WITH CHECK OPTION bug [was RLS Design]

On 20 September 2014 06:13, Andrew Gierth <andrew(at)tao11(dot)riddles(dot)org(dot)uk> wrote:
>>>>>> "Adam" == Brightwell, Adam <adam(dot)brightwell(at)crunchydatasolutions(dot)com> writes:
>
> Adam> At any rate, this appears to be a previously existing issue
> Adam> with WITH CHECK OPTION. Thoughts?
>
> It's definitely an existing issue; you can reproduce it more simply,
> no need to mess with different users.
>
> The issue as far as I can tell is that the withCheckOption exprs are
> not being processed anywhere in setrefs.c, so it only works at all by
> pure fluke: for most operators, the opfuncid is also filled in by
> eval_const_expressions, but for whatever reason SAOPs escape this
> treatment. Same goes for other similar cases:
>
> create table colors (name text);
> create view vc1 as select * from colors where name is distinct from 'grue' with check option;
> create view vc2 as select * from colors where name in ('red','green','blue') with check option;
> create view vc3 as select * from colors where nullif(name,'grue') is null with check option;
>
> insert into vc1 values ('red'); -- fails
> insert into vc2 values ('red'); -- fails
> insert into vc3 values ('red'); -- fails
>

Oh dear. I remember thinking at the time I wrote the WITH CHECK OPTION
stuff that I needed to check all the places that did returningLists
processing, because they would probably need similar processing for
withCheckOptionLists, but somehow I missed that one place.

Fortunately it looks pretty trivial though. The patch attached fixes
the above test cases.

Obviously this needs to be fixed in 9.4 and HEAD.

Regards,
Dean