Re: [PATCH] Reworks for Access Control facilities (r2311)

>>>>> I don't find the comment regarding what happened with FindConversion to
>>>>> be nearly descriptive enough. Can you elaborate on why the check wasn't
>>>>> necessary and has now been removed? If it really isn't needed, why have
>>>>> that function at all?
>>> http://archives.postgresql.org/message-id/26499.1250706473@sss.pgh.pa.us
>>>
>>> I'll add a comment about the reason why this check was simply eliminated.
>> Could these kind of changes be done as a separate patch? Perhaps one
>> which should be applied first? It's alot easier to review if we can
>> have:
>>
>> - patches to fix things in PG (such as the above)
>> - patch to add in ac_* model
>
> I think we can apply these kind of eliminations earlier or later.
> These checks might be redundant or unnecessary, but harmless.
> As far as the reworks patch does not touch them, it does not affect
> to our discussion.

The attached patch eliminates permission checks in FindConversion()
and EnableDisableRule(), because these are nonsense or redundant.

It is an separated issue from the ac_*() routines.
For now, we decided not to touch these stuffs in the access control
reworks patch. So, we can discuss about these fixes as a different
topic.

See the corresponding messages:
http://archives.postgresql.org/message-id/26499.1250706473@sss.pgh.pa.us
http://archives.postgresql.org/message-id/4ABC136A.90903@ak.jp.nec.com

Thanks,
--
OSS Platform Development Division, NEC
KaiGai Kohei <kaigai(at)ak(dot)jp(dot)nec(dot)com>