On 09/03/2016 08:59 PM, Tom Lane wrote:
> Andreas Seltenreich <seltenreich@gmx.de> writes:
>> Digging in the coredumps, it looks like set_var_from_num() is invoked on
>> an uninitialized NumericVar. Sample gdb session below.
>
> Hm, yeah, looks like numeric_poly_deserialize is missing a required
> init_var() step. Slightly astonishing that this got through even
> minimal testing.
>
> For the archives' sake: I could not reproduce this at default settings,
> because the aggregate calls don't get parallelized. But it fails fairly
> quickly with
> set parallel_tuple_cost = 0;
> set parallel_setup_cost = 0;
> set min_parallel_relation_size = 0;
>
> Thanks for the report!
Thanks for fixing this so quickly. For the record, I did test
parallelization at one point, but maybe it was with an older patch
version or I screwed up something in the tests - otherwise I clearly
would've hit that.
- Heikki