Date: Mon, 12 Apr 2021 11:13:02 +0800
From: Julien Rouhaud <rjuju123@gmail.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: t.larionov@postgrespro.ru, pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #16953: OOB access while converting "interval" to char
Message-ID: <20210412031302.tquch5rg5j3eiut3@nol>
References: <16953-f255a18f8c51f1d5@postgresql.org>
 <20210407120856.3rjxxnnzzxn74vyv@nol>
 <YG7YTjuZXylrjGAv@paquier.xyz>
 <20210408110403.wr3d552xilqsx4us@nol>
 <YG7q8YcvFcXCAbAH@paquier.xyz>
 <20210409104246.vkeef6rbyevq46f3@nol>
 <YHAza+N74LHygGnB@paquier.xyz>
 <YHO4t+MnSNPcMPAU@paquier.xyz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YHO4t+MnSNPcMPAU@paquier.xyz>
Archived-At: 
 <https://www.postgresql.org/message-id/20210412031302.tquch5rg5j3eiut3%40nol>
Precedence: bulk

On Mon, Apr 12, 2021 at 12:04:23PM +0900, Michael Paquier wrote:
> On Fri, Apr 09, 2021 at 07:58:51PM +0900, Michael Paquier wrote:
> > Yep.  The regression tests show what I would expect.  I'll check in
> > details later.
> 
> I have spent some time on that today and applied this patch down to
> 9.6, after adding more comments and simplifying a bit the calculation
> method used.  Instead of using the "mon" variable to store an
> intermediate result, I have simplified things so as this only uses
> tm->tm_mon, and "mon" for the position in the roman-numeral array is
> adjusted only once.

Thanks Michael!  I thought we'd want to keep the original calculation for
consistency, but directly computing the needed offset is clearly simpler to
read!