github pytest-dev/pytest 8.3.0

latest releases: 8.3.3, 8.3.2, 8.4.0.dev0...
3 months ago

pytest 8.3.0 (2024-07-20)

New features

  • #12231: Added [--xfail-tb]{.title-ref} flag, which turns on traceback output for XFAIL results.

    • If the [--xfail-tb]{.title-ref} flag is not given, tracebacks for XFAIL results are NOT shown.
    • The style of traceback for XFAIL is set with [--tb]{.title-ref}, and can be [auto|long|short|line|native|no]{.title-ref}.
    • Note: Even if you have [--xfail-tb]{.title-ref} set, you won't see them if [--tb=no]{.title-ref}.

    Some history:

    With pytest 8.0, [-rx]{.title-ref} or [-ra]{.title-ref} would not only turn on summary reports for xfail, but also report the tracebacks for xfail results. This caused issues with some projects that utilize xfail, but don't want to see all of the xfail tracebacks.

    This change detaches xfail tracebacks from [-rx]{.title-ref}, and now we turn on xfail tracebacks with [--xfail-tb]{.title-ref}. With this, the default [-rx]{.title-ref}/ [-ra]{.title-ref} behavior is identical to pre-8.0 with respect to xfail tracebacks. While this is a behavior change, it brings default behavior back to pre-8.0.0 behavior, which ultimately was considered the better course of action.

  • #12281: Added support for keyword matching in marker expressions.

    Now tests can be selected by marker keyword arguments.
    Supported values are int{.interpreted-text role="class"}, (unescaped) str{.interpreted-text role="class"}, bool{.interpreted-text role="class"} & None{.interpreted-text role="data"}.

    See marker examples <marker_keyword_expression_example>{.interpreted-text role="ref"} for more information.

    -- by lovetheguitar{.interpreted-text role="user"}

  • #12567: Added --no-fold-skipped command line option.

    If this option is set, then skipped tests in short summary are no longer grouped
    by reason but all tests are printed individually with their nodeid in the same
    way as other statuses.

    -- by pbrezina{.interpreted-text role="user"}

Improvements in existing functionality

  • #12469: The console output now uses the "third-party plugins" terminology,
    replacing the previously established but confusing and outdated
    reference to setuptools <setuptools:index>{.interpreted-text role="std:doc"}
    -- by webknjaz{.interpreted-text role="user"}.

  • #12544, #12545: Python virtual environment detection was improved by
    checking for a pyvenv.cfg{.interpreted-text role="file"} file, ensuring reliable detection on
    various platforms -- by zachsnickers{.interpreted-text role="user"}.

  • #2871: Do not truncate arguments to functions in output when running with [-vvv]{.title-ref}.

  • #389: The readability of assertion introspection of bound methods has been enhanced
    -- by farbodahm{.interpreted-text role="user"}, webknjaz{.interpreted-text role="user"}, obestwalter{.interpreted-text role="user"}, flub{.interpreted-text role="user"}
    and glyphack{.interpreted-text role="user"}.

    Earlier, it was like:

    =================================== FAILURES ===================================
    _____________________________________ test _____________________________________
    
        def test():
    >       assert Help().fun() == 2
    E       assert 1 == 2
    E        +  where 1 = <bound method Help.fun of <example.Help instance at 0x256a830>>()
    E        +    where <bound method Help.fun of <example.Help instance at 0x256a830>> = <example.Help instance at 0x256a830>.fun
    E        +      where <example.Help instance at 0x256a830> = Help()
    
    example.py:7: AssertionError
    =========================== 1 failed in 0.03 seconds ===========================

    And now it's like:

    =================================== FAILURES ===================================
    _____________________________________ test _____________________________________
    
        def test():
    >       assert Help().fun() == 2
    E       assert 1 == 2
    E        +  where 1 = fun()
    E        +    where fun = <test_local.Help object at 0x1074be230>.fun
    E        +      where <test_local.Help object at 0x1074be230> = Help()
    
    test_local.py:13: AssertionError
    =========================== 1 failed in 0.03 seconds ===========================
  • #7662: Added timezone information to the testsuite timestamp in the JUnit XML report.

Bug fixes

  • #11706: Fixed reporting of teardown errors in higher-scoped fixtures when using [--maxfail]{.title-ref} or [--stepwise]{.title-ref}.

    Originally added in pytest 8.0.0, but reverted in 8.0.2 due to a regression in pytest-xdist.
    This regression was fixed in pytest-xdist 3.6.1.

  • #11797: pytest.approx{.interpreted-text role="func"} now correctly handles Sequence <collections.abc.Sequence>{.interpreted-text role="class"}-like objects.

  • #12204, #12264: Fixed a regression in pytest 8.0 where tracebacks get longer and longer when multiple
    tests fail due to a shared higher-scope fixture which raised -- by bluetech{.interpreted-text role="user"}.

    Also fixed a similar regression in pytest 5.4 for collectors which raise during setup.

    The fix necessitated internal changes which may affect some plugins:

    • FixtureDef.cached_result[2] is now a tuple (exc, tb)
      instead of exc.
    • SetupState.stack failures are now a tuple (exc, tb)
      instead of exc.
  • #12275: Fixed collection error upon encountering an abstract <abc>{.interpreted-text role="mod"} class, including abstract [unittest.TestCase]{.title-ref} subclasses.

  • #12328: Fixed a regression in pytest 8.0.0 where package-scoped parameterized items were not correctly reordered to minimize setups/teardowns in some cases.

  • #12424: Fixed crash with [assert testcase is not None]{.title-ref} assertion failure when re-running unittest tests using plugins like pytest-rerunfailures. Regressed in 8.2.2.

  • #12472: Fixed a crash when returning category "error" or "failed" with a custom test status from pytest_report_teststatus{.interpreted-text role="hook"} hook -- pbrezina{.interpreted-text role="user"}.

  • #12505: Improved handling of invalid regex patterns in pytest.raises(match=r'...') <pytest.raises>{.interpreted-text role="func"} by providing a clear error message.

  • #12580: Fixed a crash when using the cache class on Windows and the cache directory was created concurrently.

  • #6962: Parametrization parameters are now compared using [==]{.title-ref} instead of [is]{.title-ref} ([is]{.title-ref} is still used as a fallback if the parameter does not support [==]{.title-ref}).
    This fixes use of parameters such as lists, which have a different [id]{.title-ref} but compare equal, causing fixtures to be re-computed instead of being cached.

  • #7166: Fixed progress percentages (the [ 87%] at the edge of the screen) sometimes not aligning correctly when running with pytest-xdist -n.

Improved documentation

  • #12153: Documented using PYTEST_VERSION{.interpreted-text role="envvar"} to detect if code is running from within a pytest run.

  • #12469: The external plugin mentions in the documentation now avoid mentioning
    setuptools entry-points <setuptools:index>{.interpreted-text role="std:doc"} as the concept is
    much more generic nowadays. Instead, the terminology of "external",
    "installed", or "third-party" plugins (or packages) replaces that.

    -- by webknjaz{.interpreted-text role="user"}

  • #12577: [CI]{.title-ref} and [BUILD_NUMBER]{.title-ref} environment variables role is discribed in
    the reference doc. They now also appear when doing [pytest -h]{.title-ref}
    -- by MarcBresson{.interpreted-text role="user"}.

Contributor-facing changes

  • #12467: Migrated all internal type-annotations to the python3.10+ style by using the [annotations]{.title-ref} future import.

    -- by RonnyPfannschmidt{.interpreted-text role="user"}

  • #11771, #12557: The PyPy runtime version has been updated to 3.9 from 3.8 that introduced
    a flaky bug at the garbage collector which was not expected to fix there
    as the 3.8 is EoL.

    -- by x612skm{.interpreted-text role="user"}

  • #12493: The change log draft preview integration has been refactored to use a
    third party extension sphinxcontib-towncrier. The previous in-repo
    script was putting the change log preview file at
    doc/en/_changelog_towncrier_draft.rst{.interpreted-text role="file"}. Said file is no longer
    ignored in Git and might show up among untracked files in the
    development environments of the contributors. To address that, the
    contributors can run the following command that will clean it up:

    $ git clean -x -i -- doc/en/_changelog_towncrier_draft.rst

    -- by webknjaz{.interpreted-text role="user"}

  • #12498: All the undocumented tox environments now have descriptions.
    They can be listed in one's development environment by invoking
    tox -av in a terminal.

    -- by webknjaz{.interpreted-text role="user"}

  • #12501: The changelog configuration has been updated to introduce more accurate
    audience-tailored categories. Previously, there was a trivial
    change log fragment type with an unclear and broad meaning. It was
    removed and we now have contrib, misc and packaging in
    place of it.

    The new change note types target the readers who are downstream
    packagers and project contributors. Additionally, the miscellaneous
    section is kept for unspecified updates that do not fit anywhere else.

    -- by webknjaz{.interpreted-text role="user"}

  • #12502: The UX of the GitHub automation making pull requests to update the
    plugin list has been updated. Previously, the maintainers had to close
    the automatically created pull requests and re-open them to trigger the
    CI runs. From now on, they only need to click the [Ready for review]{.title-ref}
    button instead.

    -- by webknjaz{.interpreted-text role="user"}

  • #12522: The :pull: RST role has been replaced with a shorter
    :pr: due to starting to use the implementation from
    the third-party sphinx-issues{.interpreted-text role="pypi"} Sphinx extension
    -- by webknjaz{.interpreted-text role="user"}.

  • #12531: The coverage reporting configuration has been updated to exclude
    pytest's own tests marked as expected to fail from the coverage
    report. This has an effect of reducing the influence of flaky
    tests on the resulting number.

    -- by webknjaz{.interpreted-text role="user"}

  • #12533: The extlinks Sphinx extension is no longer enabled. The :bpo:
    role it used to declare has been removed with that. BPO itself has
    migrated to GitHub some years ago and it is possible to link the
    respective issues by using their GitHub issue numbers and the
    :issue: role that the sphinx-issues extension implements.

    -- by webknjaz{.interpreted-text role="user"}

  • #12562: Possible typos in using the :user: RST role is now being linted
    through the pre-commit tool integration -- by webknjaz{.interpreted-text role="user"}.

Don't miss a new pytest release

NewReleases is sending notifications on new releases.