Re: set role issue - Mailing list pgsql-hackers

From Fabrice Chapuis
Subject Re: set role issue
Date
Msg-id CAA5-nLByWL++=WZEC52Espbgd6k4O27RO3b+wEBKaTkrL2bquw@mail.gmail.com
Whole thread Raw
In response to Re: set role issue  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Thank you for clarifying this point.
And yes of course grantor, sorry for the confusion.

Regards,

Fabrice




On Mon, Aug 4, 2025 at 4:30 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
Fabrice Chapuis <fabrice636861@gmail.com> writes:
> I do the following command when connecting locally with user postgres
> SET ROLE asuperrole; -- asuperrole has superuser privilege
> GRANT SELECT ON pg_statistic TO test_role;

> why the grantee is postgres?

You mean "grantor", no?  When a superuser does grant/revoke on some
object, it's recorded as though the object owner issued the command.
And postgres is the owner of the pg_statistic table.

                        regards, tom lane

pgsql-hackers by date:

Previous
From: Fujii Masao
Date:
Subject: Re: Bug in pg_dump --filter? - Invalid object types can be misinterpreted as valid
Next
From: Jacob Champion
Date:
Subject: Re: [PATCH] OAuth: fix performance bug with stuck multiplexer events