From: Tom Lane <tgl@sss.pgh.pa.us>
To: Michael Paquier <michael@paquier.xyz>
cc: starbugs@qq.com, pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #16190: The usage of NULL pointer in refint.c
In-reply-to: <20200106051800.GL3598@paquier.xyz>
References: <16190-70181c803641c3dc@postgresql.org>
 <20200106051800.GL3598@paquier.xyz>
Comments: In-reply-to Michael Paquier <michael@paquier.xyz>
	message dated "Mon, 06 Jan 2020 14:18:00 +0900"
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-ID: <1233.1578291695.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Mon, 06 Jan 2020 01:21:35 -0500
Message-ID: <1234.1578291695@sss.pgh.pa.us>
Precedence: bulk

Michael Paquier <michael@paquier.xyz> writes:
> On Mon, Jan 06, 2020 at 03:39:36AM +0000, PG Bug reporting form wrote:
>> We checked the code in file =E2=80=9Crefint.c=E2=80=9D and there is one=
 error occurring in
>> line 636.

> It could be better to switch all that to not use directly system
> calls, and rely properly on a high-level memory context with
> palloc-like allocations.

Yeah, if somebody wanted to fix this, the right way is to replace
these malloc calls with pallocs.  Some investigation would be needed
about which context to use.

> ... There could be also an argument to just
> remove the module per the lack of attention it is getting, though it
> is still useful as an example of use for SPI, and the docs mention
> it for that.

The regression tests use refint too.  Still, it's not production-grade
code by any means, and I'm not sure if there's much point in making
it so.  I won't stand in the way if somebody else wants to ;-)

			regards, tom lane