This PR contains the following updates:

Package	Change	Age	Confidence
aiohttp	==3.11.18 -> ==3.12.14
cryptography (changelog)	==45.0.2 -> ==45.0.5
google-auth	==2.40.1 -> ==2.40.3
pytest (changelog)	==8.3.5 -> ==8.4.1
pytest-cov (changelog)	==6.1.1 -> ==6.2.1

Release Notes

v3.12.14: 3.12.14

Compare Source

Bug fixes

• Fixed file uploads failing with HTTP 422 errors when encountering 307/308 redirects, and 301/302 redirects for non-POST methods, by preserving the request body when appropriate per :rfc:9110#section-15.4.3-3.1 -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  #​11270.

• Fixed :py:meth:ClientSession.close() <aiohttp.ClientSession.close> hanging indefinitely when using HTTPS requests through HTTP proxies -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  #​11273.

• Bumped minimum version of aiosignal to 1.4+ to resolve typing issues -- by :user:Dreamsorcerer.

  Related issues and pull requests on GitHub:
  #​11280.

Features

• Added initial trailer parsing logic to Python HTTP parser -- by :user:Dreamsorcerer.

  Related issues and pull requests on GitHub:
  #​11269.

Improved documentation

• Clarified exceptions raised by WebSocketResponse.send_frame et al.
  -- by :user:DoctorJohn.

  Related issues and pull requests on GitHub:
  #​11234.

v3.12.13

Compare Source

====================

Bug fixes

• Fixed auto-created :py:class:~aiohttp.TCPConnector not using the session's event loop when :py:class:~aiohttp.ClientSession is created without an explicit connector -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:11147.

v3.12.12

Compare Source

v3.12.11

Compare Source

====================

Features

• Improved SSL connection handling by changing the default ssl_shutdown_timeout
  from 0.1 to 0 seconds. SSL connections now use Python's default graceful
  shutdown during normal operation but are aborted immediately when the connector
  is closed, providing optimal behavior for both cases. Also added support for
  ssl_shutdown_timeout=0 on all Python versions. Previously, this value was
  rejected on Python 3.11+ and ignored on earlier versions. Non-zero values on
  Python < 3.11 now trigger a RuntimeWarning -- by :user:bdraco.

  The ssl_shutdown_timeout parameter is now deprecated and will be removed in
  aiohttp 4.0 as there is no clear use case for changing the default.

  Related issues and pull requests on GitHub:
  :issue:11148.

Deprecations (removal in next major release)

• Improved SSL connection handling by changing the default ssl_shutdown_timeout
  from 0.1 to 0 seconds. SSL connections now use Python's default graceful
  shutdown during normal operation but are aborted immediately when the connector
  is closed, providing optimal behavior for both cases. Also added support for
  ssl_shutdown_timeout=0 on all Python versions. Previously, this value was
  rejected on Python 3.11+ and ignored on earlier versions. Non-zero values on
  Python < 3.11 now trigger a RuntimeWarning -- by :user:bdraco.

  The ssl_shutdown_timeout parameter is now deprecated and will be removed in
  aiohttp 4.0 as there is no clear use case for changing the default.

  Related issues and pull requests on GitHub:
  :issue:11148.

v3.12.10

Compare Source

====================

Bug fixes

• Fixed leak of aiodns.DNSResolver when :py:class:~aiohttp.TCPConnector is closed and no resolver was passed when creating the connector -- by :user:Tasssadar.

  This was a regression introduced in version 3.12.0 (:pr:10897).

  Related issues and pull requests on GitHub:
  :issue:11150.

v3.12.9: 3.12.9

Compare Source

Bug fixes

• Fixed IOBasePayload and TextIOPayload reading entire files into memory when streaming large files -- by :user:bdraco.

  When using file-like objects with the aiohttp client, the entire file would be read into memory if the file size was provided in the Content-Length header. This could cause out-of-memory errors when uploading large files. The payload classes now correctly read data in chunks of READ_SIZE (64KB) regardless of the total content length.

  Related issues and pull requests on GitHub:
  #​11138.

v3.12.8

Compare Source

===================

Features

• Added preemptive digest authentication to :class:~aiohttp.DigestAuthMiddleware -- by :user:bdraco.

  The middleware now reuses authentication credentials for subsequent requests to the same
  protection space, improving efficiency by avoiding extra authentication round trips.
  This behavior matches how web browsers handle digest authentication and follows
  :rfc:7616#section-3.6.

  Preemptive authentication is enabled by default but can be disabled by passing
  preemptive=False to the middleware constructor.

  Related issues and pull requests on GitHub:
  :issue:11128, :issue:11129.

v3.12.7

Compare Source

===================

.. warning::

This release fixes an issue where the quote_cookie parameter was not being properly
respected for shared cookies (domain="", path=""). If your server does not handle quoted
cookies correctly, you may need to disable cookie quoting by setting quote_cookie=False
when creating your :class:~aiohttp.ClientSession or :class:~aiohttp.CookieJar.
See :ref:aiohttp-client-cookie-quoting-routine for details.

Bug fixes

• Fixed cookie parsing to be more lenient when handling cookies with special characters
  in names or values. Cookies with characters like {, }, and / in names are now
  accepted instead of causing a :exc:~http.cookies.CookieError and 500 errors. Additionally,
  cookies with mismatched quotes in values are now parsed correctly, and quoted cookie
  values are now handled consistently whether or not they include special attributes
  like Domain. Also fixed :class:~aiohttp.CookieJar to ensure shared cookies (domain="", path="")
  respect the quote_cookie parameter, making cookie quoting behavior consistent for
  all cookies -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:2683, :issue:5397, :issue:7993, :issue:11112.

• Fixed an issue where cookies with duplicate names but different domains or paths
  were lost when updating the cookie jar. The :class:~aiohttp.ClientSession
  cookie jar now correctly stores all cookies even if they have the same name but
  different domain or path, following the :rfc:6265#section-5.3 storage model -- by :user:bdraco.

  Note that :attr:ClientResponse.cookies <aiohttp.ClientResponse.cookies> returns
  a :class:~http.cookies.SimpleCookie which uses the cookie name as a key, so
  only the last cookie with each name is accessible via this interface. All cookies
  can be accessed via :meth:ClientResponse.headers.getall('Set-Cookie') <multidict.MultiDictProxy.getall> if needed.

  Related issues and pull requests on GitHub:
  :issue:4486, :issue:11105, :issue:11106.

Miscellaneous internal changes

• Avoided creating closed futures in ResponseHandler that will never be awaited -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:11107.

• Downgraded the logging level for connector close errors from ERROR to DEBUG, as these are expected behavior with TLS 1.3 connections -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:11114.

v3.12.6

Compare Source

===================

Bug fixes

• Fixed spurious "Future exception was never retrieved" warnings for connection lost errors when the connector is not closed -- by :user:bdraco.

  When connections are lost, the exception is now marked as retrieved since it is always propagated through other means, preventing unnecessary warnings in logs.

  Related issues and pull requests on GitHub:
  :issue:11100.

v3.12.4

Compare Source

===================

Bug fixes

• Fixed connector not waiting for connections to close before returning from :meth:~aiohttp.BaseConnector.close (partial backport of :pr:3733) -- by :user:atemate and :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:1925, :issue:11074.

v3.12.3

Compare Source

===================

Bug fixes

• Fixed memory leak in :py:meth:~aiohttp.CookieJar.filter_cookies that caused unbounded memory growth
  when making requests to different URL paths -- by :user:bdraco and :user:Cycloctane.

  Related issues and pull requests on GitHub:
  :issue:11052, :issue:11054.

v3.12.2

Compare Source

===================

Bug fixes

• Fixed Content-Length header not being set to 0 for non-GET requests with None body -- by :user:bdraco.

  Non-GET requests (POST, PUT, PATCH, DELETE) with None as the body now correctly set the Content-Length header to 0, matching the behavior of requests with empty bytes (b""). This regression was introduced in aiohttp 3.12.1.

  Related issues and pull requests on GitHub:
  :issue:11035.

v3.12.1

Compare Source

====================

Bug fixes

• Fixed file uploads failing with HTTP 422 errors when encountering 307/308 redirects, and 301/302 redirects for non-POST methods, by preserving the request body when appropriate per :rfc:9110#section-15.4.3-3.1 -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:11270.

• Fixed :py:meth:ClientSession.close() <aiohttp.ClientSession.close> hanging indefinitely when using HTTPS requests through HTTP proxies -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:11273.

• Bumped minimum version of aiosignal to 1.4+ to resolve typing issues -- by :user:Dreamsorcerer.

  Related issues and pull requests on GitHub:
  :issue:11280.

Features

• Added initial trailer parsing logic to Python HTTP parser -- by :user:Dreamsorcerer.

  Related issues and pull requests on GitHub:
  :issue:11269.

Improved documentation

• Clarified exceptions raised by WebSocketResponse.send_frame et al.
  -- by :user:DoctorJohn.

  Related issues and pull requests on GitHub:
  :issue:11234.

v3.12.0

Compare Source

===================

Bug fixes

• Fixed :py:attr:~aiohttp.web.WebSocketResponse.prepared property to correctly reflect the prepared state, especially during timeout scenarios -- by :user:bdraco

  Related issues and pull requests on GitHub:
  :issue:6009, :issue:10988.

• Response is now always True, instead of using MutableMapping behaviour (False when map is empty)

  Related issues and pull requests on GitHub:
  :issue:10119.

• Fixed connection reuse for file-like data payloads by ensuring buffer
  truncation respects content-length boundaries and preventing premature
  connection closure race -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:10325, :issue:10915, :issue:10941, :issue:10943.

• Fixed pytest plugin to not use deprecated :py:mod:asyncio policy APIs.

  Related issues and pull requests on GitHub:
  :issue:10851.

• Fixed :py:class:~aiohttp.resolver.AsyncResolver not using the loop argument in versions 3.x where it should still be supported -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:10951.

Features

• Added a comprehensive HTTP Digest Authentication client middleware (DigestAuthMiddleware)
  that implements RFC 7616. The middleware supports all standard hash algorithms
  (MD5, SHA, SHA-256, SHA-512) with session variants, handles both 'auth' and
  'auth-int' quality of protection options, and automatically manages the
  authentication flow by intercepting 401 responses and retrying with proper
  credentials -- by :user:feus4177, :user:TimMenninger, and :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:2213, :issue:10725.

• Added client middleware support -- by :user:bdraco and :user:Dreamsorcerer.

  This change allows users to add middleware to the client session and requests, enabling features like
  authentication, logging, and request/response modification without modifying the core
  request logic. Additionally, the session attribute was added to ClientRequest,
  allowing middleware to access the session for making additional requests.

  Related issues and pull requests on GitHub:
  :issue:9732, :issue:10902, :issue:10945, :issue:10952, :issue:10959, :issue:10968.

• Allow user setting zlib compression backend -- by :user:TimMenninger

  This change allows the user to call :func:aiohttp.set_zlib_backend() with the
  zlib compression module of their choice. Default behavior continues to use
  the builtin zlib library.

  Related issues and pull requests on GitHub:
  :issue:9798.

• Added support for overriding the base URL with an absolute one in client sessions
  -- by :user:vivodi.

  Related issues and pull requests on GitHub:
  :issue:10074.

• Added host parameter to aiohttp_server fixture -- by :user:christianwbrock.

  Related issues and pull requests on GitHub:
  :issue:10120.

• Detect blocking calls in coroutines using BlockBuster -- by :user:cbornet.

  Related issues and pull requests on GitHub:
  :issue:10433.

• Added socket_factory to :py:class:aiohttp.TCPConnector to allow specifying custom socket options
  -- by :user:TimMenninger.

  Related issues and pull requests on GitHub:
  :issue:10474, :issue:10520, :issue:10961, :issue:10962.

• Started building armv7l manylinux wheels -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:10797.

• Implemented shared DNS resolver management to fix excessive resolver object creation
  when using multiple client sessions. The new _DNSResolverManager singleton ensures
  only one DNSResolver object is created for default configurations, significantly
  reducing resource usage and improving performance for applications using multiple
  client sessions simultaneously -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:10847, :issue:10923, :issue:10946.

• Upgraded to LLHTTP 9.3.0 -- by :user:Dreamsorcerer.

  Related issues and pull requests on GitHub:
  :issue:10972.

• Optimized small HTTP requests/responses by coalescing headers and body into a single TCP packet -- by :user:bdraco.

  This change enhances network efficiency by reducing the number of packets sent for small HTTP payloads, improving latency and reducing overhead. Most importantly, this fixes compatibility with memory-constrained IoT devices that can only perform a single read operation and expect HTTP requests in one packet. The optimization uses zero-copy writelines when coalescing data and works with both regular and chunked transfer encoding.

  When aiohttp uses client middleware to communicate with an aiohttp server, connection reuse is more likely to occur since complete responses arrive in a single packet for small payloads.

  This aligns aiohttp with other popular HTTP clients that already coalesce small requests.

  Related issues and pull requests on GitHub:
  :issue:10991.

Improved documentation

• Improved documentation for middleware by adding warnings and examples about
  request body stream consumption. The documentation now clearly explains that
  request body streams can only be read once and provides best practices for
  sharing parsed request data between middleware and handlers -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:2914.

Packaging updates and notes for downstreams

• Removed non SPDX-license description from setup.cfg -- by :user:devanshu-ziphq.

  Related issues and pull requests on GitHub:
  :issue:10662.

• Added support for building against system llhttp library -- by :user:mgorny.

  This change adds support for :envvar:AIOHTTP_USE_SYSTEM_DEPS environment variable that
  can be used to build aiohttp against the system install of the llhttp library rather
  than the vendored one.

  Related issues and pull requests on GitHub:
  :issue:10759.

• aiodns is now installed on Windows with speedups extra -- by :user:bdraco.

  As of aiodns 3.3.0, SelectorEventLoop is no longer required when using pycares 4.7.0 or later.

  Related issues and pull requests on GitHub:
  :issue:10823.

• Fixed compatibility issue with Cython 3.1.1 -- by :user:bdraco

  Related issues and pull requests on GitHub:
  :issue:10877.

Contributor-facing changes

• Sped up tests by disabling blockbuster fixture for test_static_file_huge and test_static_file_huge_cancel tests -- by :user:dikos1337.

  Related issues and pull requests on GitHub:
  :issue:9705, :issue:10761.

• Updated tests to avoid using deprecated :py:mod:asyncio policy APIs and
  make it compatible with Python 3.14.

  Related issues and pull requests on GitHub:
  :issue:10851.

• Added Winloop to test suite to support in the future -- by :user:Vizonex.

  Related issues and pull requests on GitHub:
  :issue:10922.

Miscellaneous internal changes

• Added support for the partitioned attribute in the set_cookie method.

  Related issues and pull requests on GitHub:
  :issue:9870.

• Setting :attr:aiohttp.web.StreamResponse.last_modified to an unsupported type will now raise :exc:TypeError instead of silently failing -- by :user:bdraco.

  Related issues and pull requests on GitHub:
  :issue:10146.

v45.0.5

Compare Source

v45.0.4

Compare Source

v45.0.3

Compare Source

v2.40.3

Compare Source

Bug Fixes

• Auth fetch token from default endpoint (#​1779) (88891cc)
• Remove unnecessary call to mds service (#​1769) (7c61c7d)
• Retry 504 errors (#​1767) (554f967)

v2.40.2

Compare Source

Bug Fixes

• Remove sync response logs in AuthorizedSession (97ed1c8)
• Update test to consider new error message from cryptography (#​1765) (44e38b6)

v8.4.1

Compare Source

pytest 8.4.1 (2025-06-17)

Bug fixes

• #​13461: Corrected _pytest.terminal.TerminalReporter.isatty to support
  being called as a method. Before it was just a boolean which could
  break correct code when using -o log_cli=true).

• #​13477: Reintroduced pytest.PytestReturnNotNoneWarning{.interpreted-text role="class"} which was removed by accident in pytest [8.4]{.title-ref}.

  This warning is raised when a test functions returns a value other than None, which is often a mistake made by beginners.

  See return-not-none{.interpreted-text role="ref"} for more information.

• #​13497: Fixed compatibility with Twisted 25+.

Improved documentation

• #​13492: Fixed outdated warning about faulthandler not working on Windows.

v8.4.0

Compare Source

pytest 8.4.0 (2025-06-02)

Removals and backward incompatible breaking changes

• #​11372: Async tests will now fail, instead of warning+skipping, if you don't have any suitable plugin installed.

• #​12346: Tests will now fail, instead of raising a warning, if they return any value other than None.

• #​12874: We dropped support for Python 3.8 following its end of life (2024-10-07).

• #​12960: Test functions containing a yield now cause an explicit error. They have not been run since pytest 4.0, and were previously marked as an expected failure and deprecation warning.

  See the docs <yield tests deprecated>{.interpreted-text role="ref"} for more information.

Deprecations (removal in next major release)

• #​10839: Requesting an asynchronous fixture without a [pytest_fixture_setup]{.title-ref} hook that resolves it will now give a DeprecationWarning. This most commonly happens if a sync test requests an async fixture. This should have no effect on a majority of users with async tests or fixtures using async pytest plugins, but may affect non-standard hook setups or autouse=True. For guidance on how to work around this warning see sync-test-async-fixture{.interpreted-text role="ref"}.

New features

• #​11538: Added pytest.RaisesGroup{.interpreted-text role="class"} as an equivalent to pytest.raises{.interpreted-text role="func"} for expecting ExceptionGroup{.interpreted-text role="exc"}. Also adds pytest.RaisesExc{.interpreted-text role="class"} which is now the logic behind pytest.raises{.interpreted-text role="func"} and used as parameter to pytest.RaisesGroup{.interpreted-text role="class"}. RaisesGroup includes the ability to specify multiple different expected exceptions, the structure of nested exception groups, and flags for emulating except* <except_star>{.interpreted-text role="ref"}. See assert-matching-exception-groups{.interpreted-text role="ref"} and docstrings for more information.

• #​12081: Added capteesys{.interpreted-text role="fixture"} to capture AND pass output to next handler set by --capture=.

• #​12504: pytest.mark.xfail{.interpreted-text role="func"} now accepts pytest.RaisesGroup{.interpreted-text role="class"} for the raises parameter when you expect an exception group. You can also pass a pytest.RaisesExc{.interpreted-text role="class"} if you e.g. want to make use of the check parameter.

• #​12713: New [--force-short-summary]{.title-ref} option to force condensed summary output regardless of verbosity level.

  This lets users still see condensed summary output of failures for quick reference in log files from job outputs, being especially useful if non-condensed output is very verbose.

• #​12749: pytest traditionally collects classes/functions in the test module namespace even if they are imported from another file.

  For example:

contents of src/domain.py

class Testament: ...

contents of tests/test_testament.py

from domain import Testament

def test_testament(): ...
```

In this scenario with the default options, pytest will collect the class [Testament]{.title-ref} from [tests/test_testament.py]{.title-ref} because it starts with [Test]{.title-ref}, even though in this case it is a production class being imported in the test module namespace.

This behavior can now be prevented by setting the new `collect_imported_tests`{.interpreted-text role="confval"} configuration option to `false`, which will make pytest collect classes/functions from test files **only** if they are defined in that file.

\-- by `FreerGit`{.interpreted-text role="user"}

• #​12765: Thresholds to trigger snippet truncation can now be set with truncation_limit_lines{.interpreted-text role="confval"} and truncation_limit_chars{.interpreted-text role="confval"}.

  See truncation-params{.interpreted-text role="ref"} for more information.

• #​13125: console_output_style{.interpreted-text role="confval"} now supports times to show execution time of each test.

• #​13192: pytest.raises{.interpreted-text role="func"} will now raise a warning when passing an empty string to match, as this will match against any value. Use match="^$" if you want to check that an exception has no message.

• #​13192: pytest.raises{.interpreted-text role="func"} will now print a helpful string diff if matching fails and the match parameter has ^ and $ and is otherwise escaped.

• #​13192: You can now pass with pytest.raises(check=fn): <pytest.raises>{.interpreted-text role="func"}, where fn is a function which takes a raised exception and returns a boolean. The raises fails if no exception was raised (as usual), passes if an exception is raised and fn returns True (as well as match and the type matching, if specified, which are checked before), and propagates the exception if fn returns False (which likely also fails the test).

• #​13228: hidden-param{.interpreted-text role="ref"} can now be used in id of pytest.param{.interpreted-text role="func"} or in
  ids of Metafunc.parametrize <pytest.Metafunc.parametrize>{.interpreted-text role="py:func"}.
  It hides the parameter set from the test name.

• #​13253: New flag: --disable-plugin-autoload <disable_plugin_autoload>{.interpreted-text role="ref"} which works as an alternative to PYTEST_DISABLE_PLUGIN_AUTOLOAD{.interpreted-text role="envvar"} when setting environment variables is inconvenient; and allows setting it in config files with addopts{.interpreted-text role="confval"}.

Improvements in existing functionality

• #​10224: pytest's short and long traceback styles (how-to-modifying-python-tb-printing{.interpreted-text role="ref"})
  now have partial 657{.interpreted-text role="pep"} support and will show specific code segments in the
  traceback.

  ================================= FAILURES =================================
  _______________________ test_gets_correct_tracebacks _______________________
  
  test_tracebacks.py:12: in test_gets_correct_tracebacks
      assert manhattan_distance(p1, p2) == 1
             ^^^^^^^^^^^^^^^^^^^^^^^^^^
  test_tracebacks.py:6: in manhattan_distance
      return abs(point_1.x - point_2.x) + abs(point_1.y - point_2.y)
                             ^^^^^^^^^
  E   AttributeError: 'NoneType' object has no attribute 'x'
  

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

• #​11118: Now pythonpath{.interpreted-text role="confval"} configures [$PYTHONPATH]{.title-ref} earlier than before during the initialization process, which now also affects plugins loaded via the [-p]{.title-ref} command-line option.

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

• #​11381: The type parameter of the parser.addini method now accepts ["int"]{.title-ref} and "float" parameters, facilitating the parsing of configuration values in the configuration file.

  Example:

  def pytest_addoption(parser):
      parser.addini("int_value", type="int", default=2, help="my int value")
      parser.addini("float_value", type="float", default=4.2, help="my float value")

  The [pytest.ini]{.title-ref} file:

  [pytest]
  int_value = 3
  float_value = 5.4

• #​11525: Fixtures are now clearly represented in the output as a "fixture object", not as a normal function as before, making it easy for beginners to catch mistakes such as referencing a fixture declared in the same module but not requested in the test function.

  -- by the-compiler{.interpreted-text role="user"} and glyphack{.interpreted-text role="user"}

• #​12426: A warning is now issued when pytest.mark.usefixtures ref{.interpreted-text role="ref"} is used without specifying any fixtures. Previously, empty usefixtures markers were silently ignored.

• #​12707: Exception chains can be navigated when dropped into Pdb in Python 3.13+.

• #​12736: Added a new attribute [name]{.title-ref} with the fixed value ["pytest tests"]{.title-ref} to the root tag [testsuites]{.title-ref} of the junit-xml generated by pytest.

  This attribute is part of many junit-xml specifications and is even part of the [junit-10.xsd]{.title-ref} specification that pytest's implementation is based on.

• #​12943: If a test fails with an exceptiongroup with a single exception, the contained exception will now be displayed in the short test summary info.

• #​12958: A number of unraisable <unraisable>{.interpreted-text role="ref"} enhancements:

  • Set the unraisable hook as early as possible and unset it as late as possible, to collect the most possible number of unraisable exceptions.
  • Call the garbage collector just before unsetting the unraisable hook, to collect any straggling exceptions.
  • Collect multiple unraisable exceptions per test phase.
  • Report the tracemalloc{.interpreted-text role="mod"} allocation traceback (if available).
  • Avoid using a generator based hook to allow handling StopIteration{.interpreted-text role="class"} in test failures.
  • Report the unraisable exception as the cause of the pytest.PytestUnraisableExceptionWarning{.interpreted-text role="class"} exception if raised.
  • Compute the repr of the unraisable object in the unraisable hook so you get the latest information if available, and should help with resurrection of the object.

• #​13010: pytest.approx{.interpreted-text role="func"} now can compare collections that contain numbers and non-numbers mixed.

• #​13016: A number of threadexception <unraisable>{.interpreted-text role="ref"} enhancements:

  • Set the excepthook as early as possible and unset it as late as possible, to collect the most possible number of unhandled exceptions from threads.
  • Collect multiple thread exceptions per test phase.
  • Report the tracemalloc{.interpreted-text role="mod"} allocation traceback (if available).
  • Avoid using a generator based hook to allow handling StopIteration{.interpreted-text role="class"} in test failures.
  • Report the thread exception as the cause of the pytest.PytestUnhandledThreadExceptionWarning{.interpreted-text role="class"} exception if raised.
  • Extract the name of the thread object in the excepthook which should help with resurrection of the thread.

• #​13031: An empty parameter set as in pytest.mark.parametrize([], ids=idfunc) will no longer trigger a call to idfunc with internal objects.

• #​13115: Allows supplying ExceptionGroup[Exception] and BaseExceptionGroup[BaseException] to pytest.raises to keep full typing on ExceptionInfo <pytest.ExceptionInfo>{.interpreted-text role="class"}:

  with pytest.raises(ExceptionGroup[Exception]) as exc_info:
      some_function()

  Parametrizing with other exception types remains an error - we do not check the types of child exceptions and thus do not permit code that might look like we do.

• #​13122: The --stepwise mode received a number of improvements:

  • It no longer forgets the last failed test in case pytest is executed later without the flag.

    This enables the following workflow:

    1. Execute pytest with --stepwise, pytest then stops at the first failing test;
    2. Iteratively update the code and run the test in isolation, without the --stepwise flag (for example in an IDE), until it is fixed.
    3. Execute pytest with --stepwise again and pytest will continue from the previously failed test, and if it passes, continue on to the next tests.

    Previously, at step 3, pytest would start from the beginning, forgetting the previously failed test.

    This change however might cause issues if the --stepwise mode is used far apart in time, as the state might get stale, so the internal state will be reset automatically in case the test suite changes (for now only the number of tests are considered for this, we might change/improve this on the future).

  • New --stepwise-reset/--sw-reset flag, allowing the user to explicitly reset the stepwise state and restart the workflow from the beginning.

• #​13308: Added official support for Python 3.14.

• #​13380: Fix ExceptionGroup{.interpreted-text role="class"} traceback filtering to exclude pytest internals.

• #​13415: The author metadata of the BibTex example is now correctly formatted with last names following first names.
  An example of BibLaTex has been added.
  BibTex and BibLaTex examples now clearly indicate that what is cited is software.

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

• #​13420: Improved test collection performance by optimizing path resolution used in FSCollector.

• #​13457: The error message about duplicate parametrization no longer displays an internal stack trace.

• #​4112: Using pytest.mark.usefixtures <pytest.mark.usefixtures ref>{.interpreted-text role="ref"} on pytest.param{.interpreted-text role="func"} now produces an error instead of silently doing nothing.

• #​5473: Replace [:]{.title-ref} with [;]{.title-ref} in the assertion rewrite warning message so it can be filtered using standard Python warning filters before calling pytest.main{.interpreted-text role="func"}.

[#​6985](https://redirect.github.com/pytest-dev/pytest/issues/6985): Improved `pytest.approx`{.interpreted-text role="func"} to enhance the readability of value ranges and tolerances between 0.001 and 1000.

:   -   The [repr]{.title-ref} method now provides clearer output for values within those ranges, making it easier to interpret the results.

    -   Previously, the output for those ranges of values and tolerances was displayed in scientific notation (e.g., [42 ± 1.0e+00]{.title-ref}). The updated method now presents the tolerance as a decimal for better readability (e.g., [42 ± 1]{.title-ref}).

        Example:

        **Previous Output:**

        ``` console
        >>> pytest.approx(42, abs=1)
        42 ± 1.0e+00
        ```

        **Current Output:**

        ``` console
        >>> pytest.approx(42, abs=1)
        42 ± 1
        ```

    \-- by `fazeelghafoor`{.interpreted-text role="user"}

• #​7683: The formerly optional pygments dependency is now required, causing output always to be source-highlighted (unless disabled via the --code-highlight=no CLI option).

Bug fixes

• #​10404: Apply filterwarnings from config/cli as soon as possible, and revert them as late as possible
  so that warnings as errors are collected throughout the pytest run and before the
  unraisable and threadexcept hooks are removed.

  This allows very late warnings and unraisable/threadexcept exceptions to fail the test suite.

  This also changes the warning that the lsof plugin issues from PytestWarning to the new warning PytestFDWarning so it can be more easily filtered.

• #​11067: The test report is now consistent regardless if the test xfailed via pytest.mark.xfail <pytest.mark.xfail ref>{.interpreted-text role="ref"} or pytest.fail{.interpreted-text role="func"}.

  Previously, xfailed tests via the marker would have the string "reason: " prefixed to the message, while those xfailed via the function did not. The prefix has been removed.

• #​12008: In 11220{.interpreted-text role="pr"}, an unintended change in reordering was introduced by changing the way indices were assigned to direct params. More specifically, before that change, the indices of direct params to metafunc's callspecs were assigned after all parametrizations took place. Now, that change is reverted.

• #​12863: Fix applying markers, including pytest.mark.parametrize <pytest.mark.parametrize ref>{.interpreted-text role="ref"} when placed above [@​staticmethod]{.title-ref} or [@​classmethod]{.title-ref}.

• #​12929: Handle StopIteration from test cases, setup and teardown correctly.

• #​12938: Fixed --durations-min argument not respected if -vv is used.

• #​12946: Fixed missing help for pdb{.interpreted-text role="mod"} commands wrapped by pytest -- by adamchainz{.interpreted-text role="user"}.

• #​12981: Prevent exceptions in pytest.Config.add_cleanup{.interpreted-text role="func"} callbacks preventing further cleanups.

• #​13047: Restore pytest.approx{.interpreted-text role="func"} handling of equality checks between [bool]{.title-ref} and [numpy.bool_]{.title-ref} types.

  Comparing [bool]{.title-ref} and [numpy.bool_]{.title-ref} using pytest.approx{.interpreted-text role="func"} accidentally changed in version [8.3.4]{.title-ref} and [8.3.5]{.title-ref} to no longer match:

  >>> import numpy as np
  >>> from pytest import approx
  >>> [np.True_, np.True_] == pytest.approx([True, True])
  False

  This has now been fixed:

  >>> [np.True_, np.True_] == pytest.approx([True, True])
  True

• #​13119: Improved handling of invalid regex patterns for filter warnings by providing a clear error message.

• #​13175: The diff is now also highlighted correctly when comparing two strings.

• #​13248: Fixed an issue where passing a scope in Metafunc.parametrize <pytest.Metafunc.parametrize>{.interpreted-text role="py:func"} with indirect=True
  could result in other fixtures being unable to depend on the parametrized fixture.

• #​13291: Fixed repr of attrs objects in assertion failure messages when using attrs>=25.2.

• #​13312: Fixed a possible KeyError crash on PyPy during collection of tests involving higher-scoped parameters.

• #​13345: Fix type hints for pytest.TestReport.when{.interpreted-text role="attr"} and pytest.TestReport.location{.interpreted-text role="attr"}.

• #​13377: Fixed handling of test methods with positional-only parameter syntax.

  Now, methods are supported that formally define self as positional-only
  and/or fixture parameters as keyword-only, e.g.:

  class TestClass:
  
      def test_method(self, /, *, fixture): ...

  Before, this caused an internal error in pytest.

• #​13384: Fixed an issue where pytest could report negative durations.

• #​13420: Added lru_cache to nodes._check_initialpaths_for_relpath.

• #​9037: Honor disable_test_id_escaping_and_forfeit_all_rights_to_community_support{.interpreted-text role="confval"} when escaping ids in parametrized tests.

Improved documentation

• #​12535: [This
  example]{.title-ref}<https://docs.pytest.org/en/latest/example/simple.html#making-test-result-information-available-in-fixtures>
  showed print statements that do not exactly reflect what the
  different branches actually do. The fix makes the example more precise.

• #​13218: Pointed out in the pytest.approx{.interpreted-text role="func"} documentation that it considers booleans unequal to numeric zero or one.

• #​13221: Improved grouping of CLI options in the --help output.

• #​6649: Added ~pytest.TerminalReporter{.interpreted-text role="class"} to the api-reference{.interpreted-text role="ref"} documentation page.

• #​8612: Add a recipe for handling abstract test classes in the documentation.

  A new example has been added to the documentation to demonstrate how to use a mixin class to handle abstract
  test classes without manually setting the __test__ attribute for subclasses.
  This ensures that subclasses of abstract test classes are automatically collected by pytest.

Packaging updates and notes for downstreams

• #​13317: Specified minimum allowed versions of colorama, iniconfig,
  and packaging; and bumped the minimum allowed version
  of exceptiongroup for python_version<'3.11' from a release
  candidate to a full release.

Contributor-facing changes

• #​12017: Mixed internal improvements:

  • Migrate formatting to f-strings in some tests.
  • Use type-safe constructs in JUnitXML tests.
  • MovedMockTiming into _pytest.timing.

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

• #​12647: Fixed running the test suite with the hypothesis pytest plugin.

Miscellaneous internal changes

• #​6649: Added ~pytest.TerminalReporter{.interpreted-text role="class"} to the public pytest API, as it is part of the signature of the pytest_terminal_summary{.interpreted-text role="hook"} hook.

v6.2.1

Compare Source

• Added a version requirement for pytest's pluggy dependency (1.2.0, released 2023-06-21) that has the required new-style hookwrapper API.

• Removed deprecated license classifier (packaging).

• Disabled coverage warnings in two more situations where they have no value:

  • "module-not-measured" in workers
  • "already-imported" in subprocesses

v6.2.0

Compare Source

• The plugin now adds 3 rules in the filter warnings configuration to prevent common coverage warnings being raised as obscure errors::

  default:unclosed database in <sqlite3.Connection object at:ResourceWarning
  once::PytestCovWarning
  once::CoverageWarning

  This fixes most of the bad interactions that are occurring on pytest 8.4 with filterwarnings=error.

  The plugin will check if there already matching rules for the 3 categories
  (ResourceWarning, PytestCovWarning, CoverageWarning) and message (unclosed database in <sqlite3.Connection object at) before adding the filters.

  This means you can have this in your pytest configuration for complete oblivion (not recommended, if that is not clear)::

  filterwarnings = [
  "error",
  "ignore:unclosed database in <sqlite3.Connection object at:ResourceWarning",
  "ignore::PytestCovWarning",
  "ignore::CoverageWarning",
  ]

Configuration

📅 Schedule: Branch creation - At any time (no schedule defined), Automerge - At any time (no schedule defined).

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR is behind base branch, or you tick the rebase/retry checkbox.

👻 Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.

• If you want to rebase/retry this PR, check this box

This PR was generated by Mend Renovate. View the repository job log.