Re: replacing role-level NOINHERIT with a grant-level option

Поиск
Список
Период
Сортировка
От Robert Haas
Тема Re: replacing role-level NOINHERIT with a grant-level option
Дата
Msg-id CA+Tgmoa8Y2s9EO7kBtXp=Pn8DVk+fyjNMSpq=fmFj83jTJ7GMQ@mail.gmail.com
обсуждение исходный текст
Ответ на Re: replacing role-level NOINHERIT with a grant-level option  (Peter Eisentraut <peter.eisentraut@enterprisedb.com>)
Список pgsql-hackers
On Wed, Jun 15, 2022 at 5:23 AM Peter Eisentraut
<peter.eisentraut@enterprisedb.com> wrote:
> > Consider a user who in general prefers the NOINHERIT behavior but also
> > wants to use predefined roles. Perhaps user 'peter' is to be granted
> > both 'paul' and 'pg_execute_server_programs'. If role 'peter' is set
> > to INHERIT, Peter will be sad, because his love for NOINHERIT probably
> > means that he doesn't want to exercise Paul's privileges
> > automatically. However, he needs to inherit the privileges of
> > 'pg_execute_server_programs' or they are of no use to him. Peter
> > presumably wants to use COPY TO/FROM program to put data into a table
> > owned by 'peter', not a table owned by 'pg_execute_server_programs'.
> > If so, being able to SET ROLE to 'pg_execute_server_programs' is of no
> > use to him at all, but inheriting the privilege is useful.
>
> That's because our implementation of SET ROLE is bogus.  We should have
> a SET ROLE that is separate from SET SESSION AUTHORIZATION, where the
> current user can keep their current user-ness and additionally enable
> (non-inherited) roles.

It would help me to have a better description of what you think the
behavior ought to be. I've always thought there was something funny
about SET ROLE and SET SESSION AUTHORIZATION, because it seems like
they are too similar to each other. But it would surprise me if SET
ROLE added additional privileges to my session while leaving the old
ones intact, too, much as I'd be surprised if SET work_mem = '8MB'
followed by SET work_mem = '1GB' somehow left both values partly in
effect at the same time. It feels to me like SET is describing an
action that changes the session state, rather than adding to it.

> I'm mainly concerned that (AAIU), you propose to remove the current
> INHERIT/NOINHERIT attribute of roles.  I wouldn't like that.  If you
> want a feature that allows overriding that per-grant, maybe that's okay.

Yeah, I want to remove it and replace it with something more
fine-grained. I don't yet understand why that's a problem for anything
you want to do.

-- 
Robert Haas
EDB: http://www.enterprisedb.com



В списке pgsql-hackers по дате отправления:

Предыдущее
От: Aleksander Alekseev
Дата:
Сообщение: Re: [PATCH] Compression dictionaries for JSONB
Следующее
От: Andrey Borodin
Дата:
Сообщение: Re: [PATCH] Add sortsupport for range types and btree_gist