Potential issue with combinations

[jmasonlee]

I have a method that wraps subprocess.run and returns the result:

def run(command, capture_output = False, text = True, shell=True, record=None, playback=None):
    output = subprocess.run(command, capture_output=capture_output, text=text, shell=shell)
    return output

I have a test that verifies all combinations of command, capture_output, text and shell:

def test_subprocess_run():
    commands = ["echo 'foo'", "echo 'foo' 1>&2", ["echo", "'foo'"], ["echo", "'foo'", "1>&2"]]
    text = [True, False]
    capture_output = [True, False]
    shell = [True,False]
    verify_all_combinations(run, [commands, capture_output, text, shell])
    verify_all_combinations_with_kwargs(run, command = commands, text = text, capture_output= capture_output, shell = shell)

I am trying to check that my verify_all_combinations_with_kwargs method works as well as verify_all_combinations, but when I run the two tests on an approved file generated by verify_all_combinations, I get an error for the second verify method:

received (from verify_all_combinations_with_kwargs)
args: ({'command': "echo 'foo'", 'text': True, 'capture_output': False, 'shell': True}) => CompletedProcess(args="echo 'foo'", returncode=0)

approved (from verify_all_combinations)
args: ("echo 'foo'", True, False, True) => CompletedProcess(args="echo 'foo'", returncode=0, stdout=b'foo\n', stderr=b'')

I want the keywords printed out with the arguments, so if I ignore that difference, you can see that the result from running verify_all_combinations adds stdout=b'foo\n', stderr=b''

received (from verify_all_combinations_with_kwargs)
args: ("echo 'foo'", True, False, True) => CompletedProcess(args="echo 'foo'", returncode=0)

approved (from verify_all_combinations)
args: ("echo 'foo'", True, False, True) => CompletedProcess(args="echo 'foo'", returncode=0, stdout=b'foo\n', stderr=b'')

I wrote a second test:

def test_bug_with_combinations():
    input = {'command': "echo 'foo'", 'capture_output': True, 'text': False, 'shell': True}
    verify(run(capture_output=input["capture_output"], command=input["command"], text=input["text"], shell=input["shell"]))
    verify(run(input["command"], input["capture_output"], input["text"], input["shell"]))

In both cases, the output is: CompletedProcess(args="echo 'foo'", returncode=0)

I wrote another test:

def test_bug_with_combinations():
    input = {'command': "echo 'foo'", 'capture_output': True, 'text': False, 'shell': True}
    verify_all_combinations(run, [[input['command']], [input['capture_output']], [input['text']], [input['shell']]])

This returns:

args: ("echo 'foo'", True, False, True) => CompletedProcess(args="echo 'foo'", returncode=0, stdout=b'foo\n', stderr=b'')

I think something is up with verify_all_combinations

[nitsanavni]

Hey! We just did some work around running subprocess.
Did you know there's a subprocess.check_output method?

[jmasonlee]

Yes thank you! This was an issue with my code :-)