Re: track generic and custom plans in pg_stat_statements

On Wed, Jul 30, 2025 at 02:05:09PM -0500, Sami Imseih wrote:
> > > The term "NOT_SET" makes me itch a little bit, even if there is an
> > > existing parallel with OverridingKind. Perhaps your proposal is OK,
> > > still how about "UNKNOWN" instead to use as term for the default?
> > +1 to "UNKNOWN".
>
> We currently use both UNKNOWN and NOT_SET in different places.
> However, I'm okay with using UNKNOWN, and I've updated it in v16.

Okay, applied both patches to get all that done, then.

Patch 0002 was not without turbulences:
- Order of the tests in meson.build and Makefile, where plancache
should be placed before the cleanup. The squashing tests get that
wrong, actually..
- Some tweaks to the style of the SQL queries in the tests, two tweaks
in the wording of the docs.
- Lack of coverage for oldextversions for the transfer from 1.12 to
1.13.
- PGSS_FILE_HEADER should be bumped. We have been very bad at that
since 2022, myself included, better later than never.
- I have decided to remove the cases with plan_cache_mode = auto, due
to a lack of predictibility depending on what the backend may decide
to use.
- For the last case with EXPLAIN + procedures, I have added an extra
"toplevel" in the ORDER BY clause, to ensure the order of the results.
"toplevel" was not required for the first three cases, where we don't
track the non-top-level queries.
- Added one round of code indentation.
--
Michael