Re: Using views for row-level access control is leaky

2009/10/22 Richard Huxton <dev(at)archonet(dot)com>:
> Pavel Stehule wrote:
>> 2009/10/22 Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>:
>>> That example I ran on CVS HEAD, but it's a generic problem on all versions.
>> postgres=# select version();
>>                                                    version
>> ────────────────────────────────────────────────────────────────────────────────────
>>  PostgreSQL 8.5devel on i686-pc-linux-gnu, compiled by GCC gcc (GCC)
>> 4.4.1 20090725
>> (1 row)
>>
>> postgres=# select * from x;
>>  a  │ b
>> ────┼────
>>  10 │ 20
>> (1 row)
>>
>> postgres=# create view v as select * from x where b <> 20;
>                                                    ^^^^^^^
> This is the expression that needs to be expensive. Then the exposing
> function needs to be cheap. That makes the planner run the exposing
> function first.
>

postgres=# create or replace function vv(int, int) returns bool as
$$begin raise notice '% %', $1, $2; return true; end$$ language
plpgsql COST 0.000001;
CREATE FUNCTION
postgres=# select * from v where vv(a,b);NOTICE: 10 20
a │ b
───┼───
(0 rows)

still I have not bad result, but, yes, I see what I could not to see.

Pavel

> --
>  Richard Huxton
>  Archonet Ltd
>