Empty query with SUM and HAVING, no GROUP BY returns NULL instead of empty result.

[Timour Katchaounov]

The following test case (for LP BUG#609121) extracted from subselect_mat.test:

create table t1 (c1 int);
create table t2 (c2 int);
insert into t1 values (1);
insert into t2 values (2);

SELECT SUM(c1) c1_sum FROM t1 WHERE c1 IN (SELECT c2 FROM t2) HAVING c1_sum;

incorrectly produces a NULL instead of an empty result.

[Timour Katchaounov (timour)]

The bug is present both with MATERIALIZATION and with IN-TO-EXISTS,
but not with SEMIJOIN:
set @@optimizer_switch='materialization=on,in_to_exists=on,semijoin=off';
set @@optimizer_switch='materialization=off,in_to_exists=on,semijoin=off';
set @@optimizer_switch='materialization=on,in_to_exists=off,semijoin=off';

[Timour Katchaounov (timour)]

Analysis:
Close to its end JOIN::optimize() assigns having to tmp_having, and
sets the having clause to NULL:

  tmp_having= having;
  if (select_options & SELECT_DESCRIBE)
  {
    error= 0;
    DBUG_RETURN(0);
  }
  having= 0;

At the same time, this query detects an empty result set, and calls
return_zero_rows(), which must check the HAVING clause by:

    if (having && having->val_int() == 0)
      send_row=0;

However having has been already set to NULL, so return_zero_rows
doesn't check the having clause, hence the wrong result.

Solution:
There are two ways to solve this problem:
a) check join->tmp_having in addition to join->having, or
b) do not set having= 0 in JOIN::optimize.