Horton

Re: allowing for control over SET ROLE

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Robert Haas <robertmhaas@gmail.com>
Cc: Erik Rijkers <er@xs4all.nl>, Jeff Davis <pgsql@j-davis.com>, Stephen Frost <sfrost@snowman.net>, Nathan Bossart <nathandbossart@gmail.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2022-12-14T02:44:14Z
Lists: pgsql-hackers 
On Mon, Nov 21, 2022 at 10:45:53AM -0500, Robert Haas wrote:
> Seems like a good idea but I'm not sure about this hunk:
> 
>   TailMatches("GRANT|REVOKE", "ALTER", "SYSTEM") ||
> - TailMatches("REVOKE", "GRANT", "OPTION", "FOR", "ALTER", "SYSTEM"))
> + TailMatches("REVOKE", "GRANT", "OPTION", "FOR", "ALTER", "SYSTEM") ||
> + TailMatches("REVOKE", "GRANT", "OPTION", "FOR", "SET"))
> 
> That might be a correct change for other reasons, but it doesn't seem
> related to this patch. The rest looks good.

(Forgot to press "Send" a few days ago..)

Hmm, right, I see your point.  I have just moved that to reorder the
terms alphabetically, but moving the check on REVOKE GRANT OPTION FOR
SET is not mandatory.  I have moved it back in its previous
position, leading to less noise in the diffs, and applied the rest as
of 9d0cf57.
Thanks!
--
Michael

Commits

  1. More documentation update for GRANT ... WITH SET OPTION.

  2. Restrict the privileges of CREATEROLE users.

  3. Add support for GRANT SET in psql tab completion

  4. Add a SET option to the GRANT command.

  5. Allow grant-level control of role inheritance behavior.