Re: [v9.2] Fix Leaky View Problem

Robert Haas <robertmhaas@gmail.com>

From: Robert Haas <robertmhaas@gmail.com>

To: Kohei KaiGai <kaigai@kaigai.gr.jp>

Cc: Jeff Janes <jeff.janes@gmail.com>, Tom Lane <tgl@sss.pgh.pa.us>, Noah Misch <noah@leadboat.com>, Kevin Grittner <Kevin.Grittner@wicourts.gov>, Kohei.Kaigai@emea.nec.com, thom@linux.com, pgsql-hackers@postgresql.org

Date: 2012-01-24T17:26:58Z

Lists: pgsql-hackers

On Sun, Jan 22, 2012 at 5:57 AM, Kohei KaiGai <kaigai@kaigai.gr.jp> wrote:
>> This passes installcheck initially.  Then upon second invocation of
>> installcheck, it fails.
>>
>> It creates the role "alice", and doesn't clean it up.  On next
>> invocation alice already exists and cases a failure in test
>> select_views.
>>
> Thanks for your pointing out.
>
> The attached patch adds cleaning-up part of object being defined
> within this test;
> includes user "alice".

Urp.  I failed to notice this patch and committed a different fix for
the problem pointed out by Jeff.  I'm inclined to think it's OK to
leave the non-shared objects behind - among other things, if people
are testing pg_upgrade using the regression database, this will help
to ensure that pg_dump is handling security_barrier views correctly.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company