Float << and >> should be more compatible to Fixnum

test/t/float.rb

@@ -178,3 +178,25 @@ assert('Float#nan?') do
   assert_false (1.0/0.0).nan?
   assert_false (-1.0/0.0).nan?
 end
+
+assert('Float#<<') do
+  # Left Shift by one
+  assert_equal 46, 23.0 << 1
+
+  # Left Shift by a negative is Right Shift
+  assert_equal 23, 46.0 << -1
+end
+
+assert('Float#>>') do
+  # Right Shift by one
+  assert_equal 23, 46.0 >> 1
+
+  # Right Shift by a negative is Left Shift
+  assert_equal 46, 23.0 >> -1
+
+  # Don't raise on large Right Shift
+  assert_equal 0, 23.0 >> 128
+
+  # Don't raise on large Right Shift
+  assert_equal -1, -23.0 >> 128
+end