Re: addFkRecurseReferencing use unassigned fkconstraint->fk_with_period value

On Thu, Oct 8, 2026 at 10:22 PM Paul A Jungwirth
<pj(at)illuminatedcomputing(dot)com> wrote:
>
> On Mon, Sep 28, 2026 at 11:18 PM Haibo Yan <tristan(dot)yim(at)gmail(dot)com> wrote:
> >
> changed the wording to "with pre-existing
> > rows" and moved the NOT VALID / NOT ENFORCED tests after the
> > pg_get_constraintdef checks.
> >
> > There are no code changes from v1.
>
> This looks good to me too. Andres reported the VALIDATE CONSTRAINT
> case in [0], but it's nice to have fixes to all three cases in one
> place.
>
> If you want, here is a v3 with a less verbose commit message. Also I
> added the note suggested by Andrew, which seems worth including in
> release notes.
>
> [0] https://www.postgresql.org/message-id/sesqxzfcujkcnrwpir5xvhy7iamhsbjiqt6r2xqmb7hqkulte6%40bbuc42qzanzs
>
> Yours,
>
> --
> Paul ~{:-)
> pj(at)illuminatedcomputing(dot)com

Hi Paul,

Thanks for picking this up and preparing v3.

The changes look good to me. One minor wording nit in the commit
message: "despite covering violating rows" might be clearer as
"despite containing violating rows".

Otherwise, v3 looks good to me.

Thanks,
Haibo