Re: [PATCH] Add regress test for pg_read_all_stats role

From: Michael Paquier <michael(at)paquier(dot)xyz>
To: Alexandra Ryzhevich <aryzhevich(at)google(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org, Vladimir Rusinov <vrusinov(at)google(dot)com>, Dmitriy Potapov <atomsk(at)google(dot)com>
Subject: Re: [PATCH] Add regress test for pg_read_all_stats role
Date: 2018-09-26 06:29:16
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

On Fri, Aug 24, 2018 at 03:37:17PM +0100, Alexandra Ryzhevich wrote:
> CONNECT ON DATABASE privilege is granted to public by default (so
> all users by default can select pg_database_size()) and
> REVOKE CONNECT ... FROM <username> command doesn't impact
> user's privileges. The only way to revoke connect privilege from user
> is to revoke it from public. That's why maybe it will be less harmful
> just to remove pg_database_size tests at all. But will it be better to
> remove pg_tablespace_size() tests also? if possible costs are more
> harmful than not testing this code path then I'll remove these tests
> also.

Sorry for the delay, Alexandra, and thanks for the new version. I have
removed the tests related to pg_tablespace_size as they are costly and
those for vacuum_cost_delay as we didn't gain more with those tests.
The rest has been committed, which makes a good first cut.

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Michael Paquier 2018-09-26 06:44:08 Re: file cloning in pg_upgrade and CREATE DATABASE
Previous Message Thomas Munro 2018-09-26 04:49:09 Re: pgbench - add pseudo-random permutation function