Tests: Cope with different invocation methods

Depending on how the tests are run (`pytest`, `python -m pytest`, ...),
the program name can be e.g. `pytest` or `pytest.py`. Relax the relevant
tests a bit, this way we don't have to update them every time we change
something related to CLI arguments either.

Signed-off-by: Nils Philippsen <nils@redhat.com>

tests/plugins/test_packager_bugzilla_sync.py

@@ -118,12 +118,15 @@ class TestPackagerBugzillaSyncToddler:
     def test_main_no_args(self, capsys):
         with pytest.raises(SystemExit):
             toddlers.plugins.packager_bugzilla_sync.main([])
+
         out, err = capsys.readouterr()
-        exp = """usage: pytest [-h] [--dry-run] [-q | --debug] conf [username]
-pytest: error: the following arguments are required: conf
-"""
         assert out == ""
-        assert err == exp
+        # Expecting something along these lines, but don't make the test too tight:
+        #
+        # usage: pytest [-h] [--dry-run] [-q | --debug] conf [username]
+        # pytest: error: the following arguments are required: conf
+        assert err.startswith("usage:")
+        assert "error: the following arguments are required:" in err
 
     @patch("toml.load", new=Mock(return_value={}))
     def test_main_debug(self, capsys):