|
|
1 | |
|
|
2 | - fixed hash calculation in C♯ version (Tiago Freitas Leal). |
|
|
3 | - unroll copy for small sizes, use memcpy for larger sizes, |
|
|
4 | greatly speeding up decompression in most cases. |
|
|
5 | - finally disable rep movsb - it's a big loss on modern intel cpus, |
|
|
6 | and only a small win on amd cpus. |
|
|
7 | - improve c++ compatibility of the code. |
|
|
8 | - slightly improve compressor speed. |
|
|
9 | |
|
|
10 | 3.5 Fri May 1 02:28:42 CEST 2009 |
|
|
11 | - lzf_compress did sometimes write one octet past the given output |
|
|
12 | buffer (analyzed and nice testcase by Salvatore Sanfilippo). |
|
|
13 | |
|
|
14 | 3.4 Tue Sep 2 06:45:00 CEST 2008 |
|
|
15 | - the fix from 3.3 introduced a compression bug, which is fixed in |
|
|
16 | this release (which explains the mysterious prerelease...). Thanks |
|
|
17 | once more to Clément Calmels. |
|
|
18 | |
|
|
19 | 3.3 Mon Aug 25 03:17:42 CEST 2008 |
|
|
20 | - lzf_compress could access memory after the given input buffer |
|
|
21 | when outputting back references. reported with nice testcase |
|
|
22 | by Clément Calmels. |
|
|
23 | |
|
|
24 | 3.2 Fri May 9 18:52:23 CEST 2008 |
|
|
25 | - include a workaround for failing POSIX and real-world compliance |
|
|
26 | on 64 bit windows (microsoft claims to support POSIX, but is far |
|
|
27 | from it). (bug found and analysed nicely by John Lilley). |
|
|
28 | |
|
|
29 | 3.1 Fri Nov 30 11:33:04 CET 2007 |
|
|
30 | - IMPORTANT BUGFIX: a too long final literal run would corrupt data |
|
|
31 | in the encoder (this was introduced in 3.0 only, earlier versions |
|
|
32 | are safe). |
|
|
33 | |
|
|
34 | 3.0 Tue Nov 13 22:13:09 CET 2007 |
|
|
35 | - switched to 2-clause bsd with "GPL v2 or any later version" option. |
|
|
36 | - speed up compression by ~10-15% in common cases |
|
|
37 | by some manual unrolling. |
|
|
38 | - import some compiler tricks from JSON::XS, for further speed-ups. |
|
|
39 | - tune hash functions depending on ULTRA_FAST or VERY_FAST settings. |
|
|
40 | - for typical binary data (e.g. /bin/bash, memory dumps, |
|
|
41 | canterbury corpus etc.), speed is now comparable to fastlz, but |
|
|
42 | with better compression ratio. with ULTRA_FAST, it's typically |
|
|
43 | 3-15% faster than fastlz while still maintaining a similar ratio. |
|
|
44 | (amd64 and core 2 duo, ymmv). thanks a lot for the competition :) |
|
|
45 | - undo inline assembly in compressor, it is no longer helpful. |
|
|
46 | - no changes to the decompressor. |
|
|
47 | - use a HLOG of 16 by default now (formerly 15). |
|
|
48 | |
| 1 | 2.1 Fri Nov 2 13:34:42 CET 2007 |
49 | 2.1 Fri Nov 2 13:34:42 CET 2007 |
| 2 | - switched to a 2-clause bsd license with GPL exception. |
50 | - switched to a 2-clause bsd license with GPL exception. |
| 3 | - get rid of memcpy. |
51 | - get rid of memcpy. |
| 4 | - tentatively use rep movsb on x86 and x86_64 (gcc only) for a |
52 | - tentatively use rep movsb on x86 and x86_64 (gcc only) for a |
| 5 | moderate speed improvement. |
53 | moderate speed improvement. |
| … | |
… | |
| 27 | - default HLOG size is now 15 (cpu caches have increased). |
75 | - default HLOG size is now 15 (cpu caches have increased). |
| 28 | - documentation fixes. |
76 | - documentation fixes. |
| 29 | |
77 | |
| 30 | 1.51 Thu Apr 14 22:15:46 CEST 2005 |
78 | 1.51 Thu Apr 14 22:15:46 CEST 2005 |
| 31 | - incorporated C♯ implementation of both the en- and decoder, |
79 | - incorporated C♯ implementation of both the en- and decoder, |
| 32 | written by "Oren J. Maurice <oymaurice@hazorea.org.il>". |
80 | written by "Oren J. Maurice". |
| 33 | You can find it in the cs/ subdirectory. |
81 | You can find it in the cs/ subdirectory. |
| 34 | - make FRST, NEXT IDX overridable if lzf_c.c is directly included |
82 | - make FRST, NEXT IDX overridable if lzf_c.c is directly included |
| 35 | in the code. |
83 | in the code. |
| 36 | |
84 | |
| 37 | 1.5 Tue Mar 8 20:23:23 CET 2005 |
85 | 1.5 Tue Mar 8 20:23:23 CET 2005 |
