From: "Hiroshi Inoue" <Inoue@tpf.co.jp>
To: "Constantin Teodorescu" <teo@flex.ro>, "Tom Lane" <tgl@sss.pgh.pa.us>
Cc: "Ed Loehr" <eloehr@austin.rr.com>,
	"PostgreSQL Interfaces" <pgsql-interfaces@postgreSQL.org>
Subject: RE: [INTERFACES] A question on triggers
Date: Sun, 20 Feb 2000 00:15:29 +0900
Message-ID: <NDBBIJLOILGIKBGDINDFMENLCCAA.Inoue@tpf.co.jp>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: 7bit
In-reply-to: <38AC3840.EB320B34@flex.ro>
Importance: Normal

> -----Original Message-----
> From: owner-pgsql-interfaces@postgreSQL.org 
> [mailto:owner-pgsql-interfaces@postgreSQL.org]On Behalf Of 
> Constantin Teodorescu
> 
> Tom Lane wrote:
> > 
> > > But wouldn't be nice if PostgreSQL could offer a more elegant 
> solution?
> > 
> > Can't you do it with the normal GRANT/REVOKE access-control mechanism?
> 
> No. It didn't worked, that was my first idea!
> 
> > I'm pretty sure that for rules (views), the access rights for queries
> > issued within the rule are checked based on the owner of the rule, not
> > the user who invoked the rule.  Triggers ought to work the same way,
> > though I haven't tried it.  So you could make the trigger function and
> > the protected table owned by the same user, and then not grant write
> > permission on that table to anyone else.
> 
> It seems it didn't work!
>

Hmm,it seems to work for views(rules) but it doesn't work for functions.
It should work for stored procedures,shouldn't it ?

Regards.

Hiroshi Inoue
Inoue@tpf.co.jp