Re: [PATCH] Fix pg_upgrade test from v10

On Tue, Jul 05, 2022 at 02:08:24PM -0500, Justin Pryzby wrote:
> ..since it tries to apply all the *.patch files to the master branch, one after
> another.  For branches other than master, I suggest to name the patches *.txt
> or similar.  Or, just focus for now on allowing upgrades *to* master.  I'm not
> sure if anyone is interested in patching test.sh in backbranches.  I'm not
> sure, but there may be more interest to backpatch the conversion to TAP
> (322becb60).

I am fine to do something for v15 and HEAD, as TAP makes that slightly
easier.  Now, this patch is far from being complete, as it would still
generate a lot of diffs.  Some of them cannot be easily avoided, but
there are areas where it is straightforward to do so:
- pg_dump needs to use --extra-float-digits=0 when dumping from a
version strictly older than v12.
- This does nothing for the headers ond footers of the logical dumps
that are version-dependent.  One simple thing that can be done here is
to remove the comments from the logical dumps, something that the
buildfarm code already does.

That's the kind of things I already proposed on this thread, aimed at
improving the coverage, and this takes care of more issues than what's
proposed here:
https://www.postgresql.org/message-id/flat/Yox1ME99GhAemMq1@paquier.xyz

I'll rebase my patch to include fixes for --wal-segsize and
--allow-group-access when using versions older than v11.
--
Michael