Re: Auditing extension for PostgreSQL (Take 2) - Mailing list pgsql-hackers

From David Steele
Subject Re: Auditing extension for PostgreSQL (Take 2)
Date
Msg-id 552EA272.1040906@pgmasters.net
Whole thread Raw
In response to Re: Auditing extension for PostgreSQL (Take 2)  (Tatsuo Ishii <ishii@postgresql.org>)
List pgsql-hackers
On 4/14/15 8:37 PM, Tatsuo Ishii wrote:
> BTW, in my understanding pg_audit allows to track a table access even
> if it's used in a view. I think this is a nice feature and it would be
> better explicitly stated in the document and the test case is better
> included in the regression test.
>
> Here is a sample session:
>
> CREATE TABLE test2 (id INT);
> CREATE VIEW vtest2 AS SELECT * FROM test2;
> GRANT SELECT ON TABLE public.test2 TO auditor;
> GRANT SELECT ON TABLE public.vtest2 TO auditor;
> SELECT * FROM vtest2;
> NOTICE:  AUDIT: SESSION,1,1,READ,SELECT,,,SELECT * FROM vtest2;
> NOTICE:  AUDIT: OBJECT,1,1,READ,SELECT,VIEW,public.vtest2,SELECT * FROM vtest2;
> NOTICE:  AUDIT: OBJECT,1,1,READ,SELECT,TABLE,public.test2,SELECT * FROM vtest2;

That's the idea!  In the documentation I throw around the word
"relation" pretty liberally, but you are right that some clarification
would be helpful.

I have added a few parenthetical statements to the docs that should make
them clearer.  I also took your suggestion and added a view regression test.

Both are in patch v9 which I attached to my previous email on this thread.

Thank you for taking the time to have a look.

--
- David Steele
david@pgmasters.net


pgsql-hackers by date:

Previous
From: David Steele
Date:
Subject: Re: Auditing extension for PostgreSQL (Take 2)
Next
From: Qingqing Zhou
Date:
Subject: Re: Assert there is no duplicated exit callbacks