Обсуждение: Re: refactor AlterDomainAddConstraint (alter domain add constraint)

Hello,

On 2025-Jan-15, jian he wrote:

> we cannot error out AlterDomainAddConstraint for cases like ALTER
> DOMAIN ADD CHECK NO INHERIT.
> because "NO INHERIT" is actually a separate Constraint Node, and
> AlterDomainAddConstraint
> can only handle one Constraint node.

I had forgotten this thread, and I ended up implementing a different
solution for this issue, which I just posted at
  https://postgr.es/m/202503101758.ipn3g64twfye@alvherre.pgsql

I like my patch better than this approach because it allows us to solve
the same problem as it appears in other parts of the grammar, and also
because it avoids the bit fiddling which is harder to maintain later on.
If you'd care to have a look at it, I'd appreciate it.

Thanks

-- 
Álvaro Herrera        Breisgau, Deutschland  —  https://www.EnterpriseDB.com/

On 10.03.25 19:37, Alvaro Herrera wrote:
> On 2025-Jan-15, jian he wrote:
>> we cannot error out AlterDomainAddConstraint for cases like ALTER
>> DOMAIN ADD CHECK NO INHERIT.
>> because "NO INHERIT" is actually a separate Constraint Node, and
>> AlterDomainAddConstraint
>> can only handle one Constraint node.
> 
> I had forgotten this thread, and I ended up implementing a different
> solution for this issue, which I just posted at
>    https://postgr.es/m/202503101758.ipn3g64twfye@alvherre.pgsql
> 
> I like my patch better than this approach because it allows us to solve
> the same problem as it appears in other parts of the grammar, and also
> because it avoids the bit fiddling which is harder to maintain later on.
> If you'd care to have a look at it, I'd appreciate it.

Where are we on this?  Which of the two patches should we pursue?