Horton

Re: pg_dump does not dump domain not-null constraint's comments

Álvaro Herrera <alvherre@kurilemu.de>

View thread
From: Álvaro Herrera <alvherre@kurilemu.de>
To: jian he <jian.universality@gmail.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2025-07-21T09:48:52Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Fix pg_dump COMMENT dependency for separate domain constraints.

  2. pg_dump: include comments on not-null constraints on domains, too

  3. Fix dumping of comments on invalid constraints on domains

  4. Catalog domain not-null constraints

  5. In pg_dump, don't dump a stats object unless dumping underlying table.

  6. Enable CHECK constraints to be declared NOT VALID

On 2025-Jul-18, jian he wrote:

> one minor issue in getDomainConstraints:
> 
>  for (int i = 0, j = 0; i < ntups; i++)
> {
>         char        contype = (PQgetvalue(res, i, i_contype))[0];
>         ....
>         constraint->contype = *(PQgetvalue(res, i, i_contype));
> }
> 
> for the same code simplicity,
> ``constraint->contype = *(PQgetvalue(res, i, i_contype));``
> can change to
> `` constraint->contype = contype;  ``
> ?
> other than that, looks good to me.

Thanks, changed it that way.

I noticed some other issues however.  First, you had removed the contype
comparisons in repairDependencyLoop; I put them back several versions of
the patch ago, but I had mistakenly made them reference the wrong array
item -- in all three places.  Doh.

Second, the name comparisons to determine whether to list the constraint
name in the "CREATE DOMAIN .. NOT NULL" syntax was wrong: it was using
fmtId() around the constraint name, but of course that would have never
worked, because we're comparing to the original string, not a quoted
name.  So if you had a domain called, say
  CREATE DOMAIN "my domain" AS int NOT NULL;

then pg_dump would have said
  CREATE DOMAIN "my domain" AS int CONSTRAINT "my domain_not_null" NOT NULL;

even though listing the name is unnecessary.  This wouldn't have made
the dump fail, so it's borderline; but it would be /slightly/ ugly.

Lastly, I made dumpDomain print unadorned "NOT NULL" if a pg_constraint
row for the constraint is not found and yet we observe typnotnull set.
This can be considered catalog corruption (the row should be there), but
I think this is better than having pg_dump just crash if that row is not
there.

-- 
Álvaro Herrera               48°01'N 7°57'E  —  https://www.EnterpriseDB.com/
Voy a acabar con todos los humanos / con los humanos yo acabaré
voy a acabar con todos (bis) / con todos los humanos acabaré ¡acabaré! (Bender)