Re: Eliminating SPI / SQL from some RI triggers - take 3

On Wed, Apr 1, 2026 at 5:51 PM Amit Langote <amitlangote09(at)gmail(dot)com> wrote:
> On Wed, Apr 1, 2026 at 12:54 AM Junwang Zhao <zhjwpku(at)gmail(dot)com> wrote:
> > + if (riinfo->fpmeta == NULL)
> > + {
> > + /* Reload to ensure it's valid. */
> > + riinfo = ri_LoadConstraintInfo(riinfo->constraint_id);
> >
> > I was thinking of wrapping the reload in a conditional check like
> > `!riinfo->valid`, since `riinfo` can be valid even when `fpmeta == NULL`.
> > However, `if (riinfo->fpmeta == NULL)` should rarely be true, so the
> > unconditional reload is harmless, and the code is cleaner.
> >
> > +1 to the fix.
>
> Thanks for checking.
>
> I have just pushed a slightly modified version of that.
>
> > > 0002 is the rebased batching patch.
> >
> > The change of RI_FastPathEntry from storing riinfo to fk_relid
> > makes sense to me. I'll do another review on 0002 tomorrow.
>
> Here's another version.
>
> This time, I have another fixup patch (0001) to make FastPathMeta
> self-contained by copying the FmgrInfo structs it needs out of
> RI_CompareHashEntry rather than storing pointers into it. This avoids
> any dependency on those cache entries remaining stable. I'll push
> that once the just committed patch has seen enough BF animals.

Pushed.

> 0002 is rebased over that.

Rebased again.

--
Thanks, Amit Langote