test_division test case difference between Python and C++

[TomyYamy]

Python

rtamt/tests/python/api/test_stl_discrete_time_online_specification.py

Lines 112 to 131 in 9a0ddb1

	def test_division(self):
	spec = rtamt.STLDiscreteTimeSpecification()
	spec.declare_var('req', 'float')
	spec.declare_var('gnt', 'float')
	spec.declare_var('out', 'float')
	spec.spec = 'out = req / gnt'
	spec.parse()
	out1 = spec.update(0, [('req', self.left1), ('gnt', self.right1)])
	out2 = spec.update(1, [('req', self.left2), ('gnt', self.right2)])
	out3 = spec.update(2, [('req', self.left3), ('gnt', self.right3)])
	out4 = spec.update(3, [('req', self.left4), ('gnt', self.right4)])
	out5 = spec.update(4, [('req', self.left5), ('gnt', self.right5)])
	self.assertEqual(out1, 100 / 20, "input 1")
	self.assertEqual(out2, -1 / -2, "input 2")
	self.assertEqual(out3, -2 / 10, "input 3")
	self.assertEqual(out4, 5 / 4, "input 4")
	self.assertEqual(out5, -1 / -1, "input 5")

C++

rtamt/tests/python/api/test_stl_discrete_time_online_specification_cpp.py

Lines 112 to 131 in 9a0ddb1

	def test_division(self):
	spec = rtamt.StlDiscreteTimeOnlineSpecificationCpp()
	spec.declare_var('req', 'float')
	spec.declare_var('gnt', 'float')
	spec.declare_var('out', 'float')
	spec.spec = 'out = req / gnt'
	spec.parse()
	out1 = spec.update(0, [('req', self.left1), ('gnt', self.right1)])
	out2 = spec.update(1, [('req', self.left2), ('gnt', self.right2)])
	out3 = spec.update(2, [('req', self.left3), ('gnt', self.right3)])
	out4 = spec.update(3, [('req', self.left4), ('gnt', self.right4)])
	out5 = spec.update(4, [('req', self.left5), ('gnt', self.right5)])
	self.assertEqual(out1, 100.0 / 20.0, "input 1")
	self.assertEqual(out2, -1.0 / -2.0, "input 2")
	self.assertEqual(out3, -2.0 / 10, "input 3")
	self.assertEqual(out4, 5.0 / 4.0, "input 4")
	self.assertEqual(out5, -1.0 / -1.0, "input 5")

Is that OK?

[nickovic]

Can you please check the branch 154_division? All tests pass with Python 2.7 for me.

[TomyYamy]

@nickovic The test us passed, but the problem is the test case itself different between cpp and python.

Python

     self.assertEqual(out1, 100 / 20, "input 1") 
     self.assertEqual(out2, -1 / -2, "input 2") 
     self.assertEqual(out3, -2 / 10, "input 3") 
     self.assertEqual(out4, 5 / 4, "input 4") 
     self.assertEqual(out5, -1 / -1, "input 5") 

CPP

     self.assertEqual(out1, 100.0 / 20.0, "input 1") 
     self.assertEqual(out2, -1.0 / -2.0, "input 2") 
     self.assertEqual(out3, -2.0 / 10, "input 3") 
     self.assertEqual(out4,  5.0 / 4.0, "input 4") 
     self.assertEqual(out5, -1.0 / -1.0, "input 5") 

Ideally it should be same.