Re: BUG #1145: silent REVOKE failures

Dear Tom,

> "PostgreSQL Bugs List" <pgsql-bugs@postgresql.org> writes:
> > The REVOKE failure should be reported.
>
> What failure?  This looks perfectly fine to me.

"Ex nihilo dixit quod libet", as we used to say in latin and in maths.

Sorry if say something stupid, but I cannot see why it is fine.

Well, if I issue a "REVOKE" and the rights are not revoked and could never
have been because I have no right to issue such statement on the object, I
tend to call this deep absence of success a "failure".

If I do the very opposite GRANT, I have a clear "permission denied".
I wish I had the very same error on REVOKE, because for both operations
you should need to be either a super user, the owner or to have a relevant
grant options?

Look at the very same with unix: sh> chmod o-r /tmp/
chmod: changing permissions of `/tmp/': Operation not permitted

If you want to call that a "feature", I disagree without further strong
argument, and anyway the documentation should be clear about that.

Have a nice day,

--
Fabien Coelho - coelho@cri.ensmp.fr