Re: [PATCH] pg_permissions - Mailing list pgsql-hackers

From Joel Jacobson
Subject Re: [PATCH] pg_permissions
Date
Msg-id 0c5a6b79-408c-4910-9b2e-4aa9a7b30f3a@app.fastmail.com
Whole thread Raw
In response to Re: [PATCH] pg_permissions  (Chapman Flack <chap@anastigmatix.net>)
Responses Re: [PATCH] pg_permissions
List pgsql-hackers
On Thu, Mar 10, 2022, at 22:02, Chapman Flack wrote:
> It looked to me as if the -hackers messages of 25 and 26 March 2021 had
> found a consensus that a pg_get_acl() function would be a good thing,
> with the views to be implemented over that.
>
> I'm just not seeing any later patch that adds such a function.

My apologies for late reply. Here it comes.

Recap: This patch is about adding two new system views: pg_privileges and pg_ownerships.

Changes since patch 0005 from 2021-03-25:

- Implement SQL-callable pg_get_acl()
This is a stripped down version of AlterObjectOwner_internal() from alter.c.

- Rename pg_permissions -> pg_privileges

- Use pg_shdepend + pg_get_acl() in pg_privileges, to avoid slow UNION ALL.

- Fix indentation of the new system views to be consistent with the other views.

- Add documentation of pg_get_acl() to func.sgml

- Move documentation of system views from catalogs.sgml to system-views.sgml

- Much smaller patch, thanks to getting rid of the long UNION ALL view definition:
 1 file changed, 195 insertions(+), 460 deletions(-)

/Joel
Attachment

pgsql-hackers by date:

Previous
From: Peter Geoghegan
Date:
Subject: Re: Harmonizing pg_bsd_indent parameter names
Next
From: Melanie Plageman
Date:
Subject: Re: RFC: adding pytest as a supported test framework