Re: enable/disable broken for statement triggers on partitioned tables

From: Dmitry Koval <d(dot)koval(at)postgrespro(dot)ru>
To: Amit Langote <amitlangote09(at)gmail(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Cc: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>, Simon Riggs <simon(dot)riggs(at)enterprisedb(dot)com>
Subject: Re: enable/disable broken for statement triggers on partitioned tables
Date: 2022-07-07 18:44:35
Message-ID: ea25d7d5-0454-8d09-9936-12ee9bddf942@postgrespro.ru
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi!

I've looked through the code and everything looks good.
But there is one thing I doubt.
Patch changes result of test:
----

create function trig_nothing() returns trigger language plpgsql
as $$ begin return null; end $$;
create table parent (a int) partition by list (a);
create table child1 partition of parent for values in (1);

create trigger tg after insert on parent
for each row execute procedure trig_nothing();
select tgrelid::regclass, tgname, tgenabled from pg_trigger
where tgrelid in ('parent'::regclass, 'child1'::regclass)
order by tgrelid::regclass::text;
alter table only parent enable always trigger tg; -- no recursion
select tgrelid::regclass, tgname, tgenabled from pg_trigger
where tgrelid in ('parent'::regclass, 'child1'::regclass)
order by tgrelid::regclass::text;
alter table parent enable always trigger tg; -- recursion
select tgrelid::regclass, tgname, tgenabled from pg_trigger
where tgrelid in ('parent'::regclass, 'child1'::regclass)
order by tgrelid::regclass::text;

drop table parent, child1;
drop function trig_nothing();

----
Results of vanilla + patch:
----
CREATE FUNCTION
CREATE TABLE
CREATE TABLE
CREATE TRIGGER
tgrelid | tgname | tgenabled
---------+--------+-----------
child1 | tg | O
parent | tg | O
(2 rows)

ALTER TABLE
tgrelid | tgname | tgenabled
---------+--------+-----------
child1 | tg | O
parent | tg | A
(2 rows)

ALTER TABLE
tgrelid | tgname | tgenabled
---------+--------+-----------
child1 | tg | O
parent | tg | A
(2 rows)

DROP TABLE
DROP FUNCTION

----
Results of vanilla:
----
CREATE FUNCTION
CREATE TABLE
CREATE TABLE
CREATE TRIGGER
tgrelid | tgname | tgenabled
---------+--------+-----------
child1 | tg | O
parent | tg | O
(2 rows)

ALTER TABLE
tgrelid | tgname | tgenabled
---------+--------+-----------
child1 | tg | O
parent | tg | A
(2 rows)

ALTER TABLE
tgrelid | tgname | tgenabled
---------+--------+-----------
child1 | tg | A
parent | tg | A
(2 rows)

DROP TABLE
DROP FUNCTION
----
The patch doesn't start recursion in case 'tgenabled' flag of parent
table is not changes.
Probably vanilla result is more correct.

--
With best regards,
Dmitry Koval

Postgres Professional: http://postgrespro.com

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Justin Pryzby 2022-07-07 18:58:06 Re: [PATCH] fix wait_event of pg_stat_activity in case of high amount of connections
Previous Message Justin Pryzby 2022-07-07 18:44:10 Re: pg15b2: large objects lost on upgrade