github.com/grailbio/base@v0.0.11/compress/libdeflate/NEWS (about)

     1  Version 1.0:
     2  	Added support for multi-member gzip files.
     3  
     4  	Moved architecture-specific code into subdirectories.  If you
     5  	aren't using the provided Makefile to build libdeflate, you now
     6  	need to compile lib/*.c and lib/*/*.c instead of just lib/*.c.
     7  
     8  	Added an ARM PMULL implementation of CRC-32, which speeds up
     9  	gzip compression and decompression on 32-bit and 64-bit ARM
    10  	processors that have the Cryptography Extensions.
    11  
    12  	Improved detection of CPU features, resulting in accelerated
    13  	functions being used in more cases.  This includes:
    14  
    15  	   - Detect CPU features on 32-bit x86, not just 64-bit as was
    16  	     done previously.
    17  
    18  	   - Detect CPU features on ARM, both 32 and 64-bit.
    19  	     (Limited to Linux only currently.)
    20  
    21  Version 0.8:
    22  	Build fixes for certain platforms and compilers.
    23  
    24  	libdeflate now produces the same output on all CPU architectures.
    25  
    26  	Improved documentation for building libdeflate on Windows.
    27  
    28  Version 0.7:
    29  	Fixed a very rare bug that caused data to be compressed incorrectly.
    30  	The bug affected compression levels 7 and below since libdeflate v0.2.
    31  	Although there have been no user reports of the bug, and I believe it
    32  	would have been highly unlikely to encounter on realistic data, it could
    33  	occur on data specially crafted to reproduce it.
    34  
    35  	Fixed a compilation error when building with clang 3.7.
    36  
    37  Version 0.6:
    38  	Various improvements to the gzip program's behavior.
    39  
    40  	Faster CRC-32 on AVX-capable processors.
    41  
    42  	Other minor changes.
    43  
    44  Version 0.5:
    45  	The CRC-32 checksum algorithm has been optimized with carryless
    46  	multiplication instructions for x86_64 (PCLMUL).  This speeds up gzip
    47  	compression and decompression.
    48  
    49  	Build fixes for certain platforms and compilers.
    50  
    51  	Added more test programs and scripts.
    52  
    53  	libdeflate is now entirely MIT-licensed.
    54  
    55  Version 0.4:
    56  	The Adler-32 checksum algorithm has been optimized with vector
    57  	instructions for x86_64 (SSE2 and AVX2) and ARM (NEON).  This speeds up
    58  	zlib compression and decompression.
    59  
    60  	To avoid naming collisions, functions and definitions in libdeflate's
    61  	API have been renamed to be prefixed with "libdeflate_" or
    62  	"LIBDEFLATE_".  Programs using the old API will need to be updated.
    63  
    64  	Various bug fixes and other improvements.
    65  
    66  Version 0.3:
    67  	Some bug fixes and other minor changes.
    68  
    69  Version 0.2:
    70  	Implemented a new block splitting algorithm which typically improves the
    71  	compression ratio slightly at all compression levels.
    72  
    73  	The compressor now outputs each block using the cheapest type (dynamic
    74  	Huffman, static Huffman, or uncompressed).
    75  
    76  	The gzip program has received an overhaul and now behaves more like the
    77  	standard version.
    78  
    79  	Build system updates, including: some build options were changed and
    80  	some build options were removed, and the default 'make' target now
    81  	includes the gzip program as well as the library.
    82  
    83  Version 0.1:
    84  	Initial official release.