Re: NOT NULL NOT ENFORCED

From: solai v <solai(dot)cdac(at)gmail(dot)com>
To: jian he <jian(dot)universality(at)gmail(dot)com>
Cc: Álvaro Herrera <alvherre(at)kurilemu(dot)de>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: NOT NULL NOT ENFORCED
Date: 2026-09-28 04:34:00
Message-ID: CAF0whuc--DUH=J7aCpbC77id1Vq1CyFbK_xe4qPr8GAVRKuh8g@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi

I tested the v5 patch on a fresh cluster and verified the following cases:
- NOT NULL NOT ENFORCED constraint creation and catalog state
- NULL insertion with NOT ENFORCED constraints
- ALTER CONSTRAINT and VALIDATE CONSTRAINT behavior
- Duplicate NOT NULL constraint handling
- NOT VALID, ENFORCED and NOT ENFORCED combinations
- NO INHERIT behavior
- Inheritance and multiple inheritance cases
- CREATE TABLE LIKE INCLUDING CONSTRAINTS
- Partitioned table and partition constraint behavior
- pg_dump preservation of NOT NULL NOT ENFORCED constraints

Some of the test results were:
- Created a NOT NULL NOT ENFORCED constraint successfully and verified
that it was stored as not enforced and not validated.
- Successfully inserted a NULL value into a column having a NOT NULL
NOT ENFORCED constraint.
- Verified that attempting to change the constraint to ENFORCED or
validate it resulted in the expected errors.- Verified that duplicate
NOT NULL constraints on the same column were handled as expected.
- Tested inheritance cases with both enforced and not-enforced NOT
NULL constraints. The resulting child constraints and attnotnull state
were as expected.
- Verified that CREATE TABLE ... LIKE INCLUDING CONSTRAINTS preserved
the NOT ENFORCED state, and NULL insertion into the copied table was
allowed.
- Tested partitioned table cases with NOT ENFORCED constraints and
verified that the constraint state was preserved correctly.
- Verified that pg_dump preserved the NOT NULL NOT ENFORCED definition
in the generated dump.

I also ran the complete regression test suite successfully. All tests passed.
Based on the above testing, the tested cases behaved as expected with
the v5 patch.

Regards,
Solai

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Alexandre Felipe 2026-09-28 04:45:55 Re: aio: worker: Free SMGR objects when idle
Previous Message Jelte Fennema-Nio 2026-09-28 03:42:04 Re: pgindent to ignore build directories