Date: Mon, 24 Oct 2022 17:05:44 -0500
From: Justin Pryzby <pryzby@telsasoft.com>
To: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Cc: michael@paquier.xyz, andres@anarazel.de, nathandbossart@gmail.com,
	pgsql-hackers@postgresql.org, Peter Smith <smithpb2250@gmail.com>
Subject: Re: fix stats_fetch_consistency value in postgresql.conf.sample
Message-ID: <20221024220544.GJ16921@telsasoft.com>
References: <20220728222734.GA1913@telsasoft.com>
 <20220913004303.GM31833@telsasoft.com>
 <20220918045306.GT31833@telsasoft.com>
 <20220926.172958.1790193605025533956.horikyota.ntt@gmail.com>
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="kr14OxHsRwZHHqxS"
Content-Disposition: inline
In-Reply-To: <20220926.172958.1790193605025533956.horikyota.ntt@gmail.com>
User-Agent: Mutt/1.9.4 (2018-02-28)
Archived-At: 
 <https://www.postgresql.org/message-id/20221024220544.GJ16921%40telsasoft.com>
Precedence: bulk


--kr14OxHsRwZHHqxS
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

On Mon, Sep 26, 2022 at 05:29:58PM +0900, Kyotaro Horiguchi wrote:
> At Sat, 17 Sep 2022 23:53:07 -0500, Justin Pryzby <pryzby@telsasoft.com> wrote in 
> > This is an alternative implementation, which still relies on adding the
> > GUC_DYNAMIC, flag but doesn't require adding a new, sql-accessible
> > function to convert the GUC to a pretty/human display value.
> 
> Thanks!
> 
> I'm not sure shared_buffer is GUC_DYNAMIC_DEFAULT, and we need to read

It's set during initdb.

> postgresql.conf.sample using SQL, but +1 for the direction.
> 
> +	AND NOT (sc.sample_value ~ '^0' AND current_setting(name) ~ '^0') -- zeros may be written differently
> +	AND NOT (sc.sample_value='60s' AND current_setting(name) = '1min') -- two ways to write 1min
> 
> Mmm. Couldn't we get away from that explicit exceptions?

Suggestions are welcomed.

Rebased the patch.

I also split the flag into DEFAULTS_COMPILE and DEFAULTS_INITDB, since
that makes it easier to understand what the flags mean and the intent of
the patch.  And maybe allows fewer exclusions in patches like Peter's,
which I think would only want to exclude compile-time defaults.

-- 
Justin

--kr14OxHsRwZHHqxS
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
 filename="0001-add-DYNAMIC_DEFAULT-for-settings-which-vary-by-.-con.patch"