Enable PathLib tests under pypy-3.10

[pyxattr.git] / NEWS.md

# News

## Version 0.8.1

*Mon, 17 Apr 2023*

Very minor release:

* Fix the pypi declared python versions
* Add some more documentation - a security policy, and a contributing
  guide.
* Restore yet again the CI environment :/

## Version 0.8.0

*Mon, 12 Dec 2022*

Minor bugfix release, but due to lack of CI environments supporting old
Python versions, only Python 3.7+ is supported. Otherwise:

* Snape3058@ found some refcount issues (see issue #35), these have been
  fixed.
* Tested and enabled Python 3.10 support, dropped < 3.7.
* Various improvements to the CI environments.
* Move fully to `setuptools` (where available), in preparation for 3.12
  dropping `distutils` support.

## Version 0.7.2

*Sun, 29 Nov 2020*

Minor release:

* Expand testing by adding better mixed-access checks (e.g. set via
  symlink and read on file) and by not leaking resources during tests.
* Enable testing with Python 3.9 and confirm compatibility with it.
* Fix documentation building with Sphinx 3.0+.

## Version 0.7.1

*released Tue, 26 Nov 2019*

Typo fix release in the bug tracker link :/

## Version 0.7.0

*released Tue, 26 Nov 2019*

Major change: drop compatibility with Python 2, which allows significant
code cleanups.

Other changes:

* Switch internal implementation of argument parsing to a built-in one
  (`PyUnicode_FSConverter`), which brings automatic support for
  path-like objects in Python 3.6+ (#20), and also a more uniform
  handling of Unicode path arguments with respect to other Python code.
* Fix missing error check in list operations in `get_all` (#17).
* Switch test library to pytest; not that a reasonable recent version is
  needed. Additionally, expand test coverage, although not directly
  visible in actual coverage reports…

## Version 0.6.1

*released Tue, 24 Jul 2018*

Minor bugfix, performance and compatibility release.

* Minor compatibility fix: on Linux, drop the use of the `attr` library,
  and instead switch to the glibc header `sys/xattr.h`, which is
  provided for a really long time (since glibc 2.3). The formerly used
  header `attr/xattr.h` has been removed from the `attr` library in
  version 2.4.48. Fix provided by Lars Wendler, many thanks!
* Release the GIL when performing I/O. Patch proposed by xwhuang, many
  thanks. I tested this a long while back it seemed to impact
  performance on local filesystems, but upon further inspection, the
  downsides are minor (between 0 and 5%, in many cases negligible). For
  remote or slow filesystems, this should allow much increased
  parallelism.
* Fix symlink set operation on MacOS X; bugfix provided by adamlin, much
  appreciated! This also uncovered testing problems related to symlinks,
  which are now fixed (the bug would be caught by the updated tests).

## Version 0.6.0

*released Mon, 23 Jan 2017*

Bugfix and feature release (hence the version bump).

The main change is to the implementation of how attributes are listed
and read. This was done due to existing race issues when attributes are
modified while being read (github issue #12), but basically all various
internal paths that dealt with retrieving an attribute value or listing
attributes were unified in a single helper function that does handle
such concurrent modifications. As a side effect, the size of the buffers
used for such reads have changed, which (depending on attribute value)
might change the trade-off between number of syscalls done and memory
usage.

As feature release, OSX support was contributed by Adam Knight
<adam@movq.us>, thanks a lot! I don't have access to OSX so the testing
for it is done via Travis builds; please report any issues.

## Version 0.5.6

*released Sat, 09 Apr 2016*

Small bugfix release:

* Fixes some sign-compare warnings
* Fixes potential name truncation in merge_ns()
* Fixes building on systems which don't have ENODATA

Tested with Python 2.7.11, Python 3.5.1 and PyPy 5.0.1.

## Version 0.5.5

*released Fri, 01 May 2015*

Bugfix release:

* fixes some more memory leaks when handling out-of-memory in get_all()
  function
* improve error reporting when an attribute disappears after we asked
  for its length but before we managed to read it
* fix int/size_t issues found by RedHat/Fedora,
  https://bugzilla.redhat.com/show_bug.cgi?id=1127310; the fix is
  different than their fix, but it should accomplish the same thing
* convert all code to only do explicit casts after checking boundaries,
  making the code `-Wconversion`-clean (although that warning is not
  enabled by default)

## Version 0.5.4

*released Thu, 30 Apr 2015*

Fix memory leaks on some of the error-handling paths of the `get()`
function.

## Version 0.5.3

*released Fri, 23 May 2014*

Small optimisations release:

* ari edelkind contributed a speed-up optimisation for handling of files
  without xattrs (which is, in general, the expected case)
* Jonas Borgström contributed a behaviour change to the handling of file
  names: under Python 3 and up, unicode paths are encoded/decoded using
  the 'surogatee' handler, instead of the 'strict' handler; while this
  can hide encoding errors, it mirrors what Python libraries do
  (e.g. see os.fsencode/fsdecode)
* Sean Patrick Santos contributed improvements to the test suite so that
  it can be used even on files systems which have built-in attributes
  (e.g. when using SELinux, or NFSv4); to enable this, define the
  attributes in the TEST_IGNORE_XATTRS environment variable

## Version 0.5.2

*released Thu, 03 Jan 2013*

Bug-fix release. Thanks to Michał Górny, it looked like the library had
problem running under pypy, but actually there was a bug in the
PyArg_ParseTuple use of et# (signed vs. unsigned, and lack of compiler
warnings). This was fixed, and now the test suite passed with many
CPython versions and PyPy (version 1.9).

## Version 0.5.1

*released Wed, 16 May 2012*

Bug-fix release. Thanks to Dave Malcolm and his cpychecker tool, a
number of significant bugs (refcount leaks and potential NULL-pointer
dereferences) have been fixed.

Furthermore, compatibility with Python 3 has been improved; this however
required changing the meaning of the ``namespace`` argument to the
functions: if passed, None is no longer a valid value; pass an empty
string if (due to the structure of your program) you have to pass this
argument but want to specify no namespace.

Also, the project home page has changed from SourceForge to GitHub, and
the documentation has been converted from epydoc-based to sphinx.


## Version 0.5

*released Sun, 27 Dec 2009*

Implemented support for Python 3. This required a significant change to
the C module, hence the new version number.

## Version 0.4

*released Mon, 30 Jun 2008*

### API

The old functions ({get,set,list,remove}xattr) are deprecated and replaced with
a new API that is namespace-aware and hopefully will allow other OSes (e.g.
FreeBSD) to be supported more naturally.

Both the old and the new API are supported in the 0.4 versions, however users
are encouraged to migrate to the new API.

### New features

A new bulk get function called get_all() has been added that should be somewhat
faster in case of querying files which have many attributes.

### License

Since LGPLv3 is not compatible with GPLv2 (which unfortunately I didn't realize
before), the license was changed to LGPLv2.1 or later.

### Internals

Unittest coverage was improved.

## Version 0.3

*released Sun, 09 Mar 2008*

* changed licence from GPL to LGPL (3 or later)
* changed listxattr return type from tuple to a list
* developer-related: added unittests

## Version 0.2.2

*released Sun, 01 Jul 2007*

* fixed listing symlink xattrs

## Version 0.2.1

*released Sat, 11 Feb 2006*

* fixed a bug when reading symlink EAs (you weren't able to
  do it, actually)
* fixed a possible memory leak when the actual read of the EA
  failed but the call to get the length of the EA didn't