Re: Support EXCEPT for TABLES IN SCHEMA publications

On Thu, Jul 30, 2026 at 1:45 PM shveta malik <shveta(dot)malik(at)gmail(dot)com> wrote:
>
> On Fri, Jul 17, 2026 at 9:50 AM Peter Smith <smithpb2250(at)gmail(dot)com> wrote:
> >
> > Hi Nisha.
> >
> > I noticed multiple places where these patches want to discover the
> > root partition table OID, but they are not interested in the ancestor
> > tree.
> >
> > I've created another thread/patch [1] to introduce a
> > `get_partition_root` C function. I think this could simplify some of
> > your code by removing the requirement to declare and free the unwanted
> > ancestor lists.
> >
> > Please take a look and give your +1 there if you think it would be useful.
> >
> > ======
> > [1] https://www.postgresql.org/message-id/CAHut%2BPtc73UpWSn4teMyMxQoHwGJjRwhWiu%3DhGJqAj5k2AOXEw%40mail.gmail.com
> >
>
> Nisha, Peter,
>
> Do you think we should make get_partition_root's patch as 0001 in this
> thread and rebase rest of patches atop it? That way duplicated code in
> many places will be removed. Another idea is to make
> get_partition_root's patch as last patch here, but I prefer it as
> first patch. Thoughts?
>

FWIW, I don't think anything needs to be shuffled around -- we just
need things to happen in the right order.

The other patch was deliberately separated because it includes some
changes which IMO if put into patch 0001 of this thread would've just
received complaints like "these changes are not in scope".

Meanwhile, AFAIK that other patch is already completed (can you please
mark the CF entry as "ready-for-committer"). So, it can be
independently pushed any time -- the sooner the better.

Then, this EXCEPT code can be rebased to remove some duplicated code.

So,
1. Yes, the get_partition_root patch should come first
2. No, don't make it patch 0001 of this thread; Just try to get it pushed first

======
Kind Regards,
Peter Smith.
Fujitsu Australia