Date: Mon, 6 Dec 2021 07:36:55 -0600
From: Justin Pryzby <pryzby@telsasoft.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: pgsql-hackers@lists.postgresql.org
Subject: Re: GUC flags
Message-ID: <20211206133655.GH17618@telsasoft.com>
References: <YaSJcceKORqKS+cM@paquier.xyz>
 <YaXGfc+tRYFFN+bo@paquier.xyz>
 <20211201075904.GW17618@telsasoft.com>
 <YadjSDYGxYEO5k/R@paquier.xyz>
 <20211202033439.GL17618@telsasoft.com>
 <YahVin9880J9W45Z@paquier.xyz>
 <20211202051734.GN17618@telsasoft.com>
 <Yaltp7EjWny7Gl4e@paquier.xyz>
 <20211206053805.GG17618@telsasoft.com>
 <Ya20nxtRYfkg2jd6@paquier.xyz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Ya20nxtRYfkg2jd6@paquier.xyz>
User-Agent: Mutt/1.9.4 (2018-02-28)
Archived-At: 
 <https://www.postgresql.org/message-id/20211206133655.GH17618%40telsasoft.com>
Precedence: bulk

On Mon, Dec 06, 2021 at 03:58:39PM +0900, Michael Paquier wrote:
> On Sun, Dec 05, 2021 at 11:38:05PM -0600, Justin Pryzby wrote:
> > Thanks.  One more item.  The check_guc script currently outputs 68 false
> > positives - even though it includes a list of 20 exceptions.  This is not
> > useful.
> 
> Indeed.  Hmm.  This script does a couple of things:
> 1) Check the format of the options defined in the various lists of
> guc.c, which is something people format well, and pgindent also does 
> a part of this job.
> 2) Check that options in the hardcoded list of GUCs in
> INTENTIONALLY_NOT_INCLUDED are not included in
> postgresql.conf.sample
> 3) Check that nothing considered as a parameter in
> postgresql.conf.sample is listed in guc.c.
> 
> Your patch removes 1) and 2), but keeps 3) to check for dead
> parameter references in postgresql.conf.sample.

The script checks that guc.c and sample config are consistent.

I think your undertanding of INTENTIONALLY_NOT_INCLUDED is not right.
That's a list of stuff it "avoids reporting" as an suspected error, not an
additional list of stuff to checks.  INTENTIONALLY_NOT_INCLUDED is a list of
stuff like NOT_IN_SAMPLE, which is better done by parsing /NOT_IN_SAMPLE/.

> Is check_guc actually run on a periodic basis by somebody?  Based on
> the amount of false positives that has accumulated over the years, and
> what `git grep` can already do for 3), it seems to me that we have
> more arguments in favor of just removing it entirely.

I saw that Tom updated it within the last 12 months, which I took to mean that
it was still being maintained. But I'm okay with removing it.

-- 
Justin