The test was skipped,although the dependency test is success.

[jianchao00]

the link: https://pytest-dependency.readthedocs.io/en/stable/scope.html#explicitely-specifying-the-scope.

test_mod_01.py

import pytest

@pytest.mark.dependency()
def test_a():
print("test_a")
pass

@pytest.mark.dependency()
@pytest.mark.xfail(reason="deliberate fail")
def test_b():
print("test_b")
assert False

@pytest.mark.dependency(depends=["test_a"])
def test_c():
print("test_c")
pass

class TestClass(object):

@pytest.mark.dependency()
def test_b(self):
    pass

test_mod_02.py
import pytest

@pytest.mark.dependency()
@pytest.mark.xfail(reason="deliberate fail")
def test_a():
print("test_a")
assert False

@pytest.mark.dependency(
depends=["tests/test_mod_01.py::test_a", "tests/test_mod_01.py::test_c"],
scope='session'
)
def test_e():
print("test_e")
pass

@pytest.mark.dependency(
depends=["tests/test_mod_01.py::test_b", "tests/test_mod_02.py::test_e"],
scope='session'
)
def test_f():
print("test_f")
pass

@pytest.mark.dependency(
depends=["tests/test_mod_01.py::TestClass::test_b"],
scope='session'
)
def test_g():
print("test_g")
pass

"E:\Program Files\Python37\python.exe" "E:\Program Files\JetBrains\PyCharm 2021.3\plugins\python\helpers\pycharm_jb_pytest_runner.py" --path E:/PycharmProjects/SelniumPOM-master/tests/test_mod_02.py
Testing started at 19:23 ...
Launching pytest with arguments E:/PycharmProjects/SelniumPOM-master/tests/test_mod_02.py --no-header --no-summary -q in E:\PycharmProjects\SelniumPOM-master\tests

============================= test session starts =============================
collecting ... collected 4 items

test_mod_02.py::test_a XFAIL (deliberate fail) [ 25%]test_a

@pytest.mark.dependency()
@pytest.mark.xfail(reason="deliberate fail")
def test_a():
print("test_a")

  assert False

E assert False

test_mod_02.py:7: AssertionError

test_mod_02.py::test_e SKIPPED (test_e depends on tests/test_mod_01....) [ 50%]
Skipped: test_e depends on tests/test_mod_01.py::test_a

test_mod_02.py::test_f SKIPPED (test_f depends on tests/test_mod_01....) [ 75%]
Skipped: test_f depends on tests/test_mod_01.py::test_b

test_mod_02.py::test_g SKIPPED (test_g depends on tests/test_mod_01....) [100%]
Skipped: test_g depends on tests/test_mod_01.py::TestClass::test_b

======================== 3 skipped, 1 xfailed in 0.07s ========================

Process finished with exit code 0

the expect result is test_e and test_g is success!

[wmetryka]

I'm having the same issue. I've used this plugin successfuly in the past, but on my current setup it just skips every test that depends on another.

Code sample:

@pytest.mark.dependency()
def test_ping_host(host):
    """Pinging provided host."""
    try:
        response = requests.get(host)
    except (urllib3.exceptions.MaxRetryError, requests.exceptions.ConnectionError):
        raise AssertionError("Host didn't respond.")
    
    assert response is not None

@pytest.mark.dependency(depends=['test_ping_host'])
def test_authentication(user_auth):
    """Asserting that user authentication in conftest was successful."""

    # Checking for error key in the authenticate_user object; it existing implies authentication failed
    assert not user_auth.get("error")
    # Checking to make sure a proper token was returned by our auth function in conftest
    assert user_auth.get("token") is not ""


(venv) PS C:\Users\wmetr\Documents\xxx\pytest_demo> pytest ../pytest_demo/ --host http://127.0.0.1:5000
================================================================ test session starts =================================================================
platform win32 -- Python 3.8.0, pytest-7.1.1, pluggy-1.0.0
rootdir: C:\Users\wmetr\Documents\xxx\pytest_demo
plugins: dependency-0.5.1
collected 2 items

test_sample.py .s                                                                                                                              [100%]

============================================================ 1 passed, 1 skipped in 0.06s ============================================================```

[johncuyle]

I'm also seeing this issue.

testsuite/v1/login_test.py::TestIdentifier::test_login_help[edge_driver] PASSED testsuite/v1/login_test.py::TestIdentifier::test_login_with_security_device[edge_driver] PASSED testsuite/v1/login_test.py::TestIdentifier::test_unregistered_email[edge_driver] PASSED testsuite/v1/login_test.py::TestIdentifier::test_invalid_email[edge_driver] PASSED testsuite/v1/login_test.py::TestIdentifier::test_input_email_enter[edge_driver] PASSED testsuite/v1/login_test.py::TestIdentifier::test_input_email_click_next[edge_driver] PASSED testsuite/v1/login_test.py::TestIdentifier::test_input_email_tab_enter_next[edge_driver] PASSED testsuite/v1/login_test.py::TestPassword::test_login_help[edge_driver] PASSED testsuite/v1/login_test.py::TestPassword::test_login_with_security_device[edge_driver] PASSED testsuite/v1/login_test.py::TestPassword::test_invalid_password_enter[edge_driver] SKIPPED (test_invalid_password_enter[edge_driver] depends on TestIdentifier::test_input_email_enter) testsuite/v1/login_test.py::TestPassword::test_input_password_enter[edge_driver] SKIPPED (test_input_password_enter[edge_driver] depends on TestIdentifier::test_input_email_enter) testsuite/v1/login_test.py::TestPassword::test_input_password_click_login[edge_driver] SKIPPED (test_input_password_click_login[edge_driver] depends on TestIdentifier::test_input_email_enter) testsuite/v1/login_test.py::TestPassword::test_input_password_tab_enter_login[edge_driver] SKIPPED (test_input_password_tab_enter_login[edge_driver] depends on TestIdentifier::test_input_email_enter) testsuite/v1/login_test.py::TestPassword::test_input_password_tab_enter_cancel[edge_driver] SKIPPED (test_input_password_tab_enter_cancel[edge_driver] depends on TestIdentifier::test_input_email_enter)

[jfoust-zymergen]

Any workaround discovered?

[mrbean-bremen]

I just had a look at these (as this issue has been mentioned to me elsewhere), and saw that the original issue by @jianchao00 behaves as expected: the tests have only been run for the second test module (test_mod_02), and they are skipped because they depend on tests that have not been run.

The third example by @johncuyle has not enough information about the tests and how they have been run.

The second example by @wmetryka is weird though - I tried to reproduce it without success; this is a trivial case that has always worked, and I suspect something is different in the environment, but I have no idea what it is. @wmetryka: have you resolved this? Maybe you can point to a repo where this happens?

[autosquash]

I had the issue of dependant tests being skipped always, but only when their module name was alphabetically before their dependencies. Solved using pytest-order plugin with the flag --order-dependencies.

[Atituiset]

I had the issue of dependant tests being skipped always, but only when their module name was alphabetically before their dependencies. Solved using pytest-order plugin with the flag --order-dependencies.

same as with me, so weird

[mrbean-bremen]

Note that this is documented behavior.

[mohit1337]

I am observing this issue even when the test on which other test is dependent is passing and executed before the dependent test. Run this on a fresh pytest and pytest-dependency

import pytest
class TestABC:
    def test_abc(self):
        pass
    def test_pqr(self):
        raise
    @pytest.mark.dependency(depends=['test_abc'])
    def test_xyz(self):
        pass

Ouput:

============================= test session starts ==============================
platform darwin -- Python 3.10.8, pytest-8.4.1, pluggy-1.6.0
rootdir: /Users/mohit.chandak
plugins: anyio-4.3.0, dependency-0.6.0
collected 3 items                                                              

test_abc.py .Fs                                                          [100%]

=================================== FAILURES ===================================
_______________________________ TestABC.test_pqr _______________________________

self = <test_abc.TestABC object at 0x104660f10>

    def test_pqr(self):
>       raise
E       RuntimeError: No active exception to reraise

test_abc.py:8: RuntimeError
=========================== short test summary info ============================
FAILED test_abc.py::TestABC::test_pqr - RuntimeError: No active exception to reraise
==================== 1 failed, 1 passed, 1 skipped in 0.05s ====================

What I observed is that it skips even if there is any other failure, if all tests pass then it doesnt skip. Also it works fine if I name the tests

[mrbean-bremen]

@mohit1337 - check the documentation.
The correct way to write this is:

class TestABC:

    @pytest.mark.dependency()
    def test_abc(self):
        pass

    def test_pqr(self):
        raise

    @pytest.mark.dependency(depends=["TestABC::test_abc"])
    def test_xyz(self):
        pass