----------------------------

- Now that 4.5 properly separated the ``[run] omit`` and ``[report] omit``
settings, an old bug has become apparent.  If you specified a package name
for ``[run] source``, then omit patterns weren't matched inside that package.
This bug (`issue 638`_) is now fixed.

- On Python 3.7, reporting about a decorated function with no body other than a
docstring would crash coverage.py with an IndexError (`issue 640`_).  This is
now fixed.

- Configurer plugins are now reported in the output of ``--debug=sys``.

.. _issue 638: https://bitbucket.org/ned/coveragepy/issues/638/run-omit-is-ignored-since-45
.. _issue 640: https://bitbucket.org/ned/coveragepy/issues/640/indexerror-reporting-on-an-empty-decorated


.. _changes_45:

--------------------------

- A new kind of plugin is supported: configurators are invoked at start-up to
allow more complex configuration than the .coveragerc file can easily do.
See :ref:`api_plugin` for details.  This solves the complex configuration
problem described in `issue 563`_.

- The ``fail_under`` option can now be a float.  Note that you must specify the
``[report] precision`` configuration option for the fractional part to be
used.  Thanks to Lars Hupfeldt Nielsen for help with the implementation.
Fixes `issue 631`_.

- The ``include`` and ``omit`` options can be specified for both the ``[run]``
and ``[report]`` phases of execution.  4.4.2 introduced some incorrect
interactions between those phases, where the options for one were confused
for the other.  This is now corrected, fixing `issue 621`_ and `issue 622`_.
Thanks to Daniel Hahler for seeing more clearly than I could.

- The ``coverage combine`` command used to always overwrite the data file, even
when no data had been read from apparently combinable files.  Now, an error
is raised if we thought there were files to combine, but in fact none of them
could be used.  Fixes `issue 629`_.

- The ``coverage combine`` command could get confused about path separators
when combining data collected on Windows with data collected on Linux, as
described in `issue 618`_.  This is now fixed: the result path always uses
the path separator specified in the ``[paths]`` result.

- On Windows, the HTML report could fail when source trees are deeply nested,
due to attempting to create HTML filenames longer than the 250-character
maximum.  Now filenames will never get much larger than 200 characters,
fixing `issue 627`_.  Thanks to Alex Sandro for helping with the fix.

.. _issue 563: https://bitbucket.org/ned/coveragepy/issues/563/platform-specific-configuration
.. _issue 618: https://bitbucket.org/ned/coveragepy/issues/618/problem-when-combining-windows-generated
.. _issue 621: https://bitbucket.org/ned/coveragepy/issues/621/include-ignored-warning-when-using
.. _issue 622: https://bitbucket.org/ned/coveragepy/issues/622/report-omit-overwrites-run-omit
.. _issue 627: https://bitbucket.org/ned/coveragepy/issues/627/failure-generating-html-reports-when-the
.. _issue 629: https://bitbucket.org/ned/coveragepy/issues/629/multiple-use-of-combine-leads-to-empty
.. _issue 631: https://bitbucket.org/ned/coveragepy/issues/631/precise-coverage-percentage-value


.. _changes_442:

----------------------------

- Support for Python 3.7.  In some cases, class and module docstrings are no
longer counted in statement totals, which could slightly change your total
results.

- Specifying both ``--source`` and ``--include`` no longer silently ignores the
include setting, instead it displays a warning. Thanks, Loïc Dachary.  Closes
`issue 265`_ and `issue 101`_.

- Fixed a race condition when saving data and multiple threads are tracing
(`issue 581`_). It could produce a "dictionary changed size during iteration"
RuntimeError.  I believe this mostly but not entirely fixes the race
condition.  A true fix would likely be too expensive.  Thanks, Peter Baughman
for the debugging, and Olivier Grisel for the fix with tests.

- Configuration values which are file paths will now apply tilde-expansion,
closing `issue 589`_.

- Now secondary config files like tox.ini and setup.cfg can be specified
explicitly, and prefixed sections like `[coverage:run]` will be read. Fixes
`issue 588`_.

- Be more flexible about the command name displayed by help, fixing
`issue 600`_. Thanks, Ben Finney.

.. _issue 101: https://bitbucket.org/ned/coveragepy/issues/101/settings-under-report-affect-running
.. _issue 581: https://bitbucket.org/ned/coveragepy/issues/581/race-condition-when-saving-data-under
.. _issue 588: https://bitbucket.org/ned/coveragepy/issues/588/using-rcfile-path-to-toxini-uses-run
.. _issue 589: https://bitbucket.org/ned/coveragepy/issues/589/allow-expansion-in-coveragerc
.. _issue 600: https://bitbucket.org/ned/coveragepy/issues/600/get-program-name-from-command-line-when


.. _changes_441:

----------------------------

- No code changes: just corrected packaging for Python 2.7 Linux wheels.


.. _changes_44:

--------------------------

- Reports could produce the wrong file names for packages, reporting ``pkg.py``
instead of the correct ``pkg/__init__.py``.  This is now fixed.  Thanks, Dirk
Thomas.

- XML reports could produce ``<source>`` and ``<class>`` lines that together
didn't specify a valid source file path.  This is now fixed. (`issue 526`_)

- Namespace packages are no longer warned as having no code. (`issue 572`_)

- Code that uses ``sys.settrace(sys.gettrace())`` in a file that wasn't being
coverage-measured would prevent correct coverage measurement in following
code. An example of this was running doctests programmatically. This is now
fixed. (`issue 575`_)

- Errors printed by the ``coverage`` command now go to stderr instead of
stdout.

- Running ``coverage xml`` in a directory named with non-ASCII characters would
fail under Python 2. This is now fixed. (`issue 573`_)

.. _issue 526: https://bitbucket.org/ned/coveragepy/issues/526/generated-xml-invalid-paths-for-cobertura
.. _issue 572: https://bitbucket.org/ned/coveragepy/issues/572/no-python-source-warning-for-namespace
.. _issue 573: https://bitbucket.org/ned/coveragepy/issues/573/cant-generate-xml-report-if-some-source
.. _issue 575: https://bitbucket.org/ned/coveragepy/issues/575/running-doctest-prevents-complete-coverage

----------------------------

- Some warnings can now be individually disabled.  Warnings that can be
disabled have a short name appended.  The ``[run] disable_warnings`` setting
takes a list of these warning names to disable. Closes both `issue 96`_ and
`issue 355`_.

- The XML report now includes attributes from version 4 of the Cobertura XML
format, fixing `issue 570`_.

- In previous versions, calling a method that used collected data would prevent
further collection.  For example, `save()`, `report()`, `html_report()`, and
others would all stop collection.  An explicit `start()` was needed to get it
going again.  This is no longer true.  Now you can use the collected data and
also continue measurement. Both `issue 79`_ and `issue 448`_ described this
problem, and have been fixed.

- Plugins can now find unexecuted files if they choose, by implementing the
`find_executable_files` method.  Thanks, Emil Madsen.

- Minimal IronPython support. You should be able to run IronPython programs
under ``coverage run``, though you will still have to do the reporting phase
with CPython.

- Coverage.py has long had a special hack to support CPython's need to measure
the coverage of the standard library tests. This code was not installed by
kitted versions of coverage.py.  Now it is.

.. _issue 79: https://bitbucket.org/ned/coveragepy/issues/79/save-prevents-harvesting-on-stop
.. _issue 96: https://bitbucket.org/ned/coveragepy/issues/96/unhelpful-warnings-produced-when-using
.. _issue 355: https://bitbucket.org/ned/coveragepy/issues/355/warnings-should-be-suppressable
.. _issue 448: https://bitbucket.org/ned/coveragepy/issues/448/save-and-html_report-prevent-further
.. _issue 570: https://bitbucket.org/ned/coveragepy/issues/570/cobertura-coverage-04dtd-support


.. _changes_434:

----------------------------

- Fixing 2.6 in version 4.3.3 broke other things, because the too-tricky
exception wasn't properly derived from Exception, described in `issue 556`_.
A newb mistake; it hasn't been a good few days.

.. _issue 556: https://bitbucket.org/ned/coveragepy/issues/556/43-fails-if-there-are-html-files-in-the


.. _changes_433:

----------------------------

- Python 2.6 support was broken due to a testing exception imported for the
benefit of the coverage.py test suite.  Properly conditionalizing it fixed
`issue 554`_ so that Python 2.6 works again.

.. _issue 554: https://bitbucket.org/ned/coveragepy/issues/554/traceback-on-python-26-starting-with-432


.. _changes_432:

----------------------------

- Using the ``--skip-covered`` option on an HTML report with 100% coverage
would cause a "No data to report" error, as reported in `issue 549`_. This is
now fixed; thanks, Loïc Dachary.

- If-statements can be optimized away during compilation, for example, `if 0:`
or `if __debug__:`.  Coverage.py had problems properly understanding these
statements which existed in the source, but not in the compiled bytecode.
This problem, reported in `issue 522`_, is now fixed.

- If you specified ``--source`` as a directory, then coverage.py would look for
importable Python files in that directory, and could identify ones that had
never been executed at all.  But if you specified it as a package name, that
detection wasn't performed.  Now it is, closing `issue 426`_. Thanks to Loïc
Dachary for the fix.

- If you started and stopped coverage measurement thousands of times in your
process, you could crash Python with a "Fatal Python error: deallocating
None" error.  This is now fixed.  Thanks to Alex Groce for the bug report.

- On PyPy, measuring coverage in subprocesses could produce a warning: "Trace
function changed, measurement is likely wrong: None".  This was spurious, and
has been suppressed.

- Previously, coverage.py couldn't start on Jython, due to that implementation
missing the multiprocessing module (`issue 551`_). This problem has now been
fixed. Also, `issue 322`_ about not being able to invoke coverage
conveniently, seems much better: ``jython -m coverage run myprog.py`` works
properly.

- Let's say you ran the HTML report over and over again in the same output
directory, with ``--skip-covered``. And imagine due to your heroic
test-writing efforts, a file just achieved the goal of 100% coverage. With
coverage.py 4.3, the old HTML file with the less-than-100% coverage would be
left behind.  This file is now properly deleted.

.. _issue 322: https://bitbucket.org/ned/coveragepy/issues/322/cannot-use-coverage-with-jython
.. _issue 426: https://bitbucket.org/ned/coveragepy/issues/426/difference-between-coverage-results-with
.. _issue 522: https://bitbucket.org/ned/coveragepy/issues/522/incorrect-branch-reporting-with-__debug__
.. _issue 549: https://bitbucket.org/ned/coveragepy/issues/549/skip-covered-with-100-coverage-throws-a-no
.. _issue 551: https://bitbucket.org/ned/coveragepy/issues/551/coveragepy-cannot-be-imported-in-jython27


.. _changes_431:

----------------------------

- Some environments couldn't install 4.3, as described in `issue 540`_. This is
now fixed.

- The check for conflicting ``--source`` and ``--include`` was too simple in a
few different ways, breaking a few perfectly reasonable use cases, described
in `issue 541`_.  The check has been reverted while we re-think the fix for
`issue 265`_.

.. _issue 540: https://bitbucket.org/ned/coveragepy/issues/540/cant-install-coverage-v43-into-under
.. _issue 541: https://bitbucket.org/ned/coveragepy/issues/541/coverage-43-breaks-nosetest-with-coverage


.. _changes_43:

--------------------------

Special thanks to **Loïc Dachary**, who took an extraordinary interest in
coverage.py and contributed a number of improvements in this release.

- Subprocesses that are measured with `automatic subprocess measurement`_ used
to read in any pre-existing data file.  This meant data would be incorrectly
carried forward from run to run.  Now those files are not read, so each
subprocess only writes its own data. Fixes `issue 510`_.

- The ``coverage combine`` command will now fail if there are no data files to
combine. The combine changes in 4.2 meant that multiple combines could lose
data, leaving you with an empty .coverage data file. Fixes
`issue 525`_, `issue 412`_, `issue 516`_, and probably `issue 511`_.

- Coverage.py wouldn't execute `sys.excepthook`_ when an exception happened in
your program.  Now it does, thanks to Andrew Hoos.  Closes `issue 535`_.

- Branch coverage fixes:

- Branch coverage could misunderstand a finally clause on a try block that
 never continued on to the following statement, as described in `issue
 493`_.  This is now fixed. Thanks to Joe Doherty for the report and Loïc
 Dachary for the fix.

- A while loop with a constant condition (while True) and a continue
 statement would be mis-analyzed, as described in `issue 496`_. This is now
 fixed, thanks to a bug report by Eli Skeggs and a fix by Loïc Dachary.

- While loops with constant conditions that were never executed could result
 in a non-zero coverage report.  Artem Dayneko reported this in `issue
 502`_, and Loïc Dachary provided the fix.

- The HTML report now supports a ``--skip-covered`` option like the other
reporting commands.  Thanks, Loïc Dachary for the implementation, closing
`issue 433`_.

- Options can now be read from a tox.ini file, if any. Like setup.cfg, sections
are prefixed with "coverage:", so ``[run]`` options will be read from the
``[coverage:run]`` section of tox.ini. Implements part of `issue 519`_.
Thanks, Stephen Finucane.

- Specifying both ``--source`` and ``--include`` no longer silently ignores the
include setting, instead it fails with a message. Thanks, Nathan Land and
Loïc Dachary. Closes `issue 265`_.

- The ``Coverage.combine`` method has a new parameter, ``strict=False``, to
support failing if there are no data files to combine.

- When forking subprocesses, the coverage data files would have the same random
number appended to the file name. This didn't cause problems, because the
file names had the process id also, making collisions (nearly) impossible.
But it was disconcerting.  This is now fixed.

- The text report now properly sizes headers when skipping some files, fixing
`issue 524`_. Thanks, Anthony Sottile and Loïc Dachary.

- Coverage.py can now search .pex files for source, just as it can .zip and
.egg.  Thanks, Peter Ebden.

- Data files are now about 15% smaller.

- Improvements in the ``[run] debug`` setting:

- The "dataio" debug setting now also logs when data files are deleted during
 combining or erasing.

- A new debug option, "multiproc", for logging the behavior of
 ``concurrency=multiprocessing``.

- If you used the debug options "config" and "callers" together, you'd get a
 call stack printed for every line in the multi-line config output. This is
 now fixed.

- Fixed an unusual bug involving multiple coding declarations affecting code
containing code in multi-line strings: `issue 529`_.

- Coverage.py will no longer be misled into thinking that a plain file is a
package when interpreting ``--source`` options.  Thanks, Cosimo Lupo.

- If you try to run a non-Python file with coverage.py, you will now get a more
useful error message. `Issue 514`_.

- The default pragma regex changed slightly, but this will only matter to you
if you are deranged and use mixed-case pragmas.

- Deal properly with non-ASCII file names in an ASCII-only world, `issue 533`_.

- Programs that set Unicode configuration values could cause UnicodeErrors when
generating HTML reports.  Pytest-cov is one example.  This is now fixed.

- Prevented deprecation warnings from configparser that happened in some
circumstances, closing `issue 530`_.

- Corrected the name of the jquery.ba-throttle-debounce.js library. Thanks,
Ben Finney.  Closes `issue 505`_.

- Testing against PyPy 5.6 and PyPy3 5.5.

- Switched to pytest from nose for running the coverage.py tests.

- Renamed AUTHORS.txt to CONTRIBUTORS.txt, since there are other ways to
contribute than by writing code. Also put the count of contributors into the
author string in setup.py, though this might be too cute.

.. _sys.excepthook: https://docs.python.org/3/library/sys.htmlsys.excepthook
.. _issue 265: https://bitbucket.org/ned/coveragepy/issues/265/when-using-source-include-is-silently
.. _issue 412: https://bitbucket.org/ned/coveragepy/issues/412/coverage-combine-should-error-if-no
.. _issue 433: https://bitbucket.org/ned/coveragepy/issues/433/coverage-html-does-not-suport-skip-covered
.. _issue 493: https://bitbucket.org/ned/coveragepy/issues/493/confusing-branching-failure
.. _issue 496: https://bitbucket.org/ned/coveragepy/issues/496/incorrect-coverage-with-branching-and
.. _issue 502: https://bitbucket.org/ned/coveragepy/issues/502/incorrect-coverage-report-with-cover
.. _issue 505: https://bitbucket.org/ned/coveragepy/issues/505/use-canonical-filename-for-debounce
.. _issue 514: https://bitbucket.org/ned/coveragepy/issues/514/path-to-problem-file-not-reported-when
.. _issue 510: https://bitbucket.org/ned/coveragepy/issues/510/erase-still-needed-in-42
.. _issue 511: https://bitbucket.org/ned/coveragepy/issues/511/version-42-coverage-combine-empties
.. _issue 516: https://bitbucket.org/ned/coveragepy/issues/516/running-coverage-combine-twice-deletes-all
.. _issue 519: https://bitbucket.org/ned/coveragepy/issues/519/coverage-run-sections-in-toxini-or-as
.. _issue 524: https://bitbucket.org/ned/coveragepy/issues/524/coverage-report-with-skip-covered-column
.. _issue 525: https://bitbucket.org/ned/coveragepy/issues/525/coverage-combine-when-not-in-parallel-mode
.. _issue 529: https://bitbucket.org/ned/coveragepy/issues/529/encoding-marker-may-only-appear-on-the
.. _issue 530: https://bitbucket.org/ned/coveragepy/issues/530/deprecationwarning-you-passed-a-bytestring
.. _issue 533: https://bitbucket.org/ned/coveragepy/issues/533/exception-on-unencodable-file-name
.. _issue 535: https://bitbucket.org/ned/coveragepy/issues/535/sysexcepthook-is-not-called


.. _changes_42: