At Tue, 29 Aug 2023 19:28:28 +0200, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote in
> On 2023-Aug-29, Kyotaro Horiguchi wrote:
>
> > Attached is the initial version of the patch. It prevents "CREATE
> > TABLE" from executing if there is an inconsisntent not-null
> > constraint. Also I noticed that "ALTER TABLE t ADD NOT NULL c NO
> > INHERIT" silently ignores the "NO INHERIT" part and fixed it.
>
> Great, thank you. I pushed it after modifying it a bit -- instead of
> throwing the error in MergeAttributes, I did it in
> AddRelationNotNullConstraints(). It seems cleaner this way, mostly
> because we already have to match these two constraints there. (I guess
I agree that it is cleaner.
> you could argue that we waste catalog-insertion work before the error is
> reported and the whole thing is aborted; but I don't think this is a
> serious problem in practice.)
Given the rarity and the speed required, I agree that early-catching
is not that crucial here. Thanks for clearing that up.
regardes.
--
Kyotaro Horiguchi
NTT Open Source Software Center