Message-ID: <462cd5de-3ade-989b-0798-7e1cab562991@pgmasters.net>
Date: Fri, 9 Jun 2023 17:46:48 +0300
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:102.0)
 Gecko/20100101 Thunderbird/102.11.1
Subject: Re: Views no longer in rangeTabls?
Content-Language: en-US
To: Amit Langote <amitlangote09@gmail.com>
Cc: Pg Hackers <pgsql-hackers@postgresql.org>
References: <3953179e-9540-e5d1-a743-4bef368785b0@pgmasters.net>
 <CA+HiwqFC0xDCCiXbUjb-xisxo9PPuA8_ykn8Fz+6RbBJBE8mVA@mail.gmail.com>
From: David Steele <david@pgmasters.net>
In-Reply-To: <CA+HiwqFC0xDCCiXbUjb-xisxo9PPuA8_ykn8Fz+6RbBJBE8mVA@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <https://www.postgresql.org/message-id/462cd5de-3ade-989b-0798-7e1cab562991%40pgmasters.net>
Precedence: bulk

Hi Amit,

On 6/9/23 14:25, Amit Langote wrote:
> On Fri, Jun 9, 2023 at 17:28 David Steele <david@pgmasters.net 
> <mailto:david@pgmasters.net>> wrote:
> 
>     In prior versions of Postgres, views were listed in rangeTabls when
>     ExecutorCheckPerms_hook() was called but in PG16 the views are no
>     longer
>     in this list.
> 
> I’m not exactly sure how pgAudit’s code is searching for view relations 
> in the range table, but if the code involves filtering on rtekind == 
> RTE_RELATION, then yes, such code won’t find views anymore. That’s 
> because the rewriter no longer adds extraneous RTE_RELATION RTEs for 
> views into the range table. Views are still there, it’s just that their 
> RTEs are of kind RTE_SUBQUERY, but they do contain some RELATION fields 
> like relid, rellockmode, etc.  So an extension hook’s relation RTE 
> filtering code should also consider relid, not just rtekind.

Thank you, this was very helpful. I am able to get the expected result 
now with:

/* We only care about tables/views and can ignore subqueries, etc. */
if (!(rte->rtekind == RTE_RELATION ||
      (rte->rtekind == RTE_SUBQUERY && OidIsValid(rte->relid))))
     continue;

One thing, though, rte->relkind is not set for views, so I still need to 
call get_rel_relkind(rte->relid). Not a big deal, but do you think it 
would make sense to set rte->relkind for views?

> Perhaps, we are missing a comment near the hook definition mentioning 
> this detail about views.

I don't see any meaningful comments near the hook definition. That would 
certainly be helpful.

Thanks!
-David