zstd-ruby 1.5.2.2 → 1.5.4.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/README.md +15 -3
- data/ext/zstdruby/common.h +7 -0
- data/ext/zstdruby/libzstd/common/bits.h +175 -0
- data/ext/zstdruby/libzstd/common/bitstream.h +18 -59
- data/ext/zstdruby/libzstd/common/compiler.h +22 -3
- data/ext/zstdruby/libzstd/common/cpu.h +1 -1
- data/ext/zstdruby/libzstd/common/debug.c +1 -1
- data/ext/zstdruby/libzstd/common/debug.h +1 -1
- data/ext/zstdruby/libzstd/common/entropy_common.c +12 -40
- data/ext/zstdruby/libzstd/common/error_private.c +9 -2
- data/ext/zstdruby/libzstd/common/error_private.h +1 -1
- data/ext/zstdruby/libzstd/common/fse.h +5 -83
- data/ext/zstdruby/libzstd/common/fse_decompress.c +7 -99
- data/ext/zstdruby/libzstd/common/huf.h +65 -156
- data/ext/zstdruby/libzstd/common/mem.h +39 -46
- data/ext/zstdruby/libzstd/common/pool.c +26 -10
- data/ext/zstdruby/libzstd/common/pool.h +7 -1
- data/ext/zstdruby/libzstd/common/portability_macros.h +22 -3
- data/ext/zstdruby/libzstd/common/threading.c +68 -14
- data/ext/zstdruby/libzstd/common/threading.h +5 -10
- data/ext/zstdruby/libzstd/common/xxhash.c +2 -2
- data/ext/zstdruby/libzstd/common/xxhash.h +8 -8
- data/ext/zstdruby/libzstd/common/zstd_common.c +1 -1
- data/ext/zstdruby/libzstd/common/zstd_deps.h +1 -1
- data/ext/zstdruby/libzstd/common/zstd_internal.h +17 -113
- data/ext/zstdruby/libzstd/common/zstd_trace.h +3 -3
- data/ext/zstdruby/libzstd/compress/clevels.h +1 -1
- data/ext/zstdruby/libzstd/compress/fse_compress.c +7 -124
- data/ext/zstdruby/libzstd/compress/hist.c +1 -1
- data/ext/zstdruby/libzstd/compress/hist.h +1 -1
- data/ext/zstdruby/libzstd/compress/huf_compress.c +234 -169
- data/ext/zstdruby/libzstd/compress/zstd_compress.c +1055 -455
- data/ext/zstdruby/libzstd/compress/zstd_compress_internal.h +165 -145
- data/ext/zstdruby/libzstd/compress/zstd_compress_literals.c +115 -39
- data/ext/zstdruby/libzstd/compress/zstd_compress_literals.h +16 -8
- data/ext/zstdruby/libzstd/compress/zstd_compress_sequences.c +3 -3
- data/ext/zstdruby/libzstd/compress/zstd_compress_sequences.h +1 -1
- data/ext/zstdruby/libzstd/compress/zstd_compress_superblock.c +25 -21
- data/ext/zstdruby/libzstd/compress/zstd_compress_superblock.h +1 -1
- data/ext/zstdruby/libzstd/compress/zstd_cwksp.h +5 -3
- data/ext/zstdruby/libzstd/compress/zstd_double_fast.c +95 -33
- data/ext/zstdruby/libzstd/compress/zstd_double_fast.h +3 -2
- data/ext/zstdruby/libzstd/compress/zstd_fast.c +433 -148
- data/ext/zstdruby/libzstd/compress/zstd_fast.h +3 -2
- data/ext/zstdruby/libzstd/compress/zstd_lazy.c +306 -283
- data/ext/zstdruby/libzstd/compress/zstd_lazy.h +4 -2
- data/ext/zstdruby/libzstd/compress/zstd_ldm.c +5 -5
- data/ext/zstdruby/libzstd/compress/zstd_ldm.h +1 -1
- data/ext/zstdruby/libzstd/compress/zstd_ldm_geartab.h +1 -1
- data/ext/zstdruby/libzstd/compress/zstd_opt.c +104 -80
- data/ext/zstdruby/libzstd/compress/zstd_opt.h +1 -1
- data/ext/zstdruby/libzstd/compress/zstdmt_compress.c +12 -5
- data/ext/zstdruby/libzstd/compress/zstdmt_compress.h +1 -1
- data/ext/zstdruby/libzstd/decompress/huf_decompress.c +434 -441
- data/ext/zstdruby/libzstd/decompress/huf_decompress_amd64.S +30 -39
- data/ext/zstdruby/libzstd/decompress/zstd_ddict.c +3 -4
- data/ext/zstdruby/libzstd/decompress/zstd_ddict.h +1 -1
- data/ext/zstdruby/libzstd/decompress/zstd_decompress.c +164 -42
- data/ext/zstdruby/libzstd/decompress/zstd_decompress_block.c +186 -65
- data/ext/zstdruby/libzstd/decompress/zstd_decompress_block.h +1 -1
- data/ext/zstdruby/libzstd/decompress/zstd_decompress_internal.h +4 -2
- data/ext/zstdruby/libzstd/dictBuilder/cover.c +19 -15
- data/ext/zstdruby/libzstd/dictBuilder/cover.h +1 -1
- data/ext/zstdruby/libzstd/dictBuilder/fastcover.c +2 -2
- data/ext/zstdruby/libzstd/dictBuilder/zdict.c +9 -87
- data/ext/zstdruby/libzstd/zdict.h +53 -31
- data/ext/zstdruby/libzstd/zstd.h +489 -90
- data/ext/zstdruby/libzstd/zstd_errors.h +27 -8
- data/ext/zstdruby/main.c +4 -0
- data/ext/zstdruby/streaming_compress.c +1 -7
- data/ext/zstdruby/zstdruby.c +110 -26
- data/lib/zstd-ruby/version.rb +1 -1
- data/lib/zstd-ruby.rb +0 -1
- metadata +7 -6
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 5025a83454e6841492b428638df26d5bbacfae564e3e0dcaabce5955e2bd3688
|
4
|
+
data.tar.gz: f50bf1bb09958e116917344fddaa8bf1d45088894769d791bbe3664a74a3b907
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: c754293689a65269f3f262605ad07cd72c0e162dde804dab2d0e144b67c251cbeaf5306db451ac15878388113617d09aaef5c065349c893b233804d6f0883812
|
7
|
+
data.tar.gz: f2bc576e57766984c1d81e05f69a499461ffc437383f084d6cd83f22f3621f368a9ee668ad786165d537780112b42ade0c7d3e43b8fb35ac77634fbbad12d5c8
|
data/README.md
CHANGED
@@ -41,8 +41,14 @@ compressed_data = Zstd.compress(data)
|
|
41
41
|
compressed_data = Zstd.compress(data, complession_level) # default compression_level is 0
|
42
42
|
```
|
43
43
|
|
44
|
-
###
|
44
|
+
### Compression using Dictionary
|
45
|
+
```ruby
|
46
|
+
# dictionary is supposed to have been created using `zstd --train`
|
47
|
+
compressed_using_dict = Zstd.compress_using_dict("", IO.read('dictionary_file'))
|
45
48
|
```
|
49
|
+
|
50
|
+
### Streaming Compression
|
51
|
+
```ruby
|
46
52
|
stream = Zstd::StreamingCompress.new
|
47
53
|
stream << "abc" << "def"
|
48
54
|
res = stream.flush
|
@@ -52,7 +58,7 @@ res << stream.finish
|
|
52
58
|
|
53
59
|
or
|
54
60
|
|
55
|
-
```
|
61
|
+
```ruby
|
56
62
|
stream = Zstd::StreamingCompress.new
|
57
63
|
res = stream.compress("abc")
|
58
64
|
res << stream.flush
|
@@ -66,8 +72,14 @@ res << stream.finish
|
|
66
72
|
data = Zstd.decompress(compressed_data)
|
67
73
|
```
|
68
74
|
|
69
|
-
###
|
75
|
+
### Decomporession using Dictionary
|
76
|
+
```ruby
|
77
|
+
# dictionary is supposed to have been created using `zstd --train`
|
78
|
+
Zstd.decompress_using_dict(compressed_using_dict, IO.read('dictionary_file'))
|
70
79
|
```
|
80
|
+
|
81
|
+
### Streaming Decompression
|
82
|
+
```ruby
|
71
83
|
cstr = "" # Compressed data
|
72
84
|
stream = Zstd::StreamingDecompress.new
|
73
85
|
result = ''
|
data/ext/zstdruby/common.h
CHANGED
@@ -4,5 +4,12 @@
|
|
4
4
|
#include "ruby.h"
|
5
5
|
#include "./libzstd/zstd.h"
|
6
6
|
|
7
|
+
static int convert_compression_level(VALUE compression_level_value)
|
8
|
+
{
|
9
|
+
if (NIL_P(compression_level_value)) {
|
10
|
+
return ZSTD_CLEVEL_DEFAULT;
|
11
|
+
}
|
12
|
+
return NUM2INT(compression_level_value);
|
13
|
+
}
|
7
14
|
|
8
15
|
#endif /* ZSTD_RUBY_H */
|
@@ -0,0 +1,175 @@
|
|
1
|
+
/*
|
2
|
+
* Copyright (c) Meta Platforms, Inc. and affiliates.
|
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_BITS_H
|
12
|
+
#define ZSTD_BITS_H
|
13
|
+
|
14
|
+
#include "mem.h"
|
15
|
+
|
16
|
+
MEM_STATIC unsigned ZSTD_countTrailingZeros32_fallback(U32 val)
|
17
|
+
{
|
18
|
+
assert(val != 0);
|
19
|
+
{
|
20
|
+
static const int DeBruijnBytePos[32] = {0, 1, 28, 2, 29, 14, 24, 3,
|
21
|
+
30, 22, 20, 15, 25, 17, 4, 8,
|
22
|
+
31, 27, 13, 23, 21, 19, 16, 7,
|
23
|
+
26, 12, 18, 6, 11, 5, 10, 9};
|
24
|
+
return DeBruijnBytePos[((U32) ((val & -(S32) val) * 0x077CB531U)) >> 27];
|
25
|
+
}
|
26
|
+
}
|
27
|
+
|
28
|
+
MEM_STATIC unsigned ZSTD_countTrailingZeros32(U32 val)
|
29
|
+
{
|
30
|
+
assert(val != 0);
|
31
|
+
# if defined(_MSC_VER)
|
32
|
+
# if STATIC_BMI2 == 1
|
33
|
+
return _tzcnt_u32(val);
|
34
|
+
# else
|
35
|
+
if (val != 0) {
|
36
|
+
unsigned long r;
|
37
|
+
_BitScanForward(&r, val);
|
38
|
+
return (unsigned)r;
|
39
|
+
} else {
|
40
|
+
/* Should not reach this code path */
|
41
|
+
__assume(0);
|
42
|
+
}
|
43
|
+
# endif
|
44
|
+
# elif defined(__GNUC__) && (__GNUC__ >= 4)
|
45
|
+
return (unsigned)__builtin_ctz(val);
|
46
|
+
# else
|
47
|
+
return ZSTD_countTrailingZeros32_fallback(val);
|
48
|
+
# endif
|
49
|
+
}
|
50
|
+
|
51
|
+
MEM_STATIC unsigned ZSTD_countLeadingZeros32_fallback(U32 val) {
|
52
|
+
assert(val != 0);
|
53
|
+
{
|
54
|
+
static const U32 DeBruijnClz[32] = {0, 9, 1, 10, 13, 21, 2, 29,
|
55
|
+
11, 14, 16, 18, 22, 25, 3, 30,
|
56
|
+
8, 12, 20, 28, 15, 17, 24, 7,
|
57
|
+
19, 27, 23, 6, 26, 5, 4, 31};
|
58
|
+
val |= val >> 1;
|
59
|
+
val |= val >> 2;
|
60
|
+
val |= val >> 4;
|
61
|
+
val |= val >> 8;
|
62
|
+
val |= val >> 16;
|
63
|
+
return 31 - DeBruijnClz[(val * 0x07C4ACDDU) >> 27];
|
64
|
+
}
|
65
|
+
}
|
66
|
+
|
67
|
+
MEM_STATIC unsigned ZSTD_countLeadingZeros32(U32 val)
|
68
|
+
{
|
69
|
+
assert(val != 0);
|
70
|
+
# if defined(_MSC_VER)
|
71
|
+
# if STATIC_BMI2 == 1
|
72
|
+
return _lzcnt_u32(val);
|
73
|
+
# else
|
74
|
+
if (val != 0) {
|
75
|
+
unsigned long r;
|
76
|
+
_BitScanReverse(&r, val);
|
77
|
+
return (unsigned)(31 - r);
|
78
|
+
} else {
|
79
|
+
/* Should not reach this code path */
|
80
|
+
__assume(0);
|
81
|
+
}
|
82
|
+
# endif
|
83
|
+
# elif defined(__GNUC__) && (__GNUC__ >= 4)
|
84
|
+
return (unsigned)__builtin_clz(val);
|
85
|
+
# else
|
86
|
+
return ZSTD_countLeadingZeros32_fallback(val);
|
87
|
+
# endif
|
88
|
+
}
|
89
|
+
|
90
|
+
MEM_STATIC unsigned ZSTD_countTrailingZeros64(U64 val)
|
91
|
+
{
|
92
|
+
assert(val != 0);
|
93
|
+
# if defined(_MSC_VER) && defined(_WIN64)
|
94
|
+
# if STATIC_BMI2 == 1
|
95
|
+
return _tzcnt_u64(val);
|
96
|
+
# else
|
97
|
+
if (val != 0) {
|
98
|
+
unsigned long r;
|
99
|
+
_BitScanForward64(&r, val);
|
100
|
+
return (unsigned)r;
|
101
|
+
} else {
|
102
|
+
/* Should not reach this code path */
|
103
|
+
__assume(0);
|
104
|
+
}
|
105
|
+
# endif
|
106
|
+
# elif defined(__GNUC__) && (__GNUC__ >= 4) && defined(__LP64__)
|
107
|
+
return (unsigned)__builtin_ctzll(val);
|
108
|
+
# else
|
109
|
+
{
|
110
|
+
U32 mostSignificantWord = (U32)(val >> 32);
|
111
|
+
U32 leastSignificantWord = (U32)val;
|
112
|
+
if (leastSignificantWord == 0) {
|
113
|
+
return 32 + ZSTD_countTrailingZeros32(mostSignificantWord);
|
114
|
+
} else {
|
115
|
+
return ZSTD_countTrailingZeros32(leastSignificantWord);
|
116
|
+
}
|
117
|
+
}
|
118
|
+
# endif
|
119
|
+
}
|
120
|
+
|
121
|
+
MEM_STATIC unsigned ZSTD_countLeadingZeros64(U64 val)
|
122
|
+
{
|
123
|
+
assert(val != 0);
|
124
|
+
# if defined(_MSC_VER) && defined(_WIN64)
|
125
|
+
# if STATIC_BMI2 == 1
|
126
|
+
return _lzcnt_u64(val);
|
127
|
+
# else
|
128
|
+
if (val != 0) {
|
129
|
+
unsigned long r;
|
130
|
+
_BitScanReverse64(&r, val);
|
131
|
+
return (unsigned)(63 - r);
|
132
|
+
} else {
|
133
|
+
/* Should not reach this code path */
|
134
|
+
__assume(0);
|
135
|
+
}
|
136
|
+
# endif
|
137
|
+
# elif defined(__GNUC__) && (__GNUC__ >= 4)
|
138
|
+
return (unsigned)(__builtin_clzll(val));
|
139
|
+
# else
|
140
|
+
{
|
141
|
+
U32 mostSignificantWord = (U32)(val >> 32);
|
142
|
+
U32 leastSignificantWord = (U32)val;
|
143
|
+
if (mostSignificantWord == 0) {
|
144
|
+
return 32 + ZSTD_countLeadingZeros32(leastSignificantWord);
|
145
|
+
} else {
|
146
|
+
return ZSTD_countLeadingZeros32(mostSignificantWord);
|
147
|
+
}
|
148
|
+
}
|
149
|
+
# endif
|
150
|
+
}
|
151
|
+
|
152
|
+
MEM_STATIC unsigned ZSTD_NbCommonBytes(size_t val)
|
153
|
+
{
|
154
|
+
if (MEM_isLittleEndian()) {
|
155
|
+
if (MEM_64bits()) {
|
156
|
+
return ZSTD_countTrailingZeros64((U64)val) >> 3;
|
157
|
+
} else {
|
158
|
+
return ZSTD_countTrailingZeros32((U32)val) >> 3;
|
159
|
+
}
|
160
|
+
} else { /* Big Endian CPU */
|
161
|
+
if (MEM_64bits()) {
|
162
|
+
return ZSTD_countLeadingZeros64((U64)val) >> 3;
|
163
|
+
} else {
|
164
|
+
return ZSTD_countLeadingZeros32((U32)val) >> 3;
|
165
|
+
}
|
166
|
+
}
|
167
|
+
}
|
168
|
+
|
169
|
+
MEM_STATIC unsigned ZSTD_highbit32(U32 val) /* compress, dictBuilder, decodeCorpus */
|
170
|
+
{
|
171
|
+
assert(val != 0);
|
172
|
+
return 31 - ZSTD_countLeadingZeros32(val);
|
173
|
+
}
|
174
|
+
|
175
|
+
#endif /* ZSTD_BITS_H */
|
@@ -1,7 +1,7 @@
|
|
1
1
|
/* ******************************************************************
|
2
2
|
* bitstream
|
3
3
|
* Part of FSE library
|
4
|
-
* Copyright (c)
|
4
|
+
* Copyright (c) Meta Platforms, Inc. and affiliates.
|
5
5
|
*
|
6
6
|
* You can contact the author at :
|
7
7
|
* - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
|
@@ -30,14 +30,15 @@ extern "C" {
|
|
30
30
|
#include "compiler.h" /* UNLIKELY() */
|
31
31
|
#include "debug.h" /* assert(), DEBUGLOG(), RAWLOG() */
|
32
32
|
#include "error_private.h" /* error codes and messages */
|
33
|
+
#include "bits.h" /* ZSTD_highbit32 */
|
33
34
|
|
34
35
|
|
35
36
|
/*=========================================
|
36
37
|
* Target specific
|
37
38
|
=========================================*/
|
38
39
|
#ifndef ZSTD_NO_INTRINSICS
|
39
|
-
# if defined(__BMI__) && defined(__GNUC__)
|
40
|
-
# include <immintrin.h> /* support for bextr (experimental) */
|
40
|
+
# if (defined(__BMI__) || defined(__BMI2__)) && defined(__GNUC__)
|
41
|
+
# include <immintrin.h> /* support for bextr (experimental)/bzhi */
|
41
42
|
# elif defined(__ICCARM__)
|
42
43
|
# include <intrinsics.h>
|
43
44
|
# endif
|
@@ -132,48 +133,6 @@ MEM_STATIC void BIT_flushBitsFast(BIT_CStream_t* bitC);
|
|
132
133
|
MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits);
|
133
134
|
/* faster, but works only if nbBits >= 1 */
|
134
135
|
|
135
|
-
|
136
|
-
|
137
|
-
/*-**************************************************************
|
138
|
-
* Internal functions
|
139
|
-
****************************************************************/
|
140
|
-
MEM_STATIC unsigned BIT_highbit32 (U32 val)
|
141
|
-
{
|
142
|
-
assert(val != 0);
|
143
|
-
{
|
144
|
-
# if defined(_MSC_VER) /* Visual */
|
145
|
-
# if STATIC_BMI2 == 1
|
146
|
-
return _lzcnt_u32(val) ^ 31;
|
147
|
-
# else
|
148
|
-
if (val != 0) {
|
149
|
-
unsigned long r;
|
150
|
-
_BitScanReverse(&r, val);
|
151
|
-
return (unsigned)r;
|
152
|
-
} else {
|
153
|
-
/* Should not reach this code path */
|
154
|
-
__assume(0);
|
155
|
-
}
|
156
|
-
# endif
|
157
|
-
# elif defined(__GNUC__) && (__GNUC__ >= 3) /* Use GCC Intrinsic */
|
158
|
-
return __builtin_clz (val) ^ 31;
|
159
|
-
# elif defined(__ICCARM__) /* IAR Intrinsic */
|
160
|
-
return 31 - __CLZ(val);
|
161
|
-
# else /* Software version */
|
162
|
-
static const unsigned DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29,
|
163
|
-
11, 14, 16, 18, 22, 25, 3, 30,
|
164
|
-
8, 12, 20, 28, 15, 17, 24, 7,
|
165
|
-
19, 27, 23, 6, 26, 5, 4, 31 };
|
166
|
-
U32 v = val;
|
167
|
-
v |= v >> 1;
|
168
|
-
v |= v >> 2;
|
169
|
-
v |= v >> 4;
|
170
|
-
v |= v >> 8;
|
171
|
-
v |= v >> 16;
|
172
|
-
return DeBruijnClz[ (U32) (v * 0x07C4ACDDU) >> 27];
|
173
|
-
# endif
|
174
|
-
}
|
175
|
-
}
|
176
|
-
|
177
136
|
/*===== Local Constants =====*/
|
178
137
|
static const unsigned BIT_mask[] = {
|
179
138
|
0, 1, 3, 7, 0xF, 0x1F,
|
@@ -203,6 +162,16 @@ MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC,
|
|
203
162
|
return 0;
|
204
163
|
}
|
205
164
|
|
165
|
+
MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits)
|
166
|
+
{
|
167
|
+
#if defined(STATIC_BMI2) && STATIC_BMI2 == 1 && !defined(ZSTD_NO_INTRINSICS)
|
168
|
+
return _bzhi_u64(bitContainer, nbBits);
|
169
|
+
#else
|
170
|
+
assert(nbBits < BIT_MASK_SIZE);
|
171
|
+
return bitContainer & BIT_mask[nbBits];
|
172
|
+
#endif
|
173
|
+
}
|
174
|
+
|
206
175
|
/*! BIT_addBits() :
|
207
176
|
* can add up to 31 bits into `bitC`.
|
208
177
|
* Note : does not check for register overflow ! */
|
@@ -212,7 +181,7 @@ MEM_STATIC void BIT_addBits(BIT_CStream_t* bitC,
|
|
212
181
|
DEBUG_STATIC_ASSERT(BIT_MASK_SIZE == 32);
|
213
182
|
assert(nbBits < BIT_MASK_SIZE);
|
214
183
|
assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8);
|
215
|
-
bitC->bitContainer |= (value
|
184
|
+
bitC->bitContainer |= BIT_getLowerBits(value, nbBits) << bitC->bitPos;
|
216
185
|
bitC->bitPos += nbBits;
|
217
186
|
}
|
218
187
|
|
@@ -291,7 +260,7 @@ MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, si
|
|
291
260
|
bitD->ptr = (const char*)srcBuffer + srcSize - sizeof(bitD->bitContainer);
|
292
261
|
bitD->bitContainer = MEM_readLEST(bitD->ptr);
|
293
262
|
{ BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1];
|
294
|
-
bitD->bitsConsumed = lastByte ? 8 -
|
263
|
+
bitD->bitsConsumed = lastByte ? 8 - ZSTD_highbit32(lastByte) : 0; /* ensures bitsConsumed is always set */
|
295
264
|
if (lastByte == 0) return ERROR(GENERIC); /* endMark not present */ }
|
296
265
|
} else {
|
297
266
|
bitD->ptr = bitD->start;
|
@@ -319,7 +288,7 @@ MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, si
|
|
319
288
|
default: break;
|
320
289
|
}
|
321
290
|
{ BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1];
|
322
|
-
bitD->bitsConsumed = lastByte ? 8 -
|
291
|
+
bitD->bitsConsumed = lastByte ? 8 - ZSTD_highbit32(lastByte) : 0;
|
323
292
|
if (lastByte == 0) return ERROR(corruption_detected); /* endMark not present */
|
324
293
|
}
|
325
294
|
bitD->bitsConsumed += (U32)(sizeof(bitD->bitContainer) - srcSize)*8;
|
@@ -350,16 +319,6 @@ MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getMiddleBits(size_t bitContainer, U32 c
|
|
350
319
|
#endif
|
351
320
|
}
|
352
321
|
|
353
|
-
MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits)
|
354
|
-
{
|
355
|
-
#if defined(STATIC_BMI2) && STATIC_BMI2 == 1
|
356
|
-
return _bzhi_u64(bitContainer, nbBits);
|
357
|
-
#else
|
358
|
-
assert(nbBits < BIT_MASK_SIZE);
|
359
|
-
return bitContainer & BIT_mask[nbBits];
|
360
|
-
#endif
|
361
|
-
}
|
362
|
-
|
363
322
|
/*! BIT_lookBits() :
|
364
323
|
* Provides next n bits from local register.
|
365
324
|
* local register is not modified.
|
@@ -406,7 +365,7 @@ MEM_STATIC FORCE_INLINE_ATTR size_t BIT_readBits(BIT_DStream_t* bitD, unsigned n
|
|
406
365
|
}
|
407
366
|
|
408
367
|
/*! BIT_readBitsFast() :
|
409
|
-
* unsafe version; only works
|
368
|
+
* unsafe version; only works if nbBits >= 1 */
|
410
369
|
MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits)
|
411
370
|
{
|
412
371
|
size_t const value = BIT_lookBitsFast(bitD, nbBits);
|
@@ -1,5 +1,5 @@
|
|
1
1
|
/*
|
2
|
-
* Copyright (c)
|
2
|
+
* Copyright (c) Meta Platforms, Inc. and affiliates.
|
3
3
|
* All rights reserved.
|
4
4
|
*
|
5
5
|
* This source code is licensed under both the BSD-style license (found in the
|
@@ -165,6 +165,12 @@
|
|
165
165
|
#define UNLIKELY(x) (x)
|
166
166
|
#endif
|
167
167
|
|
168
|
+
#if __has_builtin(__builtin_unreachable) || (defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 5)))
|
169
|
+
# define ZSTD_UNREACHABLE { assert(0), __builtin_unreachable(); }
|
170
|
+
#else
|
171
|
+
# define ZSTD_UNREACHABLE { assert(0); }
|
172
|
+
#endif
|
173
|
+
|
168
174
|
/* disable warnings */
|
169
175
|
#ifdef _MSC_VER /* Visual Studio */
|
170
176
|
# include <intrin.h> /* For Visual 2005 */
|
@@ -181,6 +187,8 @@
|
|
181
187
|
# ifdef __AVX2__ //MSVC does not have a BMI2 specific flag, but every CPU that supports AVX2 also supports BMI2
|
182
188
|
# define STATIC_BMI2 1
|
183
189
|
# endif
|
190
|
+
# elif defined(__BMI2__) && defined(__x86_64__) && defined(__GNUC__)
|
191
|
+
# define STATIC_BMI2 1
|
184
192
|
# endif
|
185
193
|
#endif
|
186
194
|
|
@@ -273,7 +281,18 @@
|
|
273
281
|
* Sanitizer
|
274
282
|
*****************************************************************/
|
275
283
|
|
276
|
-
#
|
284
|
+
/* Issue #3240 reports an ASAN failure on an llvm-mingw build. Out of an
|
285
|
+
* abundance of caution, disable our custom poisoning on mingw. */
|
286
|
+
#ifdef __MINGW32__
|
287
|
+
#ifndef ZSTD_ASAN_DONT_POISON_WORKSPACE
|
288
|
+
#define ZSTD_ASAN_DONT_POISON_WORKSPACE 1
|
289
|
+
#endif
|
290
|
+
#ifndef ZSTD_MSAN_DONT_POISON_WORKSPACE
|
291
|
+
#define ZSTD_MSAN_DONT_POISON_WORKSPACE 1
|
292
|
+
#endif
|
293
|
+
#endif
|
294
|
+
|
295
|
+
#if ZSTD_MEMORY_SANITIZER && !defined(ZSTD_MSAN_DONT_POISON_WORKSPACE)
|
277
296
|
/* Not all platforms that support msan provide sanitizers/msan_interface.h.
|
278
297
|
* We therefore declare the functions we need ourselves, rather than trying to
|
279
298
|
* include the header file... */
|
@@ -294,7 +313,7 @@ void __msan_poison(const volatile void *a, size_t size);
|
|
294
313
|
intptr_t __msan_test_shadow(const volatile void *x, size_t size);
|
295
314
|
#endif
|
296
315
|
|
297
|
-
#if ZSTD_ADDRESS_SANITIZER
|
316
|
+
#if ZSTD_ADDRESS_SANITIZER && !defined(ZSTD_ASAN_DONT_POISON_WORKSPACE)
|
298
317
|
/* Not all platforms that support asan provide sanitizers/asan_interface.h.
|
299
318
|
* We therefore declare the functions we need ourselves, rather than trying to
|
300
319
|
* include the header file... */
|
@@ -1,7 +1,7 @@
|
|
1
1
|
/* ******************************************************************
|
2
2
|
* debug
|
3
3
|
* Part of FSE library
|
4
|
-
* Copyright (c)
|
4
|
+
* Copyright (c) Meta Platforms, Inc. and affiliates.
|
5
5
|
*
|
6
6
|
* You can contact the author at :
|
7
7
|
* - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
|
@@ -1,7 +1,7 @@
|
|
1
1
|
/* ******************************************************************
|
2
2
|
* debug
|
3
3
|
* Part of FSE library
|
4
|
-
* Copyright (c)
|
4
|
+
* Copyright (c) Meta Platforms, Inc. and affiliates.
|
5
5
|
*
|
6
6
|
* You can contact the author at :
|
7
7
|
* - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
|
@@ -1,6 +1,6 @@
|
|
1
1
|
/* ******************************************************************
|
2
2
|
* Common functions of New Generation Entropy library
|
3
|
-
* Copyright (c)
|
3
|
+
* Copyright (c) Meta Platforms, Inc. and affiliates.
|
4
4
|
*
|
5
5
|
* You can contact the author at :
|
6
6
|
* - FSE+HUF source repository : https://github.com/Cyan4973/FiniteStateEntropy
|
@@ -19,8 +19,8 @@
|
|
19
19
|
#include "error_private.h" /* ERR_*, ERROR */
|
20
20
|
#define FSE_STATIC_LINKING_ONLY /* FSE_MIN_TABLELOG */
|
21
21
|
#include "fse.h"
|
22
|
-
#define HUF_STATIC_LINKING_ONLY /* HUF_TABLELOG_ABSOLUTEMAX */
|
23
22
|
#include "huf.h"
|
23
|
+
#include "bits.h" /* ZSDT_highbit32, ZSTD_countTrailingZeros32 */
|
24
24
|
|
25
25
|
|
26
26
|
/*=== Version ===*/
|
@@ -38,34 +38,6 @@ const char* HUF_getErrorName(size_t code) { return ERR_getErrorName(code); }
|
|
38
38
|
/*-**************************************************************
|
39
39
|
* FSE NCount encoding-decoding
|
40
40
|
****************************************************************/
|
41
|
-
static U32 FSE_ctz(U32 val)
|
42
|
-
{
|
43
|
-
assert(val != 0);
|
44
|
-
{
|
45
|
-
# if defined(_MSC_VER) /* Visual */
|
46
|
-
if (val != 0) {
|
47
|
-
unsigned long r;
|
48
|
-
_BitScanForward(&r, val);
|
49
|
-
return (unsigned)r;
|
50
|
-
} else {
|
51
|
-
/* Should not reach this code path */
|
52
|
-
__assume(0);
|
53
|
-
}
|
54
|
-
# elif defined(__GNUC__) && (__GNUC__ >= 3) /* GCC Intrinsic */
|
55
|
-
return __builtin_ctz(val);
|
56
|
-
# elif defined(__ICCARM__) /* IAR Intrinsic */
|
57
|
-
return __CTZ(val);
|
58
|
-
# else /* Software version */
|
59
|
-
U32 count = 0;
|
60
|
-
while ((val & 1) == 0) {
|
61
|
-
val >>= 1;
|
62
|
-
++count;
|
63
|
-
}
|
64
|
-
return count;
|
65
|
-
# endif
|
66
|
-
}
|
67
|
-
}
|
68
|
-
|
69
41
|
FORCE_INLINE_TEMPLATE
|
70
42
|
size_t FSE_readNCount_body(short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
|
71
43
|
const void* headerBuffer, size_t hbSize)
|
@@ -113,7 +85,7 @@ size_t FSE_readNCount_body(short* normalizedCounter, unsigned* maxSVPtr, unsigne
|
|
113
85
|
* repeat.
|
114
86
|
* Avoid UB by setting the high bit to 1.
|
115
87
|
*/
|
116
|
-
int repeats =
|
88
|
+
int repeats = ZSTD_countTrailingZeros32(~bitStream | 0x80000000) >> 1;
|
117
89
|
while (repeats >= 12) {
|
118
90
|
charnum += 3 * 12;
|
119
91
|
if (LIKELY(ip <= iend-7)) {
|
@@ -124,7 +96,7 @@ size_t FSE_readNCount_body(short* normalizedCounter, unsigned* maxSVPtr, unsigne
|
|
124
96
|
ip = iend - 4;
|
125
97
|
}
|
126
98
|
bitStream = MEM_readLE32(ip) >> bitCount;
|
127
|
-
repeats =
|
99
|
+
repeats = ZSTD_countTrailingZeros32(~bitStream | 0x80000000) >> 1;
|
128
100
|
}
|
129
101
|
charnum += 3 * repeats;
|
130
102
|
bitStream >>= 2 * repeats;
|
@@ -189,7 +161,7 @@ size_t FSE_readNCount_body(short* normalizedCounter, unsigned* maxSVPtr, unsigne
|
|
189
161
|
* know that threshold > 1.
|
190
162
|
*/
|
191
163
|
if (remaining <= 1) break;
|
192
|
-
nbBits =
|
164
|
+
nbBits = ZSTD_highbit32(remaining) + 1;
|
193
165
|
threshold = 1 << (nbBits - 1);
|
194
166
|
}
|
195
167
|
if (charnum >= maxSV1) break;
|
@@ -264,7 +236,7 @@ size_t HUF_readStats(BYTE* huffWeight, size_t hwSize, U32* rankStats,
|
|
264
236
|
const void* src, size_t srcSize)
|
265
237
|
{
|
266
238
|
U32 wksp[HUF_READ_STATS_WORKSPACE_SIZE_U32];
|
267
|
-
return HUF_readStats_wksp(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, wksp, sizeof(wksp), /*
|
239
|
+
return HUF_readStats_wksp(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, wksp, sizeof(wksp), /* flags */ 0);
|
268
240
|
}
|
269
241
|
|
270
242
|
FORCE_INLINE_TEMPLATE size_t
|
@@ -312,14 +284,14 @@ HUF_readStats_body(BYTE* huffWeight, size_t hwSize, U32* rankStats,
|
|
312
284
|
if (weightTotal == 0) return ERROR(corruption_detected);
|
313
285
|
|
314
286
|
/* get last non-null symbol weight (implied, total must be 2^n) */
|
315
|
-
{ U32 const tableLog =
|
287
|
+
{ U32 const tableLog = ZSTD_highbit32(weightTotal) + 1;
|
316
288
|
if (tableLog > HUF_TABLELOG_MAX) return ERROR(corruption_detected);
|
317
289
|
*tableLogPtr = tableLog;
|
318
290
|
/* determine last weight */
|
319
291
|
{ U32 const total = 1 << tableLog;
|
320
292
|
U32 const rest = total - weightTotal;
|
321
|
-
U32 const verif = 1 <<
|
322
|
-
U32 const lastWeight =
|
293
|
+
U32 const verif = 1 << ZSTD_highbit32(rest);
|
294
|
+
U32 const lastWeight = ZSTD_highbit32(rest) + 1;
|
323
295
|
if (verif != rest) return ERROR(corruption_detected); /* last value must be a clean power of 2 */
|
324
296
|
huffWeight[oSize] = (BYTE)lastWeight;
|
325
297
|
rankStats[lastWeight]++;
|
@@ -356,13 +328,13 @@ size_t HUF_readStats_wksp(BYTE* huffWeight, size_t hwSize, U32* rankStats,
|
|
356
328
|
U32* nbSymbolsPtr, U32* tableLogPtr,
|
357
329
|
const void* src, size_t srcSize,
|
358
330
|
void* workSpace, size_t wkspSize,
|
359
|
-
int
|
331
|
+
int flags)
|
360
332
|
{
|
361
333
|
#if DYNAMIC_BMI2
|
362
|
-
if (
|
334
|
+
if (flags & HUF_flags_bmi2) {
|
363
335
|
return HUF_readStats_body_bmi2(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, workSpace, wkspSize);
|
364
336
|
}
|
365
337
|
#endif
|
366
|
-
(void)
|
338
|
+
(void)flags;
|
367
339
|
return HUF_readStats_body_default(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, workSpace, wkspSize);
|
368
340
|
}
|
@@ -1,5 +1,5 @@
|
|
1
1
|
/*
|
2
|
-
* Copyright (c)
|
2
|
+
* Copyright (c) Meta Platforms, Inc. and affiliates.
|
3
3
|
* All rights reserved.
|
4
4
|
*
|
5
5
|
* This source code is licensed under both the BSD-style license (found in the
|
@@ -27,9 +27,11 @@ const char* ERR_getErrorString(ERR_enum code)
|
|
27
27
|
case PREFIX(version_unsupported): return "Version not supported";
|
28
28
|
case PREFIX(frameParameter_unsupported): return "Unsupported frame parameter";
|
29
29
|
case PREFIX(frameParameter_windowTooLarge): return "Frame requires too much memory for decoding";
|
30
|
-
case PREFIX(corruption_detected): return "
|
30
|
+
case PREFIX(corruption_detected): return "Data corruption detected";
|
31
31
|
case PREFIX(checksum_wrong): return "Restored data doesn't match checksum";
|
32
|
+
case PREFIX(literals_headerWrong): return "Header of Literals' block doesn't respect format specification";
|
32
33
|
case PREFIX(parameter_unsupported): return "Unsupported parameter";
|
34
|
+
case PREFIX(parameter_combination_unsupported): return "Unsupported combination of parameters";
|
33
35
|
case PREFIX(parameter_outOfBound): return "Parameter is out of bound";
|
34
36
|
case PREFIX(init_missing): return "Context should be init first";
|
35
37
|
case PREFIX(memory_allocation): return "Allocation error : not enough memory";
|
@@ -38,17 +40,22 @@ const char* ERR_getErrorString(ERR_enum code)
|
|
38
40
|
case PREFIX(tableLog_tooLarge): return "tableLog requires too much memory : unsupported";
|
39
41
|
case PREFIX(maxSymbolValue_tooLarge): return "Unsupported max Symbol Value : too large";
|
40
42
|
case PREFIX(maxSymbolValue_tooSmall): return "Specified maxSymbolValue is too small";
|
43
|
+
case PREFIX(stabilityCondition_notRespected): return "pledged buffer stability condition is not respected";
|
41
44
|
case PREFIX(dictionary_corrupted): return "Dictionary is corrupted";
|
42
45
|
case PREFIX(dictionary_wrong): return "Dictionary mismatch";
|
43
46
|
case PREFIX(dictionaryCreation_failed): return "Cannot create Dictionary from provided samples";
|
44
47
|
case PREFIX(dstSize_tooSmall): return "Destination buffer is too small";
|
45
48
|
case PREFIX(srcSize_wrong): return "Src size is incorrect";
|
46
49
|
case PREFIX(dstBuffer_null): return "Operation on NULL destination buffer";
|
50
|
+
case PREFIX(noForwardProgress_destFull): return "Operation made no progress over multiple calls, due to output buffer being full";
|
51
|
+
case PREFIX(noForwardProgress_inputEmpty): return "Operation made no progress over multiple calls, due to input being empty";
|
47
52
|
/* following error codes are not stable and may be removed or changed in a future version */
|
48
53
|
case PREFIX(frameIndex_tooLarge): return "Frame index is too large";
|
49
54
|
case PREFIX(seekableIO): return "An I/O error occurred when reading/seeking";
|
50
55
|
case PREFIX(dstBuffer_wrong): return "Destination buffer is wrong";
|
51
56
|
case PREFIX(srcBuffer_wrong): return "Source buffer is wrong";
|
57
|
+
case PREFIX(sequenceProducer_failed): return "Block-level external sequence producer returned an error code";
|
58
|
+
case PREFIX(externalSequences_invalid): return "External sequences are not valid";
|
52
59
|
case PREFIX(maxCode):
|
53
60
|
default: return notErrorCode;
|
54
61
|
}
|