Date: Thu, 8 Apr 2021 19:04:03 +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: <20210408110403.wr3d552xilqsx4us@nol>
References: <16953-f255a18f8c51f1d5@postgresql.org>
 <20210407120856.3rjxxnnzzxn74vyv@nol>
 <YG7YTjuZXylrjGAv@paquier.xyz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <YG7YTjuZXylrjGAv@paquier.xyz>
Archived-At: 
 <https://www.postgresql.org/message-id/20210408110403.wr3d552xilqsx4us%40nol>
Precedence: bulk

On Thu, Apr 08, 2021 at 07:17:50PM +0900, Michael Paquier wrote:
> On Wed, Apr 07, 2021 at 08:08:56PM +0800, Julien Rouhaud wrote:
> > PFA a naive patch to fix this problem with some regression tests.  I'm assuming
> > that -1 month should be january and not december.  I had a quick look at the
> > rest of formatting.c and didn't spot any similar problem, but another pair of
> > eyes wouldn't hurt.
> 
> Assuming an absolute number is not really intuitive when it comes to a
> negative number here, while counting backward feels more natural, so I
> would vote for making -1 be December, -2 November, etc.

Honestly I might as well have flipped a coin here, so I'm fine either way.

> Let's also make the tests more extended.  I would suggest a single
> query with generate_series() from say -12 to 12, that checks the
> output of both RM and rm for the full range of values supported by
> tm_mon.

I'm fine with it too, although I'd probably go with [-13, 13] just to make sure
that there's isn't silly off-by-one mistake.

I'll just wait a bit to see if anyone else has any opinion on whether -1 month
should be January or December.