zstdlib 0.2.0 → 0.3.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGES.md +5 -0
- data/ext/zstdlib/extconf.rb +1 -1
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/bitstream.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/compiler.h +7 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/cpu.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/debug.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/debug.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/entropy_common.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/error_private.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/error_private.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/fse.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/fse_decompress.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/huf.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/mem.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/pool.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/pool.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/threading.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/threading.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/xxhash.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/xxhash.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/zstd_common.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/zstd_errors.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/common/zstd_internal.h +58 -6
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/fse_compress.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/hist.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/hist.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/huf_compress.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_compress.c +178 -691
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_compress_internal.h +98 -30
- data/ext/zstdlib/zstd-1.4.2/lib/compress/zstd_compress_literals.c +149 -0
- data/ext/zstdlib/zstd-1.4.2/lib/compress/zstd_compress_literals.h +29 -0
- data/ext/zstdlib/zstd-1.4.2/lib/compress/zstd_compress_sequences.c +415 -0
- data/ext/zstdlib/zstd-1.4.2/lib/compress/zstd_compress_sequences.h +47 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_double_fast.c +56 -36
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_double_fast.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_fast.c +35 -14
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_fast.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_lazy.c +10 -5
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_lazy.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_ldm.c +1 -1
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_ldm.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_opt.c +45 -32
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstd_opt.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstdmt_compress.c +18 -7
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/compress/zstdmt_compress.h +1 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/decompress/huf_decompress.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/decompress/zstd_ddict.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/decompress/zstd_ddict.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/decompress/zstd_decompress.c +14 -9
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/decompress/zstd_decompress_block.c +20 -9
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/decompress/zstd_decompress_block.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/decompress/zstd_decompress_internal.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/lib/zstd.h +53 -21
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/zlibWrapper/gzclose.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/zlibWrapper/gzcompatibility.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/zlibWrapper/gzguts.h +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/zlibWrapper/gzlib.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/zlibWrapper/gzread.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/zlibWrapper/gzwrite.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/zlibWrapper/zstd_zlibwrapper.c +0 -0
- data/ext/zstdlib/{zstd-1.4.0 → zstd-1.4.2}/zlibWrapper/zstd_zlibwrapper.h +0 -0
- metadata +62 -59
@@ -0,0 +1,47 @@
|
|
1
|
+
/*
|
2
|
+
* Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
|
3
|
+
* All rights reserved.
|
4
|
+
*
|
5
|
+
* This source code is licensed under both the BSD-style license (found in the
|
6
|
+
* LICENSE file in the root directory of this source tree) and the GPLv2 (found
|
7
|
+
* in the COPYING file in the root directory of this source tree).
|
8
|
+
* You may select, at your option, one of the above-listed licenses.
|
9
|
+
*/
|
10
|
+
|
11
|
+
#ifndef ZSTD_COMPRESS_SEQUENCES_H
|
12
|
+
#define ZSTD_COMPRESS_SEQUENCES_H
|
13
|
+
|
14
|
+
#include "fse.h" /* FSE_repeat, FSE_CTable */
|
15
|
+
#include "zstd_internal.h" /* symbolEncodingType_e, ZSTD_strategy */
|
16
|
+
|
17
|
+
typedef enum {
|
18
|
+
ZSTD_defaultDisallowed = 0,
|
19
|
+
ZSTD_defaultAllowed = 1
|
20
|
+
} ZSTD_defaultPolicy_e;
|
21
|
+
|
22
|
+
symbolEncodingType_e
|
23
|
+
ZSTD_selectEncodingType(
|
24
|
+
FSE_repeat* repeatMode, unsigned const* count, unsigned const max,
|
25
|
+
size_t const mostFrequent, size_t nbSeq, unsigned const FSELog,
|
26
|
+
FSE_CTable const* prevCTable,
|
27
|
+
short const* defaultNorm, U32 defaultNormLog,
|
28
|
+
ZSTD_defaultPolicy_e const isDefaultAllowed,
|
29
|
+
ZSTD_strategy const strategy);
|
30
|
+
|
31
|
+
size_t
|
32
|
+
ZSTD_buildCTable(void* dst, size_t dstCapacity,
|
33
|
+
FSE_CTable* nextCTable, U32 FSELog, symbolEncodingType_e type,
|
34
|
+
unsigned* count, U32 max,
|
35
|
+
const BYTE* codeTable, size_t nbSeq,
|
36
|
+
const S16* defaultNorm, U32 defaultNormLog, U32 defaultMax,
|
37
|
+
const FSE_CTable* prevCTable, size_t prevCTableSize,
|
38
|
+
void* workspace, size_t workspaceSize);
|
39
|
+
|
40
|
+
size_t ZSTD_encodeSequences(
|
41
|
+
void* dst, size_t dstCapacity,
|
42
|
+
FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
|
43
|
+
FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
|
44
|
+
FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
|
45
|
+
seqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2);
|
46
|
+
|
47
|
+
#endif /* ZSTD_COMPRESS_SEQUENCES_H */
|
@@ -43,8 +43,7 @@ void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms,
|
|
43
43
|
/* Only load extra positions for ZSTD_dtlm_full */
|
44
44
|
if (dtlm == ZSTD_dtlm_fast)
|
45
45
|
break;
|
46
|
-
|
47
|
-
}
|
46
|
+
} }
|
48
47
|
}
|
49
48
|
|
50
49
|
|
@@ -63,7 +62,10 @@ size_t ZSTD_compressBlock_doubleFast_generic(
|
|
63
62
|
const BYTE* const istart = (const BYTE*)src;
|
64
63
|
const BYTE* ip = istart;
|
65
64
|
const BYTE* anchor = istart;
|
66
|
-
const U32
|
65
|
+
const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
|
66
|
+
const U32 lowestValid = ms->window.dictLimit;
|
67
|
+
const U32 maxDistance = 1U << cParams->windowLog;
|
68
|
+
const U32 prefixLowestIndex = (endIndex - lowestValid > maxDistance) ? endIndex - maxDistance : lowestValid;
|
67
69
|
const BYTE* const prefixLowest = base + prefixLowestIndex;
|
68
70
|
const BYTE* const iend = istart + srcSize;
|
69
71
|
const BYTE* const ilimit = iend - HASH_READ_SIZE;
|
@@ -95,8 +97,15 @@ size_t ZSTD_compressBlock_doubleFast_generic(
|
|
95
97
|
dictCParams->chainLog : hBitsS;
|
96
98
|
const U32 dictAndPrefixLength = (U32)(ip - prefixLowest + dictEnd - dictStart);
|
97
99
|
|
100
|
+
DEBUGLOG(5, "ZSTD_compressBlock_doubleFast_generic");
|
101
|
+
|
98
102
|
assert(dictMode == ZSTD_noDict || dictMode == ZSTD_dictMatchState);
|
99
103
|
|
104
|
+
/* if a dictionary is attached, it must be within window range */
|
105
|
+
if (dictMode == ZSTD_dictMatchState) {
|
106
|
+
assert(lowestValid + maxDistance >= endIndex);
|
107
|
+
}
|
108
|
+
|
100
109
|
/* init */
|
101
110
|
ip += (dictAndPrefixLength == 0);
|
102
111
|
if (dictMode == ZSTD_noDict) {
|
@@ -138,7 +147,7 @@ size_t ZSTD_compressBlock_doubleFast_generic(
|
|
138
147
|
const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
|
139
148
|
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
|
140
149
|
ip++;
|
141
|
-
ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
|
150
|
+
ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
|
142
151
|
goto _match_stored;
|
143
152
|
}
|
144
153
|
|
@@ -147,7 +156,7 @@ size_t ZSTD_compressBlock_doubleFast_generic(
|
|
147
156
|
&& ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1)))) {
|
148
157
|
mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
|
149
158
|
ip++;
|
150
|
-
ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
|
159
|
+
ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
|
151
160
|
goto _match_stored;
|
152
161
|
}
|
153
162
|
|
@@ -170,8 +179,7 @@ size_t ZSTD_compressBlock_doubleFast_generic(
|
|
170
179
|
offset = (U32)(current - dictMatchIndexL - dictIndexDelta);
|
171
180
|
while (((ip>anchor) & (dictMatchL>dictStart)) && (ip[-1] == dictMatchL[-1])) { ip--; dictMatchL--; mLength++; } /* catch up */
|
172
181
|
goto _match_found;
|
173
|
-
|
174
|
-
}
|
182
|
+
} }
|
175
183
|
|
176
184
|
if (matchIndexS > prefixLowestIndex) {
|
177
185
|
/* check prefix short match */
|
@@ -186,16 +194,14 @@ size_t ZSTD_compressBlock_doubleFast_generic(
|
|
186
194
|
|
187
195
|
if (match > dictStart && MEM_read32(match) == MEM_read32(ip)) {
|
188
196
|
goto _search_next_long;
|
189
|
-
|
190
|
-
}
|
197
|
+
} }
|
191
198
|
|
192
199
|
ip += ((ip-anchor) >> kSearchStrength) + 1;
|
193
200
|
continue;
|
194
201
|
|
195
202
|
_search_next_long:
|
196
203
|
|
197
|
-
{
|
198
|
-
size_t const hl3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
|
204
|
+
{ size_t const hl3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
|
199
205
|
size_t const dictHLNext = ZSTD_hashPtr(ip+1, dictHBitsL, 8);
|
200
206
|
U32 const matchIndexL3 = hashLong[hl3];
|
201
207
|
const BYTE* matchL3 = base + matchIndexL3;
|
@@ -221,9 +227,7 @@ _search_next_long:
|
|
221
227
|
offset = (U32)(current + 1 - dictMatchIndexL3 - dictIndexDelta);
|
222
228
|
while (((ip>anchor) & (dictMatchL3>dictStart)) && (ip[-1] == dictMatchL3[-1])) { ip--; dictMatchL3--; mLength++; } /* catch up */
|
223
229
|
goto _match_found;
|
224
|
-
|
225
|
-
}
|
226
|
-
}
|
230
|
+
} } }
|
227
231
|
|
228
232
|
/* if no long +1 match, explore the short match we found */
|
229
233
|
if (dictMode == ZSTD_dictMatchState && matchIndexS < prefixLowestIndex) {
|
@@ -242,7 +246,7 @@ _match_found:
|
|
242
246
|
offset_2 = offset_1;
|
243
247
|
offset_1 = offset;
|
244
248
|
|
245
|
-
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
249
|
+
ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
246
250
|
|
247
251
|
_match_stored:
|
248
252
|
/* match found */
|
@@ -250,11 +254,14 @@ _match_stored:
|
|
250
254
|
anchor = ip;
|
251
255
|
|
252
256
|
if (ip <= ilimit) {
|
253
|
-
/*
|
254
|
-
|
255
|
-
|
256
|
-
|
257
|
-
|
257
|
+
/* Complementary insertion */
|
258
|
+
/* done after iLimit test, as candidates could be > iend-8 */
|
259
|
+
{ U32 const indexToInsert = current+2;
|
260
|
+
hashLong[ZSTD_hashPtr(base+indexToInsert, hBitsL, 8)] = indexToInsert;
|
261
|
+
hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
|
262
|
+
hashSmall[ZSTD_hashPtr(base+indexToInsert, hBitsS, mls)] = indexToInsert;
|
263
|
+
hashSmall[ZSTD_hashPtr(ip-1, hBitsS, mls)] = (U32)(ip-1-base);
|
264
|
+
}
|
258
265
|
|
259
266
|
/* check immediate repcode */
|
260
267
|
if (dictMode == ZSTD_dictMatchState) {
|
@@ -278,8 +285,7 @@ _match_stored:
|
|
278
285
|
continue;
|
279
286
|
}
|
280
287
|
break;
|
281
|
-
|
282
|
-
}
|
288
|
+
} }
|
283
289
|
|
284
290
|
if (dictMode == ZSTD_noDict) {
|
285
291
|
while ( (ip <= ilimit)
|
@@ -294,14 +300,15 @@ _match_stored:
|
|
294
300
|
ip += rLength;
|
295
301
|
anchor = ip;
|
296
302
|
continue; /* faster when present ... (?) */
|
297
|
-
|
303
|
+
} } }
|
304
|
+
} /* while (ip < ilimit) */
|
298
305
|
|
299
306
|
/* save reps for next block */
|
300
307
|
rep[0] = offset_1 ? offset_1 : offsetSaved;
|
301
308
|
rep[1] = offset_2 ? offset_2 : offsetSaved;
|
302
309
|
|
303
310
|
/* Return the last literals size */
|
304
|
-
return iend - anchor;
|
311
|
+
return (size_t)(iend - anchor);
|
305
312
|
}
|
306
313
|
|
307
314
|
|
@@ -360,10 +367,15 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
|
|
360
367
|
const BYTE* anchor = istart;
|
361
368
|
const BYTE* const iend = istart + srcSize;
|
362
369
|
const BYTE* const ilimit = iend - 8;
|
363
|
-
const U32 prefixStartIndex = ms->window.dictLimit;
|
364
370
|
const BYTE* const base = ms->window.base;
|
371
|
+
const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
|
372
|
+
const U32 maxDistance = 1U << cParams->windowLog;
|
373
|
+
const U32 lowestValid = ms->window.lowLimit;
|
374
|
+
const U32 lowLimit = (endIndex - lowestValid > maxDistance) ? endIndex - maxDistance : lowestValid;
|
375
|
+
const U32 dictStartIndex = lowLimit;
|
376
|
+
const U32 dictLimit = ms->window.dictLimit;
|
377
|
+
const U32 prefixStartIndex = (dictLimit > lowLimit) ? dictLimit : lowLimit;
|
365
378
|
const BYTE* const prefixStart = base + prefixStartIndex;
|
366
|
-
const U32 dictStartIndex = ms->window.lowLimit;
|
367
379
|
const BYTE* const dictBase = ms->window.dictBase;
|
368
380
|
const BYTE* const dictStart = dictBase + dictStartIndex;
|
369
381
|
const BYTE* const dictEnd = dictBase + prefixStartIndex;
|
@@ -371,6 +383,10 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
|
|
371
383
|
|
372
384
|
DEBUGLOG(5, "ZSTD_compressBlock_doubleFast_extDict_generic (srcSize=%zu)", srcSize);
|
373
385
|
|
386
|
+
/* if extDict is invalidated due to maxDistance, switch to "regular" variant */
|
387
|
+
if (prefixStartIndex == dictStartIndex)
|
388
|
+
return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, mls, ZSTD_noDict);
|
389
|
+
|
374
390
|
/* Search Loop */
|
375
391
|
while (ip < ilimit) { /* < instead of <=, because (ip+1) */
|
376
392
|
const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls);
|
@@ -396,7 +412,7 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
|
|
396
412
|
const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
|
397
413
|
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
|
398
414
|
ip++;
|
399
|
-
ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
|
415
|
+
ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
|
400
416
|
} else {
|
401
417
|
if ((matchLongIndex > dictStartIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {
|
402
418
|
const BYTE* const matchEnd = matchLongIndex < prefixStartIndex ? dictEnd : iend;
|
@@ -407,7 +423,7 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
|
|
407
423
|
while (((ip>anchor) & (matchLong>lowMatchPtr)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */
|
408
424
|
offset_2 = offset_1;
|
409
425
|
offset_1 = offset;
|
410
|
-
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
426
|
+
ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
411
427
|
|
412
428
|
} else if ((matchIndex > dictStartIndex) && (MEM_read32(match) == MEM_read32(ip))) {
|
413
429
|
size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
|
@@ -432,23 +448,27 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
|
|
432
448
|
}
|
433
449
|
offset_2 = offset_1;
|
434
450
|
offset_1 = offset;
|
435
|
-
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
451
|
+
ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
436
452
|
|
437
453
|
} else {
|
438
454
|
ip += ((ip-anchor) >> kSearchStrength) + 1;
|
439
455
|
continue;
|
440
456
|
} }
|
441
457
|
|
442
|
-
/*
|
458
|
+
/* move to next sequence start */
|
443
459
|
ip += mLength;
|
444
460
|
anchor = ip;
|
445
461
|
|
446
462
|
if (ip <= ilimit) {
|
447
|
-
/*
|
448
|
-
|
449
|
-
|
450
|
-
|
451
|
-
|
463
|
+
/* Complementary insertion */
|
464
|
+
/* done after iLimit test, as candidates could be > iend-8 */
|
465
|
+
{ U32 const indexToInsert = current+2;
|
466
|
+
hashLong[ZSTD_hashPtr(base+indexToInsert, hBitsL, 8)] = indexToInsert;
|
467
|
+
hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
|
468
|
+
hashSmall[ZSTD_hashPtr(base+indexToInsert, hBitsS, mls)] = indexToInsert;
|
469
|
+
hashSmall[ZSTD_hashPtr(ip-1, hBitsS, mls)] = (U32)(ip-1-base);
|
470
|
+
}
|
471
|
+
|
452
472
|
/* check immediate repcode */
|
453
473
|
while (ip <= ilimit) {
|
454
474
|
U32 const current2 = (U32)(ip-base);
|
@@ -475,7 +495,7 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
|
|
475
495
|
rep[1] = offset_2;
|
476
496
|
|
477
497
|
/* Return the last literals size */
|
478
|
-
return iend - anchor;
|
498
|
+
return (size_t)(iend - anchor);
|
479
499
|
}
|
480
500
|
|
481
501
|
|
File without changes
|
@@ -13,7 +13,8 @@
|
|
13
13
|
|
14
14
|
|
15
15
|
void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
|
16
|
-
void const
|
16
|
+
const void* const end,
|
17
|
+
ZSTD_dictTableLoadMethod_e dtlm)
|
17
18
|
{
|
18
19
|
const ZSTD_compressionParameters* const cParams = &ms->cParams;
|
19
20
|
U32* const hashTable = ms->hashTable;
|
@@ -41,6 +42,7 @@ void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
|
|
41
42
|
} } } }
|
42
43
|
}
|
43
44
|
|
45
|
+
|
44
46
|
FORCE_INLINE_TEMPLATE
|
45
47
|
size_t ZSTD_compressBlock_fast_generic(
|
46
48
|
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
|
@@ -58,7 +60,10 @@ size_t ZSTD_compressBlock_fast_generic(
|
|
58
60
|
const BYTE* ip0 = istart;
|
59
61
|
const BYTE* ip1;
|
60
62
|
const BYTE* anchor = istart;
|
61
|
-
const U32
|
63
|
+
const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
|
64
|
+
const U32 maxDistance = 1U << cParams->windowLog;
|
65
|
+
const U32 validStartIndex = ms->window.dictLimit;
|
66
|
+
const U32 prefixStartIndex = (endIndex - validStartIndex > maxDistance) ? endIndex - maxDistance : validStartIndex;
|
62
67
|
const BYTE* const prefixStart = base + prefixStartIndex;
|
63
68
|
const BYTE* const iend = istart + srcSize;
|
64
69
|
const BYTE* const ilimit = iend - HASH_READ_SIZE;
|
@@ -165,7 +170,7 @@ _match: /* Requires: ip0, match0, offcode */
|
|
165
170
|
rep[1] = offset_2 ? offset_2 : offsetSaved;
|
166
171
|
|
167
172
|
/* Return the last literals size */
|
168
|
-
return iend - anchor;
|
173
|
+
return (size_t)(iend - anchor);
|
169
174
|
}
|
170
175
|
|
171
176
|
|
@@ -222,8 +227,15 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
|
|
222
227
|
const U32 dictAndPrefixLength = (U32)(ip - prefixStart + dictEnd - dictStart);
|
223
228
|
const U32 dictHLog = dictCParams->hashLog;
|
224
229
|
|
225
|
-
/*
|
226
|
-
*
|
230
|
+
/* if a dictionary is still attached, it necessarily means that
|
231
|
+
* it is within window size. So we just check it. */
|
232
|
+
const U32 maxDistance = 1U << cParams->windowLog;
|
233
|
+
const U32 endIndex = (U32)((size_t)(ip - base) + srcSize);
|
234
|
+
assert(endIndex - prefixStartIndex <= maxDistance);
|
235
|
+
(void)maxDistance; (void)endIndex; /* these variables are not used when assert() is disabled */
|
236
|
+
|
237
|
+
/* ensure there will be no no underflow
|
238
|
+
* when translating a dict index into a local index */
|
227
239
|
assert(prefixStartIndex >= (U32)(dictEnd - dictBase));
|
228
240
|
|
229
241
|
/* init */
|
@@ -251,7 +263,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
|
|
251
263
|
const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
|
252
264
|
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
|
253
265
|
ip++;
|
254
|
-
ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
|
266
|
+
ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
|
255
267
|
} else if ( (matchIndex <= prefixStartIndex) ) {
|
256
268
|
size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls);
|
257
269
|
U32 const dictMatchIndex = dictHashTable[dictHash];
|
@@ -271,7 +283,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
|
|
271
283
|
} /* catch up */
|
272
284
|
offset_2 = offset_1;
|
273
285
|
offset_1 = offset;
|
274
|
-
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
286
|
+
ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
275
287
|
}
|
276
288
|
} else if (MEM_read32(match) != MEM_read32(ip)) {
|
277
289
|
/* it's not a match, and we're not going to check the dictionary */
|
@@ -286,7 +298,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
|
|
286
298
|
&& (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
|
287
299
|
offset_2 = offset_1;
|
288
300
|
offset_1 = offset;
|
289
|
-
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
301
|
+
ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
290
302
|
}
|
291
303
|
|
292
304
|
/* match found */
|
@@ -327,7 +339,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
|
|
327
339
|
rep[1] = offset_2 ? offset_2 : offsetSaved;
|
328
340
|
|
329
341
|
/* Return the last literals size */
|
330
|
-
return iend - anchor;
|
342
|
+
return (size_t)(iend - anchor);
|
331
343
|
}
|
332
344
|
|
333
345
|
size_t ZSTD_compressBlock_fast_dictMatchState(
|
@@ -366,15 +378,24 @@ static size_t ZSTD_compressBlock_fast_extDict_generic(
|
|
366
378
|
const BYTE* const istart = (const BYTE*)src;
|
367
379
|
const BYTE* ip = istart;
|
368
380
|
const BYTE* anchor = istart;
|
369
|
-
const U32
|
381
|
+
const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
|
382
|
+
const U32 maxDistance = 1U << cParams->windowLog;
|
383
|
+
const U32 validLow = ms->window.lowLimit;
|
384
|
+
const U32 lowLimit = (endIndex - validLow > maxDistance) ? endIndex - maxDistance : validLow;
|
385
|
+
const U32 dictStartIndex = lowLimit;
|
370
386
|
const BYTE* const dictStart = dictBase + dictStartIndex;
|
371
|
-
const U32
|
387
|
+
const U32 dictLimit = ms->window.dictLimit;
|
388
|
+
const U32 prefixStartIndex = dictLimit < lowLimit ? lowLimit : dictLimit;
|
372
389
|
const BYTE* const prefixStart = base + prefixStartIndex;
|
373
390
|
const BYTE* const dictEnd = dictBase + prefixStartIndex;
|
374
391
|
const BYTE* const iend = istart + srcSize;
|
375
392
|
const BYTE* const ilimit = iend - 8;
|
376
393
|
U32 offset_1=rep[0], offset_2=rep[1];
|
377
394
|
|
395
|
+
/* switch to "regular" variant if extDict is invalidated due to maxDistance */
|
396
|
+
if (prefixStartIndex == dictStartIndex)
|
397
|
+
return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, mls);
|
398
|
+
|
378
399
|
/* Search Loop */
|
379
400
|
while (ip < ilimit) { /* < instead of <=, because (ip+1) */
|
380
401
|
const size_t h = ZSTD_hashPtr(ip, hlog, mls);
|
@@ -394,7 +415,7 @@ static size_t ZSTD_compressBlock_fast_extDict_generic(
|
|
394
415
|
const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
|
395
416
|
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
|
396
417
|
ip++;
|
397
|
-
ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
|
418
|
+
ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
|
398
419
|
} else {
|
399
420
|
if ( (matchIndex < dictStartIndex) ||
|
400
421
|
(MEM_read32(match) != MEM_read32(ip)) ) {
|
@@ -410,7 +431,7 @@ static size_t ZSTD_compressBlock_fast_extDict_generic(
|
|
410
431
|
offset = current - matchIndex;
|
411
432
|
offset_2 = offset_1;
|
412
433
|
offset_1 = offset;
|
413
|
-
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
434
|
+
ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
|
414
435
|
} }
|
415
436
|
|
416
437
|
/* found a match : store it */
|
@@ -445,7 +466,7 @@ static size_t ZSTD_compressBlock_fast_extDict_generic(
|
|
445
466
|
rep[1] = offset_2;
|
446
467
|
|
447
468
|
/* Return the last literals size */
|
448
|
-
return iend - anchor;
|
469
|
+
return (size_t)(iend - anchor);
|
449
470
|
}
|
450
471
|
|
451
472
|
|
File without changes
|
@@ -83,7 +83,10 @@ ZSTD_insertDUBT1(ZSTD_matchState_t* ms,
|
|
83
83
|
U32* largerPtr = smallerPtr + 1;
|
84
84
|
U32 matchIndex = *smallerPtr; /* this candidate is unsorted : next sorted candidate is reached through *smallerPtr, while *largerPtr contains previous unsorted candidate (which is already saved and can be overwritten) */
|
85
85
|
U32 dummy32; /* to be nullified at the end */
|
86
|
-
U32 const
|
86
|
+
U32 const windowValid = ms->window.lowLimit;
|
87
|
+
U32 const maxDistance = 1U << cParams->windowLog;
|
88
|
+
U32 const windowLow = (current - windowValid > maxDistance) ? current - maxDistance : windowValid;
|
89
|
+
|
87
90
|
|
88
91
|
DEBUGLOG(8, "ZSTD_insertDUBT1(%u) (dictLimit=%u, lowLimit=%u)",
|
89
92
|
current, dictLimit, windowLow);
|
@@ -239,7 +242,9 @@ ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms,
|
|
239
242
|
|
240
243
|
const BYTE* const base = ms->window.base;
|
241
244
|
U32 const current = (U32)(ip-base);
|
242
|
-
U32 const
|
245
|
+
U32 const maxDistance = 1U << cParams->windowLog;
|
246
|
+
U32 const windowValid = ms->window.lowLimit;
|
247
|
+
U32 const windowLow = (current - windowValid > maxDistance) ? current - maxDistance : windowValid;
|
243
248
|
|
244
249
|
U32* const bt = ms->chainTable;
|
245
250
|
U32 const btLog = cParams->chainLog - 1;
|
@@ -490,8 +495,10 @@ size_t ZSTD_HcFindBestMatch_generic (
|
|
490
495
|
const U32 dictLimit = ms->window.dictLimit;
|
491
496
|
const BYTE* const prefixStart = base + dictLimit;
|
492
497
|
const BYTE* const dictEnd = dictBase + dictLimit;
|
493
|
-
const U32 lowLimit = ms->window.lowLimit;
|
494
498
|
const U32 current = (U32)(ip-base);
|
499
|
+
const U32 maxDistance = 1U << cParams->windowLog;
|
500
|
+
const U32 lowValid = ms->window.lowLimit;
|
501
|
+
const U32 lowLimit = (current - lowValid > maxDistance) ? current - maxDistance : lowValid;
|
495
502
|
const U32 minChain = current > chainSize ? current - chainSize : 0;
|
496
503
|
U32 nbAttempts = 1U << cParams->searchLog;
|
497
504
|
size_t ml=4-1;
|
@@ -653,7 +660,6 @@ size_t ZSTD_compressBlock_lazy_generic(
|
|
653
660
|
|
654
661
|
/* init */
|
655
662
|
ip += (dictAndPrefixLength == 0);
|
656
|
-
ms->nextToUpdate3 = ms->nextToUpdate;
|
657
663
|
if (dictMode == ZSTD_noDict) {
|
658
664
|
U32 const maxRep = (U32)(ip - prefixLowest);
|
659
665
|
if (offset_2 > maxRep) savedOffset = offset_2, offset_2 = 0;
|
@@ -933,7 +939,6 @@ size_t ZSTD_compressBlock_lazy_extDict_generic(
|
|
933
939
|
U32 offset_1 = rep[0], offset_2 = rep[1];
|
934
940
|
|
935
941
|
/* init */
|
936
|
-
ms->nextToUpdate3 = ms->nextToUpdate;
|
937
942
|
ip += (ip == prefixStart);
|
938
943
|
|
939
944
|
/* Match Loop */
|