test_has_version fails

[opoplawski]

I'm working on updating the python-toolz package in Fedora to 1.1.0 but getting the following test failure:

+ /usr/bin/pytest -v -k 'not test_shakespeare'
/usr/lib/python3.14/site-packages/pytest_asyncio/plugin.py:211: PytestDeprecationWarning: The configuration option "asyncio_default_fixture_loop_scope" is unset.
The event loop scope for asynchronous fixtures will default to the fixture caching scope. Future versions of pytest-asyncio will default the loop scope for asynchronous fixtures to function scope. Set the default fixture loop scope explicitly in order to avoid unexpected behavior in the future. Valid fixture loop scopes are: "function", "class", "module", "package", "session"

  warnings.warn(PytestDeprecationWarning(_DEFAULT_FIXTURE_LOOP_SCOPE_UNSET))
============================= test session starts ==============================
platform linux -- Python 3.14.0, pytest-8.3.5, pluggy-1.6.0 -- /usr/bin/python3
cachedir: .pytest_cache
hypothesis profile 'default' -> database=DirectoryBasedExampleDatabase(PosixPath('/home/orion/fedora/python-toolz/python-toolz-1.1.0-build/toolz-1.1.0/.hypothesis/examples'))
benchmark: 5.1.0 (defaults: timer=time.perf_counter disable_gc=False min_rounds=5 min_time=0.000005 max_time=1.0 calibration_precision=10 warmup=False warmup_iterations=100000)
Matplotlib: 3.10.6
Freetype: 2.14.1
rootdir: /home/orion/fedora/python-toolz/python-toolz-1.1.0-build/toolz-1.1.0
configfile: pyproject.toml
testpaths: toolz
plugins: typeguard-4.4.3, hypothesis-6.123.0, flaky-3.8.1, xvfb-3.1.1, pyfakefs-5.9.2, openfiles-0.6.0, libtmux-0.42.0, forked-1.6.0, requests-mock-1.12.1, benchmark-5.1.0, mpl-0.13, doctestplus-1.5.0, anyio-4.8.0, arraydiff-0.6.1, mpi-0.6, vcr-1.0.2, xprocess-1.0.2, remotedata-0.4.1, xdist-3.7.0, split-0.10.0, rerunfailures-15.0, mock-3.14.1, asyncio-1.1.0, timeout-2.4.0, localserver-0.9.0.post0
asyncio: mode=Mode.STRICT, asyncio_default_fixture_loop_scope=None, asyncio_default_test_loop_scope=function
collecting ... collected 186 items
...
=================================== FAILURES ===================================
_______________________________ test_has_version _______________________________

    def test_has_version():
        # If this test fails, then toolz probably isn't installed properly.
        # For local development, try `pip install -e .` from the project directory.
        version = toolz.__version__
        assert isinstance(version, str)
>       assert version.startswith("1.")
E       AssertionError: assert False
E        +  where False = <built-in method startswith of str object at 0x7f18621baa60>('1.')
E        +    where <built-in method startswith of str object at 0x7f18621baa60> = '0.0.1'.startswith

toolz/tests/test_package.py:9: AssertionError
=========================== short test summary info ============================
FAILED toolz/tests/test_package.py::test_has_version - AssertionError: assert...
======================== 1 failed, 185 passed in 0.89s =========================

[eriknw]

Thanks!

How do you install toolz? Not having version info implies toolz is not installed properly. Even e.g. pip install -e . should work. Does your build workflow typically not install packages before testing? Can you provide more context or details? Should we improve the error message for this?

[opoplawski]

So, this is part of the RPM build - we build a wheel, install it into the buildroot for the package. Then run the tests with PYTHONPATH pointing to the buildroot location. This is working fine for 1.0.0, but not 1.1.0. Full build log is here: https://kojipkgs.fedoraproject.org//work/tasks/4555/138484555/build.log

toolz does seem to think it's version is 0.0.1. I'm getting the source from github.

[eriknw]

Thanks for working on this @opoplawski.

I would recommend to not get the source from github. Maybe we should remove those assets from the release?

Instead, I would recommend using the source *.tar.gz from PyPI. See: https://pypi.org/project/toolz/#files

If you strongly prefer to use GitHub as the source of truth, then perhaps try something such as the following:

TAG=$(curl -s https://api.github.com/repos/pytoolz/toolz/releases/latest | grep '"tag_name":' | sed -E 's/.*"([^"]+)".*/\1/')
git clone --branch $TAG --single-branch --depth 1 https://github.com/pytoolz/toolz.git
pip install toolz

Note that we just changed to use setuptools-git-versioning to manage versions. I wonder if there's a better way that would fit both of our workflows.