Horton

Re: Valgrind - showing memory leaks

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Álvaro Herrera <alvherre@kurilemu.de>
Cc: Yasir <yasir.hussain.shah@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2025-05-11T15:13:18Z
Lists: pgsql-hackers 
=?utf-8?Q?=C3=81lvaro?= Herrera <alvherre@kurilemu.de> writes:
> On 2025-May-08, Tom Lane wrote:
>> So it seems like valgrind is wrong here, or else we're leaking the
>> whole rd_att structure later on somehow.

> Well, the problem is that if num_check is zero, FreeTupleDesc() doesn't
> free ->check.

Ah-hah.

> Done that way, thanks.

Thanks! I've been putting together a list of fixes to make
Valgrind less noisy, and it just got one shorter.

			regards, tom lane

Commits

  1. relcache: Avoid memory leak on tables with no CHECK constraints

  2. Silence compilers about extractNotNullColumn()