On Mon, Sep 14, 2020 at 11:44 PM Jeff Davis <pgsql@j-davis.com> wrote:
> Attached. HashAgg seems to accurately reflect the filesize, as does
> Sort.
For the avoidance of doubt: I think that this is the right way to go,
and that it should be committed shortly, before we stamp 13.0. The
same goes for hashagg-release-write-buffers.patch.
Thanks
--
Peter Geoghegan