On Tue, Nov 13, 2018 at 10:59:15AM -0300, Alvaro Herrera wrote:
> I gave the patch a read and it looks reasonable to me.
>
> Memory management in RelationBuildPartitionDesc is crummy, but I don't
> think it's this patch's fault.
I agree, and that shows up pretty clearly after the refactoring is done.
The mess is partially caused by the handling around the case where there
is no partition data to attach to PartitionDescData. I would personally
much prefer if we could also avoid using partition_bounds_copy.
It does not prevent the first refactoring step, so I have committed the
patch as it is already doing a lot.
--
Michael