[omniORB] Running Tests

[Antonio Beamud Montero]

Hi all:
I've a problem trying to test my corba app. If I run each test isolated,
all works fine, but trying to run the suite, the second test fails. For
example:

class CorbaTest(unittest.TestCase):
     def test_simple1(self):
            sys.argv += ['-ORBendPoint', ':unix:/tmp/corba.sock']

            orb = CORBA.ORB_init(sys.argv, CORBA.ORB_ID)   
            inspoa = orb.resolve_initial_references("omniINSPOA")
            servant = Echo_i()
            inspoa.activate_object_with_id('echo_1', servant)
            inspoa._get_the_POAManager().activate()

            assert os.path.exists('/tmp/corba.sock')

     def test_simple2(self):
            sys.argv += ['-ORBendPoint', ':unix:/tmp/corba2.sock']

            orb = CORBA.ORB_init(sys.argv, CORBA.ORB_ID)   
            inspoa = orb.resolve_initial_references("omniINSPOA")
            servant = Echo_i()
            inspoa.activate_object_with_id('echo_2', servant)
            inspoa._get_the_POAManager().activate()

            assert os.path.exists('/tmp/corba2.sock')

Running this test suite, the second socket ('/tmp/corba2.sock'), isn't
created. If I run only this test, all works fine.

I've tried to reinit the Omniorb environment, but without success.
How can I reset the environment for each test?

Thanks.