Re: pg_signal_backend() asymmetry - Mailing list pgsql-hackers

From Alvaro Herrera
Subject Re: pg_signal_backend() asymmetry
Date
Msg-id 1348774364-sup-5810@alvh.no-ip.org
Whole thread Raw
In response to Re: pg_signal_backend() asymmetry  (Noah Misch <noah@leadboat.com>)
List pgsql-hackers
Excerpts from Noah Misch's message of mié sep 26 17:54:43 -0300 2012:
> I'm marking this patch Ready for Committer.  I suggest backpatching it to 9.2;
> the patch corrects an oversight in 9.2 changes.  There's more compatibility
> value in backpatching than in retaining distinct behavior for 9.2 only.
>
> On Thu, Jun 28, 2012 at 09:32:41AM -0700, Josh Kupershmidt wrote:
> > !     if (!superuser())
> >       {
> >           /*
> > !          * Since the user is not superuser, check for matching roles.
> >            */
> > !         if (proc->roleId != GetUserId())
> > !             return SIGNAL_BACKEND_NOPERMISSION;
> >       }
>
> I would have collapsed the conditionals and deleted the obvious comment:
>
>     if (!(superuser() || proc->roleId == GetUserId()))
>         return SIGNAL_BACKEND_NOPERMISSION;
>
> The committer can do that if desired; no need for another version.

Thanks, I pushed to HEAD and 9.2 with the suggested tweaks.

Documentation doesn't specify the behavior of the non-superuser case
when there's trouble, so I too think the behavior change in 9.2 is okay.
I am less sure about it needing documenting.

--
Álvaro Herrera                http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: psql, remove include of psqlscan.c
Next
From: Tom Lane
Date:
Subject: Re: psql, remove include of psqlscan.c