Test is skipped without a skip mark added to the test

[201dreamers]

• a detailed description of the bug or problem you are having
• output of pip list from the virtual environment you are using
• pytest and operating system versions
• minimal example if possible

Tree of the project:

$ tree -I venv -I __pycache__
.
├── conftest.py
└── test.py

I have 2 test classes that are based on the same class (they use Template Method pattern).
When I add a skip mark to only one test function from the pytest_generate_test hook, both tests are skipped.

test.py contents

class Base:
    def step_1(self):
        pass

    def test_func(self):
        self.step_1()
        assert True


class TestCls1(Base):
    def step_1():
        return 1


class TestCls2(Base):
    def step_1():
        return 2

conftest.py contents:

import pytest


def pytest_generate_tests(metafunc):
    print(f"\n{metafunc}")
    print(f"{metafunc.function=}")

    if  metafunc.cls.__name__ == "TestCls1":
        metafunc.function.pytestmark = pytest.mark.skip("Skip 1")

Output of the run:

(venv) [15:20:05] check_pytest_issue $ pytest test.py -vv
===================================================================== test session starts =====================================================================
platform linux -- Python 3.12.3, pytest-8.3.4, pluggy-1.5.0 -- /mnt/c/Users/dhakman/check_pytest_issue/venv/bin/python
cachedir: .pytest_cache
rootdir: /mnt/c/Users/dhakman/check_pytest_issue
collecting ...
<_pytest.python.Metafunc object at 0x7fb9d4de7c20>
metafunc.function=<function Base.test_func at 0x7fb9d4e404a0>

<_pytest.python.Metafunc object at 0x7fb9d4df5400>
metafunc.function=<function Base.test_func at 0x7fb9d4e404a0>
collected 2 items

test.py::TestCls1::test_func SKIPPED (Skip 1)                                                                                                           [ 50%]
test.py::TestCls2::test_func SKIPPED (Skip 1)                                                                                                           [100%]

===================================================================== 2 skipped in 0.03s ======================================================================

If you take a look at the output, despite metafunc object is different for each test class, the underlying function object is the same object in both cases (id = 0x7fb9d4e404a0).

If we take a look at plain classes in python, in similar structure they have different ids:

Python 3.12.3 (main, Nov  6 2024, 18:32:19) [GCC 13.2.0]
Type 'copyright', 'credits' or 'license' for more information
IPython 8.30.0 -- An enhanced Interactive Python. Type '?' for help.

In [1]: class Base:
   ...:     def test(self):
   ...:         pass
   ...:

In [2]: class A(Base):
   ...:     pass
   ...:

In [3]: class B(Base):
   ...:     pass
   ...:

In [4]: a = A()

In [5]: b = B()

In [6]: id(a.test)
Out[6]: 140631076810816

In [7]: id(b.test)
Out[7]: 140631076791872

Environment:

System:
Ubuntu 24.04 in wsl2 under Windows 11
Also reproduces in original Ubuntu 24.04

(venv) [21:30:49] check_pytest_issue $ uname -a
Linux dllwn 5.15.167.4-microsoft-standard-WSL2 #1 SMP Tue Nov 5 00:21:55 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux

Python & packages:

(venv) [21:30:52] check_pytest_issue $ python --version
Python 3.12.3
(venv) [21:31:24] check_pytest_issue $ pip list
Package               Version
--------------------- -------
asttokens             3.0.0
decorator             5.1.1
docstring-to-markdown 0.15
executing             2.1.0
iniconfig             2.0.0
ipdb                  0.13.13
ipython               8.30.0
jedi                  0.19.2
matplotlib-inline     0.1.7
packaging             24.2
parso                 0.8.4
pexpect               4.9.0
pip                   24.0
pluggy                1.5.0
prompt_toolkit        3.0.48
ptyprocess            0.7.0
pure_eval             0.2.3
Pygments              2.18.0
pytest                8.3.4
python-lsp-jsonrpc    1.1.2
python-lsp-server     1.12.0
ruff                  0.8.4
stack-data            0.6.3
traitlets             5.14.3
ujson                 5.10.0
wcwidth               0.2.13

[RonnyPfannschmidt]

this is expected - and a user error

direct modification of a function object will assure the net time its collected from a different class

[201dreamers]

okay, thanks for such a quick response!
Could you please tell me how can I properly add marks in such cases, so the mark isn't applied to all cases? (Of course if you have time. If not, I'll try to look myself)

[RonnyPfannschmidt]

more context is needed

the minimal example you posted is best served by using the mark as decorator in the target class