public inbox for [email protected]
help / color / mirror / Atom feedFrom: Junwang Zhao <[email protected]>
To: [email protected]
To: [email protected]
Subject: Re: BUG #19442: PL/pgSQL: domain over composite type bypasses type validation when assigning NULL (PostgreSQL 18.3)
Date: Mon, 30 Mar 2026 22:39:23 +0800
Message-ID: <CAEG8a3KMOS2dGzOYaZqDPZL+HOrYoAab50T4VYmWFGUWx7YZjg@mail.gmail.com> (raw)
In-Reply-To: <[email protected]>
References: <[email protected]>
On Mon, Mar 30, 2026 at 7:52 PM PG Bug reporting form
<[email protected]> wrote:
>
> The following bug has been logged on the website:
>
> Bug reference: 19442
> Logged by: Xianghang Zheng
> Email address: [email protected]
> PostgreSQL version: 18.3
> Operating system: Linux x86_64
> Description:
>
> I found an issue on PostgreSQL 18.3 related to domain handling over
> composite types in PL/pgSQL.
> After the fix for BUG #18735, the "cache lookup failed for type 0" error no
> longer occurs, but a new correctness problem arises:
> When assigning a NULL composite variable to a domain over composite type in
> PL/pgSQL,
> the domain's type validation and coercion logic is bypassed.
> This indicates an incomplete fix and may lead to incorrect enforcement of
> domain constraints (CHECK, NOT NULL, etc.) in the future.
> Version: PostgreSQL 18.3
> Test case:
> CREATE TYPE comp AS (a int, b text);
> CREATE DOMAIN dcomp AS comp;
> CREATE OR REPLACE FUNCTION test_domain_bug()
> RETURNS dcomp
> AS $$
> DECLARE
> v_comp comp := NULL;
> v_dcomp dcomp;
> BEGIN
> v_dcomp := v_comp;
> RETURN v_dcomp;
> END;
> $$ LANGUAGE plpgsql;
> SELECT test_domain_bug();
> Actual result: NULL (no error)
> Expected result: Domain type coercion should be properly executed even for
> NULL values.
dcomp doesn't have the not null constraint, so I think this is ok?
I tried `CREATE DOMAIN dcomp AS comp not null;`, it does report error:
[local] zhjwpku@postgres:5432-20309=# SELECT test_domain_bug();
ERROR: domain dcomp does not allow null values
CONTEXT: PL/pgSQL function test_domain_bug() line 4 during statement
block local variable initialization
Time: 2.730 ms
> This is a regression caused by incomplete backpatch of BUG #18735.
>
>
>
>
--
Regards
Junwang Zhao
reply
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: [email protected]
Cc: [email protected], [email protected], [email protected]
Subject: Re: BUG #19442: PL/pgSQL: domain over composite type bypasses type validation when assigning NULL (PostgreSQL 18.3)
In-Reply-To: <CAEG8a3KMOS2dGzOYaZqDPZL+HOrYoAab50T4VYmWFGUWx7YZjg@mail.gmail.com>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox