def setUpClass(cls):
# Run a launch so we can get some real looking proc_info and proc_output objects to test
# against
proc_command = os.path.join(
ament_index_python.get_package_prefix('apex_launchtest'),
'lib/apex_launchtest',
'good_proc',
)
proc_env = os.environ.copy()
proc_env['PYTHONUNBUFFERED'] = '1'
def generate_test_description(ready_fn):
no_arg_proc = launch.actions.ExecuteProcess(cmd=[proc_command],
env=proc_env)
one_arg_proc = launch.actions.ExecuteProcess(
cmd=[proc_command, '--one-arg'], env=proc_env)
two_arg_proc = launch.actions.ExecuteProcess(
cmd=[proc_command, '--two-arg', 'arg_two'], env=proc_env)
ld = launch.LaunchDescription([
no_arg_proc, one_arg_proc, two_arg_proc,
launch.actions.OpaqueFunction(function=lambda ctx: ready_fn())
])
return (ld, locals())
arr = []
class PreShutdownTests(unittest.TestCase):
def test_wait_self(self, proc_output, proc_info, no_arg_proc,
one_arg_proc, two_arg_proc):
proc_output.assertWaitFor('--one-arg')
proc_output.assertWaitFor('--two-arg')
proc_output.assertWaitFor('arg_two')
arr.append(proc_info)
# Set up a fake module containing the test data:
test_module = types.ModuleType('test_module')
test_module.generate_test_description = generate_test_description
test_module.FakePreShutdownTests = PreShutdownTests
# Run the test:
runner = ApexRunner(LoadTestsFromPythonModule(test_module))
runner.run()
# Grab the very realistic proc_info object to use for tests below
cls.proc_info = arr[0]
def test_dut_that_has_exception(capsys):
# This is the same as above, but we also want to check we get extra output from processes
# that had an exit code
def generate_test_description(ready_fn):
TEST_PROC_PATH = os.path.join(
ament_index_python.get_package_prefix('apex_launchtest'),
'lib/apex_launchtest', 'terminating_proc')
EXIT_PROC_PATH = os.path.join(
ament_index_python.get_package_prefix('apex_launchtest'),
'lib/apex_launchtest', 'exit_code_proc')
return launch.LaunchDescription([
launch.actions.ExecuteProcess(cmd=[TEST_PROC_PATH, '--exception']),
# This process makes sure we can handle processes that exit with a code but don't
# generate any output. This is a regression test for an issue fixed in PR31
launch.actions.ExecuteProcess(cmd=[EXIT_PROC_PATH, '--silent']),
launch.actions.OpaqueFunction(function=lambda context: ready_fn()),
])
with mock.patch('apex_launchtest.apex_runner._RunnerWorker._run_test'):
runner = ApexRunner([TR(generate_test_description, {}, [], [])])
results = runner.run()
for result in results.values():
assert not result.wasSuccessful()
# Make sure some information about WHY the process died shows up in the output
out, err = capsys.readouterr()
assert 'Starting Up' in out
assert 'Process had a pretend error' in out # This is the exception text from exception_node
def test_skipped_launch_description():
@unittest.skip('skip reason string')
def generate_test_description(ready_fn):
raise Exception('This should never be invoked') # pragma: no cover
class FakePreShutdownTests(unittest.TestCase):
def test_fail_always(self):
assert False # pragma: no cover
@apex_launchtest.post_shutdown_test()
class FakePostShutdownTests(unittest.TestCase):
def test_fail_always(self):
assert False # pragma: no cover
test_module = types.ModuleType('test_module')
test_module.generate_test_description = generate_test_description
test_module.FakePreShutdownTests = FakePreShutdownTests
test_module.FakePostShutdownTests = FakePostShutdownTests
# Run the test:
runner = ApexRunner(LoadTestsFromPythonModule(test_module))
results = runner.run()
# Should just get one result, even though there were multiple tests
assert len(results.values()) == 1
def test_dut_that_shuts_down(capsys):
def generate_test_description(ready_fn):
TEST_PROC_PATH = os.path.join(
ament_index_python.get_package_prefix('apex_launchtest'),
'lib/apex_launchtest', 'terminating_proc')
return launch.LaunchDescription([
launch.actions.ExecuteProcess(cmd=[TEST_PROC_PATH]),
launch.actions.OpaqueFunction(function=lambda context: ready_fn()),
])
with mock.patch('apex_launchtest.apex_runner.ApexRunner._run_test'):
runner = ApexRunner(
gen_launch_description_fn=generate_test_description,
test_module=None)
pre_result, post_result = runner.run()
assert not pre_result.wasSuccessful()
assert not post_result.wasSuccessful()
# This is the negative version of the test below. If no exit code, no extra output
# is generated
out, err = capsys.readouterr()
assert "Starting Up" not in out
def test_parametrized_run_with_one_failure():
# Test Data
@apex_launchtest.parametrize('arg_val', [1, 2, 3, 4, 5])
def generate_test_description(arg_val, ready_fn):
TEST_PROC_PATH = os.path.join(
ament_index_python.get_package_prefix('apex_launchtest'),
'lib/apex_launchtest', 'good_proc')
# This is necessary to get unbuffered output from the process under test
proc_env = os.environ.copy()
proc_env['PYTHONUNBUFFERED'] = '1'
return launch.LaunchDescription([
launch.actions.ExecuteProcess(
cmd=[TEST_PROC_PATH],
env=proc_env,
),
launch.actions.OpaqueFunction(function=lambda context: ready_fn())
])
class FakePreShutdownTests(unittest.TestCase):
def test_fail_on_two(self, proc_output, arg_val):
proc_output.assertWaitFor('Starting Up')
assert arg_val != 2
@apex_launchtest.post_shutdown_test()
class FakePostShutdownTests(unittest.TestCase):
def test_fail_on_three(self, arg_val):
assert arg_val != 3
# Set up a fake module containing the test data:
test_module = types.ModuleType('test_module')
test_module.generate_test_description = generate_test_description
test_module.FakePreShutdownTests = FakePreShutdownTests
test_module.FakePostShutdownTests = FakePostShutdownTests
# Run the test:
runner = ApexRunner(LoadTestsFromPythonModule(test_module))
results = runner.run()
passes = [result for result in results.values() if result.wasSuccessful()]
fails = [
result for result in results.values() if not result.wasSuccessful()
]
assert len(passes) == 3 # 1, 4, and 5 should pass
assert len(
fails
) == 2 # 2 fails in an active test, 3 fails in a post-shutdown test
def test_nominally_good_dut():
# The following is test setup nonsense to turn a string into a python module that can be
# passed to the apex runner. You can skip over this. It does not add to your understanding
# of the test.
test_code = """
import unittest
from apex_launchtest import post_shutdown_test
class PreTest(unittest.TestCase):
def test_pre_ok(self):
pass
@post_shutdown_test()
class PostTest(unittest.TestCase):
def test_post_ok(self):
pass
"""
module = imp.new_module('test_module')
exec(test_code, module.__dict__)
# Here's the actual 'test' part of the test:
TEST_PROC_PATH = os.path.join(
ament_index_python.get_package_prefix('apex_launchtest'),
'lib/apex_launchtest', 'good_proc')
def generate_test_description(ready_fn):
return launch.LaunchDescription([
launch.actions.ExecuteProcess(cmd=[TEST_PROC_PATH]),
launch.actions.OpaqueFunction(function=lambda context: ready_fn()),
])
module.generate_test_description = generate_test_description
runner = ApexRunner(LoadTestsFromPythonModule(module))
results = runner.run()
for result in results.values():
assert result.wasSuccessful()