Environment details
- OS type and version: Linux (gLinux / Debian based)
- Python version:
3.13.12 (also affects all supported Python versions 3.10+)
- pip version:
26.0.1
google-api-python-client version: 2.197.0 (at commit 6e471c075039dfef24e28d11658e03d5c949c7c3)
Steps to reproduce
- Check out the
main branch of google-api-python-client.
- Set up a clean virtual environment and install test dependencies (
pytest, mock, parameterized).
- Run the strict positional parameters enforcement test directly using
pytest:
pytest tests/test_discovery.py -k test_tests_should_be_run_with_strict_positional_enforcement
Code example
The bug lies in how strict argument enforcement is configured for the test suite in tests/__init__.py.
Historically, the enforcement was configured inside the nose-legacy setup_package() hook:
# tests/__init__.py
from googleapiclient import _helpers as util
def setup_package():
"""Run on testing package."""
util.positional_parameters_enforcement = "EXCEPTION"
Because modern test runs in this repository (including via nox sessions) run under pytest, the setup_package() package-level hook is ignored. As a result, the tests run in warning-only mode, causing strict parameter assertion tests to be bypassed and fail with an AssertionError instead of raising TypeError.
Since the repository has migrated to pytest, this legacy hook is dead code. We should configure the variable directly at the package module level and completely remove the unused setup_package hook:
# tests/__init__.py
from googleapiclient import _helpers as util
+util.positional_parameters_enforcement = "EXCEPTION"
-
-def setup_package():
- """Run on testing package."""
- util.positional_parameters_enforcement = "EXCEPTION"
Stack trace
When running the tests using pytest, the test fails with the following output:
=================================== FAILURES ===================================
_ DiscoveryErrors.test_tests_should_be_run_with_strict_positional_enforcement __
self = <tests.test_discovery.DiscoveryErrors testMethod=test_tests_should_be_run_with_strict_positional_enforcement>
def test_tests_should_be_run_with_strict_positional_enforcement(self):
try:
plus = build("plus", "v1", None, static_discovery=False)
> self.fail("should have raised a TypeError exception over missing http=.")
E AssertionError: should have raised a TypeError exception over missing http=.
tests/test_discovery.py:509: AssertionError
------------------------------ Captured log call -------------------------------
WARNING googleapiclient._helpers:_helpers.py:129 build() takes at most 2 positional arguments (3 given)
=========================== short test summary info ============================
FAILED tests/test_discovery.py::DiscoveryErrors::test_tests_should_be_run_with_strict_positional_enforcement
====================== 1 failed, 332 deselected in 1.65s =======================
Environment details
3.13.12(also affects all supported Python versions3.10+)26.0.1google-api-python-clientversion:2.197.0(at commit6e471c075039dfef24e28d11658e03d5c949c7c3)Steps to reproduce
mainbranch ofgoogle-api-python-client.pytest,mock,parameterized).pytest:Code example
The bug lies in how strict argument enforcement is configured for the test suite in
tests/__init__.py.Historically, the enforcement was configured inside the nose-legacy
setup_package()hook:Because modern test runs in this repository (including via
noxsessions) run underpytest, thesetup_package()package-level hook is ignored. As a result, the tests run in warning-only mode, causing strict parameter assertion tests to be bypassed and fail with anAssertionErrorinstead of raisingTypeError.Since the repository has migrated to
pytest, this legacy hook is dead code. We should configure the variable directly at the package module level and completely remove the unusedsetup_packagehook:Stack trace
When running the tests using
pytest, the test fails with the following output: