Re: INSERT ... ON CONFLICT DO SELECT [FOR ...] take 2

I’ve fixed the remaining issues mentioned here, and submitted a new thread here: https://www.postgresql.org/message-id/d631b406-13b7-433e-8c0b-c6040c4b4663%40Spark
Would appreciate any help in reviewing!

/Viktor
On 2 Sep 2025 at 20:56 +0200, v(at)viktorh(dot)net <v(at)viktorh(dot)net>, wrote:
> Hello, I was working on my own patch for the same thing, until I found this was already there.
> I think this would be very useful for a lot of people.
> Do you need any help moving this forward Anderas? I have both tests and docs written, although not for the FOR UPDATE part.
>
> > On 25 Jun 2025, at 13:39, Dean Rasheed <dean(dot)a(dot)rasheed(at)gmail(dot)com> wrote:
> >
> > On Mon, 12 May 2025 at 19:33, Andreas Karlsson <andreas(at)proxel(dot)se> wrote:
> > >
> > > I have fixed that one and some other issues locally and will submit a
> > > new version in a while after I have added more tests because you are
> > > very correct in that a big issue with my last version of the patch was
> > > the big lack of tests and lack of making sure all features which
> > > interact with UPSERT actually worked with my changes. Plus some islation
> > > tests would be nice to have.
> > >
> >
> > +1. Don't forget to move the CF entry to the next open CF.
> >
> > FYI, over on [1] I proposed more tests and doc updates for RLS. I
> > think those updates might make it easier to test and document the RLS
> > aspects of this patch.
> >
> > [1] https://www.postgresql.org/message-id/flat/CAEZATCWqnfeChjK=n1V_dYZT4rt4mnq+ybf9c0qXDYTVMsy8pg(at)mail(dot)gmail(dot)com
> >
> > Regards,
> > Dean
> >
> >
> >
>