[ViewVC] Annotation of: cvs/liblzf/Changes


TODO: try unaligned copy again in decompressor
TODO: allow size-optimised binaries by avoiding unrolling
TODO: implement lzf_c_best in lzf.
TODO: fix lzf_c_best.

3.7 (unreleased)
        - add lzf_c_best.c, a slower but better compressor.
        - switch to a multiplicative hash (developed with Steinar Gunderson),
          which is faster on modern cpus and compresses a bit better. The old
          hash function which uses only shifts is still available.
        - allow user configurable hash table slots, which makes it possible
          to use e.g. 16 bit offsets for a smaller hashtable (if your data is
          always < 64kb).
        - use _WIN32, not WIN32, when testing for windows (fails with bcc),
          patch by Tamas Tevesz.
        - try to port to win64+gcc, which is non-posix but doesn't have _int64
          (reported by Maciej Adamczyk).

3.6  Mon Feb  7 17:37:31 CET 2011
        - fixed hash calculation in C♯ version (Tiago Freitas Leal).
        - unroll copy for small sizes, use memcpy for larger sizes,
          greatly speeding up decompression in most cases.
        - finally disable rep movsb - it's a big loss on modern intel cpus,
          and only a small win on amd cpus.
        - improve C++ compatibility of the code.
        - slightly improve compressor speed.
        - halved memory requirements for compressor on 64 bit architectures,
          which can improve the speed quite a bit on older cpus.

3.5  Fri May  1 02:28:42 CEST 2009
        - lzf_compress did sometimes write one octet past the given output
          buffer (analyzed and nice testcase by Salvatore Sanfilippo).

3.4  Tue Sep  2 06:45:00 CEST 2008
        - the fix from 3.3 introduced a compression bug, which is fixed in
          this release (which explains the mysterious prerelease...). Thanks
          once more to Clément Calmels.

3.3  Mon Aug 25 03:17:42 CEST 2008
        - lzf_compress could access memory after the given input buffer
          when outputting back references. reported with nice testcase
          by Clément Calmels.

3.2  Fri May  9 18:52:23 CEST 2008
        - include a workaround for failing POSIX and real-world compliance
          on 64 bit windows (microsoft claims to support POSIX, but is far
          from it). (bug found and analysed nicely by John Lilley).

3.1  Fri Nov 30 11:33:04 CET 2007
        - IMPORTANT BUGFIX: a too long final literal run would corrupt data
          in the encoder (this was introduced in 3.0 only, earlier versions
          are safe).

3.0  Tue Nov 13 22:13:09 CET 2007
        - switched to 2-clause bsd with "GPL v2 or any later version" option.
        - speed up compression by ~10-15% in common cases
          by some manual unrolling.
        - import some compiler tricks from JSON::XS, for further speed-ups.
        - tune hash functions depending on ULTRA_FAST or VERY_FAST settings.
        - for typical binary data (e.g. /bin/bash, memory dumps,
          canterbury corpus etc.), speed is now comparable to fastlz, but
          with better compression ratio. with ULTRA_FAST, it's typically
          3-15% faster than fastlz while still maintaining a similar ratio.
          (amd64 and core 2 duo, ymmv). thanks a lot for the competition :)
        - undo inline assembly in compressor, it is no longer helpful.
        - no changes to the decompressor.
        - use a HLOG of 16 by default now (formerly 15).

2.1  Fri Nov  2 13:34:42 CET 2007
        - switched to a 2-clause bsd license with GPL exception.
        - get rid of memcpy.
        - tentatively use rep movsb on x86 and x86_64 (gcc only) for a
          moderate speed improvement.
        - applied patch by Kein-Hong Man to maske lzf.c compile under
          the crippled mingw32 environment.

2.0  Fri Feb 16 23:11:18 CET 2007
        - replaced lzf demo by industrial-strength lzf utility with behaviour
          similar other compression utilities. Thanks for Stefan Traby for
          rewriting it!
        - fix state arg prototype.

1.7  Wed Sep 27 17:29:15 CEST 2006
        - remove bogus "unlzf" patch.
          note to self: never accept well-meant patches.
        - make lzf more robust in presence of padding bytes or sudden eof.

1.6  Fri Jul  7 17:31:26 CEST 2006
        - the lzf example utility will now uncompress if invoked
          as "unlzf" (patch by Scott Feeney).
        - add CHECK_INPUT option that adds more checks for input
          data validity.
        - help applications that do not pass in the correct length
          (such as php) by returning either EINVAL or E2BIG.
        - default HLOG size is now 15 (cpu caches have increased).
        - documentation fixes.

1.51 Thu Apr 14 22:15:46 CEST 2005
        - incorporated C♯ implementation of both the en- and decoder,
          written by "Oren J. Maurice".
          You can find it in the cs/ subdirectory.
        - make FRST, NEXT IDX overridable if lzf_c.c is directly included
          in the code.

1.5  Tue Mar  8 20:23:23 CET 2005
        - incorporated improvements by Adam D. Moss,
          which includes a new VERY_FAST mode which is
          a bit slower than ULTRA_FAST but much better,
          and enabled it as default.

1.401 Thu Mar  3 18:00:52 CET 2005
        - use cstring in c++, not string.h.
        - change of contact address.

1.4  Wed Dec 15 08:08:49 CET 2004
        - very very slight tuning of the hashing function.

1.3  Thu Mar 25 15:41:17 CET 2004
        - changed license of lzf core code to explicitly allow
          relicensing under the GPLv2.
        - added VPATH support as suggested by Björn Eriksson.

1.2  Mon Dec 29 13:47:28 CET 2003
        - avoid spurious memory accesses after the to-be-compressed
          memory region. originally reported by Michal Zalewski.
        - flip LZF_STACK_ARG meaning (to be correct).

1.1  Tue Dec 23 05:48:32 CET 2003
        - removed #warn directive, it's not worth the hassle.
        - add LZF_STACK_ARG and AVOID_ERRNO configurations
          for embedded systems.
        - make it compile cleanly as c++.
        - some small documentation and code fixes.

1.0  Sun Nov 17 12:37:37 CET 2002
        - slightly better compression ratio, almost unmeasurably
          slower.
        - some documentation fixes.

0.4  Thu Jun 13 14:11:10 CEST 2002
        - typoe fix.
        - lzf demo program now properly decompresses small files.
        - fix another 64 bit issue, found by Laurent Deniel.

0.3  Tue Jan 16 13:21:14 CET 2001
        - fix silly beginners 32/64 bit mistake.

0.2  Thu Jan  4 05:56:42 CET 2001
        - now totally independent of autoconfig, for
          easy inclusion into other programs.
        - much better fine-tuning, faster and better than 0.1.

0.1  2000
        - initial release.
Revision:	1.58
Committed:	Sun Aug 25 17:53:39 2013 UTC (10 years, 8 months ago) by root
Branch:	MAIN
Changes since 1.57:	+2 -0 lines
Log Message:	* empty log message *
#	User	Rev	Content
1	root	1.53
2	root	1.55	TODO: try unaligned copy again in decompressor
3			TODO: allow size-optimised binaries by avoiding unrolling
4	root	1.57	TODO: implement lzf_c_best in lzf.
5			TODO: fix lzf_c_best.
6	root	1.55
7	root	1.58	3.7 (unreleased)
8			- add lzf_c_best.c, a slower but better compressor.
9	root	1.54	- switch to a multiplicative hash (developed with Steinar Gunderson),
10			which is faster on modern cpus and compresses a bit better. The old
11			hash function which uses only shifts is still available.
12			- allow user configurable hash table slots, which makes it possible
13			to use e.g. 16 bit offsets for a smaller hashtable (if your data is
14			always < 64kb).
15	root	1.53	- use _WIN32, not WIN32, when testing for windows (fails with bcc),
16			patch by Tamas Tevesz.
17	root	1.57	- try to port to win64+gcc, which is non-posix but doesn't have _int64
18	root	1.56	(reported by Maciej Adamczyk).
19	root	1.53
20	root	1.52	3.6 Mon Feb 7 17:37:31 CET 2011
21	root	1.47	- fixed hash calculation in C♯ version (Tiago Freitas Leal).
22	root	1.48	- unroll copy for small sizes, use memcpy for larger sizes,
23			greatly speeding up decompression in most cases.
24			- finally disable rep movsb - it's a big loss on modern intel cpus,
25			and only a small win on amd cpus.
26	root	1.52	- improve C++ compatibility of the code.
27	root	1.50	- slightly improve compressor speed.
28	root	1.51	- halved memory requirements for compressor on 64 bit architectures,
29			which can improve the speed quite a bit on older cpus.
30	root	1.47
31	root	1.46	3.5 Fri May 1 02:28:42 CEST 2009
32			- lzf_compress did sometimes write one octet past the given output
33			buffer (analyzed and nice testcase by Salvatore Sanfilippo).
34
35	root	1.44	3.4 Tue Sep 2 06:45:00 CEST 2008
36			- the fix from 3.3 introduced a compression bug, which is fixed in
37	root	1.45	this release (which explains the mysterious prerelease...). Thanks
38	root	1.44	once more to Clément Calmels.
39
40	root	1.43	3.3 Mon Aug 25 03:17:42 CEST 2008
41			- lzf_compress could access memory after the given input buffer
42			when outputting back references. reported with nice testcase
43			by Clément Calmels.
44	root	1.41
45	root	1.43	3.2 Fri May 9 18:52:23 CEST 2008
46	root	1.42	- include a workaround for failing POSIX and real-world compliance
47			on 64 bit windows (microsoft claims to support POSIX, but is far
48			from it). (bug found and analysed nicely by John Lilley).
49	root	1.41
50	root	1.39	3.1 Fri Nov 30 11:33:04 CET 2007
51			- IMPORTANT BUGFIX: a too long final literal run would corrupt data
52	root	1.40	in the encoder (this was introduced in 3.0 only, earlier versions
53			are safe).
54	root	1.39
55	root	1.38	3.0 Tue Nov 13 22:13:09 CET 2007
56	root	1.37	- switched to 2-clause bsd with "GPL v2 or any later version" option.
57	root	1.33	- speed up compression by ~10-15% in common cases
58			by some manual unrolling.
59	root	1.37	- import some compiler tricks from JSON::XS, for further speed-ups.
60	root	1.33	- tune hash functions depending on ULTRA_FAST or VERY_FAST settings.
61			- for typical binary data (e.g. /bin/bash, memory dumps,
62			canterbury corpus etc.), speed is now comparable to fastlz, but
63			with better compression ratio. with ULTRA_FAST, it's typically
64	root	1.34	3-15% faster than fastlz while still maintaining a similar ratio.
65			(amd64 and core 2 duo, ymmv). thanks a lot for the competition :)
66	root	1.37	- undo inline assembly in compressor, it is no longer helpful.
67	root	1.35	- no changes to the decompressor.
68	root	1.36	- use a HLOG of 16 by default now (formerly 15).
69	root	1.32
70	root	1.30	2.1 Fri Nov 2 13:34:42 CET 2007
71	root	1.31	- switched to a 2-clause bsd license with GPL exception.
72	root	1.28	- get rid of memcpy.
73			- tentatively use rep movsb on x86 and x86_64 (gcc only) for a
74			moderate speed improvement.
75	root	1.29	- applied patch by Kein-Hong Man to maske lzf.c compile under
76			the crippled mingw32 environment.
77	root	1.18
78	root	1.27	2.0 Fri Feb 16 23:11:18 CET 2007
79	root	1.24	- replaced lzf demo by industrial-strength lzf utility with behaviour
80			similar other compression utilities. Thanks for Stefan Traby for
81			rewriting it!
82	root	1.27	- fix state arg prototype.
83	root	1.23
84	root	1.22	1.7 Wed Sep 27 17:29:15 CEST 2006
85	root	1.26	- remove bogus "unlzf" patch.
86	root	1.20	note to self: never accept well-meant patches.
87	root	1.21	- make lzf more robust in presence of padding bytes or sudden eof.
88	root	1.20
89	root	1.18	1.6 Fri Jul 7 17:31:26 CEST 2006
90			- the lzf example utility will now uncompress if invoked
91			as "unlzf" (patch by Scott Feeney).
92			- add CHECK_INPUT option that adds more checks for input
93			data validity.
94			- help applications that do not pass in the correct length
95			(such as php) by returning either EINVAL or E2BIG.
96	root	1.19	- default HLOG size is now 15 (cpu caches have increased).
97			- documentation fixes.
98	root	1.18
99	root	1.16	1.51 Thu Apr 14 22:15:46 CEST 2005
100			- incorporated C♯ implementation of both the en- and decoder,
101	root	1.43	written by "Oren J. Maurice".
102	root	1.17	You can find it in the cs/ subdirectory.
103	root	1.16	- make FRST, NEXT IDX overridable if lzf_c.c is directly included
104			in the code.
105
106	root	1.15	1.5 Tue Mar 8 20:23:23 CET 2005
107			- incorporated improvements by Adam D. Moss,
108			which includes a new VERY_FAST mode which is
109			a bit slower than ULTRA_FAST but much better,
110			and enabled it as default.
111	root	1.12
112	root	1.13	1.401 Thu Mar 3 18:00:52 CET 2005
113	root	1.12	- use cstring in c++, not string.h.
114	root	1.14	- change of contact address.
115	root	1.12
116	root	1.11	1.4 Wed Dec 15 08:08:49 CET 2004
117	root	1.10	- very very slight tuning of the hashing function.
118
119	pcg	1.9	1.3 Thu Mar 25 15:41:17 CET 2004
120			- changed license of lzf core code to explicitly allow
121			relicensing under the GPLv2.
122	pcg	1.8	- added VPATH support as suggested by Björn Eriksson.
123
124	pcg	1.6	1.2 Mon Dec 29 13:47:28 CET 2003
125			- avoid spurious memory accesses after the to-be-compressed
126			memory region. originally reported by Michal Zalewski.
127	pcg	1.7	- flip LZF_STACK_ARG meaning (to be correct).
128	pcg	1.6
129	pcg	1.5	1.1 Tue Dec 23 05:48:32 CET 2003
130	pcg	1.4	- removed #warn directive, it's not worth the hassle.
131	pcg	1.5	- add LZF_STACK_ARG and AVOID_ERRNO configurations
132			for embedded systems.
133			- make it compile cleanly as c++.
134			- some small documentation and code fixes.
135	pcg	1.4
136	root	1.3	1.0 Sun Nov 17 12:37:37 CET 2002
137	root	1.2	- slightly better compression ratio, almost unmeasurably
138			slower.
139	root	1.3	- some documentation fixes.
140	root	1.2
141			0.4 Thu Jun 13 14:11:10 CEST 2002
142	root	1.1	- typoe fix.
143			- lzf demo program now properly decompresses small files.
144			- fix another 64 bit issue, found by Laurent Deniel.
145
146			0.3 Tue Jan 16 13:21:14 CET 2001
147			- fix silly beginners 32/64 bit mistake.
148
149			0.2 Thu Jan 4 05:56:42 CET 2001
150	root	1.3	- now totally independent of autoconfig, for
151	root	1.1	easy inclusion into other programs.
152			- much better fine-tuning, faster and better than 0.1.
153
154	root	1.3	0.1 2000
155	root	1.1	- initial release.