Re: RLS policy dump/restore failure due to elided type-casts

On Wed, Apr 20, 2016 at 6:04 PM, David G. Johnston <
david(dot)g(dot)johnston(at)gmail(dot)com> wrote:

> On Wed, Apr 20, 2016 at 5:18 PM, Karl Czajkowski <karlcz(at)isi(dot)edu> wrote:
>
>>
>> CREATE POLICY delete_stuff ON stuff
>> FOR DELETE USING ('example attribute value' = ANY ( ((SELECT
>> current_attributes()))::text[] ));
>>
>>
> The following (untested) structure should be immune to this problem...use
> the knowledge as you see ​fit.
>
> USING ('example_attribute_value' = ANY ( ARRAY( SELECT unnest(attr) FROM
> current_attributes() ca (attr) ) )
>
> I cannot imagine the ARRAY(...) being removed and its presence should
> force the scalar = ANY(array) interpretation.
>
> This does seem broken and likely to be back-patched though - and
> unnest+ARRAY is definitely inefficient so there is a trade-off involved -
> but hopefully only in the short-term (a couple of months probably...?).
>
>
​Actually, the ARRAY is pointless - just use the scalar = ANY(subquery)
form which the unnest makes work correctly.

David J.​