SUM(distinct) gives wrong result when reducing max_heap_table_size

CREATE TABLE t1 (id INTEGER);

CREATE TABLE t2 (id INTEGER);

INSERT INTO t1 VALUES (1), (1), (1),(1);

INSERT INTO t1 SELECT id FROM t1;

INSERT INTO t1 SELECT id FROM t1;

INSERT INTO t1 SELECT id FROM t1;

INSERT INTO t1 SELECT id FROM t1;

INSERT INTO t1 SELECT id FROM t1;

INSERT INTO t1 SELECT id+1 FROM t1;

INSERT INTO t1 SELECT id+2 FROM t1;

INSERT INTO t1 SELECT id+4 FROM t1;

INSERT INTO t1 SELECT id+8 FROM t1;

INSERT INTO t1 SELECT id+16 FROM t1;

INSERT INTO t1 SELECT id+32 FROM t1;

INSERT INTO t1 SELECT id+64 FROM t1;

INSERT INTO t1 SELECT id+128 FROM t1;

INSERT INTO t1 SELECT id+256 FROM t1;

INSERT INTO t1 SELECT id+512 FROM t1;

INSERT INTO t1 SELECT id+1024 FROM t1;

INSERT INTO t1 SELECT id+2048 FROM t1;

INSERT INTO t1 SELECT id+4096 FROM t1;

INSERT INTO t1 SELECT id+8192 FROM t1;

INSERT INTO t2 SELECT id FROM t1 ORDER BY id*rand();

SELECT SUM(DISTINCT id) sm FROM t2;

SET max_heap_table_size=16384;

SHOW variables LIKE 'max_heap_table_size';

SELECT SUM(DISTINCT id) sm FROM t2;

DROP TABLE t1;

DROP TABLE t2;