Re: Improve documentation regarding custom settings, placeholders, and the administrative functions

Zhang Mingli <zmlpostgres@gmail.com>

From: Zhang Mingli <zmlpostgres@gmail.com>
To: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, "David G. Johnston" <david.g.johnston@gmail.com>
Date: 2025-02-06T02:36:20Z
Lists: pgsql-hackers
On Oct 20, 2024 at 04:12 +0800, David G. Johnston <david.g.johnston@gmail.com>, wrote:
>
> Mostly I'm pointing out the fact that one can never take the null value to be the actual value of a setting.  In terms of current_setting this then establishes the fact that the null value it may return is an error-handling alternative only and not something to be relied upon as being an actual value of the setting.

> -        Returns the current value of the+        Returns the current non-null value of the         setting <parameter>setting_name</parameter>.  If there is no such         setting, <function>current_setting</function> throws an error         unless <parameter>missing_ok</parameter> is supplied and

Hi,

current_setting() could return NULL when missing_ok is passed, so is it right to say: Returns the current non-null value of the setting <parameter>setting_name</parameter>?
As the doc is for function of current_setting(), and it could return NULL actually.

gpadmin=# \pset null NULL
Null display is "NULL".
gpadmin=# select current_setting('pg_xmen', true);
 current_setting
-----------------
 NULL
(1 row)


--
Zhang Mingli
HashData

Commits

  1. docs: Clarify that NULL arg to set_config() means reset to default