public inbox for [email protected]
help / color / mirror / Atom feedFrom: Fujii Masao <[email protected]>
To: Yasuo Honda <[email protected]>
Cc: PostgreSQL-development <[email protected]>
Subject: Re: [PATCH] Fix unexpected loss of DEFERRABLE property after toggling NOT ENFORCED / ENFORCED
Date: Sat, 28 Mar 2026 00:01:03 +0900
Message-ID: <CAHGQGwEyx0McXgD3wghUVuBw=thTViEae-PAEF7d+Weo6XjXZw@mail.gmail.com> (raw)
In-Reply-To: <CAKmOUTkrgZCWq63Ab0Ff01JQ7u3ua__NyT=V8mZxcH3sSAhX3w@mail.gmail.com>
References: <CAKmOUTms2nkxEZDdcrsjq5P3b2L_PR266Hv8kW5pANwmVaRJJQ@mail.gmail.com>
<CAHGQGwFrsVLxmXrtrQD2cR1K74de6sq3eEvPOPiEDDoRDdhdSw@mail.gmail.com>
<CAKmOUTkrgZCWq63Ab0Ff01JQ7u3ua__NyT=V8mZxcH3sSAhX3w@mail.gmail.com>
On Fri, Mar 27, 2026 at 4:41 PM Yasuo Honda <[email protected]> wrote:
>
> (Resending to the list, as my previous reply went to the sender only.)
>
> Thanks for the comment.
>
> > Regarding the regression test, would it be better to verify not only catalog
> state (e.g., pg_trigger) but also the actual behavior?
>
> Agreed. In addition to inspecting pg_trigger contents, I've added test
> cases that verify the actual runtime behavior of deferred constraints.
Thanks for updating the patch!
> I've also covered INITIALLY IMMEDIATE alongside INITIALLY DEFERRED.
I'm not sure this test adds much value. Testing DEFERRABLE INITIALLY DEFERRED
seems sufficient, so I removed the DEFERRABLE INITIALLY IMMEDIATE case.
The updated patch is attached.
+-- verify that tgdeferrable/tginitdeferred are preserved after NOT
ENFORCED -> ENFORCED
+ALTER TABLE FKTABLE ALTER CONSTRAINT fktable_fk_fkey DEFERRABLE
INITIALLY DEFERRED;
+ALTER TABLE FKTABLE ALTER CONSTRAINT fktable_fk_fkey NOT ENFORCED;
It looks like this assumes the constraint is initially ENFORCED,
but in this part of the test it is already NOT ENFORCED. So I moved
this test to a section where the constraint is ENFORCED, which
seems more appropriate.
I also updated the commit log message.
Barring any objections, I will commit this and backpatch it to v18, where
NOT ENFORCED foreign keys were introduced.
Regards,
--
Fujii Masao
Attachments:
[application/octet-stream] v3-0001-Fix-FK-triggers-losing-DEFERRABLE-INITIALLY-DEFER.patch (7.4K, 2-v3-0001-Fix-FK-triggers-losing-DEFERRABLE-INITIALLY-DEFER.patch)
download | inline diff:
From b15a5183bc99a1985ae5899c5e5aa7257c602f4c Mon Sep 17 00:00:00 2001
From: Fujii Masao <[email protected]>
Date: Fri, 27 Mar 2026 18:29:07 +0900
Subject: [PATCH v3] Fix FK triggers losing DEFERRABLE/INITIALLY DEFERRED when
marked ENFORCED again
Previously, a foreign key defined as DEFERRABLE INITIALLY DEFERRED could
behave as NOT DEFERRABLE after being set to NOT ENFORCED and then back
to ENFORCED.
This happened because recreating the FK triggers on re-enabling the constraint
forgot to restore the tgdeferrable and tginitdeferred fields in pg_trigger.
Fix this bug by properly setting those fields when the foreign key constraint
is marked ENFORCED again and its triggers are recreated, so the original
DEFERRABLE and INITIALLY DEFERRED properties are preserved.
Backpatch to v18, where NOT ENFORCED foreign keys were introduced.
Author: Yasuo Honda <[email protected]>
Reviewed-by: Fujii Masao <[email protected]>
Discussion: https://postgr.es/m/CAKmOUTms2nkxEZDdcrsjq5P3b2L_PR266Hv8kW5pANwmVaRJJQ@mail.gmail.com
Backpatch-through: 18
---
src/backend/commands/tablecmds.c | 2 +
src/test/regress/expected/foreign_key.out | 53 +++++++++++++++++++++++
src/test/regress/sql/foreign_key.sql | 37 ++++++++++++++++
3 files changed, 92 insertions(+)
diff --git a/src/backend/commands/tablecmds.c b/src/backend/commands/tablecmds.c
index c69c12dc014..dd00e36c69d 100644
--- a/src/backend/commands/tablecmds.c
+++ b/src/backend/commands/tablecmds.c
@@ -12558,6 +12558,8 @@ ATExecAlterFKConstrEnforceability(List **wqueue, ATAlterConstraint *cmdcon,
fkconstraint->fk_matchtype = currcon->confmatchtype;
fkconstraint->fk_upd_action = currcon->confupdtype;
fkconstraint->fk_del_action = currcon->confdeltype;
+ fkconstraint->deferrable = currcon->condeferrable;
+ fkconstraint->initdeferred = currcon->condeferred;
/* Create referenced triggers */
if (currcon->conrelid == fkrelid)
diff --git a/src/test/regress/expected/foreign_key.out b/src/test/regress/expected/foreign_key.out
index 9ae4dbf1b0a..4c69cb999cd 100644
--- a/src/test/regress/expected/foreign_key.out
+++ b/src/test/regress/expected/foreign_key.out
@@ -1157,6 +1157,59 @@ INSERT INTO fktable VALUES (500, 1000);
ERROR: insert or update on table "fktable" violates foreign key constraint "fktable_fk_fkey"
DETAIL: Key (fk)=(1000) is not present in table "pktable".
COMMIT;
+-- Check that the existing FK trigger is both deferrable and initially deferred
+SELECT conname, tgrelid::regclass as tgrel,
+ regexp_replace(tgname, '[0-9]+', 'N') as tgname, tgtype,
+ tgdeferrable, tginitdeferred
+FROM pg_trigger t JOIN pg_constraint c ON (t.tgconstraint = c.oid)
+WHERE conrelid = 'fktable'::regclass AND conname = 'fktable_fk_fkey'
+ORDER BY tgrelid, tgtype;
+ conname | tgrel | tgname | tgtype | tgdeferrable | tginitdeferred
+-----------------+---------+--------------------------+--------+--------------+----------------
+ fktable_fk_fkey | pktable | RI_ConstraintTrigger_a_N | 9 | t | t
+ fktable_fk_fkey | pktable | RI_ConstraintTrigger_a_N | 17 | t | t
+ fktable_fk_fkey | fktable | RI_ConstraintTrigger_c_N | 5 | t | t
+ fktable_fk_fkey | fktable | RI_ConstraintTrigger_c_N | 17 | t | t
+(4 rows)
+
+-- Changing the constraint to NOT ENFORCED drops the associated FK triggers
+ALTER TABLE FKTABLE ALTER CONSTRAINT fktable_fk_fkey NOT ENFORCED;
+SELECT conname, tgrelid::regclass as tgrel,
+ regexp_replace(tgname, '[0-9]+', 'N') as tgname, tgtype,
+ tgdeferrable, tginitdeferred
+FROM pg_trigger t JOIN pg_constraint c ON (t.tgconstraint = c.oid)
+WHERE conrelid = 'fktable'::regclass AND conname = 'fktable_fk_fkey'
+ORDER BY tgrelid, tgtype;
+ conname | tgrel | tgname | tgtype | tgdeferrable | tginitdeferred
+---------+-------+--------+--------+--------------+----------------
+(0 rows)
+
+-- Changing it back to ENFORCED will recreate the necessary FK triggers
+-- that are deferrable and initially deferred
+ALTER TABLE FKTABLE ALTER CONSTRAINT fktable_fk_fkey ENFORCED;
+SELECT conname, tgrelid::regclass as tgrel,
+ regexp_replace(tgname, '[0-9]+', 'N') as tgname, tgtype,
+ tgdeferrable, tginitdeferred
+FROM pg_trigger t JOIN pg_constraint c ON (t.tgconstraint = c.oid)
+WHERE conrelid = 'fktable'::regclass AND conname = 'fktable_fk_fkey'
+ORDER BY tgrelid, tgtype;
+ conname | tgrel | tgname | tgtype | tgdeferrable | tginitdeferred
+-----------------+---------+--------------------------+--------+--------------+----------------
+ fktable_fk_fkey | pktable | RI_ConstraintTrigger_a_N | 9 | t | t
+ fktable_fk_fkey | pktable | RI_ConstraintTrigger_a_N | 17 | t | t
+ fktable_fk_fkey | fktable | RI_ConstraintTrigger_c_N | 5 | t | t
+ fktable_fk_fkey | fktable | RI_ConstraintTrigger_c_N | 17 | t | t
+(4 rows)
+
+-- Verify that a deferrable, initially deferred foreign key still works
+-- as expected after being set to NOT ENFORCED and then re-enabled
+BEGIN;
+-- doesn't match PK, but no error yet
+INSERT INTO fktable VALUES (2, 20);
+-- should catch error from INSERT at commit
+COMMIT;
+ERROR: insert or update on table "fktable" violates foreign key constraint "fktable_fk_fkey"
+DETAIL: Key (fk)=(20) is not present in table "pktable".
DROP TABLE fktable, pktable;
-- tricky behavior: according to SQL99, if a deferred constraint is set
-- to 'immediate' mode, it should be checked for validity *immediately*,
diff --git a/src/test/regress/sql/foreign_key.sql b/src/test/regress/sql/foreign_key.sql
index 3b8c95bf893..02a60d661a1 100644
--- a/src/test/regress/sql/foreign_key.sql
+++ b/src/test/regress/sql/foreign_key.sql
@@ -784,6 +784,43 @@ INSERT INTO fktable VALUES (500, 1000);
COMMIT;
+-- Check that the existing FK trigger is both deferrable and initially deferred
+SELECT conname, tgrelid::regclass as tgrel,
+ regexp_replace(tgname, '[0-9]+', 'N') as tgname, tgtype,
+ tgdeferrable, tginitdeferred
+FROM pg_trigger t JOIN pg_constraint c ON (t.tgconstraint = c.oid)
+WHERE conrelid = 'fktable'::regclass AND conname = 'fktable_fk_fkey'
+ORDER BY tgrelid, tgtype;
+
+-- Changing the constraint to NOT ENFORCED drops the associated FK triggers
+ALTER TABLE FKTABLE ALTER CONSTRAINT fktable_fk_fkey NOT ENFORCED;
+SELECT conname, tgrelid::regclass as tgrel,
+ regexp_replace(tgname, '[0-9]+', 'N') as tgname, tgtype,
+ tgdeferrable, tginitdeferred
+FROM pg_trigger t JOIN pg_constraint c ON (t.tgconstraint = c.oid)
+WHERE conrelid = 'fktable'::regclass AND conname = 'fktable_fk_fkey'
+ORDER BY tgrelid, tgtype;
+
+-- Changing it back to ENFORCED will recreate the necessary FK triggers
+-- that are deferrable and initially deferred
+ALTER TABLE FKTABLE ALTER CONSTRAINT fktable_fk_fkey ENFORCED;
+SELECT conname, tgrelid::regclass as tgrel,
+ regexp_replace(tgname, '[0-9]+', 'N') as tgname, tgtype,
+ tgdeferrable, tginitdeferred
+FROM pg_trigger t JOIN pg_constraint c ON (t.tgconstraint = c.oid)
+WHERE conrelid = 'fktable'::regclass AND conname = 'fktable_fk_fkey'
+ORDER BY tgrelid, tgtype;
+
+-- Verify that a deferrable, initially deferred foreign key still works
+-- as expected after being set to NOT ENFORCED and then re-enabled
+BEGIN;
+
+-- doesn't match PK, but no error yet
+INSERT INTO fktable VALUES (2, 20);
+
+-- should catch error from INSERT at commit
+COMMIT;
+
DROP TABLE fktable, pktable;
-- tricky behavior: according to SQL99, if a deferred constraint is set
--
2.51.2
view thread (6+ messages) latest in thread
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]
Subject: Re: [PATCH] Fix unexpected loss of DEFERRABLE property after toggling NOT ENFORCED / ENFORCED
In-Reply-To: <CAHGQGwEyx0McXgD3wghUVuBw=thTViEae-PAEF7d+Weo6XjXZw@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