pytest中setup()函数异常行为原因及冲突函数名咨询

Let's break down what's happening here and how to avoid similar issues.

The Root Cause #

Pytest maintains compatibility with x-unit style test fixtures, which includes automatically recognizing certain function names as setup/teardown hooks at different scopes.

In your case, you named your function setup() at the module level. Pytest interprets this as the legacy module-level x-unit setup hook, which expects no arguments. But your setup() function requires a conf_func parameter. When pytest tries to run this hook automatically (without passing any arguments) during test setup, it hits a parameter mismatch error—this is exactly the exception you're seeing.

When you rename the function to setup_func() or something else, pytest no longer treats it as a built-in hook, so your @pytest.mark.parametrize call can use it normally to generate test parameters.

Your Code for Reference #

Problematic Code #

import pytest
def config():
    login = 'John'
    password = '123'
    env = 'staging'
    return login, password, env
def setup(conf_func):
    return [conf_func()]
@pytest.mark.parametrize('login, password, env', setup(config))
def test_func(login, password, env):
    print(login, password, env)
    assert False

Error Snippet #

============================= test session starts =============================
platform win32 -- Python 3.6.8, pytest-5.0.1, py-1.8.0, pluggy-0.12.0
rootdir: C:\python\experiments\fixtures
collected 1 item

test_fix6.py E
test setup failed

cls = <class '_pytest.runner.CallInfo'>
func = <function call_runtest_hook.<locals>.<lambda> at 0x0000000003AA6B70>
when = 'setup'
reraise = (<class '_pytest.outcomes.Exit'>, <class 'KeyboardInterrupt'>)

@classmethod
def from_call(cls, func, when, reraise=None):
    #: context of invocation: one of "setup", "call",
    #: "teardown", "memocollect"
    start = time()
    excinfo = None
    try:
>       result = func()
venv\lib\site-packages\_pytest\runner.py:220:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
venv\lib\site-packages\_pytest\runner.py:192: in <lambda>
    lambda: ihook(item=item, **kwds), when=when, reraise=reraise
venv\lib\site-packages\pluggy\hooks.py:289: in __call__
    return self._hookexec(self, self.get_hookimpls(), kwargs)
venv\lib\site-packages\pluggy\manager.py:87: in _hookexec
    return self._inner_hookexec(hook, methods, kwargs)
venv\lib\site-packages\pluggy\manager.py:81: in <lambda>
    firstresult=hook.spec.opts.get("firstresult") if hook.spec else False,
venv\lib\site-packages\_pytest\nose.py:34: in pytest_runtest_setup
    call_optional(item.parent.obj, "setup")

Other Function Names to Avoid #

To prevent similar conflicts, avoid using these x-unit style hook names in your test modules (unless you intend to use them as setup/teardown hooks):

Module Scope #

• setup() (legacy module setup, no arguments)
• teardown() (legacy module teardown, no arguments)
• setup_module() (modern module setup, accepts a module parameter)
• teardown_module() (modern module teardown, accepts a module parameter)

Class Scope (inside test classes) #

• setup_class() (runs once before all methods in the class, accepts cls parameter)
• teardown_class() (runs once after all methods in the class, accepts cls parameter)
• setup_method() (runs before each test method, accepts self parameter)
• teardown_method() (runs after each test method, accepts self parameter)

Function Scope (for standalone test functions) #

• setup_function() (runs before each test function, accepts a function parameter)
• teardown_function() (runs after each test function, accepts a function parameter)

Sticking to non-reserved names for your helper functions will keep pytest from misinterpreting them as built-in hooks.

内容的提问来源于stack exchange，提问作者Andrey Glazkov