Hey!
I was investigating a leak reported in the PostGIS issues tracker [1] which led me to the Postgres side where the problem really is. The leak is reproducible with query from original ticket [1]:
WITH latitudes AS ( SELECT generate_series AS latitude FROM generate_series(-90, 90, 0.1)
), longitudes AS ( SELECT generate_series AS longitude FROM generate_series(-180, 180, 0.1)
), points AS ( SELECT ST_SetSRID(ST_Point(longitude, latitude), 4326)::geography AS geog FROM latitudes CROSS JOIN longitudes
)
SELECT geog, ( SELECT name FROM ne_110m_admin_0_countries AS ne ORDER BY p.geog <-> ne.geog LIMIT 1 )
FROM points AS p
;
The leak is only noticeable when index scan with reorder happens as part of subquery plan which is explained by the fact that heap tuples cloned in reorderqueue_push are not freed during flush of reorder queue in ExecReScanIndex.
[1]
https://trac.osgeo.org/postgis/ticket/4720