Re: pgsql: Enforce foreign key correctly during cross-partition updates - Mailing list pgsql-committers

From Alvaro Herrera
Subject Re: pgsql: Enforce foreign key correctly during cross-partition updates
Date
Msg-id 202203202126.5susqfvj43q5@alvherre.pgsql
Whole thread Raw
In response to Re: pgsql: Enforce foreign key correctly during cross-partition updates  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-committers
On 2022-Mar-20, Tom Lane wrote:

> ==2663594== Conditional jump or move depends on uninitialised value(s)
> ==2663594==    at 0x421526: ExecUpdateAct (nodeModifyTable.c:1855)

> It reproduces easily for me under valgrind.  I guess
> ExecCrossPartitionUpdate must be failing to set
> inserted_tuple or insert_destrel.

Yeah, I was trying to reproduce it -- but it doesn't for me.  Strange.
Anyway, I think you're right: insert_destrel is used in the test when
ExecCrossPartitionUpdate returns true, and clearly it can do that and
not set insert_destrel.

-- 
Álvaro Herrera              Valdivia, Chile  —  https://www.EnterpriseDB.com/
"La grandeza es una experiencia transitoria.  Nunca es consistente.
Depende en gran parte de la imaginación humana creadora de mitos"
(Irulan)



pgsql-committers by date:

Previous
From: Alvaro Herrera
Date:
Subject: pgsql: Blind fix for uninitialized memory bug in ba9a7e392171
Next
From: Andres Freund
Date:
Subject: pgsql: pgstat: introduce pgstat_relation_should_count().