Pryzby <pryzby@telsasoft.com> wrote:
>
> On Mon, Nov 07, 2022 at 05:01:01PM -0600, David Christensen wrote:
> > Hi Justin et al,
> >
> > Enclosed is v5 of this patch which now passes the CirrusCI checks for
> > all supported OSes. I went ahead and reworked the test a bit so it's a
> > little more amenable to the OS-agnostic approach for testing.
>
> Great, thanks.
>
> This includes the changes that I'd started a few months ago.
> Plus adding the test which was missing for meson.

Cool, will review, thanks.

> +        format:
<literal><replaceable>LSN</replaceable>.<replaceable>TSOID</replaceable>.<replaceable>DBOID</replaceable>.<replaceable>RELNODE</replaceable>.<replaceable>BLKNO</replaceable></literal>
>
> I'd prefer if the abbreviations were "reltablespace" and "datoid"

Sure, no issues there.

> Also, should the test case call pg_relation_filenode() rather than using
> relfilenode directly ?  Is it a problem that the test code assumes
> pagesize=8192 ?

Both good points. Is pagesize just exposed via
`current_setting('block_size')` or is there a different approach?

David