Adjust cost for re-creating a row from the JOIN CACHE

Creating a record from the join cache is faster than getting a row from
the engine (less and simpler code to execute).

Added JOIN_CACHE_ROW_COPY_COST_FACTOR (0.5 for now) as the factor to
take this into account. This is multiplied with ROW_COPY_COST.

Other things:
- Added cost of copying rows to hash join, similar to join_cache joins.

mysql-test/main/subselect_no_exists_to_in.result

@@ -7046,8 +7046,8 @@ SELECT * FROM t1
 WHERE a = (SELECT MAX(b) FROM t2 WHERE c IN (SELECT MAX(d) FROM t3)) OR a = 10;
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
 1	PRIMARY	t1	index	a	a	5	NULL	2	Using where; Using index
-2	SUBQUERY	t2	ALL	NULL	NULL	NULL	NULL	2	Using where
-2	SUBQUERY	<subquery3>	eq_ref	distinct_key	distinct_key	4	test.t2.c	1	
+2	SUBQUERY	<subquery3>	ALL	distinct_key	NULL	NULL	NULL	1	
+2	SUBQUERY	t2	ALL	NULL	NULL	NULL	NULL	2	Using where; Using join buffer (flat, BNL join)
 3	MATERIALIZED	t3	ALL	NULL	NULL	NULL	NULL	2	
 SELECT * FROM t1
 WHERE a = (SELECT MAX(b) FROM t2 WHERE c IN (SELECT MAX(d) FROM t3)) OR a = 10;