ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/liblzf/Changes
Revision: 1.46
Committed: Fri May 1 00:30:53 2009 UTC (15 years ago) by root
Branch: MAIN
CVS Tags: rel-3_5
Changes since 1.45: +4 -0 lines
Log Message:
3.5

File Contents

# Content
1 3.5 Fri May 1 02:28:42 CEST 2009
2 - lzf_compress did sometimes write one octet past the given output
3 buffer (analyzed and nice testcase by Salvatore Sanfilippo).
4
5 3.4 Tue Sep 2 06:45:00 CEST 2008
6 - the fix from 3.3 introduced a compression bug, which is fixed in
7 this release (which explains the mysterious prerelease...). Thanks
8 once more to Clément Calmels.
9
10 3.3 Mon Aug 25 03:17:42 CEST 2008
11 - lzf_compress could access memory after the given input buffer
12 when outputting back references. reported with nice testcase
13 by Clément Calmels.
14
15 3.2 Fri May 9 18:52:23 CEST 2008
16 - include a workaround for failing POSIX and real-world compliance
17 on 64 bit windows (microsoft claims to support POSIX, but is far
18 from it). (bug found and analysed nicely by John Lilley).
19
20 3.1 Fri Nov 30 11:33:04 CET 2007
21 - IMPORTANT BUGFIX: a too long final literal run would corrupt data
22 in the encoder (this was introduced in 3.0 only, earlier versions
23 are safe).
24
25 3.0 Tue Nov 13 22:13:09 CET 2007
26 - switched to 2-clause bsd with "GPL v2 or any later version" option.
27 - speed up compression by ~10-15% in common cases
28 by some manual unrolling.
29 - import some compiler tricks from JSON::XS, for further speed-ups.
30 - tune hash functions depending on ULTRA_FAST or VERY_FAST settings.
31 - for typical binary data (e.g. /bin/bash, memory dumps,
32 canterbury corpus etc.), speed is now comparable to fastlz, but
33 with better compression ratio. with ULTRA_FAST, it's typically
34 3-15% faster than fastlz while still maintaining a similar ratio.
35 (amd64 and core 2 duo, ymmv). thanks a lot for the competition :)
36 - undo inline assembly in compressor, it is no longer helpful.
37 - no changes to the decompressor.
38 - use a HLOG of 16 by default now (formerly 15).
39
40 2.1 Fri Nov 2 13:34:42 CET 2007
41 - switched to a 2-clause bsd license with GPL exception.
42 - get rid of memcpy.
43 - tentatively use rep movsb on x86 and x86_64 (gcc only) for a
44 moderate speed improvement.
45 - applied patch by Kein-Hong Man to maske lzf.c compile under
46 the crippled mingw32 environment.
47
48 2.0 Fri Feb 16 23:11:18 CET 2007
49 - replaced lzf demo by industrial-strength lzf utility with behaviour
50 similar other compression utilities. Thanks for Stefan Traby for
51 rewriting it!
52 - fix state arg prototype.
53
54 1.7 Wed Sep 27 17:29:15 CEST 2006
55 - remove bogus "unlzf" patch.
56 note to self: never accept well-meant patches.
57 - make lzf more robust in presence of padding bytes or sudden eof.
58
59 1.6 Fri Jul 7 17:31:26 CEST 2006
60 - the lzf example utility will now uncompress if invoked
61 as "unlzf" (patch by Scott Feeney).
62 - add CHECK_INPUT option that adds more checks for input
63 data validity.
64 - help applications that do not pass in the correct length
65 (such as php) by returning either EINVAL or E2BIG.
66 - default HLOG size is now 15 (cpu caches have increased).
67 - documentation fixes.
68
69 1.51 Thu Apr 14 22:15:46 CEST 2005
70 - incorporated C♯ implementation of both the en- and decoder,
71 written by "Oren J. Maurice".
72 You can find it in the cs/ subdirectory.
73 - make FRST, NEXT IDX overridable if lzf_c.c is directly included
74 in the code.
75
76 1.5 Tue Mar 8 20:23:23 CET 2005
77 - incorporated improvements by Adam D. Moss,
78 which includes a new VERY_FAST mode which is
79 a bit slower than ULTRA_FAST but much better,
80 and enabled it as default.
81
82 1.401 Thu Mar 3 18:00:52 CET 2005
83 - use cstring in c++, not string.h.
84 - change of contact address.
85
86 1.4 Wed Dec 15 08:08:49 CET 2004
87 - very very slight tuning of the hashing function.
88
89 1.3 Thu Mar 25 15:41:17 CET 2004
90 - changed license of lzf core code to explicitly allow
91 relicensing under the GPLv2.
92 - added VPATH support as suggested by Björn Eriksson.
93
94 1.2 Mon Dec 29 13:47:28 CET 2003
95 - avoid spurious memory accesses after the to-be-compressed
96 memory region. originally reported by Michal Zalewski.
97 - flip LZF_STACK_ARG meaning (to be correct).
98
99 1.1 Tue Dec 23 05:48:32 CET 2003
100 - removed #warn directive, it's not worth the hassle.
101 - add LZF_STACK_ARG and AVOID_ERRNO configurations
102 for embedded systems.
103 - make it compile cleanly as c++.
104 - some small documentation and code fixes.
105
106 1.0 Sun Nov 17 12:37:37 CET 2002
107 - slightly better compression ratio, almost unmeasurably
108 slower.
109 - some documentation fixes.
110
111 0.4 Thu Jun 13 14:11:10 CEST 2002
112 - typoe fix.
113 - lzf demo program now properly decompresses small files.
114 - fix another 64 bit issue, found by Laurent Deniel.
115
116 0.3 Tue Jan 16 13:21:14 CET 2001
117 - fix silly beginners 32/64 bit mistake.
118
119 0.2 Thu Jan 4 05:56:42 CET 2001
120 - now totally independent of autoconfig, for
121 easy inclusion into other programs.
122 - much better fine-tuning, faster and better than 0.1.
123
124 0.1 2000
125 - initial release.