Date: Thu, 26 May 2022 11:10:18 +0900 (JST)
Message-Id: <20220526.111018.1043498429473068058.horikyota.ntt@gmail.com>
To: michael@paquier.xyz
Cc: andres@anarazel.de, nathandbossart@gmail.com,
 pgsql-hackers@postgresql.org
Subject: Re: fix stats_fetch_consistency value in postgresql.conf.sample
From: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
In-Reply-To: <Yo7Bk0tPTyKvIduo@paquier.xyz>
References: <20220525.155623.1140106707572944089.horikyota.ntt@gmail.com>
	<20220525.161207.1389849691748709887.horikyota.ntt@gmail.com>
	<Yo7Bk0tPTyKvIduo@paquier.xyz>
User-Agent: Mew version 6.8 on Emacs 26.1
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Archived-At: 
 <https://www.postgresql.org/message-id/20220526.111018.1043498429473068058.horikyota.ntt%40gmail.com>
Precedence: bulk

At Thu, 26 May 2022 08:53:55 +0900, Michael Paquier <michael@paquier.xyz> wrote in 
> On Wed, May 25, 2022 at 04:12:07PM +0900, Kyotaro Horiguchi wrote:
> > (sigh..) As the result, no need to fix in this area for now, and I
> > don't think there's any generic and reliable way to detect
> > inconsistencies of guc variable definitions.
> 
> Hmm.  Making the automation test painless in terms of maintenance
> consists in making it require zero manual filtering in the list of
> GUCs involved, while still being useful in what it can detect.  The
> units involved in a GUC make the checks between postgresql.conf.sample 
> and pg_settings.boot_value annoying because they would require extra
> calculations depending on the unit with a logic maintained in the
> test.
> 
> I may be missing something obvious, of course, but it seems to me that
> as long as you fetch the values from postgresql.conf.sample and
> cross-check them with pg_settings.boot_value for GUCs that do not have
> units, the maintenance would be painless, while still being useful (it
> would cover the case of enums, for one).  The values need to be
> lower-cased for consistency, similarly to the GUC names.

Yeah, "boot_val" is appropreate here.  And I noticed that pg_settings
has the "unit" field.  I'll try using them.

Thanks for the suggestion!

regards.

-- 
Kyotaro Horiguchi
NTT Open Source Software Center