From: Tom Lane <tgl@sss.pgh.pa.us>
To: Michael Paquier <michael@paquier.xyz>
cc: "Anton A. Melnikov" <a.melnikov@postgrespro.ru>,
        PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: 003_check_guc.pl crashes if some extensions were loaded.
In-reply-to: <ZULW3m1h5K5PvSoD@paquier.xyz>
References: <fc5509ce-5144-4dac-8d13-21793da44fc5@postgrespro.ru> <ZULW3m1h5K5PvSoD@paquier.xyz>
Comments: In-reply-to Michael Paquier <michael@paquier.xyz>
	message dated "Thu, 02 Nov 2023 07:53:18 +0900"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <2885774.1698881391.1@sss.pgh.pa.us>
Date: Wed, 01 Nov 2023 19:29:51 -0400
Message-ID: <2885775.1698881391@sss.pgh.pa.us>
Archived-At: <https://www.postgresql.org/message-id/2885775.1698881391%40sss.pgh.pa.us>
Precedence: bulk

Michael Paquier <michael@paquier.xyz> writes:
> On Thu, Nov 02, 2023 at 12:28:05AM +0300, Anton A. Melnikov wrote:
>> "SELECT name
>> FROM pg_settings
>> WHERE NOT 'NOT_IN_SAMPLE' = ANY (pg_settings_get_flags(name)) AND
>> -       name <> 'config_file'
>> +       name <> 'config_file' AND name NOT LIKE '%.%'
>> ORDER BY 1");

> Wouldn't it be better to add a qual as of "category <> 'Customized
> Options'"?

+1, seems like a cleaner answer.

> That's something arbitrarily assigned for all custom GUCs
> and we are sure that none of them will exist in
> postgresql.conf.sample.  There's also no guarantee that out-of-core
> custom GUCs will include a dot in their name (even if I know that
> maintainers close to the community adopt this convention and are
> rather careful about that).

Actually we do force that, see valid_custom_variable_name().
But I think your idea is better.

			regards, tom lane