Re: [HACKERS] Permissions on copy

Bruce Momjian <maillist@candle.pha.pa.us>

From: Bruce Momjian <maillist@candle.pha.pa.us>
To: scrappy@hub.org (The Hermit Hacker)
Cc: Andreas.Zeugswetter@telecom.at, pgsql-hackers@hub.org
Date: 1998-02-20T17:22:03Z
Lists: pgsql-hackers
> 
> On Fri, 20 Feb 1998, Bruce Momjian wrote:
> 
> > Two things.  First was a separate COPY priviledge, which I vote against.
> > I see no real value to it, except to work around the problem that COPY
> > doesn't use rules.
> 
> 	Okay, I may be totally out in left field here (ie. unrelated), but
> what stops a user from doing a 'COPY out' on a table that they don't have
> SELECT privileges on?  Kind of negates 'REVOKE ALL...', no?

Yes I think a separate COPY permission makes no sense.

> 

> > Second, there was the idea of making copy allow a real select statement
> > and not just a table name.  If we do that, all goes through the
> > executor, and you get view and rules working properly.  May have some
> > performance penalty, though it probabably will be minor.
> 
> 	This sounds reasonable...
> 
> 
> 


-- 
Bruce Momjian
maillist@candle.pha.pa.us