Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Coverage regressions on 3.10.rc1 #1205

skirpichev opened this issue Aug 5, 2021 · 19 comments

Coverage regressions on 3.10.rc1 #1205

skirpichev opened this issue Aug 5, 2021 · 19 comments
bug Something isn't working cpython Reported as a bug in CPython fixed


Copy link

skirpichev commented Aug 5, 2021

See up to date codecov status and the build artifact. Coveragepy at 57a691f.

I'll select some cases (codecov shows more regressions, but only ones in the and seems to be in the build artifact, perhaps it's a codecov issue).

  1. L1032, L1037 of no jumps to exit. These lines could be covered, for example, by: exp(Symbol('', real=True, transcendental=True)).is_algebraic and exp(Symbol('', integer=True)).is_algebraic.
  2. L309 of and similar lines with the DecrementLevel context manager. These can be triggered by tests in the diofant/tests/integrals/

PS: You can see more "green" lines counted in the codecov statistics. Maybe it's a codecov issue. I've reopened the pr request as diofant/diofant#1162 to see if this and mentioned above disagreement with the codecov.xml in the build artifact will disappear.

@skirpichev skirpichev added the bug Something isn't working label Aug 5, 2021
Copy link

nedbat commented Aug 5, 2021

Please provide specific instructions about how I can reproduce the problem. Specific commits of your repo to clone, commands to run, and specific lines that you believe are incorrectly measured.

@nedbat nedbat added this to the 6.0 milestone Aug 5, 2021
Copy link

I though I did this...

Ok, next try. You can use diofant/diofant#1162, last commit (c59bd75).

  1. L1032 and L1037 of the diofant/core/ are not covered (no jumps to exit). This should be covered by tests:
exp(Symbol('', integer=True)).is_algebraic  # L1032
exp(Symbol('', real=True, transcendental=True)).is_algebraic  # L1037

In principle, pytest --cov diofant -m 'not slow' diofant/tests/core should also cover these lines.

  1. L309 of the diofant/integrals/ (and similar lines with the DecrementLevel context manager). I'm unsure, maybe it's not a bug, but this line was marked as covered on the master, see the codecov output. Now coveragepy report "line 309 didn't except from function 'bound_degree', because the raise on line 320 wasn't executed or the raise on line 342 wasn't executed" (pytest --cov diofant -m 'not slow' diofant/tests/integrals/ should trigger all tests for the bound_degree() function).

Copy link

FYI: here is the codecov bugreport

Copy link

nedbat commented Aug 5, 2021

I've reported the L1032 issue against CPython:

@nedbat nedbat added the cpython Reported as a bug in CPython label Aug 5, 2021
Copy link

nedbat commented Aug 5, 2021

For the L309 issue, I see this in my coverage results:
It has executed line 336, but then neither line 339 or line 341. That makes it seem like some other exception was raised at line 336, which could explain why the context manager on line 309 never jumped to 371. Can you confirm this theory?

Copy link

Hmm, I did pytest -m 'not slow and not xfail and not regression' diofant/tests/integrals/. It seems, L344 (exit from with) has hit from the test_integration() of diofant/tests/integrals/

Copy link

nedbat commented Aug 5, 2021

@skirpichev I'm sorry, I don't understand what your last comment means.

I tried adding another except clause. It shows that line 336 is raising an uncaught exception:
So I think the L309 issue is not a regression.

Copy link

I don't understand what your last comment means.

Well, perhaps I misunderstood you, but it seems your guess was that the only exit from the with clause was by an exception (from L336). That's not true. E.g. you can replace L344 with print(111) (i.e. right after "with" statement) and run pytest -q -n0 -s diofant/tests/integrals/ -k test_integration - you will see some hits. Or run in console from diofant import *; integrate(exp(-log(x)**2)).

So I think the L309 issue is not a regression.

Clearly, there is an exit by an uncaught exception AND a normal exit.

I'll try to reduce this example, of course.

Copy link

PS: You can see more "green" lines counted in the codecov statistics. Maybe it's a codecov issue.

Regarding this, it seems to be a coveragepy "feature". Indeed, some new lines are counted. For instance, L1489 of, which is "while True" statement.

I'll try to reduce this example, of course.

Another example is in the cancel_primitive():

<line number="546" hits="1" branch="true" condition-coverage="66% (2/3)" missing-branches="exit"/>
IIUC, three branches are: normal exit from "with", an uncaught exception from some function inside of "with" and the NotImplementedError on L552?

Probably, you were right, this is not an issue. On another hand, in this example we have only a NotImplementedError, which is ignored by my config. Shouldn't then this branch be ignored by the coveragepy (like for if/else blocks)?

At least, the coveragepy behaviour was changed significantly, could be this documented?

nedbat added a commit that referenced this issue Aug 9, 2021
Copy link

FYI, unless I miss something, 1st issue wasn't solved in cpython:


Coveragepy at 57a691f. I've tested this issue on the recent 3.10 head, which has backport from your issue:

$ python
Python 3.10.0rc1+ (heads/3.10:d5c8ad2471, Aug  9 2021, 15:31:13) [GCC 10.2.1 20210110] on linux
Type "help", "copyright", "credits" or "license" for more information.

Copy link

nedbat commented Aug 9, 2021

@skirpichev I really appreciate your dedication to finding these problems, but please please please please please PLEASE always include the steps you used to produce the results you are showing.

And can you say specifically what about that output is incorrect?

This is what I did (with output showing versions):

$ mktmpenv -p /usr/local/cpython/bin/python3 -n
$ python -VV
Python 3.10.0rc1+ (heads/3.10:d5c8ad2471, Aug  9 2021, 10:24:02) [Clang 12.0.0 (clang-1200.0.32.29)]
$ cd diofant
$ git log --oneline | head -1
c59bd753e XXX removed all workarounds for issue2506 & skipif ground types != gmpy
$ pip install -e '.[develop,docs,tests]'
$ pip install git+
$ pytest --cov diofant -m 'not slow' diofant/tests/core

My HTML report looks like this:

We seem to be running different tests.

Copy link

include the steps you used to produce the results you are showing

Sorry, I thought I did it before.

And can you say specifically what about that output is incorrect?

Exactly same as in the issue description, 1st point.

We seem to be running different tests.

I see no difference for L1032 and L1037.

The source of discrepancy seems to be that I did tests with -m 'not slow and not xfail and not regression' instead. That's not an important thing: mentioned before steps should cover exit's from L1032 and L1037. (On the current diofant/diofant#1162 you can run pytest -k test_Pow_is_algebraic --cov diofant diofant/tests/core/ to trigger these exits.)

Copy link

nedbat commented Aug 9, 2021

I mistakenly skipped the step of generating the HTML report after running the tests. When I re-do the HTML report, I get this:


Copy link

Unfortunately, I run this step. Are you sure you didn't run tests on CPython 3.9 or more older?

I've tested by pytest -k test_Pow_is_algebraic --cov diofant diofant/tests/core/ && coverage html on the current diofant/diofant#1162 (merge commit: diofant/diofant@e62cf84). Both for 3.9 and 3.10 CPython (heads/3.10:d5c8ad2471). I still see uncovered exits on L1032 and L1037, as it shown before.

Probably, this is my fault, but I've no glue on what can be wrong on my side.

Copy link

After adding "pass" statements I see this:
for latest CPython 3.10:

$ python
Python 3.10.0rc1+ (heads/3.10:a3185da26f, Aug 10 2021, 06:42:46) [GCC 10.2.1 20210110] on linux
Type "help", "copyright", "credits" or "license" for more information.


pytest -k test_Pow_is_algebraic --cov diofant diofant/tests/core/ && coverage html

PS: coveragepy version as before.

Copy link

nedbat commented Aug 10, 2021

Are you in a virtualenv? The python command may be finding the latest Python 3.10, but the pytest command may be finding a different python. Try using python -m pytest instead.

Copy link

Are you in a virtualenv?

No, I use venv.

Try using python -m pytest instead.

No, both pytest & python -m pytest - report same version.

It seems, git clean -Xdf did the trick. I guess, some old *.pyc files were used.

Thank you very much for help! 1st issue is solved.

What about the second one? I think, the original example may be a non-issue (or a documentation issue), but another one (with a NotImplementedError, which should be ignored) - is.

Copy link

nedbat commented Aug 12, 2021

@skirpichev I am going to close this issue as solved for #1. Please open a new issue for #2, with specific instructions about the tests you ran to get the output you saw, and specific details about exactly what you think is wrong. I tried running tests to reproduce #2, and got a completely different result than you showed.

Again, I really appreciate your persistence, but we have to get a smoother flow here with reproducible instructions. Assume I know nothing. Give me exact commands to run. Thanks.

Copy link

nedbat commented Oct 3, 2021

This is now released as part of coverage 6.0.

netbsd-srcmastr pushed a commit to NetBSD/pkgsrc that referenced this issue Aug 24, 2022
Version 6.4.4 — 2022-08-16

- Wheels are now provided for Python 3.11.

.. _changes_6-4-3:

Version 6.4.3 — 2022-08-06

- Fix a failure when combining data files if the file names contained
  glob-like patterns (`pull 1405`_).  Thanks, Michael Krebs and Benjamin

- Fix a messaging failure when combining Windows data files on a different
  drive than the current directory. (`pull 1430`_, fixing `issue 1428`_).
  Thanks, Lorenzo Micò.

- Fix path calculations when running in the root directory, as you might do in
  a Docker container: `pull 1403`_, thanks Arthur Rio.

- Filtering in the HTML report wouldn't work when reloading the index page.
  This is now fixed (`pull 1413`_).  Thanks, Marc Legendre.

- Fix a problem with Cython code measurement (`pull 1347`_, fixing `issue
  972`_).  Thanks, Matus Valo.

.. _issue 972: nedbat/coveragepy#972
.. _pull 1347: nedbat/coveragepy#1347
.. _pull 1403: nedbat/coveragepy#1403
.. _pull 1405: nedbat/coveragepy#1405
.. _pull 1413: nedbat/coveragepy#1413
.. _issue 1428: nedbat/coveragepy#1428
.. _pull 1430: nedbat/coveragepy#1430

.. _changes_6-4-2:

Version 6.4.2 — 2022-07-12

- Updated for a small change in Python 3.11.0 beta 4: modules now start with a
  line with line number 0, which is ignored.  This line cannnot be executed, so
  coverage totals were thrown off.  This line is now ignored by,
  but this also means that truly empty modules (like ````) have no
  lines in them, rather than one phantom line.  Fixes `issue 1419`_.

- Internal debugging data added to sys.modules is now an actual module, to
  avoid confusing code that examines everything in sys.modules.  Thanks,
  Yilei Yang (`pull 1399`_).

.. _pull 1399: nedbat/coveragepy#1399
.. _issue 1419: nedbat/coveragepy#1419

.. _changes_6-4-1:

Version 6.4.1 — 2022-06-02

- Greatly improved performance on PyPy, and other environments that need the
  pure Python trace function.  Thanks, Carl Friedrich Bolz-Tereick (`pull
  1381`_ and `pull 1388`_).  Slightly improved performance when using the C
  trace function, as most environments do.  Closes `issue 1339`_.

- The conditions for using tomllib from the standard library have been made
  more precise, so that 3.11 alphas will continue to work. Closes `issue

.. _issue 1339: nedbat/coveragepy#1339
.. _pull 1381: nedbat/coveragepy#1381
.. _pull 1388: nedbat/coveragepy#1388
.. _issue 1390: nedbat/coveragepy#1390

.. _changes_64:

Version 6.4 — 2022-05-22

- A new setting, :ref:`config_run_sigterm`, controls whether a SIGTERM signal
  handler is used.  In 6.3, the signal handler was always installed, to capture
  data at unusual process ends.  Unfortunately, this introduced other problems
  (see `issue 1310`_).  Now the signal handler is only used if you opt-in by
  setting ``[run] sigterm = true``.

- Small changes to the HTML report:

  - Added links to next and previous file, and more keyboard shortcuts: ``[``
    and ``]`` for next file and previous file; ``u`` for up to the index; and
    ``?`` to open/close the help panel.  Thanks, `J. M. F. Tsang
    <pull 1364_>`_.

  - The timestamp and version are displayed at the top of the report.  Thanks,
    `Ammar Askar <pull 1354_>`_. Closes `issue 1351`_.

- A new debug option ``debug=sqldata`` adds more detail to ``debug=sql``,
  logging all the data being written to the database.

- Previously, running ``coverage report`` (or any of the reporting commands) in
  an empty directory would create a .coverage data file.  Now they do not,
  fixing `issue 1328`_.

- On Python 3.11, the ``[toml]`` extra no longer installs tomli, instead using
  tomllib from the standard library.  Thanks `Shantanu <pull 1359_>`_.

- In-memory CoverageData objects now properly update(), closing `issue 1323`_.

.. _issue 1310: nedbat/coveragepy#1310
.. _issue 1323: nedbat/coveragepy#1323
.. _issue 1328: nedbat/coveragepy#1328
.. _issue 1351: nedbat/coveragepy#1351
.. _pull 1354: nedbat/coveragepy#1354
.. _pull 1359: nedbat/coveragepy#1359
.. _pull 1364: nedbat/coveragepy#1364

.. _changes_633:

Version 6.3.3 — 2022-05-12

- Fix: now builds successfully on CPython 3.11 (3.11.0b1) again.
  Closes `issue 1367`_.  Some results for generators may have changed.

.. _issue 1367: nedbat/coveragepy#1367

.. _changes_632:

Version 6.3.2 — 2022-02-20

- Fix: adapt to pypy3.9's decorator tracing behavior.  It now traces function
  decorators like CPython 3.8: both the @-line and the def-line are traced.
  Fixes `issue 1326`_.

- Debug: added ``pybehave`` to the list of :ref:`coverage debug <cmd_debug>`
  and :ref:`cmd_run_debug` options.

- Fix: show an intelligible error message if ``--concurrency=multiprocessing``
  is used without a configuration file.  Closes `issue 1320`_.

.. _issue 1320: nedbat/coveragepy#1320
.. _issue 1326: nedbat/coveragepy#1326

.. _changes_631:

Version 6.3.1 — 2022-02-01

- Fix: deadlocks could occur when terminating processes.  Some of these
  deadlocks (described in `issue 1310`_) are now fixed.

- Fix: a signal handler was being set from multiple threads, causing an error:
  "ValueError: signal only works in main thread".  This is now fixed, closing
  `issue 1312`_.

- Fix: ``--precision`` on the command-line was being ignored while considering
  ``--fail-under``.  This is now fixed, thanks to
  `Marcelo Trylesinski <pull 1317_>`_.

- Fix: releases no longer provide 3.11.0-alpha wheels. uses CPython
  internal fields which are moving during the alpha phase. Fixes `issue 1316`_.

.. _issue 1310: nedbat/coveragepy#1310
.. _issue 1312: nedbat/coveragepy#1312
.. _issue 1316: nedbat/coveragepy#1316
.. _pull 1317: nedbat/coveragepy#1317

.. _changes_63:

Version 6.3 — 2022-01-25

- Feature: Added the ``lcov`` command to generate reports in LCOV format.
  Thanks, `Bradley Burns <pull 1289_>`_. Closes issues `587 <issue 587_>`_
  and `626 <issue 626_>`_.

- Feature: the coverage data file can now be specified on the command line with
  the ``--data-file`` option in any command that reads or writes data.  This is
  in addition to the existing ``COVERAGE_FILE`` environment variable.  Closes
  `issue 624`_. Thanks, `Nikita Bloshchanevich <pull 1304_>`_.

- Feature: coverage measurement data will now be written when a SIGTERM signal
  is received by the process.  This includes
  :meth:`Process.terminate <python:multiprocessing.Process.terminate>`,
  and other ways to terminate a process.  Currently this is only on Linux and
  Mac; Windows is not supported.  Fixes `issue 1307`_.

- Dropped support for Python 3.6, which reached end-of-life on 2021-12-23.

- Updated Python 3.11 support to 3.11.0a4, fixing `issue 1294`_.

- Fix: the coverage data file is now created in a more robust way, to avoid
  problems when multiple processes are trying to write data at once. Fixes
  issues `1303 <issue 1303_>`_ and `883 <issue 883_>`_.

- Fix: a .gitignore file will only be written into the HTML report output
  directory if the directory is empty.  This should prevent certain unfortunate
  accidents of writing the file where it is not wanted.

- Releases now have MacOS arm64 wheels for Apple Silicon, fixing `issue 1288`_.

.. _issue 587: nedbat/coveragepy#587
.. _issue 624: nedbat/coveragepy#624
.. _issue 626: nedbat/coveragepy#626
.. _issue 883: nedbat/coveragepy#883
.. _issue 1288: nedbat/coveragepy#1288
.. _issue 1294: nedbat/coveragepy#1294
.. _issue 1303: nedbat/coveragepy#1303
.. _issue 1307: nedbat/coveragepy#1307
.. _pull 1289: nedbat/coveragepy#1289
.. _pull 1304: nedbat/coveragepy#1304

.. _changes_62:

Version 6.2 — 2021-11-26

- Feature: Now the ``--concurrency`` setting can now have a list of values, so
  that threads and another lightweight threading package can be measured
  together, such as ``--concurrency=gevent,thread``.  Closes `issue 1012`_ and
  `issue 1082`_.

- Fix: A module specified as the ``source`` setting is imported during startup,
  before the user program imports it.  This could cause problems if the rest of
  the program isn't ready yet.  For example, `issue 1203`_ describes a Django
  setting that is accessed before settings have been configured.  Now the early
  import is wrapped in a try/except so errors then don't stop execution.

- Fix: A colon in a decorator expression would cause an exclusion to end too
  early, preventing the exclusion of the decorated function. This is now fixed.

- Fix: The HTML report now will not overwrite a .gitignore file that already
  exists in the HTML output directory (follow-on for `issue 1244`_).

- API: The exceptions raised by have been specialized, to provide
  finer-grained catching of exceptions by third-party code.

- API: Using ``suffix=False`` when constructing a Coverage object with
  multiprocessing wouldn't suppress the data file suffix (`issue 989`_).  This
  is now fixed.

- Debug: The ``coverage debug data`` command will now sniff out combinable data
  files, and report on all of them.

- Debug: The ``coverage debug`` command used to accept a number of topics at a
  time, and show all of them, though this was never documented.  This no longer
  works, to allow for command-line options in the future.

.. _issue 989: nedbat/coveragepy#989
.. _issue 1012: nedbat/coveragepy#1012
.. _issue 1082: nedbat/coveragepy#1082
.. _issue 1203: nedbat/coveragepy#1203

.. _changes_612:

Version 6.1.2 — 2021-11-10

- Python 3.11 is supported (tested with 3.11.0a2).  One still-open issue has to
  do with `exits through with-statements <issue 1270_>`_.

- Fix: When remapping file paths through the ``[paths]`` setting while
  combining, the ``[run] relative_files`` setting was ignored, resulting in
  absolute paths for remapped file names (`issue 1147`_).  This is now fixed.

- Fix: Complex conditionals over excluded lines could have incorrectly reported
  a missing branch (`issue 1271`_). This is now fixed.

- Fix: More exceptions are now handled when trying to parse source files for
  reporting.  Problems that used to terminate can now be handled
  with ``[report] ignore_errors``.  This helps with plugins failing to read
  files (`django_coverage_plugin issue 78`_).

- Fix: Removed another vestige of jQuery from the source tarball
  (`issue 840`_).

- Fix: Added a default value for a new-to-6.x argument of an internal class.
  This unsupported class is being used by coveralls (`issue 1273`_). Although
  I'd rather not "fix" unsupported interfaces, it's actually nicer with a
  default value.

.. _django_coverage_plugin issue 78: nedbat/django_coverage_plugin#78
.. _issue 1147: nedbat/coveragepy#1147
.. _issue 1270: nedbat/coveragepy#1270
.. _issue 1271: nedbat/coveragepy#1271
.. _issue 1273: nedbat/coveragepy#1273

.. _changes_611:

Version 6.1.1 — 2021-10-31

- Fix: The sticky header on the HTML report didn't work unless you had branch
  coverage enabled. This is now fixed: the sticky header works for everyone.
  (Do people still use coverage without branch measurement!? j/k)

- Fix: When using explicitly declared namespace packages, the "already imported
  a file that will be measured" warning would be issued (`issue 888`_).  This
  is now fixed.

.. _issue 888: nedbat/coveragepy#888

.. _changes_61:

Version 6.1 — 2021-10-30

- Deprecated: The ``annotate`` command and the ``Coverage.annotate`` function
  will be removed in a future version, unless people let me know that they are
  using it.  Instead, the ``html`` command gives better-looking (and more
  accurate) output, and the ``report -m`` command will tell you line numbers of
  missing lines.  Please get in touch if you have a reason to use ``annotate``
  over those better options:

- Feature: Coverage now sets an environment variable, ``COVERAGE_RUN`` when
  running your code with the ``coverage run`` command.  The value is not
  important, and may change in the future.  Closes `issue 553`_.

- Feature: The HTML report pages for Python source files now have a sticky
  header so the file name and controls are always visible.

- Feature: The ``xml`` and ``json`` commands now describe what they wrote

- Feature: The ``html``, ``combine``, ``xml``, and ``json`` commands all accept
  a ``-q/--quiet`` option to suppress the messages they write to stdout about
  what they are doing (`issue 1254`_).

- Feature: The ``html`` command writes a ``.gitignore`` file into the HTML
  output directory, to prevent the report from being committed to git.  If you
  want to commit it, you will need to delete that file.  Closes `issue 1244`_.

- Feature: Added support for PyPy 3.8.

- Fix: More generated code is now excluded from measurement.  Code such as
  `attrs`_ boilerplate, or doctest code, was being measured though the
  synthetic line numbers meant they were never reported.  Once Cython was
  involved though, the generated .so files were parsed as Python, raising
  syntax errors, as reported in `issue 1160`_.  This is now fixed.

- Fix: When sorting human-readable names, numeric components are sorted
  correctly: will appear after  This applies to file names,
  module names, environment variables, and test contexts.

- Performance: Branch coverage measurement is faster, though you might only
  notice on code that is executed many times, such as long-running loops.

- Build: jQuery is no longer used or vendored (`issue 840`_ and `issue 1118`_).
  Huge thanks to Nils Kattenbeck (septatrix) for the conversion to vanilla
  JavaScript in `pull request 1248`_.

.. _issue 553: nedbat/coveragepy#553
.. _issue 840: nedbat/coveragepy#840
.. _issue 1118: nedbat/coveragepy#1118
.. _issue 1160: nedbat/coveragepy#1160
.. _issue 1244: nedbat/coveragepy#1244
.. _pull request 1248: nedbat/coveragepy#1248
.. _issue 1254: nedbat/coveragepy#1254
.. _attrs:

.. _changes_602:

Version 6.0.2 — 2021-10-11

- Namespace packages being measured weren't properly handled by the new code
  that ignores third-party packages. If the namespace package was installed, it
  was ignored as a third-party package.  That problem (`issue 1231`_) is now

- Packages named as "source packages" (with ``source``, or ``source_pkgs``, or
  pytest-cov's ``--cov``) might have been only partially measured.  Their
  top-level statements could be marked as unexecuted, because they were
  imported by before measurement began (`issue 1232`_).  This is
  now fixed, but the package will be imported twice, once by, then
  again by your test suite.  This could cause problems if importing the package
  has side effects.

- The :meth:`.CoverageData.contexts_by_lineno` method was documented to return
  a dict, but was returning a defaultdict.  Now it returns a plain dict.  It
  also no longer returns negative numbered keys.

.. _issue 1231: nedbat/coveragepy#1231
.. _issue 1232: nedbat/coveragepy#1232

.. _changes_601:

Version 6.0.1 — 2021-10-06

- In 6.0, the exceptions moved from coverage.misc to
  coverage.exceptions. These exceptions are not part of the public supported
  API, CoverageException is. But a number of other third-party packages were
  importing the exceptions from coverage.misc, so they are now available from
  there again (`issue 1226`_).

- Changed an internal detail of how tomli is imported, so that tomli can use for their own test suite (`issue 1228`_).

- Defend against an obscure possibility under code obfuscation, where a
  function can have an argument called "self", but no local named "self"
  (`pull request 1210`_).  Thanks, Ben Carlsson.

.. _pull request 1210: nedbat/coveragepy#1210
.. _issue 1226: nedbat/coveragepy#1226
.. _issue 1228: nedbat/coveragepy#1228

.. _changes_60:

Version 6.0 — 2021-10-03

- The ``coverage html`` command now prints a message indicating where the HTML
  report was written.  Fixes `issue 1195`_.

- The ``coverage combine`` command now prints messages indicating each data
  file being combined.  Fixes `issue 1105`_.

- The HTML report now includes a sentence about skipped files due to
  ``skip_covered`` or ``skip_empty`` settings.  Fixes `issue 1163`_.

- Unrecognized options in the configuration file are no longer errors. They are
  now warnings, to ease the use of coverage across versions.  Fixes `issue

- Fix handling of exceptions through context managers in Python 3.10. A missing
  exception is no longer considered a missing branch from the with statement.
  Fixes `issue 1205`_.

- Fix another rarer instance of "Error binding parameter 0 - probably
  unsupported type." (`issue 1010`_).

- Creating a directory for the coverage data file now is safer against
  conflicts when two coverage runs happen simultaneously (`pull 1220`_).
  Thanks, Clément Pit-Claudel.

.. _issue 1035: nedbat/coveragepy#1035
.. _issue 1105: nedbat/coveragepy#1105
.. _issue 1163: nedbat/coveragepy#1163
.. _issue 1195: nedbat/coveragepy#1195
.. _issue 1205: nedbat/coveragepy#1205
.. _pull 1220: nedbat/coveragepy#1220

.. _changes_60b1:

Version 6.0b1 — 2021-07-18

- Dropped support for Python 2.7, PyPy 2, and Python 3.5.

- Added support for the Python 3.10 ``match/case`` syntax.

- Data collection is now thread-safe.  There may have been rare instances of
  exceptions raised in multi-threaded programs.

- Plugins (like the `Django coverage plugin`_) were generating "Already
  imported a file that will be measured" warnings about Django itself.  These
  have been fixed, closing `issue 1150`_.

- Warnings generated by are now real Python warnings.

- Using ``--fail-under=100`` with coverage near 100% could result in the
  self-contradictory message :code:`total of 100 is less than fail-under=100`.
  This bug (`issue 1168`_) is now fixed.

- The ``COVERAGE_DEBUG_FILE`` environment variable now accepts ``stdout`` and
  ``stderr`` to write to those destinations.

- TOML parsing now uses the `tomli`_ library.

- Some minor changes to usually invisible details of the HTML report:

  - Use a modern hash algorithm when fingerprinting, for high-security
    environments (`issue 1189`_).  When generating the HTML report, we save the
    hash of the data, to avoid regenerating an unchanged HTML page. We used to
    use MD5 to generate the hash, and now use SHA-3-256.  This was never a
    security concern, but security scanners would notice the MD5 algorithm and
    raise a false alarm.

  - Change how report file names are generated, to avoid leading underscores
    (`issue 1167`_), to avoid rare file name collisions (`issue 584`_), and to
    avoid file names becoming too long (`issue 580`_).

.. _Django coverage plugin:
.. _issue 580: nedbat/coveragepy#580
.. _issue 584: nedbat/coveragepy#584
.. _issue 1150: nedbat/coveragepy#1150
.. _issue 1167: nedbat/coveragepy#1167
.. _issue 1168: nedbat/coveragepy#1168
.. _issue 1189: nedbat/coveragepy#1189
.. _tomli:

.. _changes_56b1:

Version 5.6b1 — 2021-04-13

Note: 5.6 final was never released. These changes are part of 6.0.

- Third-party packages are now ignored in coverage reporting.  This solves a
  few problems:

  - Coverage will no longer report about other people's code (`issue 876`_).
    This is true even when using ``--source=.`` with a venv in the current

  - Coverage will no longer generate "Already imported a file that will be
    measured" warnings about coverage itself (`issue 905`_).

- The HTML report uses j/k to move up and down among the highlighted chunks of
  code.  They used to highlight the current chunk, but 5.0 broke that behavior.
  Now the highlighting is working again.

- The JSON report now includes ``percent_covered_display``, a string with the
  total percentage, rounded to the same number of decimal places as the other
  reports' totals.

.. _issue 876: nedbat/coveragepy#876
.. _issue 905: nedbat/coveragepy#905
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
bug Something isn't working cpython Reported as a bug in CPython fixed
None yet

No branches or pull requests

2 participants