Re: New Privilege model purposal - Mailing list pgsql-hackers

From Peter Eisentraut
Subject Re: New Privilege model purposal
Date
Msg-id Pine.LNX.4.21.0008050001180.1568-100000@localhost.localdomain
Whole thread Raw
In response to New Privilege model purposal  (JanWieck@t-online.de (Jan Wieck))
List pgsql-hackers
Jan Wieck writes:

>     Anyway, it's good to hear you're  still  on  it.  What's  the
>     estimated time you think it'll be ready to get patched in?

Next release. I would hope we can get the current stuff into beta in a
month or so, whereas this project would break open a lot of things.


>     The thing users actually complain about is the requirement of
>     UPDATE permissions to REFERENCE a table. This could be  fixed
>     with  making  RI  triggers setuid functions for 7.1 and check
>     that  the  user  at  least  has  SELECT  permission  on   the
>     referenced table during constraint creation.  This would also
>     remove the actual DOS problem, that a user  could  potentiall
>     create  a  referencing  table  and  not giving anyone who can
>     update the referenced one update permissions on it too.
> 
>     I think it's worth doing it now, and  couple  it  later  with
>     your general access control things.

True. I had already looked into this, it's not fundamentally difficult,
but there's a lot of code that will need to be touched.

If you want to go for it, be my guest; I agree that it is fairly
orthogonal to the rest of the privilege system. I'll put it on my priority
list if no one's taking it.


-- 
Peter Eisentraut                  Sernanders väg 10:115
peter_e@gmx.net                   75262 Uppsala
http://yi.org/peter-e/            Sweden



pgsql-hackers by date:

Previous
From: Peter Eisentraut
Date:
Subject: Installation layout idea
Next
From: Peter Eisentraut
Date:
Subject: Re: Re: [GENERAL] random() function produces wrong range