lz4-ruby 0.1.4 → 0.1.6

Sign up to get free protection for your applications and to get access to all the features.
@@ -0,0 +1,120 @@
1
+ /*
2
+ LZ4 - Fast LZ compression algorithm
3
+ Header File
4
+ Copyright (C) 2011-2012, Yann Collet.
5
+ BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
6
+
7
+ Redistribution and use in source and binary forms, with or without
8
+ modification, are permitted provided that the following conditions are
9
+ met:
10
+
11
+ * Redistributions of source code must retain the above copyright
12
+ notice, this list of conditions and the following disclaimer.
13
+ * Redistributions in binary form must reproduce the above
14
+ copyright notice, this list of conditions and the following disclaimer
15
+ in the documentation and/or other materials provided with the
16
+ distribution.
17
+
18
+ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
19
+ "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
20
+ LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
21
+ A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
22
+ OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
23
+ SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
24
+ LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
25
+ DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
26
+ THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
27
+ (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
28
+ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
29
+
30
+ You can contact the author at :
31
+ - LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
32
+ - LZ4 source repository : http://code.google.com/p/lz4/
33
+ */
34
+ #pragma once
35
+
36
+ #if defined (__cplusplus)
37
+ extern "C" {
38
+ #endif
39
+
40
+
41
+ //****************************
42
+ // Simple Functions
43
+ //****************************
44
+
45
+ int LZ4_compress (const char* source, char* dest, int isize);
46
+ int LZ4_uncompress (const char* source, char* dest, int osize);
47
+
48
+ /*
49
+ LZ4_compress() :
50
+ isize : is the input size. Max supported value is ~1.9GB
51
+ return : the number of bytes written in buffer dest
52
+ or 0 if the compression fails (if LZ4_COMPRESSMIN is set)
53
+ note : destination buffer must be already allocated.
54
+ destination buffer must be sized to handle worst cases situations (input data not compressible)
55
+ worst case size evaluation is provided by function LZ4_compressBound()
56
+
57
+ LZ4_uncompress() :
58
+ osize : is the output size, therefore the original size
59
+ return : the number of bytes read in the source buffer
60
+ If the source stream is malformed, the function will stop decoding and return a negative result, indicating the byte position of the faulty instruction
61
+ This function never writes beyond dest + osize, and is therefore protected against malicious data packets
62
+ note : destination buffer must be already allocated
63
+ */
64
+
65
+
66
+ //****************************
67
+ // Advanced Functions
68
+ //****************************
69
+
70
+ int LZ4_compressBound(int isize);
71
+
72
+ /*
73
+ LZ4_compressBound() :
74
+ Provides the maximum size that LZ4 may output in a "worst case" scenario (input data not compressible)
75
+ primarily useful for memory allocation of output buffer.
76
+
77
+ isize : is the input size. Max supported value is ~1.9GB
78
+ return : maximum output size in a "worst case" scenario
79
+ note : this function is limited by "int" range (2^31-1)
80
+ */
81
+
82
+
83
+ int LZ4_uncompress_unknownOutputSize (const char* source, char* dest, int isize, int maxOutputSize);
84
+
85
+ /*
86
+ LZ4_uncompress_unknownOutputSize() :
87
+ isize : is the input size, therefore the compressed size
88
+ maxOutputSize : is the size of the destination buffer (which must be already allocated)
89
+ return : the number of bytes decoded in the destination buffer (necessarily <= maxOutputSize)
90
+ If the source stream is malformed, the function will stop decoding and return a negative result, indicating the byte position of the faulty instruction
91
+ This function never writes beyond dest + maxOutputSize, and is therefore protected against malicious data packets
92
+ note : Destination buffer must be already allocated.
93
+ This version is slightly slower than LZ4_uncompress()
94
+ */
95
+
96
+
97
+ int LZ4_compressCtx(void** ctx, const char* source, char* dest, int isize);
98
+ int LZ4_compress64kCtx(void** ctx, const char* source, char* dest, int isize);
99
+
100
+ /*
101
+ LZ4_compressCtx() :
102
+ This function explicitly handles the CTX memory structure.
103
+ It avoids allocating/deallocating memory between each call, improving performance when malloc is heavily invoked.
104
+ This function is only useful when memory is allocated into the heap (HASH_LOG value beyond STACK_LIMIT)
105
+ Performance difference will be noticeable only when repetitively calling the compression function over many small segments.
106
+ Note : by default, memory is allocated into the stack, therefore "malloc" is not invoked.
107
+ LZ4_compress64kCtx() :
108
+ Same as LZ4_compressCtx(), but specific to small inputs (<64KB).
109
+ isize *Must* be <64KB, otherwise the output will be corrupted.
110
+
111
+ On first call : provide a *ctx=NULL; It will be automatically allocated.
112
+ On next calls : reuse the same ctx pointer.
113
+ Use different pointers for different threads when doing multi-threading.
114
+
115
+ */
116
+
117
+
118
+ #if defined (__cplusplus)
119
+ }
120
+ #endif
@@ -0,0 +1,663 @@
1
+ /*
2
+ LZ4 HC - High Compression Mode of LZ4
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
+ //**************************************
36
+ // CPU Feature Detection
37
+ //**************************************
38
+ // 32 or 64 bits ?
39
+ #if (defined(__x86_64__) || defined(__x86_64) || defined(__amd64__) || defined(__amd64) || defined(__ppc64__) || defined(_WIN64) || defined(__LP64__) || defined(_LP64) ) // Detects 64 bits mode
40
+ #define LZ4_ARCH64 1
41
+ #else
42
+ #define LZ4_ARCH64 0
43
+ #endif
44
+
45
+ // Little Endian or Big Endian ?
46
+ #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__))) )
47
+ #define LZ4_BIG_ENDIAN 1
48
+ #else
49
+ // Little Endian assumed. PDP Endian and other very rare endian format are unsupported.
50
+ #endif
51
+
52
+ // Unaligned memory access is automatically enabled for "common" CPU, such as x86.
53
+ // For others CPU, the compiler will be more cautious, and insert extra code to ensure aligned access is respected
54
+ // If you know your target CPU supports unaligned memory access, you may want to force this option manually to improve performance
55
+ #if defined(__ARM_FEATURE_UNALIGNED)
56
+ #define LZ4_FORCE_UNALIGNED_ACCESS 1
57
+ #endif
58
+
59
+
60
+ //**************************************
61
+ // Compiler Options
62
+ //**************************************
63
+ #if __STDC_VERSION__ >= 199901L // C99
64
+ /* "restrict" is a known keyword */
65
+ #else
66
+ #define restrict // Disable restrict
67
+ #endif
68
+
69
+ #ifdef _MSC_VER
70
+ #define inline __forceinline // Visual is not C99, but supports some kind of inline
71
+ #endif
72
+
73
+ #ifdef _MSC_VER // Visual Studio
74
+ #define bswap16(x) _byteswap_ushort(x)
75
+ #else
76
+ #define bswap16(x) ((unsigned short int) ((((x) >> 8) & 0xffu) | (((x) & 0xffu) << 8)))
77
+ #endif
78
+
79
+
80
+ //**************************************
81
+ // Includes
82
+ //**************************************
83
+ #include <stdlib.h> // calloc, free
84
+ #include <string.h> // memset, memcpy
85
+ #include "lz4hc.h"
86
+
87
+ #define ALLOCATOR(s) calloc(1,s)
88
+ #define FREEMEM free
89
+ #define MEM_INIT memset
90
+
91
+
92
+ //**************************************
93
+ // Basic Types
94
+ //**************************************
95
+ #if defined(_MSC_VER) // Visual Studio does not support 'stdint' natively
96
+ #define BYTE unsigned __int8
97
+ #define U16 unsigned __int16
98
+ #define U32 unsigned __int32
99
+ #define S32 __int32
100
+ #define U64 unsigned __int64
101
+ #else
102
+ #include <stdint.h>
103
+ #define BYTE uint8_t
104
+ #define U16 uint16_t
105
+ #define U32 uint32_t
106
+ #define S32 int32_t
107
+ #define U64 uint64_t
108
+ #endif
109
+
110
+ #ifndef LZ4_FORCE_UNALIGNED_ACCESS
111
+ #pragma pack(push, 1)
112
+ #endif
113
+
114
+ typedef struct _U16_S { U16 v; } U16_S;
115
+ typedef struct _U32_S { U32 v; } U32_S;
116
+ typedef struct _U64_S { U64 v; } U64_S;
117
+
118
+ #ifndef LZ4_FORCE_UNALIGNED_ACCESS
119
+ #pragma pack(pop)
120
+ #endif
121
+
122
+ #define A64(x) (((U64_S *)(x))->v)
123
+ #define A32(x) (((U32_S *)(x))->v)
124
+ #define A16(x) (((U16_S *)(x))->v)
125
+
126
+
127
+ //**************************************
128
+ // Constants
129
+ //**************************************
130
+ #define MINMATCH 4
131
+
132
+ #define DICTIONARY_LOGSIZE 16
133
+ #define MAXD (1<<DICTIONARY_LOGSIZE)
134
+ #define MAXD_MASK ((U32)(MAXD - 1))
135
+ #define MAX_DISTANCE (MAXD - 1)
136
+
137
+ #define HASH_LOG (DICTIONARY_LOGSIZE-1)
138
+ #define HASHTABLESIZE (1 << HASH_LOG)
139
+ #define HASH_MASK (HASHTABLESIZE - 1)
140
+
141
+ #define MAX_NB_ATTEMPTS 256
142
+
143
+ #define ML_BITS 4
144
+ #define ML_MASK (size_t)((1U<<ML_BITS)-1)
145
+ #define RUN_BITS (8-ML_BITS)
146
+ #define RUN_MASK ((1U<<RUN_BITS)-1)
147
+
148
+ #define COPYLENGTH 8
149
+ #define LASTLITERALS 5
150
+ #define MFLIMIT (COPYLENGTH+MINMATCH)
151
+ #define MINLENGTH (MFLIMIT+1)
152
+ #define OPTIMAL_ML (int)((ML_MASK-1)+MINMATCH)
153
+
154
+
155
+ //**************************************
156
+ // Architecture-specific macros
157
+ //**************************************
158
+ #if LZ4_ARCH64 // 64-bit
159
+ #define STEPSIZE 8
160
+ #define LZ4_COPYSTEP(s,d) A64(d) = A64(s); d+=8; s+=8;
161
+ #define LZ4_COPYPACKET(s,d) LZ4_COPYSTEP(s,d)
162
+ #define UARCH U64
163
+ #define AARCH A64
164
+ #define HTYPE U32
165
+ #define INITBASE(b,s) const BYTE* const b = s
166
+ #else // 32-bit
167
+ #define STEPSIZE 4
168
+ #define LZ4_COPYSTEP(s,d) A32(d) = A32(s); d+=4; s+=4;
169
+ #define LZ4_COPYPACKET(s,d) LZ4_COPYSTEP(s,d); LZ4_COPYSTEP(s,d);
170
+ #define UARCH U32
171
+ #define AARCH A32
172
+ #define HTYPE const BYTE*
173
+ #define INITBASE(b,s) const int b = 0
174
+ #endif
175
+
176
+ #if defined(LZ4_BIG_ENDIAN)
177
+ #define LZ4_READ_LITTLEENDIAN_16(d,s,p) { U16 v = A16(p); v = bswap16(v); d = (s) - v; }
178
+ #define LZ4_WRITE_LITTLEENDIAN_16(p,i) { U16 v = (U16)(i); v = bswap16(v); A16(p) = v; p+=2; }
179
+ #else // Little Endian
180
+ #define LZ4_READ_LITTLEENDIAN_16(d,s,p) { d = (s) - A16(p); }
181
+ #define LZ4_WRITE_LITTLEENDIAN_16(p,v) { A16(p) = v; p+=2; }
182
+ #endif
183
+
184
+
185
+ //************************************************************
186
+ // Local Types
187
+ //************************************************************
188
+ typedef struct
189
+ {
190
+ const BYTE* base;
191
+ HTYPE hashTable[HASHTABLESIZE];
192
+ U16 chainTable[MAXD];
193
+ const BYTE* nextToUpdate;
194
+ } LZ4HC_Data_Structure;
195
+
196
+
197
+ //**************************************
198
+ // Macros
199
+ //**************************************
200
+ #define LZ4_WILDCOPY(s,d,e) do { LZ4_COPYPACKET(s,d) } while (d<e);
201
+ #define LZ4_BLINDCOPY(s,d,l) { BYTE* e=d+l; LZ4_WILDCOPY(s,d,e); d=e; }
202
+ #define HASH_FUNCTION(i) (((i) * 2654435761U) >> ((MINMATCH*8)-HASH_LOG))
203
+ #define HASH_VALUE(p) HASH_FUNCTION(*(U32*)(p))
204
+ #define HASH_POINTER(p) (HashTable[HASH_VALUE(p)] + base)
205
+ #define DELTANEXT(p) chainTable[(size_t)(p) & MAXD_MASK]
206
+ #define GETNEXT(p) ((p) - (size_t)DELTANEXT(p))
207
+ #define ADD_HASH(p) { size_t delta = (p) - HASH_POINTER(p); if (delta>MAX_DISTANCE) delta = MAX_DISTANCE; DELTANEXT(p) = (U16)delta; HashTable[HASH_VALUE(p)] = (p) - base; }
208
+
209
+
210
+ //**************************************
211
+ // Private functions
212
+ //**************************************
213
+ #if LZ4_ARCH64
214
+
215
+ inline static int LZ4_NbCommonBytes (register U64 val)
216
+ {
217
+ #if defined(LZ4_BIG_ENDIAN)
218
+ #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
219
+ unsigned long r = 0;
220
+ _BitScanReverse64( &r, val );
221
+ return (int)(r>>3);
222
+ #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
223
+ return (__builtin_clzll(val) >> 3);
224
+ #else
225
+ int r;
226
+ if (!(val>>32)) { r=4; } else { r=0; val>>=32; }
227
+ if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }
228
+ r += (!val);
229
+ return r;
230
+ #endif
231
+ #else
232
+ #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
233
+ unsigned long r = 0;
234
+ _BitScanForward64( &r, val );
235
+ return (int)(r>>3);
236
+ #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
237
+ return (__builtin_ctzll(val) >> 3);
238
+ #else
239
+ 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 };
240
+ return DeBruijnBytePos[((U64)((val & -val) * 0x0218A392CDABBD3F)) >> 58];
241
+ #endif
242
+ #endif
243
+ }
244
+
245
+ #else
246
+
247
+ inline static int LZ4_NbCommonBytes (register U32 val)
248
+ {
249
+ #if defined(LZ4_BIG_ENDIAN)
250
+ #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
251
+ unsigned long r = 0;
252
+ _BitScanReverse( &r, val );
253
+ return (int)(r>>3);
254
+ #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
255
+ return (__builtin_clz(val) >> 3);
256
+ #else
257
+ int r;
258
+ if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }
259
+ r += (!val);
260
+ return r;
261
+ #endif
262
+ #else
263
+ #if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
264
+ unsigned long r = 0;
265
+ _BitScanForward( &r, val );
266
+ return (int)(r>>3);
267
+ #elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
268
+ return (__builtin_ctz(val) >> 3);
269
+ #else
270
+ 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 };
271
+ return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
272
+ #endif
273
+ #endif
274
+ }
275
+
276
+ #endif
277
+
278
+
279
+ inline static int LZ4HC_Init (LZ4HC_Data_Structure* hc4, const BYTE* base)
280
+ {
281
+ MEM_INIT((void*)hc4->hashTable, 0, sizeof(hc4->hashTable));
282
+ MEM_INIT(hc4->chainTable, 0xFF, sizeof(hc4->chainTable));
283
+ hc4->nextToUpdate = base + LZ4_ARCH64;
284
+ hc4->base = base;
285
+ return 1;
286
+ }
287
+
288
+
289
+ inline static void* LZ4HC_Create (const BYTE* base)
290
+ {
291
+ void* hc4 = ALLOCATOR(sizeof(LZ4HC_Data_Structure));
292
+
293
+ LZ4HC_Init (hc4, base);
294
+ return hc4;
295
+ }
296
+
297
+
298
+ inline static int LZ4HC_Free (void** LZ4HC_Data)
299
+ {
300
+ FREEMEM(*LZ4HC_Data);
301
+ *LZ4HC_Data = NULL;
302
+ return (1);
303
+ }
304
+
305
+
306
+ inline static void LZ4HC_Insert (LZ4HC_Data_Structure* hc4, const BYTE* ip)
307
+ {
308
+ U16* chainTable = hc4->chainTable;
309
+ HTYPE* HashTable = hc4->hashTable;
310
+ INITBASE(base,hc4->base);
311
+
312
+ while(hc4->nextToUpdate < ip)
313
+ {
314
+ ADD_HASH(hc4->nextToUpdate);
315
+ hc4->nextToUpdate++;
316
+ }
317
+ }
318
+
319
+
320
+ inline static int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* const matchlimit, const BYTE** matchpos)
321
+ {
322
+ U16* const chainTable = hc4->chainTable;
323
+ HTYPE* const HashTable = hc4->hashTable;
324
+ const BYTE* ref;
325
+ INITBASE(base,hc4->base);
326
+ int nbAttempts=MAX_NB_ATTEMPTS;
327
+ int ml=0;
328
+
329
+ // HC4 match finder
330
+ LZ4HC_Insert(hc4, ip);
331
+ ref = HASH_POINTER(ip);
332
+ while ((ref > (ip-MAX_DISTANCE)) && (nbAttempts))
333
+ {
334
+ nbAttempts--;
335
+ if (*(ref+ml) == *(ip+ml))
336
+ if (*(U32*)ref == *(U32*)ip)
337
+ {
338
+ const BYTE* reft = ref+MINMATCH;
339
+ const BYTE* ipt = ip+MINMATCH;
340
+
341
+ while (ipt<matchlimit-(STEPSIZE-1))
342
+ {
343
+ UARCH diff = AARCH(reft) ^ AARCH(ipt);
344
+ if (!diff) { ipt+=STEPSIZE; reft+=STEPSIZE; continue; }
345
+ ipt += LZ4_NbCommonBytes(diff);
346
+ goto _endCount;
347
+ }
348
+ if (LZ4_ARCH64) if ((ipt<(matchlimit-3)) && (A32(reft) == A32(ipt))) { ipt+=4; reft+=4; }
349
+ if ((ipt<(matchlimit-1)) && (A16(reft) == A16(ipt))) { ipt+=2; reft+=2; }
350
+ if ((ipt<matchlimit) && (*reft == *ipt)) ipt++;
351
+ _endCount:
352
+
353
+ if (ipt-ip > ml) { ml = ipt-ip; *matchpos = ref; }
354
+ }
355
+ ref = GETNEXT(ref);
356
+ }
357
+
358
+ return ml;
359
+ }
360
+
361
+
362
+ inline static int LZ4HC_InsertAndGetWiderMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* startLimit, const BYTE* matchlimit, int longest, const BYTE** matchpos, const BYTE** startpos)
363
+ {
364
+ U16* const chainTable = hc4->chainTable;
365
+ HTYPE* const HashTable = hc4->hashTable;
366
+ INITBASE(base,hc4->base);
367
+ const BYTE* ref;
368
+ int nbAttempts = MAX_NB_ATTEMPTS;
369
+ int delta = ip-startLimit;
370
+
371
+ // First Match
372
+ LZ4HC_Insert(hc4, ip);
373
+ ref = HASH_POINTER(ip);
374
+
375
+ while ((ref > ip-MAX_DISTANCE) && (ref >= hc4->base) && (nbAttempts))
376
+ {
377
+ nbAttempts--;
378
+ if (*(startLimit + longest) == *(ref - delta + longest))
379
+ if (*(U32*)ref == *(U32*)ip)
380
+ {
381
+ const BYTE* reft = ref+MINMATCH;
382
+ const BYTE* ipt = ip+MINMATCH;
383
+ const BYTE* startt = ip;
384
+
385
+ while (ipt<matchlimit-(STEPSIZE-1))
386
+ {
387
+ UARCH diff = AARCH(reft) ^ AARCH(ipt);
388
+ if (!diff) { ipt+=STEPSIZE; reft+=STEPSIZE; continue; }
389
+ ipt += LZ4_NbCommonBytes(diff);
390
+ goto _endCount;
391
+ }
392
+ if (LZ4_ARCH64) if ((ipt<(matchlimit-3)) && (A32(reft) == A32(ipt))) { ipt+=4; reft+=4; }
393
+ if ((ipt<(matchlimit-1)) && (A16(reft) == A16(ipt))) { ipt+=2; reft+=2; }
394
+ if ((ipt<matchlimit) && (*reft == *ipt)) ipt++;
395
+ _endCount:
396
+
397
+ reft = ref;
398
+ while ((startt>startLimit) && (reft > hc4->base) && (startt[-1] == reft[-1])) {startt--; reft--;}
399
+
400
+ if ((ipt-startt) > longest)
401
+ {
402
+ longest = ipt-startt;
403
+ *matchpos = reft;
404
+ *startpos = startt;
405
+ }
406
+ }
407
+ ref = GETNEXT(ref);
408
+ }
409
+
410
+ return longest;
411
+ }
412
+
413
+
414
+ inline static int LZ4_encodeSequence(const BYTE** ip, BYTE** op, const BYTE** anchor, int ml, const BYTE* ref)
415
+ {
416
+ int length, len;
417
+ BYTE* token;
418
+
419
+ // Encode Literal length
420
+ length = *ip - *anchor;
421
+ token = (*op)++;
422
+ if (length>=(int)RUN_MASK) { *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *(*op)++ = 255; *(*op)++ = (BYTE)len; }
423
+ else *token = (length<<ML_BITS);
424
+
425
+ // Copy Literals
426
+ LZ4_BLINDCOPY(*anchor, *op, length);
427
+
428
+ // Encode Offset
429
+ LZ4_WRITE_LITTLEENDIAN_16(*op,*ip-ref);
430
+
431
+ // Encode MatchLength
432
+ len = (int)(ml-MINMATCH);
433
+ 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; }
434
+ else *token += len;
435
+
436
+ // Prepare next loop
437
+ *ip += ml;
438
+ *anchor = *ip;
439
+
440
+ return 0;
441
+ }
442
+
443
+
444
+ //****************************
445
+ // Compression CODE
446
+ //****************************
447
+
448
+ int LZ4_compressHCCtx(LZ4HC_Data_Structure* ctx,
449
+ const char* source,
450
+ char* dest,
451
+ int isize)
452
+ {
453
+ const BYTE* ip = (const BYTE*) source;
454
+ const BYTE* anchor = ip;
455
+ const BYTE* const iend = ip + isize;
456
+ const BYTE* const mflimit = iend - MFLIMIT;
457
+ const BYTE* const matchlimit = (iend - LASTLITERALS);
458
+
459
+ BYTE* op = (BYTE*) dest;
460
+
461
+ int ml, ml2, ml3, ml0;
462
+ const BYTE* ref=NULL;
463
+ const BYTE* start2=NULL;
464
+ const BYTE* ref2=NULL;
465
+ const BYTE* start3=NULL;
466
+ const BYTE* ref3=NULL;
467
+ const BYTE* start0;
468
+ const BYTE* ref0;
469
+
470
+ ip++;
471
+
472
+ // Main Loop
473
+ while (ip < mflimit)
474
+ {
475
+ ml = LZ4HC_InsertAndFindBestMatch (ctx, ip, matchlimit, (&ref));
476
+ if (!ml) { ip++; continue; }
477
+
478
+ // saved, in case we would skip too much
479
+ start0 = ip;
480
+ ref0 = ref;
481
+ ml0 = ml;
482
+
483
+ _Search2:
484
+ if (ip+ml < mflimit)
485
+ ml2 = LZ4HC_InsertAndGetWiderMatch(ctx, ip + ml - 2, ip + 1, matchlimit, ml, &ref2, &start2);
486
+ else ml2=ml;
487
+
488
+ if (ml2 == ml) // No better match
489
+ {
490
+ LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
491
+ continue;
492
+ }
493
+
494
+ if (start0 < ip)
495
+ {
496
+ if (start2 < ip + ml0) // empirical
497
+ {
498
+ ip = start0;
499
+ ref = ref0;
500
+ ml = ml0;
501
+ }
502
+ }
503
+
504
+ // Here, start0==ip
505
+ if ((start2 - ip) < 3) // First Match too small : removed
506
+ {
507
+ ml = ml2;
508
+ ip = start2;
509
+ ref =ref2;
510
+ goto _Search2;
511
+ }
512
+
513
+ _Search3:
514
+ // Currently we have :
515
+ // ml2 > ml1, and
516
+ // ip1+3 <= ip2 (usually < ip1+ml1)
517
+ if ((start2 - ip) < OPTIMAL_ML)
518
+ {
519
+ int correction;
520
+ int new_ml = ml;
521
+ if (new_ml > OPTIMAL_ML) new_ml = OPTIMAL_ML;
522
+ if (ip+new_ml > start2 + ml2 - MINMATCH) new_ml = start2 - ip + ml2 - MINMATCH;
523
+ correction = new_ml - (start2 - ip);
524
+ if (correction > 0)
525
+ {
526
+ start2 += correction;
527
+ ref2 += correction;
528
+ ml2 -= correction;
529
+ }
530
+ }
531
+ // Now, we have start2 = ip+new_ml, with new_ml=min(ml, OPTIMAL_ML=18)
532
+
533
+ if (start2 + ml2 < mflimit)
534
+ ml3 = LZ4HC_InsertAndGetWiderMatch(ctx, start2 + ml2 - 3, start2, matchlimit, ml2, &ref3, &start3);
535
+ else ml3=ml2;
536
+
537
+ if (ml3 == ml2) // No better match : 2 sequences to encode
538
+ {
539
+ // ip & ref are known; Now for ml
540
+ if (start2 < ip+ml)
541
+ {
542
+ if ((start2 - ip) < OPTIMAL_ML)
543
+ {
544
+ int correction;
545
+ if (ml > OPTIMAL_ML) ml = OPTIMAL_ML;
546
+ if (ip+ml > start2 + ml2 - MINMATCH) ml = start2 - ip + ml2 - MINMATCH;
547
+ correction = ml - (start2 - ip);
548
+ if (correction > 0)
549
+ {
550
+ start2 += correction;
551
+ ref2 += correction;
552
+ ml2 -= correction;
553
+ }
554
+ }
555
+ else
556
+ {
557
+ ml = start2 - ip;
558
+ }
559
+ }
560
+ // Now, encode 2 sequences
561
+ LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
562
+ ip = start2;
563
+ LZ4_encodeSequence(&ip, &op, &anchor, ml2, ref2);
564
+ continue;
565
+ }
566
+
567
+ if (start3 < ip+ml+3) // Not enough space for match 2 : remove it
568
+ {
569
+ if (start3 >= (ip+ml)) // can write Seq1 immediately ==> Seq2 is removed, so Seq3 becomes Seq1
570
+ {
571
+ if (start2 < ip+ml)
572
+ {
573
+ int correction = (ip+ml) - start2;
574
+ start2 += correction;
575
+ ref2 += correction;
576
+ ml2 -= correction;
577
+ if (ml2 < MINMATCH)
578
+ {
579
+ start2 = start3;
580
+ ref2 = ref3;
581
+ ml2 = ml3;
582
+ }
583
+ }
584
+
585
+ LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
586
+ ip = start3;
587
+ ref = ref3;
588
+ ml = ml3;
589
+
590
+ start0 = start2;
591
+ ref0 = ref2;
592
+ ml0 = ml2;
593
+ goto _Search2;
594
+ }
595
+
596
+ start2 = start3;
597
+ ref2 = ref3;
598
+ ml2 = ml3;
599
+ goto _Search3;
600
+ }
601
+
602
+ // OK, now we have 3 ascending matches; let's write at least the first one
603
+ // ip & ref are known; Now for ml
604
+ if (start2 < ip+ml)
605
+ {
606
+ if ((start2 - ip) < (int)ML_MASK)
607
+ {
608
+ int correction;
609
+ if (ml > OPTIMAL_ML) ml = OPTIMAL_ML;
610
+ if (ip + ml > start2 + ml2 - MINMATCH) ml = start2 - ip + ml2 - MINMATCH;
611
+ correction = ml - (start2 - ip);
612
+ if (correction > 0)
613
+ {
614
+ start2 += correction;
615
+ ref2 += correction;
616
+ ml2 -= correction;
617
+ }
618
+ }
619
+ else
620
+ {
621
+ ml = start2 - ip;
622
+ }
623
+ }
624
+ LZ4_encodeSequence(&ip, &op, &anchor, ml, ref);
625
+
626
+ ip = start2;
627
+ ref = ref2;
628
+ ml = ml2;
629
+
630
+ start2 = start3;
631
+ ref2 = ref3;
632
+ ml2 = ml3;
633
+
634
+ goto _Search3;
635
+
636
+ }
637
+
638
+ // Encode Last Literals
639
+ {
640
+ int lastRun = iend - anchor;
641
+ if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; }
642
+ else *op++ = (lastRun<<ML_BITS);
643
+ memcpy(op, anchor, iend - anchor);
644
+ op += iend-anchor;
645
+ }
646
+
647
+ // End
648
+ return (int) (((char*)op)-dest);
649
+ }
650
+
651
+
652
+ int LZ4_compressHC(const char* source,
653
+ char* dest,
654
+ int isize)
655
+ {
656
+ void* ctx = LZ4HC_Create((const BYTE*)source);
657
+ int result = LZ4_compressHCCtx(ctx, source, dest, isize);
658
+ LZ4HC_Free (&ctx);
659
+
660
+ return result;
661
+ }
662
+
663
+