Date: Mon, 6 Jan 2020 18:26:54 -0800
From: Andres Freund <andres@anarazel.de>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Michael Paquier <michael@paquier.xyz>, starbugs@qq.com,
	pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #16190: The usage of NULL pointer in refint.c
Message-ID: <20200107022654.vgfwx7hactiokcsa@alap3.anarazel.de>
References: <16190-70181c803641c3dc@postgresql.org>
 <20200106051800.GL3598@paquier.xyz>
 <1234.1578291695@sss.pgh.pa.us>
 <20200106174443.3pwnmshbxrbwggrm@alap3.anarazel.de>
 <20200107012741.GD2386@paquier.xyz>
 <7984.1578363125@sss.pgh.pa.us>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <7984.1578363125@sss.pgh.pa.us>
Precedence: bulk

On 2020-01-06 21:12:05 -0500, Tom Lane wrote:
> Michael Paquier <michael@paquier.xyz> writes:
> > On Mon, Jan 06, 2020 at 09:44:43AM -0800, Andres Freund wrote:
> >> I think we should consider either moving this out of contrib, or fixing
> >> it up. test/example code is fine, but contrib gets installed by default
> >> for a lot of people... And yea, this isn't just about contrib/spi.
> 
> > No idea about moving that out of contrib/, but here is a patch to fix
> > things that just moves the allocations to TopMemoryContext and removes
> > the system calls.
> 
> WFM.  There are probably more elegant ways to do it than to drop this
> stuff into TopMemoryContext, but this is surely better than unchecked
> malloc calls.

Yea, it's certainly better than the current situation. An incremental
improvement would be to do the allocations in a separate contect, for easier
debugging should there ever be a leak...

- Andres