File python-blosc.changes of Package python-blosc

Thu Aug 20 18:37:32 UTC 2020 - Bernhard Wiedemann <>

- Disable AVX2 to avoid compile-time CPU-detection (boo#1100677)

Fri Jul 17 07:17:58 UTC 2020 - Guillaume GARDET <>

- Enable build on all archs

Wed Jul  8 09:37:15 UTC 2020 - Marketa Calabkova <>

- Update to 1.9.1
  * Disable the attempt to include support for SSE2 and AVX2 on non-Intel 
    platforms, allowing the build on such platforms (see #244).
  * Dropped support for Python 2.7 and 3.5.
  * Fixed the copy of the leftovers of a chunk when its size is not 
    a multiple of the typesize.

Thu Mar 26 14:20:07 UTC 2020 - Paolo Stivanin <>

- Use unittest instead of nose 

Thu Jan 30 16:30:20 UTC 2020 - Todd R <>

- Update to 1.8.3
  * Add a missing pyproject.toml to This allows
    to install the necessary skbuild module.
- Update to 1.8.2
  * Use cmake internally to build the Python extension via the
    scikit-build library. This seems enough to cope with the
    conflicting types in using XGETBV when using a recent GCC
    (>= 9.1) compiler.
  * Include C-Blosc v1.17.1.

Mon Jul 22 16:25:44 UTC 2019 - Todd R <>

- Update to 1.8.1
  * Fix a bug that prevented the source distribution from PyPi to
    be compiled. Specifcally, *.inc were not included via the
- Update to 1.8.0
  * Include C-Blosc v1.16.2
  * Fix usage on Windows.
  * Implement Python access to the C function cbuffer_validate which
    was added to c-blosc in version 1.16.0
  * Check if compiler supports CPU capabilities.
  * Many minor improvements and fixes
- Update to 1.7.0
  * Include C-Blosc v1.15.1
  * Remove Support for Python 2.6 and 3.3
  * Fix vendored
  * Rework Windows CI via Appveyor
  * Various minor bugfixes

Sat Dec 22 02:14:28 UTC 2018 - Todd R <>

- Update to 1.6.2
  * Fixed `numpy` import in ``, this makes python-blosc usable without
  numpy once again.
- Update to 1.6.1
  * Updated C-Blosc to 1.14.3
  * Adding support for aarch64
  * `unpack_array` can now accepts keyword arguments. This allows decompressing
    pickled arrays that were compressed with Python 2.x using Python 3.x. Thanks
    to Simba Nyatsanga and Juan Maree.
  * Implemented `get_blocksize()`, thanks to Alberto Sabater
  * Seperate compilation of codecs, thanks to Robert McLeod
  * Removal of Numpy dependencies, thanks to Robert McLeod
  * Allow codecs to be included by using environment variables, thanks to
    Robert McLeod
  * Don't compile snappy by default, thanks to Robert McLeod
  * Update to use dmesg.boot, thanks to Prakhar Goel
  * Allow SSE2 and AVX2 detection to be disabled via environment variables,
    thanks to Lorenzo Bolla
  * Varios minor fixes

Wed Jul 18 11:54:58 UTC 2018 -

- Stricten requirement on the blosc-devel to require version we
  actually compile with

Thu May 24 16:51:14 UTC 2018 -

- Python files were not getting included.
- Enable unit tests

Sun Jan 21 05:33:31 UTC 2018 -

- Fix building

Wed Apr 19 22:13:57 UTC 2017 -

- Source url must be https.

Wed Apr 19 18:43:04 UTC 2017 -

- Update to version 1.5.1
  * License updated from MIT to BSD.
  * Updated to C-Blosc 1.11.3.
- Update to version 1.5.0
  * Added a new `blosc.set_releasegil()` function that allows to
    release/acquire the GIL at will.  See PR #116.
    Thanks to Robert McLeod.
  * Updated to C-Blosc 1.11.2.
  * Added tests that detect possible memory leaks.
    Thanks to Robert McLeod.
- Update to version 1.4.4
  * Updated to C-Blosc 1.11.1.  Fixes #115.
- Update to version 1.4.3
  * Internal C-Blosc sources updated to 1.11.0. Among other things, this
    updates the internal Zstd codec to version 1.0.0 (i.e. it is
    officially apt for production usage!).
- Update to version 1.4.1
  * Internal C-Blosc sources updated to 1.10.1.  This fixes an outstanding issue
    with the clang compiler.  For details, see:
- Update to version 1.4.0
  * Internal C-Blosc sources updated to 1.10.0.
  * Benchmarks updated for a Skylake processor (Xeon E3-1245 v5 @
- Update to version 1.3.3
  * Internal C-Blosc sources updated to 1.9.3.
  * C-Blosc do not segfaults anymore, so -O1 flag on Linux is not the
    default anymore.
  * SSE2 and AVX2 are now auto-discovered so the internal C-Blosc will be
    compiled with maximum optimization on processors supporting them.
- Update to version 1.3.2
  * Fixed the version of the include C-Blosc library (should be 1.8.1 not
- Update to version 1.3.1
  * Use the -O1 flag for compiling the included C-Blosc sources on Linux.
    This represents slower performance, but fixes the nasty issue #110.
    Also, it prints a warning for using an external C-Blosc library.
  * Internal C-Blosc version bumped to 1.8.1 for better compatibility
    with gcc 5.3.1 in forthcoming Ubuntu Xenial.
  * Added a protection to avoid using BITSHUFLE with C-Blosc < 1.8.0.
  * Restored old symbols for backward compatibility with pre 1.3.0:
    However, these are considered deprecated and should be replaced by
    libraries using python-blosc by the ones without the BLOSC_ prefix.
- Update to version 1.3.0
  * Internal C-Blosc version bumped to 1.8.0.  As consequence, support for
    BITSHUFFLE is here.  For activating it, just pass `blosc.BITSHUFFLE`
    to the `shuffle` parameter of compression functions.
  * Added a new `as_bytearray=False` parameter to the `decompress()` function
    so that a mutable bytearray will be returned instead of a bytes one
    (inmutable).  PR #107.  Thanks to Joe Jevnik.
  * The '__all__' variable has been removed from the module.  I consider
    this good practice to avoid things like "from blosc import *".
  * For consistency, the next symbols have been renamed:
  * The `typesize` parameter is set by default to 8 in compression
    functions.  This usually behaves well for 4-bytes typesizes too.
    Nevertheless, it is advised to use the actual typesize.
  * The maximum number of threads to use by default is set to 4 (less if
    less cores are detected).  Feel free to use more or less threads
    depending on the resources you want to use for compression.
- Implement single-spec version.

Thu Mar 10 19:48:59 UTC 2016 -

- Update to version 1.2.8
  * Updated to c-blosc v1.7.0. However, the new bitshuffle filter
    has not been made public because recent reports indicate that
    it seems too green for production.
  * Support bytes-like objects that support the buffer interface 
    as input to compress and decompress. On Python 2.x this
    includes unicode, on Python 3.x it doesn't. (#80 #94 @esc)
  * Fix a memory leak in decompress. Added tests to catch memory
    leaks. (#102 #103 #104 @sdvillal)
  * Various miscellaneous fixes and improvements.

Tue May 19 12:15:06 UTC 2015 -

- Update to version 1.2.7
  * Updated to c-blosc v1.6.1.

Sun Apr 26 14:22:04 UTC 2015 -

- update to version 1.2.5:
  * Updated to c-blosc v1.5.4.
  * Added wrapper for the expert function ``set_blocksize``.
    (#72 @esc)
  * Fix to allow compilation on posix architectures
    without SSE2. (#70 @andreas-schwab)
  * Don't release the GIL on compression/decompression (#77 @esc)
  * Various miscellaneous fixes.
- drop setup.patch: fixed upstream

Wed Aug  6 15:16:57 UTC 2014 -

- setup.patch: fix use of unknown compiler option

Thu Jul 17 09:32:38 UTC 2014 -

- Update to 1.2.4
  - Updated to c-blosc 1.4.0.  This added support for non-Intel
    architectures, most specially those not supporting unaligned access.

Thu May  8 12:14:34 UTC 2014 -

- Update to 1.2.3
  - Updated to c-blosc 1.3.5.  This removed a 'pointer from integer
    without a cast' compiler warning due to a bad macro definition.
- Update to 1.2.2
  - Updated to c-blosc 1.3.4.  This fixed a false buffer overrun
    condition.  This bug made c-blosc (and hence python-blosc) to fail,
    even if the failure was not real.
- Update to 1.2.1
  - Updated to c-blosc 1.3.3.
  - Added a new `cname2clib` map for programatically determine the library
    associated to a compressor.
  - New `get_clib(cbuffer)` that tells which compression library format
    has been used to created the compressed `cbuffer`.
- Update to 1.2.0
  - This release adds support for the multiple compressors added in Blosc
    1.3 series.
  - Added new `cname` parameter in compression functions like
    `compress()`, `compress_ptr()` and `pack_array()`.
  - Added a new utility function named `compressor_list()` that returns
    the list of compressors supported in the Blosc build.
  - Added 'bench/' for comparing times of the different
    compressors in Blosc and NumPy.
- Update to 1.1.0
  - Added new `compress_ptr` and `decompress_ptr` functions that allows to
    compress and decompress from/to a data pointer.  These are low level
    calls and user must make sure that the pointer data area is safe.
  - Since Blosc (the C library) already supports to be installed as an
    standalone library (via cmake), it is also possible to link
    python-blosc against a system Blosc library.
  - The Python calls to Blosc are now thread-safe (another consequence of
    recent Blosc library supporting this at C level).
  - Many checks on types and ranges of values have been added.  Most of
    the calls will now complain when passed the wrong values.
  - Docstrings are much improved. Also, Sphinx-based docs are available
- Update to 1.0.6
  - Fix compile error with msvc compilers.  Thanks to Christoph Gohlke.
- Update to 1.0.5
  - Upgraded to latest Blosc 1.1.4.
  - Better handling of condition errors, and improved memory releasing in
    case of errors (thanks to Valentin Haenel and Han Genuit).
  - Better handling of types (should compile without warning now, at least
    with GCC).
- Update to 1.0.4
  - Optimized the amount of data copied during compression (using
    _PyBytes_Resize() now instead of old PyBytes_FromStringAndSize()).
    This leads to improvements in compression speed ranging from 1.2x for
    highly compressible chunks up to 7x for mostly uncompressible data.
    Thanks to Valentin Haenel for this nice contribution.
- Use external blosc library.  The internal one builds a bad version
  of zlib.

Fri Sep  7 14:30:54 UTC 2012 -

- Initial version