lz4-ruby 0.1.6-x86-mingw32
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/.document +5 -0
- data/Gemfile +14 -0
- data/LICENSE.txt +20 -0
- data/README.rdoc +76 -0
- data/Rakefile +97 -0
- data/VERSION +1 -0
- data/benchmarking/bench.rb +137 -0
- data/benchmarking/bench_all.sh +5 -0
- data/benchmarking/compressor.rb +71 -0
- data/benchmarking/compressor_lz4.rb +25 -0
- data/benchmarking/compressor_lzo.rb +25 -0
- data/benchmarking/compressor_snappy.rb +25 -0
- data/ext/lz4ruby/extconf.rb +6 -0
- data/ext/lz4ruby/lz4.c +819 -0
- data/ext/lz4ruby/lz4.h +120 -0
- data/ext/lz4ruby/lz4hc.c +663 -0
- data/ext/lz4ruby/lz4hc.h +60 -0
- data/ext/lz4ruby/lz4ruby.c +125 -0
- data/lib/1.8/lz4ruby.so +0 -0
- data/lib/1.9/lz4ruby.so +0 -0
- data/lib/lz4-ruby.rb +23 -0
- data/test/helper.rb +28 -0
- data/test/test_lz4-ruby.rb +87 -0
- metadata +162 -0
data/ext/lz4ruby/lz4.c
ADDED
@@ -0,0 +1,819 @@
|
|
1
|
+
/*
|
2
|
+
LZ4 - Fast LZ compression algorithm
|
3
|
+
Copyright (C) 2011-2012, Yann Collet.
|
4
|
+
BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
|
5
|
+
|
6
|
+
Redistribution and use in source and binary forms, with or without
|
7
|
+
modification, are permitted provided that the following conditions are
|
8
|
+
met:
|
9
|
+
|
10
|
+
* Redistributions of source code must retain the above copyright
|
11
|
+
notice, this list of conditions and the following disclaimer.
|
12
|
+
* Redistributions in binary form must reproduce the above
|
13
|
+
copyright notice, this list of conditions and the following disclaimer
|
14
|
+
in the documentation and/or other materials provided with the
|
15
|
+
distribution.
|
16
|
+
|
17
|
+
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
18
|
+
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
|
19
|
+
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
|
20
|
+
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
|
21
|
+
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
22
|
+
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
|
23
|
+
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
|
24
|
+
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
|
25
|
+
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
|
26
|
+
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
|
27
|
+
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
28
|
+
|
29
|
+
You can contact the author at :
|
30
|
+
- LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
|
31
|
+
- LZ4 source repository : http://code.google.com/p/lz4/
|
32
|
+
*/
|
33
|
+
|
34
|
+
//**************************************
|
35
|
+
// Tuning parameters
|
36
|
+
//**************************************
|
37
|
+
// COMPRESSIONLEVEL :
|
38
|
+
// Increasing this value improves compression ratio
|
39
|
+
// Lowering this value reduces memory usage
|
40
|
+
// Reduced memory usage typically improves speed, due to cache effect (ex : L1 32KB for Intel, L1 64KB for AMD)
|
41
|
+
// Memory usage formula : N->2^(N+2) Bytes (examples : 12 -> 16KB ; 17 -> 512KB)
|
42
|
+
#define COMPRESSIONLEVEL 12
|
43
|
+
|
44
|
+
// NOTCOMPRESSIBLE_CONFIRMATION :
|
45
|
+
// Decreasing this value will make the algorithm skip faster data segments considered "incompressible"
|
46
|
+
// This may decrease compression ratio dramatically, but will be faster on incompressible data
|
47
|
+
// Increasing this value will make the algorithm search more before declaring a segment "incompressible"
|
48
|
+
// This could improve compression a bit, but will be slower on incompressible data
|
49
|
+
// The default value (6) is recommended
|
50
|
+
#define NOTCOMPRESSIBLE_CONFIRMATION 6
|
51
|
+
|
52
|
+
// LZ4_COMPRESSMIN :
|
53
|
+
// Compression function will *fail* if it is not successful at compressing input by at least LZ4_COMPRESSMIN bytes
|
54
|
+
// Since the compression function stops working prematurely, it results in a speed gain
|
55
|
+
// The output however is unusable. Compression function result will be zero.
|
56
|
+
// Default : 0 = disabled
|
57
|
+
#define LZ4_COMPRESSMIN 0
|
58
|
+
|
59
|
+
// BIG_ENDIAN_NATIVE_BUT_INCOMPATIBLE :
|
60
|
+
// This will provide a boost to performance for big endian cpu, but the resulting compressed stream will be incompatible with little-endian CPU.
|
61
|
+
// You can set this option to 1 in situations where data will stay within closed environment
|
62
|
+
// This option is useless on Little_Endian CPU (such as x86)
|
63
|
+
//#define BIG_ENDIAN_NATIVE_BUT_INCOMPATIBLE 1
|
64
|
+
|
65
|
+
|
66
|
+
|
67
|
+
//**************************************
|
68
|
+
// CPU Feature Detection
|
69
|
+
//**************************************
|
70
|
+
// 32 or 64 bits ?
|
71
|
+
#if (defined(__x86_64__) || defined(__x86_64) || defined(__amd64__) || defined(__amd64) || defined(__ppc64__) || defined(_WIN64) || defined(__LP64__) || defined(_LP64) ) // Detects 64 bits mode
|
72
|
+
# define LZ4_ARCH64 1
|
73
|
+
#else
|
74
|
+
# define LZ4_ARCH64 0
|
75
|
+
#endif
|
76
|
+
|
77
|
+
// Little Endian or Big Endian ?
|
78
|
+
// Note : overwrite the below #define if you know your architecture endianess
|
79
|
+
#if (defined(__BIG_ENDIAN__) || defined(__BIG_ENDIAN) || defined(_BIG_ENDIAN) || defined(_ARCH_PPC) || defined(__PPC__) || defined(__PPC) || defined(PPC) || defined(__powerpc__) || defined(__powerpc) || defined(powerpc) || ((defined(__BYTE_ORDER__)&&(__BYTE_ORDER__ == __ORDER_BIG_ENDIAN__))) )
|
80
|
+
# define LZ4_BIG_ENDIAN 1
|
81
|
+
#else
|
82
|
+
// Little Endian assumed. PDP Endian and other very rare endian format are unsupported.
|
83
|
+
#endif
|
84
|
+
|
85
|
+
// Unaligned memory access is automatically enabled for "common" CPU, such as x86.
|
86
|
+
// For others CPU, the compiler will be more cautious, and insert extra code to ensure aligned access is respected
|
87
|
+
// If you know your target CPU supports unaligned memory access, you may want to force this option manually to improve performance
|
88
|
+
#if defined(__ARM_FEATURE_UNALIGNED)
|
89
|
+
# define LZ4_FORCE_UNALIGNED_ACCESS 1
|
90
|
+
#endif
|
91
|
+
|
92
|
+
// Define this parameter if your target system or compiler does not support hardware bit count
|
93
|
+
#if defined(_MSC_VER) && defined(_WIN32_WCE) // Visual Studio for Windows CE does not support Hardware bit count
|
94
|
+
# define LZ4_FORCE_SW_BITCOUNT
|
95
|
+
#endif
|
96
|
+
|
97
|
+
|
98
|
+
//**************************************
|
99
|
+
// Compiler Options
|
100
|
+
//**************************************
|
101
|
+
#if __STDC_VERSION__ >= 199901L // C99
|
102
|
+
/* "restrict" is a known keyword */
|
103
|
+
#else
|
104
|
+
# define restrict // Disable restrict
|
105
|
+
#endif
|
106
|
+
|
107
|
+
#define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
|
108
|
+
|
109
|
+
#ifdef _MSC_VER // Visual Studio
|
110
|
+
# define inline __forceinline // Visual is not C99, but supports some kind of inline
|
111
|
+
# if LZ4_ARCH64 // 64-bit
|
112
|
+
# pragma intrinsic(_BitScanForward64) // For Visual 2005
|
113
|
+
# pragma intrinsic(_BitScanReverse64) // For Visual 2005
|
114
|
+
# else
|
115
|
+
# pragma intrinsic(_BitScanForward) // For Visual 2005
|
116
|
+
# pragma intrinsic(_BitScanReverse) // For Visual 2005
|
117
|
+
# endif
|
118
|
+
#endif
|
119
|
+
|
120
|
+
#ifdef _MSC_VER
|
121
|
+
# define lz4_bswap16(x) _byteswap_ushort(x)
|
122
|
+
#else
|
123
|
+
# define lz4_bswap16(x) ((unsigned short int) ((((x) >> 8) & 0xffu) | (((x) & 0xffu) << 8)))
|
124
|
+
#endif
|
125
|
+
|
126
|
+
#if (GCC_VERSION >= 302) || (__INTEL_COMPILER >= 800) || defined(__clang__)
|
127
|
+
# define expect(expr,value) (__builtin_expect ((expr),(value)) )
|
128
|
+
#else
|
129
|
+
# define expect(expr,value) (expr)
|
130
|
+
#endif
|
131
|
+
|
132
|
+
#define likely(expr) expect((expr) != 0, 1)
|
133
|
+
#define unlikely(expr) expect((expr) != 0, 0)
|
134
|
+
|
135
|
+
|
136
|
+
//**************************************
|
137
|
+
// Includes
|
138
|
+
//**************************************
|
139
|
+
#include <stdlib.h> // for malloc
|
140
|
+
#include <string.h> // for memset
|
141
|
+
#include "lz4.h"
|
142
|
+
|
143
|
+
|
144
|
+
//**************************************
|
145
|
+
// Basic Types
|
146
|
+
//**************************************
|
147
|
+
#if defined(_MSC_VER) // Visual Studio does not support 'stdint' natively
|
148
|
+
# define BYTE unsigned __int8
|
149
|
+
# define U16 unsigned __int16
|
150
|
+
# define U32 unsigned __int32
|
151
|
+
# define S32 __int32
|
152
|
+
# define U64 unsigned __int64
|
153
|
+
#else
|
154
|
+
# include <stdint.h>
|
155
|
+
# define BYTE uint8_t
|
156
|
+
# define U16 uint16_t
|
157
|
+
# define U32 uint32_t
|
158
|
+
# define S32 int32_t
|
159
|
+
# define U64 uint64_t
|
160
|
+
#endif
|
161
|
+
|
162
|
+
#ifndef LZ4_FORCE_UNALIGNED_ACCESS
|
163
|
+
# pragma pack(push, 1)
|
164
|
+
#endif
|
165
|
+
|
166
|
+
typedef struct _U16_S { U16 v; } U16_S;
|
167
|
+
typedef struct _U32_S { U32 v; } U32_S;
|
168
|
+
typedef struct _U64_S { U64 v; } U64_S;
|
169
|
+
|
170
|
+
#ifndef LZ4_FORCE_UNALIGNED_ACCESS
|
171
|
+
# pragma pack(pop)
|
172
|
+
#endif
|
173
|
+
|
174
|
+
#define A64(x) (((U64_S *)(x))->v)
|
175
|
+
#define A32(x) (((U32_S *)(x))->v)
|
176
|
+
#define A16(x) (((U16_S *)(x))->v)
|
177
|
+
|
178
|
+
|
179
|
+
//**************************************
|
180
|
+
// Constants
|
181
|
+
//**************************************
|
182
|
+
#define MINMATCH 4
|
183
|
+
|
184
|
+
#define HASH_LOG COMPRESSIONLEVEL
|
185
|
+
#define HASHTABLESIZE (1 << HASH_LOG)
|
186
|
+
#define HASH_MASK (HASHTABLESIZE - 1)
|
187
|
+
|
188
|
+
#define SKIPSTRENGTH (NOTCOMPRESSIBLE_CONFIRMATION>2?NOTCOMPRESSIBLE_CONFIRMATION:2)
|
189
|
+
#define STACKLIMIT 13
|
190
|
+
#define HEAPMODE (HASH_LOG>STACKLIMIT) // Defines if memory is allocated into the stack (local variable), or into the heap (malloc()).
|
191
|
+
#define COPYLENGTH 8
|
192
|
+
#define LASTLITERALS 5
|
193
|
+
#define MFLIMIT (COPYLENGTH+MINMATCH)
|
194
|
+
#define MINLENGTH (MFLIMIT+1)
|
195
|
+
|
196
|
+
#define MAXD_LOG 16
|
197
|
+
#define MAX_DISTANCE ((1 << MAXD_LOG) - 1)
|
198
|
+
|
199
|
+
#define ML_BITS 4
|
200
|
+
#define ML_MASK ((1U<<ML_BITS)-1)
|
201
|
+
#define RUN_BITS (8-ML_BITS)
|
202
|
+
#define RUN_MASK ((1U<<RUN_BITS)-1)
|
203
|
+
|
204
|
+
|
205
|
+
//**************************************
|
206
|
+
// Architecture-specific macros
|
207
|
+
//**************************************
|
208
|
+
#if LZ4_ARCH64 // 64-bit
|
209
|
+
# define STEPSIZE 8
|
210
|
+
# define UARCH U64
|
211
|
+
# define AARCH A64
|
212
|
+
# define LZ4_COPYSTEP(s,d) A64(d) = A64(s); d+=8; s+=8;
|
213
|
+
# define LZ4_COPYPACKET(s,d) LZ4_COPYSTEP(s,d)
|
214
|
+
# define LZ4_SECURECOPY(s,d,e) if (d<e) LZ4_WILDCOPY(s,d,e)
|
215
|
+
# define HTYPE U32
|
216
|
+
# define INITBASE(base) const BYTE* const base = ip
|
217
|
+
#else // 32-bit
|
218
|
+
# define STEPSIZE 4
|
219
|
+
# define UARCH U32
|
220
|
+
# define AARCH A32
|
221
|
+
# define LZ4_COPYSTEP(s,d) A32(d) = A32(s); d+=4; s+=4;
|
222
|
+
# define LZ4_COPYPACKET(s,d) LZ4_COPYSTEP(s,d); LZ4_COPYSTEP(s,d);
|
223
|
+
# define LZ4_SECURECOPY LZ4_WILDCOPY
|
224
|
+
# define HTYPE const BYTE*
|
225
|
+
# define INITBASE(base) const int base = 0
|
226
|
+
#endif
|
227
|
+
|
228
|
+
#if (defined(LZ4_BIG_ENDIAN) && !defined(BIG_ENDIAN_NATIVE_BUT_INCOMPATIBLE))
|
229
|
+
# define LZ4_READ_LITTLEENDIAN_16(d,s,p) { U16 v = A16(p); v = lz4_bswap16(v); d = (s) - v; }
|
230
|
+
# define LZ4_WRITE_LITTLEENDIAN_16(p,i) { U16 v = (U16)(i); v = lz4_bswap16(v); A16(p) = v; p+=2; }
|
231
|
+
#else // Little Endian
|
232
|
+
# define LZ4_READ_LITTLEENDIAN_16(d,s,p) { d = (s) - A16(p); }
|
233
|
+
# define LZ4_WRITE_LITTLEENDIAN_16(p,v) { A16(p) = v; p+=2; }
|
234
|
+
#endif
|
235
|
+
|
236
|
+
|
237
|
+
//**************************************
|
238
|
+
// Local structures
|
239
|
+
//**************************************
|
240
|
+
struct refTables
|
241
|
+
{
|
242
|
+
HTYPE hashTable[HASHTABLESIZE];
|
243
|
+
};
|
244
|
+
|
245
|
+
|
246
|
+
//**************************************
|
247
|
+
// Macros
|
248
|
+
//**************************************
|
249
|
+
#define LZ4_HASH_FUNCTION(i) (((i) * 2654435761U) >> ((MINMATCH*8)-HASH_LOG))
|
250
|
+
#define LZ4_HASH_VALUE(p) LZ4_HASH_FUNCTION(A32(p))
|
251
|
+
#define LZ4_WILDCOPY(s,d,e) do { LZ4_COPYPACKET(s,d) } while (d<e);
|
252
|
+
#define LZ4_BLINDCOPY(s,d,l) { BYTE* e=(d)+l; LZ4_WILDCOPY(s,d,e); d=e; }
|
253
|
+
|
254
|
+
|
255
|
+
//****************************
|
256
|
+
// Private functions
|
257
|
+
//****************************
|
258
|
+
#if LZ4_ARCH64
|
259
|
+
|
260
|
+
inline static int LZ4_NbCommonBytes (register U64 val)
|
261
|
+
{
|
262
|
+
#if defined(LZ4_BIG_ENDIAN)
|
263
|
+
#if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
264
|
+
unsigned long r = 0;
|
265
|
+
_BitScanReverse64( &r, val );
|
266
|
+
return (int)(r>>3);
|
267
|
+
#elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
268
|
+
return (__builtin_clzll(val) >> 3);
|
269
|
+
#else
|
270
|
+
int r;
|
271
|
+
if (!(val>>32)) { r=4; } else { r=0; val>>=32; }
|
272
|
+
if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }
|
273
|
+
r += (!val);
|
274
|
+
return r;
|
275
|
+
#endif
|
276
|
+
#else
|
277
|
+
#if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
278
|
+
unsigned long r = 0;
|
279
|
+
_BitScanForward64( &r, val );
|
280
|
+
return (int)(r>>3);
|
281
|
+
#elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
282
|
+
return (__builtin_ctzll(val) >> 3);
|
283
|
+
#else
|
284
|
+
static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2, 0, 3, 1, 3, 1, 4, 2, 7, 0, 2, 3, 6, 1, 5, 3, 5, 1, 3, 4, 4, 2, 5, 6, 7, 7, 0, 1, 2, 3, 3, 4, 6, 2, 6, 5, 5, 3, 4, 5, 6, 7, 1, 2, 4, 6, 4, 4, 5, 7, 2, 6, 5, 7, 6, 7, 7 };
|
285
|
+
return DeBruijnBytePos[((U64)((val & -val) * 0x0218A392CDABBD3F)) >> 58];
|
286
|
+
#endif
|
287
|
+
#endif
|
288
|
+
}
|
289
|
+
|
290
|
+
#else
|
291
|
+
|
292
|
+
inline static int LZ4_NbCommonBytes (register U32 val)
|
293
|
+
{
|
294
|
+
#if defined(LZ4_BIG_ENDIAN)
|
295
|
+
#if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
296
|
+
unsigned long r = 0;
|
297
|
+
_BitScanReverse( &r, val );
|
298
|
+
return (int)(r>>3);
|
299
|
+
#elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
300
|
+
return (__builtin_clz(val) >> 3);
|
301
|
+
#else
|
302
|
+
int r;
|
303
|
+
if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }
|
304
|
+
r += (!val);
|
305
|
+
return r;
|
306
|
+
#endif
|
307
|
+
#else
|
308
|
+
#if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
309
|
+
unsigned long r = 0;
|
310
|
+
_BitScanForward( &r, val );
|
311
|
+
return (int)(r>>3);
|
312
|
+
#elif defined(__GNUC__) && (GCC_VERSION >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
313
|
+
return (__builtin_ctz(val) >> 3);
|
314
|
+
#else
|
315
|
+
static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0, 3, 2, 2, 1, 3, 2, 0, 1, 3, 3, 1, 2, 2, 2, 2, 0, 3, 1, 2, 0, 1, 0, 1, 1 };
|
316
|
+
return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
|
317
|
+
#endif
|
318
|
+
#endif
|
319
|
+
}
|
320
|
+
|
321
|
+
#endif
|
322
|
+
|
323
|
+
|
324
|
+
//****************************
|
325
|
+
// Public functions
|
326
|
+
//****************************
|
327
|
+
|
328
|
+
int LZ4_compressBound(int isize)
|
329
|
+
{
|
330
|
+
return (isize + (isize/255) + 16);
|
331
|
+
}
|
332
|
+
|
333
|
+
|
334
|
+
|
335
|
+
//******************************
|
336
|
+
// Compression functions
|
337
|
+
//******************************
|
338
|
+
|
339
|
+
int LZ4_compressCtx(void** ctx,
|
340
|
+
const char* source,
|
341
|
+
char* dest,
|
342
|
+
int isize)
|
343
|
+
{
|
344
|
+
#if HEAPMODE
|
345
|
+
struct refTables *srt = (struct refTables *) (*ctx);
|
346
|
+
HTYPE* HashTable;
|
347
|
+
#else
|
348
|
+
HTYPE HashTable[HASHTABLESIZE] = {0};
|
349
|
+
#endif
|
350
|
+
|
351
|
+
const BYTE* ip = (BYTE*) source;
|
352
|
+
INITBASE(base);
|
353
|
+
const BYTE* anchor = ip;
|
354
|
+
const BYTE* const iend = ip + isize;
|
355
|
+
const BYTE* const mflimit = iend - MFLIMIT;
|
356
|
+
#define matchlimit (iend - LASTLITERALS)
|
357
|
+
|
358
|
+
BYTE* op = (BYTE*) dest;
|
359
|
+
|
360
|
+
int len, length;
|
361
|
+
const int skipStrength = SKIPSTRENGTH;
|
362
|
+
U32 forwardH;
|
363
|
+
|
364
|
+
|
365
|
+
// Init
|
366
|
+
if (isize<MINLENGTH) goto _last_literals;
|
367
|
+
#if HEAPMODE
|
368
|
+
if (*ctx == NULL)
|
369
|
+
{
|
370
|
+
srt = (struct refTables *) malloc ( sizeof(struct refTables) );
|
371
|
+
*ctx = (void*) srt;
|
372
|
+
}
|
373
|
+
HashTable = (HTYPE*)(srt->hashTable);
|
374
|
+
memset((void*)HashTable, 0, sizeof(srt->hashTable));
|
375
|
+
#else
|
376
|
+
(void) ctx;
|
377
|
+
#endif
|
378
|
+
|
379
|
+
|
380
|
+
// First Byte
|
381
|
+
HashTable[LZ4_HASH_VALUE(ip)] = ip - base;
|
382
|
+
ip++; forwardH = LZ4_HASH_VALUE(ip);
|
383
|
+
|
384
|
+
// Main Loop
|
385
|
+
for ( ; ; )
|
386
|
+
{
|
387
|
+
int findMatchAttempts = (1U << skipStrength) + 3;
|
388
|
+
const BYTE* forwardIp = ip;
|
389
|
+
const BYTE* ref;
|
390
|
+
BYTE* token;
|
391
|
+
|
392
|
+
// Find a match
|
393
|
+
do {
|
394
|
+
U32 h = forwardH;
|
395
|
+
int step = findMatchAttempts++ >> skipStrength;
|
396
|
+
ip = forwardIp;
|
397
|
+
forwardIp = ip + step;
|
398
|
+
|
399
|
+
if unlikely(forwardIp > mflimit) { goto _last_literals; }
|
400
|
+
|
401
|
+
forwardH = LZ4_HASH_VALUE(forwardIp);
|
402
|
+
ref = base + HashTable[h];
|
403
|
+
HashTable[h] = ip - base;
|
404
|
+
|
405
|
+
} while ((ref < ip - MAX_DISTANCE) || (A32(ref) != A32(ip)));
|
406
|
+
|
407
|
+
// Catch up
|
408
|
+
while ((ip>anchor) && (ref>(BYTE*)source) && unlikely(ip[-1]==ref[-1])) { ip--; ref--; }
|
409
|
+
|
410
|
+
// Encode Literal length
|
411
|
+
length = ip - anchor;
|
412
|
+
token = op++;
|
413
|
+
if (length>=(int)RUN_MASK) { *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *op++ = 255; *op++ = (BYTE)len; }
|
414
|
+
else *token = (length<<ML_BITS);
|
415
|
+
|
416
|
+
// Copy Literals
|
417
|
+
LZ4_BLINDCOPY(anchor, op, length);
|
418
|
+
|
419
|
+
_next_match:
|
420
|
+
// Encode Offset
|
421
|
+
LZ4_WRITE_LITTLEENDIAN_16(op,ip-ref);
|
422
|
+
|
423
|
+
// Start Counting
|
424
|
+
ip+=MINMATCH; ref+=MINMATCH; // MinMatch verified
|
425
|
+
anchor = ip;
|
426
|
+
while likely(ip<matchlimit-(STEPSIZE-1))
|
427
|
+
{
|
428
|
+
UARCH diff = AARCH(ref) ^ AARCH(ip);
|
429
|
+
if (!diff) { ip+=STEPSIZE; ref+=STEPSIZE; continue; }
|
430
|
+
ip += LZ4_NbCommonBytes(diff);
|
431
|
+
goto _endCount;
|
432
|
+
}
|
433
|
+
if (LZ4_ARCH64) if ((ip<(matchlimit-3)) && (A32(ref) == A32(ip))) { ip+=4; ref+=4; }
|
434
|
+
if ((ip<(matchlimit-1)) && (A16(ref) == A16(ip))) { ip+=2; ref+=2; }
|
435
|
+
if ((ip<matchlimit) && (*ref == *ip)) ip++;
|
436
|
+
_endCount:
|
437
|
+
|
438
|
+
// Encode MatchLength
|
439
|
+
len = (ip - anchor);
|
440
|
+
if (len>=(int)ML_MASK) { *token+=ML_MASK; len-=ML_MASK; for(; len > 509 ; len-=510) { *op++ = 255; *op++ = 255; } if (len > 254) { len-=255; *op++ = 255; } *op++ = (BYTE)len; }
|
441
|
+
else *token += len;
|
442
|
+
|
443
|
+
// Test end of chunk
|
444
|
+
if (ip > mflimit) { anchor = ip; break; }
|
445
|
+
|
446
|
+
// Fill table
|
447
|
+
HashTable[LZ4_HASH_VALUE(ip-2)] = ip - 2 - base;
|
448
|
+
|
449
|
+
// Test next position
|
450
|
+
ref = base + HashTable[LZ4_HASH_VALUE(ip)];
|
451
|
+
HashTable[LZ4_HASH_VALUE(ip)] = ip - base;
|
452
|
+
if ((ref > ip - (MAX_DISTANCE + 1)) && (A32(ref) == A32(ip))) { token = op++; *token=0; goto _next_match; }
|
453
|
+
|
454
|
+
// Prepare next loop
|
455
|
+
anchor = ip++;
|
456
|
+
forwardH = LZ4_HASH_VALUE(ip);
|
457
|
+
}
|
458
|
+
|
459
|
+
_last_literals:
|
460
|
+
// Encode Last Literals
|
461
|
+
{
|
462
|
+
int lastRun = iend - anchor;
|
463
|
+
if ((LZ4_COMPRESSMIN>0) && (((op - (BYTE*)dest) + lastRun + 1 + ((lastRun-15)/255)) > isize - LZ4_COMPRESSMIN)) return 0;
|
464
|
+
if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; }
|
465
|
+
else *op++ = (lastRun<<ML_BITS);
|
466
|
+
memcpy(op, anchor, iend - anchor);
|
467
|
+
op += iend-anchor;
|
468
|
+
}
|
469
|
+
|
470
|
+
// End
|
471
|
+
return (int) (((char*)op)-dest);
|
472
|
+
}
|
473
|
+
|
474
|
+
|
475
|
+
|
476
|
+
// Note : this function is valid only if isize < LZ4_64KLIMIT
|
477
|
+
#define LZ4_64KLIMIT ((1<<16) + (MFLIMIT-1))
|
478
|
+
#define HASHLOG64K (HASH_LOG+1)
|
479
|
+
#define HASH64KTABLESIZE (1U<<HASHLOG64K)
|
480
|
+
#define LZ4_HASH64K_FUNCTION(i) (((i) * 2654435761U) >> ((MINMATCH*8)-HASHLOG64K))
|
481
|
+
#define LZ4_HASH64K_VALUE(p) LZ4_HASH64K_FUNCTION(A32(p))
|
482
|
+
int LZ4_compress64kCtx(void** ctx,
|
483
|
+
const char* source,
|
484
|
+
char* dest,
|
485
|
+
int isize)
|
486
|
+
{
|
487
|
+
#if HEAPMODE
|
488
|
+
struct refTables *srt = (struct refTables *) (*ctx);
|
489
|
+
U16* HashTable;
|
490
|
+
#else
|
491
|
+
U16 HashTable[HASH64KTABLESIZE] = {0};
|
492
|
+
#endif
|
493
|
+
|
494
|
+
const BYTE* ip = (BYTE*) source;
|
495
|
+
const BYTE* anchor = ip;
|
496
|
+
const BYTE* const base = ip;
|
497
|
+
const BYTE* const iend = ip + isize;
|
498
|
+
const BYTE* const mflimit = iend - MFLIMIT;
|
499
|
+
#define matchlimit (iend - LASTLITERALS)
|
500
|
+
|
501
|
+
BYTE* op = (BYTE*) dest;
|
502
|
+
|
503
|
+
int len, length;
|
504
|
+
const int skipStrength = SKIPSTRENGTH;
|
505
|
+
U32 forwardH;
|
506
|
+
|
507
|
+
|
508
|
+
// Init
|
509
|
+
if (isize<MINLENGTH) goto _last_literals;
|
510
|
+
#if HEAPMODE
|
511
|
+
if (*ctx == NULL)
|
512
|
+
{
|
513
|
+
srt = (struct refTables *) malloc ( sizeof(struct refTables) );
|
514
|
+
*ctx = (void*) srt;
|
515
|
+
}
|
516
|
+
HashTable = (U16*)(srt->hashTable);
|
517
|
+
memset((void*)HashTable, 0, sizeof(srt->hashTable));
|
518
|
+
#else
|
519
|
+
(void) ctx;
|
520
|
+
#endif
|
521
|
+
|
522
|
+
|
523
|
+
// First Byte
|
524
|
+
ip++; forwardH = LZ4_HASH64K_VALUE(ip);
|
525
|
+
|
526
|
+
// Main Loop
|
527
|
+
for ( ; ; )
|
528
|
+
{
|
529
|
+
int findMatchAttempts = (1U << skipStrength) + 3;
|
530
|
+
const BYTE* forwardIp = ip;
|
531
|
+
const BYTE* ref;
|
532
|
+
BYTE* token;
|
533
|
+
|
534
|
+
// Find a match
|
535
|
+
do {
|
536
|
+
U32 h = forwardH;
|
537
|
+
int step = findMatchAttempts++ >> skipStrength;
|
538
|
+
ip = forwardIp;
|
539
|
+
forwardIp = ip + step;
|
540
|
+
|
541
|
+
if (forwardIp > mflimit) { goto _last_literals; }
|
542
|
+
|
543
|
+
forwardH = LZ4_HASH64K_VALUE(forwardIp);
|
544
|
+
ref = base + HashTable[h];
|
545
|
+
HashTable[h] = ip - base;
|
546
|
+
|
547
|
+
} while (A32(ref) != A32(ip));
|
548
|
+
|
549
|
+
// Catch up
|
550
|
+
while ((ip>anchor) && (ref>(BYTE*)source) && (ip[-1]==ref[-1])) { ip--; ref--; }
|
551
|
+
|
552
|
+
// Encode Literal length
|
553
|
+
length = ip - anchor;
|
554
|
+
token = op++;
|
555
|
+
if (length>=(int)RUN_MASK) { *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *op++ = 255; *op++ = (BYTE)len; }
|
556
|
+
else *token = (length<<ML_BITS);
|
557
|
+
|
558
|
+
// Copy Literals
|
559
|
+
LZ4_BLINDCOPY(anchor, op, length);
|
560
|
+
|
561
|
+
_next_match:
|
562
|
+
// Encode Offset
|
563
|
+
LZ4_WRITE_LITTLEENDIAN_16(op,ip-ref);
|
564
|
+
|
565
|
+
// Start Counting
|
566
|
+
ip+=MINMATCH; ref+=MINMATCH; // MinMatch verified
|
567
|
+
anchor = ip;
|
568
|
+
while (ip<matchlimit-(STEPSIZE-1))
|
569
|
+
{
|
570
|
+
UARCH diff = AARCH(ref) ^ AARCH(ip);
|
571
|
+
if (!diff) { ip+=STEPSIZE; ref+=STEPSIZE; continue; }
|
572
|
+
ip += LZ4_NbCommonBytes(diff);
|
573
|
+
goto _endCount;
|
574
|
+
}
|
575
|
+
if (LZ4_ARCH64) if ((ip<(matchlimit-3)) && (A32(ref) == A32(ip))) { ip+=4; ref+=4; }
|
576
|
+
if ((ip<(matchlimit-1)) && (A16(ref) == A16(ip))) { ip+=2; ref+=2; }
|
577
|
+
if ((ip<matchlimit) && (*ref == *ip)) ip++;
|
578
|
+
_endCount:
|
579
|
+
|
580
|
+
// Encode MatchLength
|
581
|
+
len = (ip - anchor);
|
582
|
+
if (len>=(int)ML_MASK) { *token+=ML_MASK; len-=ML_MASK; for(; len > 509 ; len-=510) { *op++ = 255; *op++ = 255; } if (len > 254) { len-=255; *op++ = 255; } *op++ = (BYTE)len; }
|
583
|
+
else *token += len;
|
584
|
+
|
585
|
+
// Test end of chunk
|
586
|
+
if (ip > mflimit) { anchor = ip; break; }
|
587
|
+
|
588
|
+
// Fill table
|
589
|
+
HashTable[LZ4_HASH64K_VALUE(ip-2)] = ip - 2 - base;
|
590
|
+
|
591
|
+
// Test next position
|
592
|
+
ref = base + HashTable[LZ4_HASH64K_VALUE(ip)];
|
593
|
+
HashTable[LZ4_HASH64K_VALUE(ip)] = ip - base;
|
594
|
+
if (A32(ref) == A32(ip)) { token = op++; *token=0; goto _next_match; }
|
595
|
+
|
596
|
+
// Prepare next loop
|
597
|
+
anchor = ip++;
|
598
|
+
forwardH = LZ4_HASH64K_VALUE(ip);
|
599
|
+
}
|
600
|
+
|
601
|
+
_last_literals:
|
602
|
+
// Encode Last Literals
|
603
|
+
{
|
604
|
+
int lastRun = iend - anchor;
|
605
|
+
if ((LZ4_COMPRESSMIN>0) && (((op - (BYTE*)dest) + lastRun + 1 + ((lastRun-15)/255)) > isize - LZ4_COMPRESSMIN)) return 0;
|
606
|
+
if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; }
|
607
|
+
else *op++ = (lastRun<<ML_BITS);
|
608
|
+
memcpy(op, anchor, iend - anchor);
|
609
|
+
op += iend-anchor;
|
610
|
+
}
|
611
|
+
|
612
|
+
// End
|
613
|
+
return (int) (((char*)op)-dest);
|
614
|
+
}
|
615
|
+
|
616
|
+
|
617
|
+
|
618
|
+
int LZ4_compress(const char* source,
|
619
|
+
char* dest,
|
620
|
+
int isize)
|
621
|
+
{
|
622
|
+
#if HEAPMODE
|
623
|
+
void* ctx = malloc(sizeof(struct refTables));
|
624
|
+
int result;
|
625
|
+
if (isize < LZ4_64KLIMIT)
|
626
|
+
result = LZ4_compress64kCtx(&ctx, source, dest, isize);
|
627
|
+
else result = LZ4_compressCtx(&ctx, source, dest, isize);
|
628
|
+
free(ctx);
|
629
|
+
return result;
|
630
|
+
#else
|
631
|
+
if (isize < (int)LZ4_64KLIMIT) return LZ4_compress64kCtx(NULL, source, dest, isize);
|
632
|
+
return LZ4_compressCtx(NULL, source, dest, isize);
|
633
|
+
#endif
|
634
|
+
}
|
635
|
+
|
636
|
+
|
637
|
+
|
638
|
+
|
639
|
+
//****************************
|
640
|
+
// Decompression functions
|
641
|
+
//****************************
|
642
|
+
|
643
|
+
// Note : The decoding functions LZ4_uncompress() and LZ4_uncompress_unknownOutputSize()
|
644
|
+
// are safe against "buffer overflow" attack type.
|
645
|
+
// They will never write nor read outside of the provided output buffers.
|
646
|
+
// LZ4_uncompress_unknownOutputSize() also insures that it will never read outside of the input buffer.
|
647
|
+
// A corrupted input will produce an error result, a negative int, indicating the position of the error within input stream.
|
648
|
+
|
649
|
+
int LZ4_uncompress(const char* source,
|
650
|
+
char* dest,
|
651
|
+
int osize)
|
652
|
+
{
|
653
|
+
// Local Variables
|
654
|
+
const BYTE* restrict ip = (const BYTE*) source;
|
655
|
+
const BYTE* restrict ref;
|
656
|
+
|
657
|
+
BYTE* restrict op = (BYTE*) dest;
|
658
|
+
BYTE* const oend = op + osize;
|
659
|
+
BYTE* cpy;
|
660
|
+
|
661
|
+
BYTE token;
|
662
|
+
|
663
|
+
int len, length;
|
664
|
+
size_t dec[] ={0, 3, 2, 3, 0, 0, 0, 0};
|
665
|
+
|
666
|
+
|
667
|
+
// Main Loop
|
668
|
+
while (1)
|
669
|
+
{
|
670
|
+
// get runlength
|
671
|
+
token = *ip++;
|
672
|
+
if ((length=(token>>ML_BITS)) == RUN_MASK) { for (;(len=*ip++)==255;length+=255){} length += len; }
|
673
|
+
|
674
|
+
// copy literals
|
675
|
+
cpy = op+length;
|
676
|
+
if unlikely(cpy>oend-COPYLENGTH)
|
677
|
+
{
|
678
|
+
if (cpy > oend) goto _output_error; // Error : request to write beyond destination buffer
|
679
|
+
memcpy(op, ip, length);
|
680
|
+
ip += length;
|
681
|
+
break; // Necessarily EOF
|
682
|
+
}
|
683
|
+
LZ4_WILDCOPY(ip, op, cpy); ip -= (op-cpy); op = cpy;
|
684
|
+
|
685
|
+
// get offset
|
686
|
+
LZ4_READ_LITTLEENDIAN_16(ref,cpy,ip); ip+=2;
|
687
|
+
if (ref < (BYTE* const)dest) goto _output_error; // Error : offset create reference outside destination buffer
|
688
|
+
|
689
|
+
// get matchlength
|
690
|
+
if ((length=(token&ML_MASK)) == ML_MASK) { for (;*ip==255;length+=255) {ip++;} length += *ip++; }
|
691
|
+
|
692
|
+
// copy repeated sequence
|
693
|
+
if unlikely(op-ref<STEPSIZE)
|
694
|
+
{
|
695
|
+
#if LZ4_ARCH64
|
696
|
+
size_t dec2table[]={0, 0, 0, -1, 0, 1, 2, 3};
|
697
|
+
size_t dec2 = dec2table[op-ref];
|
698
|
+
#else
|
699
|
+
const int dec2 = 0;
|
700
|
+
#endif
|
701
|
+
*op++ = *ref++;
|
702
|
+
*op++ = *ref++;
|
703
|
+
*op++ = *ref++;
|
704
|
+
*op++ = *ref++;
|
705
|
+
ref -= dec[op-ref];
|
706
|
+
A32(op)=A32(ref); op += STEPSIZE-4;
|
707
|
+
ref -= dec2;
|
708
|
+
} else { LZ4_COPYSTEP(ref,op); }
|
709
|
+
cpy = op + length - (STEPSIZE-4);
|
710
|
+
if (cpy>oend-COPYLENGTH)
|
711
|
+
{
|
712
|
+
if (cpy > oend) goto _output_error; // Error : request to write beyond destination buffer
|
713
|
+
LZ4_SECURECOPY(ref, op, (oend-COPYLENGTH));
|
714
|
+
while(op<cpy) *op++=*ref++;
|
715
|
+
op=cpy;
|
716
|
+
if (op == oend) break; // Check EOF (should never happen, since last 5 bytes are supposed to be literals)
|
717
|
+
continue;
|
718
|
+
}
|
719
|
+
LZ4_SECURECOPY(ref, op, cpy);
|
720
|
+
op=cpy; // correction
|
721
|
+
}
|
722
|
+
|
723
|
+
// end of decoding
|
724
|
+
return (int) (((char*)ip)-source);
|
725
|
+
|
726
|
+
// write overflow error detected
|
727
|
+
_output_error:
|
728
|
+
return (int) (-(((char*)ip)-source));
|
729
|
+
}
|
730
|
+
|
731
|
+
|
732
|
+
int LZ4_uncompress_unknownOutputSize(
|
733
|
+
const char* source,
|
734
|
+
char* dest,
|
735
|
+
int isize,
|
736
|
+
int maxOutputSize)
|
737
|
+
{
|
738
|
+
// Local Variables
|
739
|
+
const BYTE* restrict ip = (const BYTE*) source;
|
740
|
+
const BYTE* const iend = ip + isize;
|
741
|
+
const BYTE* restrict ref;
|
742
|
+
|
743
|
+
BYTE* restrict op = (BYTE*) dest;
|
744
|
+
BYTE* const oend = op + maxOutputSize;
|
745
|
+
BYTE* cpy;
|
746
|
+
|
747
|
+
size_t dec[] ={0, 3, 2, 3, 0, 0, 0, 0};
|
748
|
+
|
749
|
+
|
750
|
+
// Main Loop
|
751
|
+
while (ip<iend)
|
752
|
+
{
|
753
|
+
BYTE token;
|
754
|
+
int length;
|
755
|
+
|
756
|
+
// get runlength
|
757
|
+
token = *ip++;
|
758
|
+
if ((length=(token>>ML_BITS)) == RUN_MASK) { int s=255; while ((ip<iend) && (s==255)) { s=*ip++; length += s; } }
|
759
|
+
|
760
|
+
// copy literals
|
761
|
+
cpy = op+length;
|
762
|
+
if ((cpy>oend-COPYLENGTH) || (ip+length>iend-COPYLENGTH))
|
763
|
+
{
|
764
|
+
if (cpy > oend) goto _output_error; // Error : request to write beyond destination buffer
|
765
|
+
if (ip+length > iend) goto _output_error; // Error : request to read beyond source buffer
|
766
|
+
memcpy(op, ip, length);
|
767
|
+
op += length;
|
768
|
+
ip += length;
|
769
|
+
if (ip<iend) goto _output_error; // Error : LZ4 format violation
|
770
|
+
break; // Necessarily EOF, due to parsing restrictions
|
771
|
+
}
|
772
|
+
LZ4_WILDCOPY(ip, op, cpy); ip -= (op-cpy); op = cpy;
|
773
|
+
|
774
|
+
// get offset
|
775
|
+
LZ4_READ_LITTLEENDIAN_16(ref,cpy,ip); ip+=2;
|
776
|
+
if (ref < (BYTE* const)dest) goto _output_error; // Error : offset creates reference outside of destination buffer
|
777
|
+
|
778
|
+
// get matchlength
|
779
|
+
if ((length=(token&ML_MASK)) == ML_MASK) { while (ip<iend) { int s = *ip++; length +=s; if (s==255) continue; break; } }
|
780
|
+
|
781
|
+
// copy repeated sequence
|
782
|
+
if unlikely(op-ref<STEPSIZE)
|
783
|
+
{
|
784
|
+
#if LZ4_ARCH64
|
785
|
+
size_t dec2table[]={0, 0, 0, -1, 0, 1, 2, 3};
|
786
|
+
size_t dec2 = dec2table[op-ref];
|
787
|
+
#else
|
788
|
+
const int dec2 = 0;
|
789
|
+
#endif
|
790
|
+
*op++ = *ref++;
|
791
|
+
*op++ = *ref++;
|
792
|
+
*op++ = *ref++;
|
793
|
+
*op++ = *ref++;
|
794
|
+
ref -= dec[op-ref];
|
795
|
+
A32(op)=A32(ref); op += STEPSIZE-4;
|
796
|
+
ref -= dec2;
|
797
|
+
} else { LZ4_COPYSTEP(ref,op); }
|
798
|
+
cpy = op + length - (STEPSIZE-4);
|
799
|
+
if (cpy>oend-COPYLENGTH)
|
800
|
+
{
|
801
|
+
if (cpy > oend) goto _output_error; // Error : request to write outside of destination buffer
|
802
|
+
LZ4_SECURECOPY(ref, op, (oend-COPYLENGTH));
|
803
|
+
while(op<cpy) *op++=*ref++;
|
804
|
+
op=cpy;
|
805
|
+
if (op == oend) break; // Check EOF (should never happen, since last 5 bytes are supposed to be literals)
|
806
|
+
continue;
|
807
|
+
}
|
808
|
+
LZ4_SECURECOPY(ref, op, cpy);
|
809
|
+
op=cpy; // correction
|
810
|
+
}
|
811
|
+
|
812
|
+
// end of decoding
|
813
|
+
return (int) (((char*)op)-dest);
|
814
|
+
|
815
|
+
// write overflow error detected
|
816
|
+
_output_error:
|
817
|
+
return (int) (-(((char*)ip)-source));
|
818
|
+
}
|
819
|
+
|