From: Tom Lane <tgl@sss.pgh.pa.us>
To: Albrecht =?iso-8859-1?b?RHJl3w==?= <albrecht.dress@posteo.de>
cc: pgsql-general@lists.postgresql.org
Subject: Re: Q: limit the length of log file entries?
In-reply-to: <YJI7DQ4I.3WYAVH2R.IWZZPYNU@AP5TCLKO.ZBA47T5B.KHWCFARS>
References: <YJI7DQ4I.3WYAVH2R.IWZZPYNU@AP5TCLKO.ZBA47T5B.KHWCFARS>
Comments: In-reply-to Albrecht =?iso-8859-1?b?RHJl3w==?= <albrecht.dress@posteo.de>
	message dated "Wed, 03 Sep 2025 15:40:45 -0000"
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-ID: <832140.1756926428.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Wed, 03 Sep 2025 15:07:08 -0400
Message-ID: <832141.1756926428@sss.pgh.pa.us>
Archived-At: <https://www.postgresql.org/message-id/832141.1756926428%40sss.pgh.pa.us>
Precedence: bulk

Albrecht =3D?iso-8859-1?b?RHJl3w=3D=3D?=3D <albrecht.dress@posteo.de> writ=
es:
> I have a PL/PGsql DB function 'addval(_id text, _data bytea) RETURNS VOI=
D'.  When a broken bytea is passed as 2nd parameter, something like

> albrecht@test FEHLER:  ung=C3=BCltige hexadezimale Ziffer: =C2=BBr=C2=AB=
 bei Zeichen 28
> albrecht@test ANWEISUNG:  SELECT FROM addval('hopp', '\xbroken')

> is printed in the log file.  This is fine and very helpful in this case,=
 but I have a =E2=80=9Cfriendly user=E2=80=9D who transmitted a broken hex=
 string of ~100 KBytes as second parameter, resulting in a log line of thi=
s size which is not really readable and when done frequently quickly fills=
 the logs.

There isn't any provision for limiting the length of source queries
quoted in the log.  Had your user sent the bytea value as a query
parameter, then log_parameter_max_length[_on_error] would have
applied, but this looks like the value was just inline in the query.

			regards, tom lane