Re: [PATCHES] Users/Groups -> Roles

From: Stephen Frost <sfrost(at)snowman(dot)net>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: [PATCHES] Users/Groups -> Roles
Date: 2005-06-29 18:36:51
Message-ID: 20050629183651.GY24207@ns.snowman.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers pgsql-patches

* Tom Lane (tgl(at)sss(dot)pgh(dot)pa(dot)us) wrote:
> I notice that AddRoleMems/DelRoleMems assume that ADMIN OPTION is not
> inherited indirectly; that is it must be granted directly to you.
> This seems wrong; SQL99 has under <privileges>
>
> 19) B has the WITH ADMIN OPTION on a role if a role authorization
> descriptor identifies the role as granted to B WITH ADMIN OPTION
> or a role authorization descriptor identifies it as granted WITH
> ADMIN OPTION to another applicable role for B.
>
> and in the Access Rules for <grant role statement>
>
> 1) Every role identified by <role granted> shall be contained
> in the applicable roles for A and the corresponding role
> authorization descriptors shall specify WITH ADMIN OPTION.
>
> I can't see any support in the spec for the idea that WITH ADMIN OPTION
> doesn't flow through role memberships in the same way as ordinary
> membership; can you quote someplace that implies this?

Hrm, no, sorry, I just interpreted the 'Access Rules' line for <grant
role statement> differently. That is to say:

1) Every role identified by <role granted> shall be contained
(Alright, all the roles which you're granting, right)

in the applicable roles for A and the corresponding role
(A must be in all the roles which are being granted)

authorization descriptors shall specify WITH ADMIN OPTION.
(the grants to A for those rules specify ADMIN OPTION)

This came across to me as meaning "there must exist an authorization
descriptor such that the granted-role equals <role granted>, the grantee
is A and WITH ADMIN OPTION is set". That could only be true if the
grant was done explicitly. Reading from 19 above (which I don't recall
seeing before, or at least not reading very carefully) I think you're
right. Either way is fine with me.

Thanks,

Stephen

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Josh Berkus 2005-06-29 18:37:15 Checkpoint cost, looks like it is WAL/CRC
Previous Message Pavel Stehule 2005-06-29 18:36:37 Re: Proposal: associative arrays for plpgsql (concept)

Browse pgsql-patches by date

  From Date Subject
Next Message Stephen Frost 2005-06-29 19:43:54 Re: Open items
Previous Message Tom Lane 2005-06-29 17:40:20 Re: [PATCHES] Users/Groups -> Roles