Welcome to pytest-cov’s documentation!

Contents:

Overview

docs

Documentation Status

tests

GitHub Actions Status

package

PyPI Package latest release conda-forge PyPI Wheel Supported versions Supported implementations Commits since latest release

This plugin provides coverage functionality as a pytest plugin. Compared to just using coverage run this plugin does some extras:

  • Automatic erasing and combination of .coverage files and default reporting.

  • Support for detailed coverage contexts (add --cov-context=test to have the full test name including parametrization as the context).

  • Xdist support: you can use all of pytest-xdist’s features including remote interpreters and still get coverage.

  • Consistent pytest behavior. If you run coverage run -m pytest you will have slightly different sys.path (CWD will be in it, unlike when running pytest).

All features offered by the coverage package should work, either through pytest-cov’s command line options or through coverage’s config file.

  • Free software: MIT license

Installation

Install with pip:

pip install pytest-cov

For distributed testing support install pytest-xdist:

pip install pytest-xdist

Upgrading from pytest-cov 6.3

pytest-cov 6.3 and older were using a .pth file to enable coverage measurements in subprocesses. This was removed in pytest-cov 7 - use coverage’s patch options to enable subprocess measurements.

Uninstalling

Uninstall with pip:

pip uninstall pytest-cov

Under certain scenarios a stray .pth file may be left around in site-packages.

  • pytest-cov 2.0 may leave a pytest-cov.pth if you installed without wheels (easy_install, setup.py install etc).

  • pytest-cov 1.8 or older will leave a init_cov_core.pth.

Usage

pytest --cov=myproj tests/

Would produce a report like:

-------------------- coverage: ... ---------------------
Name                 Stmts   Miss  Cover
----------------------------------------
myproj/__init__          2      0   100%
myproj/myproj          257     13    94%
myproj/feature4286      94      7    92%
----------------------------------------
TOTAL                  353     20    94%

Documentation

https://pytest-cov.readthedocs.io/en/latest/

Coverage Data File

The data file is erased at the beginning of testing to ensure clean data for each test run. If you need to combine the coverage of several test runs you can use the --cov-append option to append this coverage data to coverage data from previous test runs.

The data file is left at the end of testing so that it is possible to use normal coverage tools to examine it.

Limitations

For distributed testing the workers must have the pytest-cov package installed. This is needed since the plugin must be registered through setuptools for pytest to start the plugin on the worker.

Security

To report a security vulnerability please use the Tidelift security contact. Tidelift will coordinate the fix and disclosure.

Acknowledgements

Whilst this plugin has been built fresh from the ground up it has been influenced by the work done on pytest-coverage (Ross Lawley, James Mills, Holger Krekel) and nose-cover (Jason Pellerin) which are other coverage plugins.

Ned Batchelder for coverage and its ability to combine the coverage results of parallel runs.

Holger Krekel for pytest with its distributed testing support.

Jason Pellerin for nose.

Michael Foord for unittest2.

No doubt others have contributed to these tools as well.

Configuration

This plugin provides a clean minimal set of command line options that are added to pytest. For further control of coverage use a coverage config file.

CLI options:

--cov [SOURCE]

Path or package name to measure during execution (multi-allowed). Use --cov= to not do any source filtering and record everything.

--cov-reset

Reset cov sources accumulated in options so far.

--cov-report TYPE

Type of report to generate: term, term-missing, annotate, html, xml, json, markdown, markdown-append, lcov (multi-allowed). term, term-missing may be followed by “:skip-covered”. annotate, html, xml, json, markdown, markdown-append and lcov may be followed by “:DEST” where DEST specifies the output location. Use --cov-report= to not generate any output.

--cov-config PATH

Config file for coverage. Default: .coveragerc

--no-cov-on-fail

Do not report coverage if test run fails. Default: False

--no-cov

Disable coverage report completely (useful for debuggers). Default: False

--cov-fail-under MIN

Fail if the total coverage is less than MIN.

--cov-append

Do not delete coverage but append to current. Default: False

--cov-branch

Enable branch coverage. Can also be specified in the coverage config file [run] section.

--cov-precision COV_PRECISION

Override the reporting precision. Can also be specified in the coverage config file [report] section.

--cov-context CONTEXT

Dynamic contexts to use. “test” for now.

Note

Important Note

This plugin overrides the parallel option of coverage. Unless you also run coverage without pytest-cov it’s pointless to set those options in your .coveragerc.

If you use the --cov=something option (with a value) then coverage’s source option will also get overridden. If you have multiple sources it might be easier to set those in .coveragerc and always use --cov (without a value) instead of having a long command line with --cov=pkg1 --cov=pkg2 --cov=pkg3 ....

If you use the --cov-branch option then coverage’s branch option will also get overridden.

If you wish to always run pytest-cov with pytest, you can use addopts under the pytest or tool:pytest section of your setup.cfg, or the tool.pytest.ini_options section of your pyproject.toml file.

For example, in setup.cfg:

[tool:pytest]
addopts = --cov=<project-name> --cov-report html

Or for pyproject.toml:

[tool.pytest.ini_options]
addopts = "--cov=<project-name> --cov-report html"

Note

Important Note

The --cov option has an optional argument. If it’s your last option in addopts it might eat the next CLI argument, make sure to force it to take a blank value if that’s what you wanted by using --cov= (essentially the same as --cov="").

Caveats

An unfortunate consequence of coverage.py’s history is that .coveragerc is a magic name: it’s the default file but it also means “try to also lookup coverage configuration in tox.ini or setup.cfg”.

In practical terms this means that if you have multiple configuration files around (tox.ini, pyproject.toml or setup.cfg) you might need to use --cov-config to make coverage use the correct configuration file.

Also, if you change the working directory and also use subprocesses in a test you might also need to use --cov-config to make pytest-cov use the expected configuration file in the subprocess.

Reporting

It is possible to generate any combination of the reports for a single test run.

The available reports are terminal (with or without missing line numbers shown), HTML, XML, JSON, Markdown (either in ‘write’ or ‘append’ mode to file), LCOV and annotated source code.

The default is terminal report without line numbers:

pytest --cov=myproj tests/

-------------------- coverage: platform linux2, python 2.6.4-final-0 ---------------------
Name                 Stmts   Miss  Cover
----------------------------------------
myproj/__init__          2      0   100%
myproj/myproj          257     13    94%
myproj/feature4286      94      7    92%
----------------------------------------
TOTAL                  353     20    94%

The terminal report with line numbers:

pytest --cov-report=term-missing --cov=myproj tests/

-------------------- coverage: platform linux2, python 2.6.4-final-0 ---------------------
Name                 Stmts   Miss  Cover   Missing
--------------------------------------------------
myproj/__init__          2      0   100%
myproj/myproj          257     13    94%   24-26, 99, 149, 233-236, 297-298, 369-370
myproj/feature4286      94      7    92%   183-188, 197
--------------------------------------------------
TOTAL                  353     20    94%

The terminal report with skip covered:

pytest --cov-report term:skip-covered --cov=myproj tests/

-------------------- coverage: platform linux2, python 2.6.4-final-0 ---------------------
Name                 Stmts   Miss  Cover
----------------------------------------
myproj/myproj          257     13    94%
myproj/feature4286      94      7    92%
----------------------------------------
TOTAL                  353     20    94%

1 files skipped due to complete coverage.

You can use skip-covered with term-missing as well. e.g. --cov-report term-missing:skip-covered

If any reporting options are used then the default (--cov-report=term) is not added automatically. For example this would not show any terminal output:

pytest --cov-report html
       --cov-report xml
       --cov-report json
       --cov-report markdown
       --cov-report markdown-append:cov-append.md
       --cov-report lcov
       --cov-report annotate
       --cov=myproj tests/

You can specify output paths for reports. The output location for the XML, JSON, Markdown and LCOV report is a file. Where as the output location for the HTML and annotated source code reports are directories:

pytest --cov-report html:cov_html
       --cov-report xml:cov.xml
       --cov-report json:cov.json
       --cov-report markdown:cov.md
       --cov-report markdown-append:cov-append.md
       --cov-report lcov:cov.info
       --cov-report annotate:cov_annotate
       --cov=myproj tests/

Example for GitHub Actions with markdown-append:

pytest --cov-report markdown-append:$GITHUB_STEP_SUMMARY
       --cov=myproj tests/

To disable the default term report provide an empty report:

pytest --cov-report= --cov=myproj tests/

This mode can be especially useful on continuous integration servers, where a coverage file is needed for subsequent processing, but no local report needs to be viewed. For example, tests run on GitHub Actions could produce a .coverage file for use with Coveralls.

Debuggers and PyCharm

(or other IDEs)

When it comes to TDD one obviously would like to debug tests. Debuggers in Python use mostly the sys.settrace function to gain access to context. Coverage uses the same technique to get access to the lines executed. Coverage does not play well with other tracers simultaneously running. This manifests itself in behaviour that PyCharm might not hit a breakpoint no matter what the user does, or encountering an error like this:

PYDEV DEBUGGER WARNING:
sys.settrace() should not be used when the debugger is being used.
This may cause the debugger to stop working correctly.

Since it is common practice to have coverage configuration in the pytest.ini file and pytest does not support removeopts or similar the –no-cov flag can disable coverage completely.

At the reporting part a warning message will show on screen:

Coverage disabled via --no-cov switch!

Distributed testing (xdist)

“load” mode

Distributed testing with dist mode set to “load” will report on the combined coverage of all workers. The workers may be spread out over any number of hosts and each worker may be located anywhere on the file system. Each worker will have its subprocesses measured.

Running distributed testing with dist mode set to load:

pytest --cov=myproj -n 2 tests/

Shows a terminal report:

-------------------- coverage: platform linux2, python 2.6.4-final-0 ---------------------
Name                 Stmts   Miss  Cover
----------------------------------------
myproj/__init__          2      0   100%
myproj/myproj          257     13    94%
myproj/feature4286      94      7    92%
----------------------------------------
TOTAL                  353     20    94%

Again but spread over different hosts and different directories:

pytest --cov=myproj --dist load
        --tx ssh=memedough@host1//chdir=testenv1
        --tx ssh=memedough@host2//chdir=/tmp/testenv2//python=/tmp/env1/bin/python
        --rsyncdir myproj --rsyncdir tests --rsync examples
        tests/

Shows a terminal report:

-------------------- coverage: platform linux2, python 2.6.4-final-0 ---------------------
Name                 Stmts   Miss  Cover
----------------------------------------
myproj/__init__          2      0   100%
myproj/myproj          257     13    94%
myproj/feature4286      94      7    92%
----------------------------------------
TOTAL                  353     20    94%

“each” mode

Distributed testing with dist mode set to each will report on the combined coverage of all workers. Since each worker is running all tests this allows generating a combined coverage report for multiple environments.

Running distributed testing with dist mode set to each:

pytest --cov=myproj --dist each
        --tx popen//chdir=/tmp/testenv3//python=/usr/local/python27/bin/python
        --tx ssh=memedough@host2//chdir=/tmp/testenv4//python=/tmp/env2/bin/python
        --rsyncdir myproj --rsyncdir tests --rsync examples
        tests/

Shows a terminal report:

---------------------------------------- coverage ----------------------------------------
                          platform linux2, python 2.6.5-final-0
                          platform linux2, python 2.7.0-final-0
Name                 Stmts   Miss  Cover
----------------------------------------
myproj/__init__          2      0   100%
myproj/myproj          257     13    94%
myproj/feature4286      94      7    92%
----------------------------------------
TOTAL                  353     20    94%

Subprocess support

Subprocess support was removed in pytest-cov 7.0 due to various complexities resulting from coverage’s own subprocess support. To migrate you should change your coverage config to have at least this:

[run]
patch = subprocess

Or if you use pyproject.toml:

[tool.coverage.run]
patch = ["subprocess"]

Note that if you enable the subprocess patch then parallel = true is automatically set.

If it still doesn’t produce the same coverage as before you may need to enable more patches, see the coverage config and subprocess documentation.

Contexts

Coverage.py 5.0 can record separate coverage data for different contexts during one run of a test suite. Pytest-cov can use this feature to record coverage data for each test individually, with the --cov-context=test option.

The context name recorded in the coverage.py database is the pytest test id, and the phase of execution, one of “setup”, “run”, or “teardown”. These two are separated with a pipe symbol. You might see contexts like:

test_functions.py::test_addition|run
test_fancy.py::test_parametrized[1-101]|setup
test_oldschool.py::RegressionTests::test_error|run

Note that parameterized tests include the values of the parameters in the test id, and each set of parameter values is recorded as a separate test.

To view contexts when using --cov-report=html, add this to your .coveragerc:

[html]
show_contexts = True

The HTML report will include an annotation on each covered line, indicating the number of contexts that executed the line. Clicking the annotation displays a list of the contexts.

Tox

When using tox you can have ultra-compact configuration - you can have all of it in tox.ini:

[tox]
envlist = ...

[tool:pytest]
...

[coverage:paths]
...

[coverage:run]
...

[coverage:report]
..

[testenv]
commands = ...

An usual problem users have is that pytest-cov will erase the previous coverage data by default, thus if you run tox with multiple environments you’ll get incomplete coverage at the end.

To prevent this problem you need to use --cov-append. It’s still recommended to clean the previous coverage data to have consistent output. A tox.ini like this should be enough for sequential runs:

[tox]
envlist = clean,py27,py36,...

[testenv]
commands = pytest --cov --cov-append --cov-report=term-missing ...
deps =
    pytest
    pytest-cov

[testenv:clean]
deps = coverage
skip_install = true
commands = coverage erase

For parallel runs we need to set some dependencies and have an extra report env like so:

[tox]
envlist = clean,py27,py36,report

[testenv]
commands = pytest --cov --cov-append --cov-report=term-missing
deps =
    pytest
    pytest-cov
depends =
    {py27,py36}: clean
    report: py27,py36

[testenv:report]
deps = coverage
skip_install = true
commands =
    coverage report
    coverage html

[testenv:clean]
deps = coverage
skip_install = true
commands = coverage erase

Depending on your project layout you might need extra configuration, see the working examples at https://github.com/pytest-dev/pytest-cov/tree/master/examples for two common layouts.

Plugin coverage

Getting coverage on pytest plugins is a very particular situation. Because of how pytest implements plugins (using setuptools entrypoints) it doesn’t allow controlling the order in which the plugins load. See pytest/issues/935 for technical details.

Currently there is no way to measure your pytest plugin if you use pytest-cov. You should change your test invocations to use coverage run -m pytest ... instead.

Markers and fixtures

There are some builtin markers and fixtures in pytest-cov.

Markers

no_cover

Eg:

@pytest.mark.no_cover
def test_foobar():
    # do some stuff that needs coverage disabled

Warning

Caveat

Note that subprocess coverage will also be disabled.

Fixtures

no_cover

Eg:

def test_foobar(no_cover):
    # same as the marker ...

cov

For reasons that no one can remember there is a cov fixture that provides access to the underlying Coverage instance. Some say this is a disguised foot-gun and should be removed, and some think mysteries make life more interesting and it should be left alone.

Changelog

7.1.0 (2026-03-21)

  • Fixed total coverage computation to always be consistent, regardless of reporting settings. Previously some reports could produce different total counts, and consequently can make –cov-fail-under behave different depending on reporting options. See #641.

  • Improve handling of ResourceWarning from sqlite3.

    The plugin adds warning filter for sqlite3 ResourceWarning unclosed database (since 6.2.0). It checks if there is already existing plugin for this message by comparing filter regular expression. When filter is specified on command line the message is escaped and does not match an expected message. A check for an escaped regular expression is added to handle this case.

    With this fix one can suppress ResourceWarning from sqlite3 from command line:

    pytest -W "ignore:unclosed database in <sqlite3.Connection object at:ResourceWarning" ...

  • Various improvements to documentation. Contributed by Art Pelling in #718 and “vivodi” in #738. Also closed #736.

  • Fixed some assertions in tests. Contributed by in Markéta Machová in #722.

  • Removed unnecessary coverage configuration copying (meant as a backup because reporting commands had configuration side-effects before coverage 5.0).

7.0.0 (2025-09-09)

  • Dropped support for subprocesses measurement.

    It was a feature added long time ago when coverage lacked a nice way to measure subprocesses created in tests. It relied on a .pth file, there was no way to opt-out and it created bad interations with coverage’s new patch system added in 7.10.

    To migrate to this release you might need to enable the suprocess patch, example for .coveragerc:

    [run]
patch = subprocess

    This release also requires at least coverage 7.10.6.

  • Switched packaging to have metadata completely in pyproject.toml and use hatchling for building. Contributed by Ofek Lev in #551 with some extras in #716.

  • Removed some not really necessary testing deps like six.

6.3.0 (2025-09-06)

  • Added support for markdown reports. Contributed by Marcos Boger in #712 and #714.

  • Fixed some formatting issues in docs. Anonymous contribution in #706.

6.2.1 (2025-06-12)

  • 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

6.2.0 (2025-06-11)

  • 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",
]

6.1.1 (2025-04-05)

  • Fixed breakage that occurs when --cov-context and the no_cover marker are used together.

6.1.0 (2025-04-01)

  • Change terminal output to use full width lines for the coverage header. Contributed by Tsvika Shapira in #678.

  • Removed unnecessary CovFailUnderWarning. Fixes #675.

  • Fixed the term report not using the precision specified via --cov-precision.

6.0.0 (2024-10-29)

  • Updated various documentation inaccuracies, especially on subprocess handling.

  • Changed fail under checks to use the precision set in the coverage configuration. Now it will perform the check just like coverage report would.

  • Added a --cov-precision cli option that can override the value set in your coverage configuration.

  • Dropped support for now EOL Python 3.8.

5.0.0 (2024-03-24)

  • Removed support for xdist rsync (now deprecated). Contributed by Matthias Reichenbach in #623.

  • Switched docs theme to Furo.

  • Various legacy Python cleanup and CI improvements. Contributed by Christian Clauss and Hugo van Kemenade in #630, #631, #632 and #633.

  • Added a pyproject.toml example in the docs. Contributed by Dawn James in #626.

  • Modernized project’s pre-commit hooks to use ruff. Initial POC contributed by Christian Clauss in #584.

  • Dropped support for Python 3.7.

4.1.0 (2023-05-24)

  • Updated CI with new Pythons and dependencies.

  • Removed rsyncdir support. This makes pytest-cov compatible with xdist 3.0. Contributed by Sorin Sbarnea in #558.

  • Optimized summary generation to not be performed if no reporting is active (for example, when --cov-report='' is used without --cov-fail-under). Contributed by Jonathan Stewmon in #589.

  • Added support for JSON reporting. Contributed by Matthew Gamble in #582.

  • Refactored code to use f-strings. Contributed by Mark Mayo in #572.

  • Fixed a skip in the test suite for some old xdist. Contributed by a bunch of people in #565.

  • Dropped support for Python 3.6.

4.0.0 (2022-09-28)

Note that this release drops support for multiprocessing.

  • –cov-fail-under no longer causes pytest –collect-only to fail Contributed by Zac Hatfield-Dodds in #511.

  • Dropped support for multiprocessing (mostly because issue 82408). This feature was mostly working but very broken in certain scenarios and made the test suite very flaky and slow.

    There is builtin multiprocessing support in coverage and you can migrate to that. All you need is this in your .coveragerc:

    [run]
concurrency = multiprocessing
parallel = true
sigterm = true

  • Fixed deprecation in setup.py by trying to import setuptools before distutils. Contributed by Ben Greiner in #545.

  • Removed undesirable new lines that were displayed while reporting was disabled. Contributed by Delgan in #540.

  • Documentation fixes. Contributed by Andre Brisco in #543 and Colin O’Dell in #525.

  • Added support for LCOV output format via –cov-report=lcov. Only works with coverage 6.3+. Contributed by Christian Fetzer in #536.

  • Modernized pytest hook implementation. Contributed by Bruno Oliveira in #549 and Ronny Pfannschmidt in #550.

3.0.0 (2021-10-04)

Note that this release drops support for Python 2.7 and Python 3.5.

  • Added support for Python 3.10 and updated various test dependencies. Contributed by Hugo van Kemenade in #500.

  • Switched from Travis CI to GitHub Actions. Contributed by Hugo van Kemenade in #494 and #495.

  • Add a --cov-reset CLI option. Contributed by Danilo Šegan in #459.

  • Improved validation of --cov-fail-under CLI option. Contributed by … Ronny Pfannschmidt’s desire for skark in #480.

  • Dropped Python 2.7 support. Contributed by Thomas Grainger in #488.

  • Updated trove classifiers. Contributed by Michał Bielawski in #481.

  • Reverted change for toml requirement. Contributed by Thomas Grainger in #477.

2.12.1 (2021-06-01)

  • Changed the toml requirement to be always be directly required (instead of being required through a coverage extra). This fixes issues with pip-compile (pip-tools#1300). Contributed by Sorin Sbarnea in #472.

  • Documented show_contexts. Contributed by Brian Rutledge in #473.

2.12.0 (2021-05-14)

  • Added coverage’s toml extra to install requirements in setup.py. Contributed by Christian Riedel in #410.

  • Fixed pytest_cov.__version__ to have the right value (string with version instead of a string including __version__ =).

  • Fixed license classifier in setup.py. Contributed by Chris Sreesangkom in #467.

  • Fixed commits since badge. Contributed by Terence Honles in #470.

2.11.1 (2021-01-20)

  • Fixed support for newer setuptools (v42+). Contributed by Michał Górny in #451.

2.11.0 (2021-01-18)

  • Bumped minimum coverage requirement to 5.2.1. This prevents reporting issues. Contributed by Mateus Berardo de Souza Terra in #433.

  • Improved sample projects (from the examples directory) to support running tox -e pyXY. Now the example configures a suffixed coverage data file, and that makes the cleanup environment unnecessary. Contributed by Ganden Schaffner in #435.

  • Removed the empty console_scripts entrypoint that confused some Gentoo build script. I didn’t ask why it was so broken cause I didn’t want to ruin my day. Contributed by Michał Górny in #434.

  • Fixed the missing coverage context when using subprocesses. Contributed by Bernát Gábor in #443.

  • Updated the config section in the docs. Contributed by Pamela McA’Nulty in #429.

  • Migrated CI to travis-ci.com (from .org).

2.10.1 (2020-08-14)

  • Support for pytest-xdist 2.0, which breaks compatibility with pytest-xdist before 1.22.3 (from 2017). Contributed by Zac Hatfield-Dodds in #412.

  • Fixed the LocalPath has no attribute startswith failure that occurred when using the pytester plugin in inline mode.

2.10.0 (2020-06-12)

  • Improved the --no-cov warning. Now it’s only shown if --no-cov is present before --cov.

  • Removed legacy pytest support. Changed setup.py so that pytest>=4.6 is required.

2.9.0 (2020-05-22)

  • Fixed RemovedInPytest4Warning when using Pytest 3.10. Contributed by Michael Manganiello in #354.

  • Made pytest startup faster when plugin not active by lazy-importing. Contributed by Anders Hovmöller in #339.

  • Various CI improvements. Contributed by Daniel Hahler in #363 and #364.

  • Various Python support updates (drop EOL 3.4, test against 3.8 final). Contributed by Hugo van Kemenade in #336 and #367.

  • Changed --cov-append to always enable data_suffix (a coverage setting). Contributed by Harm Geerts in #387.

  • Changed --cov-append to handle loading previous data better (fixes various path aliasing issues).

  • Various other testing improvements, github issue templates, example updates.

  • Fixed internal failures that are caused by tests that change the current working directory by ensuring a consistent working directory when coverage is called. See #306 and coveragepy#881

2.8.1 (2019-10-05)

  • Fixed #348 - regression when only certain reports (html or xml) are used then --cov-fail-under always fails.

2.8.0 (2019-10-04)

  • Fixed RecursionError that can occur when using cleanup_on_signal or cleanup_on_sigterm. See: #294. The 2.7.x releases of pytest-cov should be considered broken regarding aforementioned cleanup API.

  • Added compatibility with future xdist release that deprecates some internals (match pytest-xdist master/worker terminology). Contributed by Thomas Grainger in #321

  • Fixed breakage that occurs when multiple reporting options are used. Contributed by Thomas Grainger in #338.

  • Changed internals to use a stub instead of os.devnull. Contributed by Thomas Grainger in #332.

  • Added support for Coverage 5.0. Contributed by Ned Batchelder in #319.

  • Added support for float values in --cov-fail-under. Contributed by Martín Gaitán in #311.

  • Various documentation fixes. Contributed by Juanjo Bazán, Andrew Murray and Albert Tugushev in #298, #299 and #307.

  • Various testing improvements. Contributed by Ned Batchelder, Daniel Hahler, Ionel Cristian Mărieș and Hugo van Kemenade in #313, #314, #315, #316, #325, #326, #334 and #335.

  • Added the --cov-context CLI options that enables coverage contexts. Only works with coverage 5.0+. Contributed by Ned Batchelder in #345.

2.7.1 (2019-05-03)

  • Fixed source distribution manifest so that garbage ain’t included in the tarball.

2.7.0 (2019-05-03)

  • Fixed AttributeError: 'NoneType' object has no attribute 'configure_node' error when --no-cov is used. Contributed by Alexander Shadchin in #263.

  • Various testing and CI improvements. Contributed by Daniel Hahler in #255, #266, #272, #271 and #269.

  • Improved pytest_cov.embed.cleanup_on_sigterm to be reentrant (signal deliveries while signal handling is running won’t break stuff).

  • Added pytest_cov.embed.cleanup_on_signal for customized cleanup.

  • Improved cleanup code and fixed various issues with leftover data files. All contributed in #265 or #262.

  • Improved examples. Now there are two examples for the common project layouts, complete with working coverage configuration. The examples have CI testing. Contributed in #267.

  • Improved help text for CLI options.

2.6.1 (2019-01-07)

  • Added support for Pytest 4.1. Contributed by Daniel Hahler and Семён Марьясин in #253 and #230.

  • Various test and docs fixes. Contributed by Daniel Hahler in #224 and #223.

  • Fixed the “Module already imported” issue (#211). Contributed by Daniel Hahler in #228.

2.6.0 (2018-09-03)

  • Dropped support for Python 3 < 3.4, Pytest < 3.5 and Coverage < 4.4.

  • Fixed some documentation formatting. Contributed by Jean Jordaan and Julian.

  • Added an example with addopts in documentation. Contributed by Samuel Giffard in #195.

  • Fixed TypeError: 'NoneType' object is not iterable in certain xdist configurations. Contributed by Jeremy Bowman in #213.

  • Added a no_cover marker and fixture. Fixes #78.

  • Fixed broken no_cover check when running doctests. Contributed by Terence Honles in #200.

  • Fixed various issues with path normalization in reports (when combining coverage data from parallel mode). Fixes #130. Contributed by Ryan Hiebert & Ionel Cristian Mărieș in #178.

  • Report generation failures don’t raise exceptions anymore. A warning will be logged instead. Fixes #161.

  • Fixed multiprocessing issue on Windows (empty env vars are not passed). Fixes #165.

2.5.1 (2017-05-11)

  • Fixed xdist breakage (regression in 2.5.0). Fixes #157.

  • Allow setting custom data_file name in .coveragerc. Fixes #145. Contributed by Jannis Leidel & Ionel Cristian Mărieș in #156.

2.5.0 (2017-05-09)

  • Always show a summary when --cov-fail-under is used. Contributed by Francis Niu in PR#141.

  • Added --cov-branch option. Fixes #85.

  • Improve exception handling in subprocess setup. Fixes #144.

  • Fixed handling when --cov is used multiple times. Fixes #151.

2.4.0 (2016-10-10)

  • Added a “disarm” option: --no-cov. It will disable coverage measurements. Contributed by Zoltan Kozma in PR#135.

    WARNING: Do not put this in your configuration files, it’s meant to be an one-off for situations where you want to disable coverage from command line.

  • Fixed broken exception handling on .pth file. See #136.

2.3.1 (2016-08-07)

  • Fixed regression causing spurious errors when xdist was used. See #124.

  • Fixed DeprecationWarning about incorrect addoption use. Contributed by Florian Bruhin in PR#127.

  • Fixed deprecated use of funcarg fixture API. Contributed by Daniel Hahler in PR#125.

2.3.0 (2016-07-05)

  • Add support for specifying output location for html, xml, and annotate report. Contributed by Patrick Lannigan in PR#113.

  • Fix bug hiding test failure when cov-fail-under failed.

  • For coverage >= 4.0, match the default behaviour of coverage report and error if coverage fails to find the source instead of just printing a warning. Contributed by David Szotten in PR#116.

  • Fixed bug occurred when bare --cov parameter was used with xdist. Contributed by Michael Elovskikh in PR#120.

  • Add support for skip_covered and added --cov-report=term-skip-covered command line options. Contributed by Saurabh Kumar in PR#115.

2.2.1 (2016-01-30)

  • Fixed incorrect merging of coverage data when xdist was used and coverage was >= 4.0.

2.2.0 (2015-10-04)

  • Added support for changing working directory in tests. Previously changing working directory would disable coverage measurements in suprocesses.

  • Fixed broken handling for --cov-report=annotate.

2.1.0 (2015-08-23)

  • Added support for coverage 4.0b2.

  • Added the --cov-append command line options. Contributed by Christian Ledermann in PR#80.

2.0.0 (2015-07-28)

  • Added --cov-fail-under, akin to the new fail_under option in coverage-4.0 (automatically activated if there’s a [report] fail_under = ... in .coveragerc).

  • Changed --cov-report=term to automatically upgrade to --cov-report=term-missing if there’s [run] show_missing = True in .coveragerc.

  • Changed --cov so it can be used with no path argument (in which case the source settings from .coveragerc will be used instead).

  • Fixed .pth installation to work in all cases (install, easy_install, wheels, develop etc).

  • Fixed .pth uninstallation to work for wheel installs.

  • Support for coverage 4.0.

  • Data file suffixing changed to use coverage’s data_suffix=True option (instead of the custom suffixing).

  • Avoid warning about missing coverage data (just like coverage.control.process_startup).

  • Fixed a race condition when running with xdist (all the workers tried to combine the files). It’s possible that this issue is not present in pytest-cov 1.8.X.

1.8.2 (2014-11-06)

  • N/A

Authors

Releasing

The process for releasing should follow these steps:

  1. Test that docs build and render properly by running tox -e docs.

    If there are bogus spelling issues add the words in spelling_wordlist.txt.

  2. Update CHANGELOG.rst and AUTHORS.rst to be up to date.

  3. Bump the version by running bumpversion [ major | minor | patch ]. This will automatically add a tag.

  4. Push changes and tags with:

    git push
git push --tags

  5. Wait GitHub Actions to give the green builds.

  6. Check that the docs on ReadTheDocs are built.

  7. Make sure you have a clean checkout, run git status to verify.

  8. Manually clean temporary files (that are ignored and won’t show up in git status):

    rm -rf dist build src/*.egg-info

    These files need to be removed to force distutils/setuptools to rebuild everything and recreate the egg-info metadata.

  9. Build the dists:

    python -m build

  10. Verify that the resulting archives (found in dist/) are good.

  11. Upload the sdist and wheel with twine:

    twine upload dist/*

Contributing

Contributions are welcome, and they are greatly appreciated! Every little bit helps, and credit will always be given.

Bug reports

When reporting a bug please include:

  • Your operating system name and version.

  • Any details about your local setup that might be helpful in troubleshooting.

  • Detailed steps to reproduce the bug.

Documentation improvements

pytest-cov could always use more documentation, whether as part of the official pytest-cov docs, in docstrings, or even on the web in blog posts, articles, and such.

Feature requests and feedback

The best way to send feedback is to file an issue at https://github.com/pytest-dev/pytest-cov/issues.

If you are proposing a feature:

  • Explain in detail how it would work.

  • Keep the scope as narrow as possible, to make it easier to implement.

  • Remember that this is a volunteer-driven project, and that code contributions are welcome :)

Development

To set up pytest-cov for local development:

  1. Fork pytest-cov (look for the “Fork” button).

  2. Clone your fork locally:

    git clone git@github.com:YOURGITHUBNAME/pytest-cov.git

  3. Create a branch for local development:

    git checkout -b name-of-your-bugfix-or-feature

    Now you can make your changes locally.

  4. When you’re done making changes run all the checks and docs builder with one command:

    tox

  5. Commit your changes and push your branch to GitHub:

    git add .
git commit -m "Your detailed description of your changes."
git push origin name-of-your-bugfix-or-feature

  6. Submit a pull request through the GitHub website.

Pull Request Guidelines

If you need some code review or feedback while you’re developing the code just make the pull request.

For merging, you should:

  1. Include passing tests (run tox).

  2. Update documentation when there’s new API, functionality etc.

  3. Add a note to CHANGELOG.rst about the changes.

  4. Add yourself to AUTHORS.rst.

Tips

To run a subset of tests:

tox -e envname -- pytest -k test_myfeature

To run all the test environments in parallel:

tox -p auto

Indices and tables