Re: Schemas: status report, call for developers - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Schemas: status report, call for developers
Date
Msg-id 18643.1020377107@sss.pgh.pa.us
Whole thread Raw
In response to Re: Schemas: status report, call for developers  (Ian Barwick <barwick@gmx.de>)
List pgsql-hackers
Ian Barwick <barwick@gmx.de> writes:
> i.e. user "joe" can see which objects exist in schema "foo2", even though
> he has no USAGE privilege. (Is this behaviour intended?)

It's open for debate I suppose.  Historically we have not worried about
preventing people from looking into the system tables, except for cases
such as pg_statistic where this might expose actual user data.

AFAICS we could only prevent this by making selective views on the
system tables and then prohibiting ordinary users from accessing the
underlying tables directly.  I'm not in a big hurry to do that myself,
if only for backward-compatibility reasons.

We still do have the option of separate databases, and I'd be inclined
to tell people to use those if they want airtight separation between
users.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Jason Earl
Date:
Subject: Re: PostgreSQL mission statement?
Next
From: Mark kirkwood
Date:
Subject: Re: PostgreSQL mission statement?