From: Daniel Gustafsson <daniel@yesql.se>
Content-Type: text/plain;
	charset=us-ascii
Content-Transfer-Encoding: quoted-printable
Mime-Version: 1.0 (Mac OS X Mail 16.0 \(3776.700.51.11.1\))
Subject: Re: BUG #18745: /src/bin/pg_waldump/compat.c
Date: Wed, 11 Dec 2024 22:18:05 +0100
References: <18745-5afc39855ce0208e@postgresql.org>
To: dan-eli@mail.ru,
 pgsql-bugs@lists.postgresql.org
In-Reply-To: <18745-5afc39855ce0208e@postgresql.org>
Message-Id: <EAB008E3-6B87-413F-BA39-22A6ED5DF0F6@yesql.se>
Archived-At: 
 <https://www.postgresql.org/message-id/EAB008E3-6B87-413F-BA39-22A6ED5DF0F6%40yesql.se>
Precedence: bulk

> On 9 Dec 2024, at 14:07, PG Bug reporting form =
<noreply@postgresql.org> wrote:
>=20
> The following bug has been logged on the website:
>=20
> Bug reference:      18745
> Logged by:          Daniel Elishakov
> Email address:      dan-eli@mail.ru
> PostgreSQL version: 16.4
> Operating system:   ubuntu 20.04
> Description:       =20
>=20
> Pointer 'ltime', returned from function 'localtime' at compat.c:55, =
may be
> NULL and is dereferenced at compat.c:57 by calling function =
'strftime'.

That's a fair point, ISTM we should mimic what the backend =
timestamptz_to_str
does with something like the following.  We could inspect errno and come =
up
with other errors as well but keeping the copies reasonably in sync for =
rare
errors seems worthwhile.

diff --git a/src/bin/pg_waldump/compat.c b/src/bin/pg_waldump/compat.c
index 1541eac611..9e1a45af81 100644
--- a/src/bin/pg_waldump/compat.c
+++ b/src/bin/pg_waldump/compat.c
@@ -54,6 +54,12 @@ timestamptz_to_str(TimestampTz t)
        time_t          result =3D (time_t) timestamptz_to_time_t(t);
        struct tm  *ltime =3D localtime(&result);

+       if (!ltime)
+       {
+               strlcpy(buf, "(timestamp out of range)", sizeof(buf));
+               return buf;
+       }
+
        strftime(ts, sizeof(ts), "%Y-%m-%d %H:%M:%S", ltime);
        strftime(zone, sizeof(zone), "%Z", ltime);

--
Daniel Gustafsson