Fix unit test failure in test_roledb.py

Ensure roledb is empty prior to running the test cases.  roledb.py may have been modified by other unit tests that did not properly clean up.

tests/unit/test_roledb.py

@@ -38,7 +38,7 @@
 
 class TestRoledb(unittest.TestCase):
   def setUp(self):
-    pass
+    pass 
 
 
 
@@ -403,6 +403,21 @@ def _test_rolename(self, test_function):
 
 
 
+def setUpModule():
+  # setUpModule is called before any test cases run.
+  # Ensure the roledb has not been modified by a previous test, which may
+  # affect assumptions (i.e., empty roledb) made by the tests cases in this
+  # unit test.
+  tuf.roledb.clear_roledb()
+
+def tearDownModule():
+  # tearDownModule is called after all the tests have run.
+  # Ensure we clean up roledb.  Courtesy is contagious, and it begins with
+  # test_roledb.py.
+  tuf.roledb.clear_roledb()
+
+
+
 # Run the unit tests.
 if __name__ == '__main__':
   unittest.main()