Re: Missing update of all_hasnulls in BRIN opclasses

[Tomas Vondra]

On 5/7/23 07:08, Julien Rouhaud wrote:
> Hi,
> 
> On Sun, May 07, 2023 at 12:13:07AM +0200, Tomas Vondra wrote:
>>
>> c) ignore the issue - AFAICS this would be an issue only for (external)
>> code accessing BrinMemTuple structs, but I don't think we're aware of
>> any out-of-core BRIN opclasses or anything like that ...
> 
> FTR there's at least postgis that implments BRIN opclasses (for geometries and
> geographies), but there's nothing fancy in the implementation and it doesn't
> access BrinMemTuple struct.

Right. I believe that should be fine, because opclasses don't access the
tuple directly - instead we pass pointers to individual pieces. But
maybe it'd be a good idea to test this.


regards

-- 
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company