sereal 0.0.3 → 0.0.4
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.
- checksums.yaml +4 -4
- data/ext/sereal/buffer.h +15 -8
- data/ext/sereal/decode.c +72 -28
- data/ext/sereal/encode.c +107 -18
- data/ext/sereal/lz4.c +822 -0
- data/ext/sereal/lz4.h +205 -0
- data/ext/sereal/lz4/lz4.c +822 -0
- data/ext/sereal/lz4/lz4.h +205 -0
- data/ext/sereal/lz4/lz4hc.c +817 -0
- data/ext/sereal/lz4/lz4hc.h +111 -0
- data/ext/sereal/lz4hc.c +817 -0
- data/ext/sereal/lz4hc.h +111 -0
- data/ext/sereal/proto.h +4 -2
- data/ext/sereal/sereal.c +30 -0
- data/ext/sereal/sereal.h +13 -1
- data/ext/sereal/snappy/csnappy_compat.h +16 -0
- data/ext/sereal/snappy/csnappy_compress.c +19 -16
- data/ext/sereal/snappy/csnappy_decompress.c +8 -5
- data/ext/sereal/snappy/csnappy_internal.h +12 -3
- data/ext/sereal/snappy/csnappy_internal_userspace.h +56 -18
- metadata +11 -2
@@ -0,0 +1,205 @@
|
|
1
|
+
/*
|
2
|
+
LZ4 - Fast LZ compression algorithm
|
3
|
+
Header File
|
4
|
+
Copyright (C) 2011-2013, 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
|
+
// Compiler Options
|
43
|
+
//**************************************
|
44
|
+
#if defined(_MSC_VER) && !defined(__cplusplus) // Visual Studio
|
45
|
+
# define inline __inline // Visual C is not C99, but supports some kind of inline
|
46
|
+
#endif
|
47
|
+
|
48
|
+
|
49
|
+
//****************************
|
50
|
+
// Simple Functions
|
51
|
+
//****************************
|
52
|
+
|
53
|
+
int LZ4_compress (const char* source, char* dest, int inputSize);
|
54
|
+
int LZ4_decompress_safe (const char* source, char* dest, int inputSize, int maxOutputSize);
|
55
|
+
|
56
|
+
/*
|
57
|
+
LZ4_compress() :
|
58
|
+
Compresses 'inputSize' bytes from 'source' into 'dest'.
|
59
|
+
Destination buffer must be already allocated,
|
60
|
+
and must be sized to handle worst cases situations (input data not compressible)
|
61
|
+
Worst case size evaluation is provided by function LZ4_compressBound()
|
62
|
+
inputSize : Max supported value is LZ4_MAX_INPUT_VALUE
|
63
|
+
return : the number of bytes written in buffer dest
|
64
|
+
or 0 if the compression fails
|
65
|
+
|
66
|
+
LZ4_decompress_safe() :
|
67
|
+
maxOutputSize : is the size of the destination buffer (which must be already allocated)
|
68
|
+
return : the number of bytes decoded in the destination buffer (necessarily <= maxOutputSize)
|
69
|
+
If the source stream is detected malformed, the function will stop decoding and return a negative result.
|
70
|
+
This function is protected against buffer overflow exploits (never writes outside of output buffer, and never reads outside of input buffer). Therefore, it is protected against malicious data packets
|
71
|
+
*/
|
72
|
+
|
73
|
+
|
74
|
+
//****************************
|
75
|
+
// Advanced Functions
|
76
|
+
//****************************
|
77
|
+
#define LZ4_MAX_INPUT_SIZE 0x7E000000 // 2 113 929 216 bytes
|
78
|
+
#define LZ4_COMPRESSBOUND(isize) ((unsigned int)(isize) > (unsigned int)LZ4_MAX_INPUT_SIZE ? 0 : (isize) + ((isize)/255) + 16)
|
79
|
+
static inline int LZ4_compressBound(int isize) { return LZ4_COMPRESSBOUND(isize); }
|
80
|
+
|
81
|
+
/*
|
82
|
+
LZ4_compressBound() :
|
83
|
+
Provides the maximum size that LZ4 may output in a "worst case" scenario (input data not compressible)
|
84
|
+
primarily useful for memory allocation of output buffer.
|
85
|
+
inline function is recommended for the general case,
|
86
|
+
macro is also provided when result needs to be evaluated at compilation (such as stack memory allocation).
|
87
|
+
|
88
|
+
isize : is the input size. Max supported value is LZ4_MAX_INPUT_SIZE
|
89
|
+
return : maximum output size in a "worst case" scenario
|
90
|
+
or 0, if input size is too large ( > LZ4_MAX_INPUT_SIZE)
|
91
|
+
*/
|
92
|
+
|
93
|
+
|
94
|
+
int LZ4_compress_limitedOutput (const char* source, char* dest, int inputSize, int maxOutputSize);
|
95
|
+
|
96
|
+
/*
|
97
|
+
LZ4_compress_limitedOutput() :
|
98
|
+
Compress 'inputSize' bytes from 'source' into an output buffer 'dest' of maximum size 'maxOutputSize'.
|
99
|
+
If it cannot achieve it, compression will stop, and result of the function will be zero.
|
100
|
+
This function never writes outside of provided output buffer.
|
101
|
+
|
102
|
+
inputSize : Max supported value is LZ4_MAX_INPUT_VALUE
|
103
|
+
maxOutputSize : is the size of the destination buffer (which must be already allocated)
|
104
|
+
return : the number of bytes written in buffer 'dest'
|
105
|
+
or 0 if the compression fails
|
106
|
+
*/
|
107
|
+
|
108
|
+
|
109
|
+
int LZ4_decompress_fast (const char* source, char* dest, int outputSize);
|
110
|
+
|
111
|
+
/*
|
112
|
+
LZ4_decompress_fast() :
|
113
|
+
outputSize : is the original (uncompressed) size
|
114
|
+
return : the number of bytes read from the source buffer (in other words, the compressed size)
|
115
|
+
If the source stream is malformed, the function will stop decoding and return a negative result.
|
116
|
+
note : This function is a bit faster than LZ4_decompress_safe()
|
117
|
+
This function never writes outside of output buffers, but may read beyond input buffer in case of malicious data packet.
|
118
|
+
Use this function preferably into a trusted environment (data to decode comes from a trusted source).
|
119
|
+
Destination buffer must be already allocated. Its size must be a minimum of 'outputSize' bytes.
|
120
|
+
*/
|
121
|
+
|
122
|
+
int LZ4_decompress_safe_partial (const char* source, char* dest, int inputSize, int targetOutputSize, int maxOutputSize);
|
123
|
+
|
124
|
+
/*
|
125
|
+
LZ4_decompress_safe_partial() :
|
126
|
+
This function decompress a compressed block of size 'inputSize' at position 'source'
|
127
|
+
into output buffer 'dest' of size 'maxOutputSize'.
|
128
|
+
The function tries to stop decompressing operation as soon as 'targetOutputSize' has been reached,
|
129
|
+
reducing decompression time.
|
130
|
+
return : the number of bytes decoded in the destination buffer (necessarily <= maxOutputSize)
|
131
|
+
Note : this number can be < 'targetOutputSize' should the compressed block to decode be smaller.
|
132
|
+
Always control how many bytes were decoded.
|
133
|
+
If the source stream is detected malformed, the function will stop decoding and return a negative result.
|
134
|
+
This function never writes outside of output buffer, and never reads outside of input buffer. It is therefore protected against malicious data packets
|
135
|
+
*/
|
136
|
+
|
137
|
+
|
138
|
+
//****************************
|
139
|
+
// Stream Functions
|
140
|
+
//****************************
|
141
|
+
|
142
|
+
void* LZ4_create (const char* inputBuffer);
|
143
|
+
int LZ4_compress_continue (void* LZ4_Data, const char* source, char* dest, int inputSize);
|
144
|
+
int LZ4_compress_limitedOutput_continue (void* LZ4_Data, const char* source, char* dest, int inputSize, int maxOutputSize);
|
145
|
+
char* LZ4_slideInputBuffer (void* LZ4_Data);
|
146
|
+
int LZ4_free (void* LZ4_Data);
|
147
|
+
|
148
|
+
/*
|
149
|
+
These functions allow the compression of dependent blocks, where each block benefits from prior 64 KB within preceding blocks.
|
150
|
+
In order to achieve this, it is necessary to start creating the LZ4 Data Structure, thanks to the function :
|
151
|
+
|
152
|
+
void* LZ4_create (const char* inputBuffer);
|
153
|
+
The result of the function is the (void*) pointer on the LZ4 Data Structure.
|
154
|
+
This pointer will be needed in all other functions.
|
155
|
+
If the pointer returned is NULL, then the allocation has failed, and compression must be aborted.
|
156
|
+
The only parameter 'const char* inputBuffer' must, obviously, point at the beginning of input buffer.
|
157
|
+
The input buffer must be already allocated, and size at least 192KB.
|
158
|
+
'inputBuffer' will also be the 'const char* source' of the first block.
|
159
|
+
|
160
|
+
All blocks are expected to lay next to each other within the input buffer, starting from 'inputBuffer'.
|
161
|
+
To compress each block, use either LZ4_compress_continue() or LZ4_compress_limitedOutput_continue().
|
162
|
+
Their behavior are identical to LZ4_compress() or LZ4_compress_limitedOutput(),
|
163
|
+
but require the LZ4 Data Structure as their first argument, and check that each block starts right after the previous one.
|
164
|
+
If next block does not begin immediately after the previous one, the compression will fail (return 0).
|
165
|
+
|
166
|
+
When it's no longer possible to lay the next block after the previous one (not enough space left into input buffer), a call to :
|
167
|
+
char* LZ4_slideInputBuffer(void* LZ4_Data);
|
168
|
+
must be performed. It will typically copy the latest 64KB of input at the beginning of input buffer.
|
169
|
+
Note that, for this function to work properly, minimum size of an input buffer must be 192KB.
|
170
|
+
==> The memory position where the next input data block must start is provided as the result of the function.
|
171
|
+
|
172
|
+
Compression can then resume, using LZ4_compress_continue() or LZ4_compress_limitedOutput_continue(), as usual.
|
173
|
+
|
174
|
+
When compression is completed, a call to LZ4_free() will release the memory used by the LZ4 Data Structure.
|
175
|
+
*/
|
176
|
+
|
177
|
+
|
178
|
+
int LZ4_decompress_safe_withPrefix64k (const char* source, char* dest, int inputSize, int maxOutputSize);
|
179
|
+
int LZ4_decompress_fast_withPrefix64k (const char* source, char* dest, int outputSize);
|
180
|
+
|
181
|
+
/*
|
182
|
+
*_withPrefix64k() :
|
183
|
+
These decoding functions work the same as their "normal name" versions,
|
184
|
+
but can use up to 64KB of data in front of 'char* dest'.
|
185
|
+
These functions are necessary to decode inter-dependant blocks.
|
186
|
+
*/
|
187
|
+
|
188
|
+
|
189
|
+
//****************************
|
190
|
+
// Obsolete Functions
|
191
|
+
//****************************
|
192
|
+
|
193
|
+
static inline int LZ4_uncompress (const char* source, char* dest, int outputSize) { return LZ4_decompress_fast(source, dest, outputSize); }
|
194
|
+
static inline int LZ4_uncompress_unknownOutputSize (const char* source, char* dest, int isize, int maxOutputSize) { return LZ4_decompress_safe(source, dest, isize, maxOutputSize); }
|
195
|
+
|
196
|
+
/*
|
197
|
+
These functions are deprecated and should no longer be used.
|
198
|
+
They are provided here for compatibility with existing user programs.
|
199
|
+
*/
|
200
|
+
|
201
|
+
|
202
|
+
|
203
|
+
#if defined (__cplusplus)
|
204
|
+
}
|
205
|
+
#endif
|
@@ -0,0 +1,817 @@
|
|
1
|
+
/*
|
2
|
+
LZ4 HC - High Compression Mode of LZ4
|
3
|
+
Copyright (C) 2011-2013, 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
|
+
// Memory routines
|
36
|
+
//**************************************
|
37
|
+
#include <stdlib.h> // calloc, free
|
38
|
+
#define ALLOCATOR(s) calloc(1,s)
|
39
|
+
#define FREEMEM free
|
40
|
+
#include <string.h> // memset, memcpy
|
41
|
+
#define MEM_INIT memset
|
42
|
+
|
43
|
+
|
44
|
+
//**************************************
|
45
|
+
// CPU Feature Detection
|
46
|
+
//**************************************
|
47
|
+
// 32 or 64 bits ?
|
48
|
+
#if (defined(__x86_64__) || defined(_M_X64) || defined(_WIN64) \
|
49
|
+
|| defined(__powerpc64__) || defined(__ppc64__) || defined(__PPC64__) \
|
50
|
+
|| defined(__64BIT__) || defined(_LP64) || defined(__LP64__) \
|
51
|
+
|| defined(__ia64) || defined(__itanium__) || defined(_M_IA64) ) // Detects 64 bits mode
|
52
|
+
# define LZ4_ARCH64 1
|
53
|
+
#else
|
54
|
+
# define LZ4_ARCH64 0
|
55
|
+
#endif
|
56
|
+
|
57
|
+
// Little Endian or Big Endian ?
|
58
|
+
// Overwrite the #define below if you know your architecture endianess
|
59
|
+
#if defined (__GLIBC__)
|
60
|
+
# include <endian.h>
|
61
|
+
# if (__BYTE_ORDER == __BIG_ENDIAN)
|
62
|
+
# define LZ4_BIG_ENDIAN 1
|
63
|
+
# endif
|
64
|
+
#elif (defined(__BIG_ENDIAN__) || defined(__BIG_ENDIAN) || defined(_BIG_ENDIAN)) && !(defined(__LITTLE_ENDIAN__) || defined(__LITTLE_ENDIAN) || defined(_LITTLE_ENDIAN))
|
65
|
+
# define LZ4_BIG_ENDIAN 1
|
66
|
+
#elif defined(__sparc) || defined(__sparc__) \
|
67
|
+
|| defined(__powerpc__) || defined(__ppc__) || defined(__PPC__) \
|
68
|
+
|| defined(__hpux) || defined(__hppa) \
|
69
|
+
|| defined(_MIPSEB) || defined(__s390__)
|
70
|
+
# define LZ4_BIG_ENDIAN 1
|
71
|
+
#else
|
72
|
+
// Little Endian assumed. PDP Endian and other very rare endian format are unsupported.
|
73
|
+
#endif
|
74
|
+
|
75
|
+
// Unaligned memory access is automatically enabled for "common" CPU, such as x86.
|
76
|
+
// For others CPU, the compiler will be more cautious, and insert extra code to ensure aligned access is respected
|
77
|
+
// If you know your target CPU supports unaligned memory access, you want to force this option manually to improve performance
|
78
|
+
#if defined(__ARM_FEATURE_UNALIGNED)
|
79
|
+
# define LZ4_FORCE_UNALIGNED_ACCESS 1
|
80
|
+
#endif
|
81
|
+
|
82
|
+
// Define this parameter if your target system or compiler does not support hardware bit count
|
83
|
+
#if defined(_MSC_VER) && defined(_WIN32_WCE) // Visual Studio for Windows CE does not support Hardware bit count
|
84
|
+
# define LZ4_FORCE_SW_BITCOUNT
|
85
|
+
#endif
|
86
|
+
|
87
|
+
|
88
|
+
//**************************************
|
89
|
+
// Compiler Options
|
90
|
+
//**************************************
|
91
|
+
#if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L // C99
|
92
|
+
/* "restrict" is a known keyword */
|
93
|
+
#else
|
94
|
+
# define restrict // Disable restrict
|
95
|
+
#endif
|
96
|
+
|
97
|
+
#ifdef _MSC_VER // Visual Studio
|
98
|
+
# define FORCE_INLINE static __forceinline
|
99
|
+
# include <intrin.h> // For Visual 2005
|
100
|
+
# if LZ4_ARCH64 // 64-bits
|
101
|
+
# pragma intrinsic(_BitScanForward64) // For Visual 2005
|
102
|
+
# pragma intrinsic(_BitScanReverse64) // For Visual 2005
|
103
|
+
# else // 32-bits
|
104
|
+
# pragma intrinsic(_BitScanForward) // For Visual 2005
|
105
|
+
# pragma intrinsic(_BitScanReverse) // For Visual 2005
|
106
|
+
# endif
|
107
|
+
# pragma warning(disable : 4127) // disable: C4127: conditional expression is constant
|
108
|
+
# pragma warning(disable : 4701) // disable: C4701: potentially uninitialized local variable used
|
109
|
+
#else
|
110
|
+
# ifdef __GNUC__
|
111
|
+
# define FORCE_INLINE static inline __attribute__((always_inline))
|
112
|
+
# else
|
113
|
+
# define FORCE_INLINE static inline
|
114
|
+
# endif
|
115
|
+
#endif
|
116
|
+
|
117
|
+
#ifdef _MSC_VER // Visual Studio
|
118
|
+
# define lz4_bswap16(x) _byteswap_ushort(x)
|
119
|
+
#else
|
120
|
+
# define lz4_bswap16(x) ((unsigned short int) ((((x) >> 8) & 0xffu) | (((x) & 0xffu) << 8)))
|
121
|
+
#endif
|
122
|
+
|
123
|
+
|
124
|
+
//**************************************
|
125
|
+
// Includes
|
126
|
+
//**************************************
|
127
|
+
#include "lz4hc.h"
|
128
|
+
#include "lz4.h"
|
129
|
+
|
130
|
+
|
131
|
+
//**************************************
|
132
|
+
// Basic Types
|
133
|
+
//**************************************
|
134
|
+
#if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L // C99
|
135
|
+
# include <stdint.h>
|
136
|
+
typedef uint8_t BYTE;
|
137
|
+
typedef uint16_t U16;
|
138
|
+
typedef uint32_t U32;
|
139
|
+
typedef int32_t S32;
|
140
|
+
typedef uint64_t U64;
|
141
|
+
#else
|
142
|
+
typedef unsigned char BYTE;
|
143
|
+
typedef unsigned short U16;
|
144
|
+
typedef unsigned int U32;
|
145
|
+
typedef signed int S32;
|
146
|
+
typedef unsigned long long U64;
|
147
|
+
#endif
|
148
|
+
|
149
|
+
#if defined(__GNUC__) && !defined(LZ4_FORCE_UNALIGNED_ACCESS)
|
150
|
+
# define _PACKED __attribute__ ((packed))
|
151
|
+
#else
|
152
|
+
# define _PACKED
|
153
|
+
#endif
|
154
|
+
|
155
|
+
#if !defined(LZ4_FORCE_UNALIGNED_ACCESS) && !defined(__GNUC__)
|
156
|
+
# ifdef __IBMC__
|
157
|
+
# pragma pack(1)
|
158
|
+
# else
|
159
|
+
# pragma pack(push, 1)
|
160
|
+
# endif
|
161
|
+
#endif
|
162
|
+
|
163
|
+
typedef struct _U16_S { U16 v; } _PACKED U16_S;
|
164
|
+
typedef struct _U32_S { U32 v; } _PACKED U32_S;
|
165
|
+
typedef struct _U64_S { U64 v; } _PACKED U64_S;
|
166
|
+
|
167
|
+
#if !defined(LZ4_FORCE_UNALIGNED_ACCESS) && !defined(__GNUC__)
|
168
|
+
# pragma pack(pop)
|
169
|
+
#endif
|
170
|
+
|
171
|
+
#define A64(x) (((U64_S *)(x))->v)
|
172
|
+
#define A32(x) (((U32_S *)(x))->v)
|
173
|
+
#define A16(x) (((U16_S *)(x))->v)
|
174
|
+
|
175
|
+
|
176
|
+
//**************************************
|
177
|
+
// Constants
|
178
|
+
//**************************************
|
179
|
+
#define MINMATCH 4
|
180
|
+
|
181
|
+
#define DICTIONARY_LOGSIZE 16
|
182
|
+
#define MAXD (1<<DICTIONARY_LOGSIZE)
|
183
|
+
#define MAXD_MASK ((U32)(MAXD - 1))
|
184
|
+
#define MAX_DISTANCE (MAXD - 1)
|
185
|
+
|
186
|
+
#define HASH_LOG (DICTIONARY_LOGSIZE-1)
|
187
|
+
#define HASHTABLESIZE (1 << HASH_LOG)
|
188
|
+
#define HASH_MASK (HASHTABLESIZE - 1)
|
189
|
+
|
190
|
+
#define MAX_NB_ATTEMPTS 256
|
191
|
+
|
192
|
+
#define ML_BITS 4
|
193
|
+
#define ML_MASK (size_t)((1U<<ML_BITS)-1)
|
194
|
+
#define RUN_BITS (8-ML_BITS)
|
195
|
+
#define RUN_MASK ((1U<<RUN_BITS)-1)
|
196
|
+
|
197
|
+
#define COPYLENGTH 8
|
198
|
+
#define LASTLITERALS 5
|
199
|
+
#define MFLIMIT (COPYLENGTH+MINMATCH)
|
200
|
+
#define MINLENGTH (MFLIMIT+1)
|
201
|
+
#define OPTIMAL_ML (int)((ML_MASK-1)+MINMATCH)
|
202
|
+
|
203
|
+
#define KB *(1U<<10)
|
204
|
+
#define MB *(1U<<20)
|
205
|
+
#define GB *(1U<<30)
|
206
|
+
|
207
|
+
|
208
|
+
//**************************************
|
209
|
+
// Architecture-specific macros
|
210
|
+
//**************************************
|
211
|
+
#if LZ4_ARCH64 // 64-bit
|
212
|
+
# define STEPSIZE 8
|
213
|
+
# define LZ4_COPYSTEP(s,d) A64(d) = A64(s); d+=8; s+=8;
|
214
|
+
# define LZ4_COPYPACKET(s,d) LZ4_COPYSTEP(s,d)
|
215
|
+
# define UARCH U64
|
216
|
+
# define AARCH A64
|
217
|
+
# define HTYPE U32
|
218
|
+
# define INITBASE(b,s) const BYTE* const b = s
|
219
|
+
#else // 32-bit
|
220
|
+
# define STEPSIZE 4
|
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 UARCH U32
|
224
|
+
# define AARCH A32
|
225
|
+
//# define HTYPE const BYTE*
|
226
|
+
//# define INITBASE(b,s) const int b = 0
|
227
|
+
# define HTYPE U32
|
228
|
+
# define INITBASE(b,s) const BYTE* const b = s
|
229
|
+
#endif
|
230
|
+
|
231
|
+
#if defined(LZ4_BIG_ENDIAN)
|
232
|
+
# define LZ4_READ_LITTLEENDIAN_16(d,s,p) { U16 v = A16(p); v = lz4_bswap16(v); d = (s) - v; }
|
233
|
+
# define LZ4_WRITE_LITTLEENDIAN_16(p,i) { U16 v = (U16)(i); v = lz4_bswap16(v); A16(p) = v; p+=2; }
|
234
|
+
#else // Little Endian
|
235
|
+
# define LZ4_READ_LITTLEENDIAN_16(d,s,p) { d = (s) - A16(p); }
|
236
|
+
# define LZ4_WRITE_LITTLEENDIAN_16(p,v) { A16(p) = v; p+=2; }
|
237
|
+
#endif
|
238
|
+
|
239
|
+
|
240
|
+
//************************************************************
|
241
|
+
// Local Types
|
242
|
+
//************************************************************
|
243
|
+
typedef struct
|
244
|
+
{
|
245
|
+
const BYTE* inputBuffer;
|
246
|
+
const BYTE* base;
|
247
|
+
const BYTE* end;
|
248
|
+
HTYPE hashTable[HASHTABLESIZE];
|
249
|
+
U16 chainTable[MAXD];
|
250
|
+
const BYTE* nextToUpdate;
|
251
|
+
} LZ4HC_Data_Structure;
|
252
|
+
|
253
|
+
|
254
|
+
//**************************************
|
255
|
+
// Macros
|
256
|
+
//**************************************
|
257
|
+
#define LZ4_WILDCOPY(s,d,e) do { LZ4_COPYPACKET(s,d) } while (d<e);
|
258
|
+
#define LZ4_BLINDCOPY(s,d,l) { BYTE* e=d+l; LZ4_WILDCOPY(s,d,e); d=e; }
|
259
|
+
#define HASH_FUNCTION(i) (((i) * 2654435761U) >> ((MINMATCH*8)-HASH_LOG))
|
260
|
+
#define HASH_VALUE(p) HASH_FUNCTION(A32(p))
|
261
|
+
#define HASH_POINTER(p) (HashTable[HASH_VALUE(p)] + base)
|
262
|
+
#define DELTANEXT(p) chainTable[(size_t)(p) & MAXD_MASK]
|
263
|
+
#define GETNEXT(p) ((p) - (size_t)DELTANEXT(p))
|
264
|
+
|
265
|
+
|
266
|
+
//**************************************
|
267
|
+
// Private functions
|
268
|
+
//**************************************
|
269
|
+
#if LZ4_ARCH64
|
270
|
+
|
271
|
+
FORCE_INLINE int LZ4_NbCommonBytes (register U64 val)
|
272
|
+
{
|
273
|
+
#if defined(LZ4_BIG_ENDIAN)
|
274
|
+
# if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
275
|
+
unsigned long r = 0;
|
276
|
+
_BitScanReverse64( &r, val );
|
277
|
+
return (int)(r>>3);
|
278
|
+
# elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
279
|
+
return (__builtin_clzll(val) >> 3);
|
280
|
+
# else
|
281
|
+
int r;
|
282
|
+
if (!(val>>32)) { r=4; } else { r=0; val>>=32; }
|
283
|
+
if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; }
|
284
|
+
r += (!val);
|
285
|
+
return r;
|
286
|
+
# endif
|
287
|
+
#else
|
288
|
+
# if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
289
|
+
unsigned long r = 0;
|
290
|
+
_BitScanForward64( &r, val );
|
291
|
+
return (int)(r>>3);
|
292
|
+
# elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
293
|
+
return (__builtin_ctzll(val) >> 3);
|
294
|
+
# else
|
295
|
+
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 };
|
296
|
+
return DeBruijnBytePos[((U64)((val & -val) * 0x0218A392CDABBD3F)) >> 58];
|
297
|
+
# endif
|
298
|
+
#endif
|
299
|
+
}
|
300
|
+
|
301
|
+
#else
|
302
|
+
|
303
|
+
FORCE_INLINE int LZ4_NbCommonBytes (register U32 val)
|
304
|
+
{
|
305
|
+
#if defined(LZ4_BIG_ENDIAN)
|
306
|
+
# if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
307
|
+
unsigned long r;
|
308
|
+
_BitScanReverse( &r, val );
|
309
|
+
return (int)(r>>3);
|
310
|
+
# elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
311
|
+
return (__builtin_clz(val) >> 3);
|
312
|
+
# else
|
313
|
+
int r;
|
314
|
+
if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; }
|
315
|
+
r += (!val);
|
316
|
+
return r;
|
317
|
+
# endif
|
318
|
+
#else
|
319
|
+
# if defined(_MSC_VER) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
320
|
+
unsigned long r;
|
321
|
+
_BitScanForward( &r, val );
|
322
|
+
return (int)(r>>3);
|
323
|
+
# elif defined(__GNUC__) && ((__GNUC__ * 100 + __GNUC_MINOR__) >= 304) && !defined(LZ4_FORCE_SW_BITCOUNT)
|
324
|
+
return (__builtin_ctz(val) >> 3);
|
325
|
+
# else
|
326
|
+
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 };
|
327
|
+
return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27];
|
328
|
+
# endif
|
329
|
+
#endif
|
330
|
+
}
|
331
|
+
|
332
|
+
#endif
|
333
|
+
|
334
|
+
|
335
|
+
FORCE_INLINE void LZ4_initHC (LZ4HC_Data_Structure* hc4, const BYTE* base)
|
336
|
+
{
|
337
|
+
MEM_INIT((void*)hc4->hashTable, 0, sizeof(hc4->hashTable));
|
338
|
+
MEM_INIT(hc4->chainTable, 0xFF, sizeof(hc4->chainTable));
|
339
|
+
hc4->nextToUpdate = base + 1;
|
340
|
+
hc4->base = base;
|
341
|
+
hc4->inputBuffer = base;
|
342
|
+
hc4->end = base;
|
343
|
+
}
|
344
|
+
|
345
|
+
|
346
|
+
void* LZ4_createHC (const char* inputBuffer)
|
347
|
+
{
|
348
|
+
void* hc4 = ALLOCATOR(sizeof(LZ4HC_Data_Structure));
|
349
|
+
LZ4_initHC ((LZ4HC_Data_Structure*)hc4, (const BYTE*)inputBuffer);
|
350
|
+
return hc4;
|
351
|
+
}
|
352
|
+
|
353
|
+
|
354
|
+
int LZ4_freeHC (void* LZ4HC_Data)
|
355
|
+
{
|
356
|
+
FREEMEM(LZ4HC_Data);
|
357
|
+
return (0);
|
358
|
+
}
|
359
|
+
|
360
|
+
|
361
|
+
// Update chains up to ip (excluded)
|
362
|
+
FORCE_INLINE void LZ4HC_Insert (LZ4HC_Data_Structure* hc4, const BYTE* ip)
|
363
|
+
{
|
364
|
+
U16* chainTable = hc4->chainTable;
|
365
|
+
HTYPE* HashTable = hc4->hashTable;
|
366
|
+
INITBASE(base,hc4->base);
|
367
|
+
|
368
|
+
while(hc4->nextToUpdate < ip)
|
369
|
+
{
|
370
|
+
const BYTE* const p = hc4->nextToUpdate;
|
371
|
+
size_t delta = (p) - HASH_POINTER(p);
|
372
|
+
if (delta>MAX_DISTANCE) delta = MAX_DISTANCE;
|
373
|
+
DELTANEXT(p) = (U16)delta;
|
374
|
+
HashTable[HASH_VALUE(p)] = (HTYPE)((p) - base);
|
375
|
+
hc4->nextToUpdate++;
|
376
|
+
}
|
377
|
+
}
|
378
|
+
|
379
|
+
|
380
|
+
char* LZ4_slideInputBufferHC(void* LZ4HC_Data)
|
381
|
+
{
|
382
|
+
LZ4HC_Data_Structure* hc4 = (LZ4HC_Data_Structure*)LZ4HC_Data;
|
383
|
+
U32 distance = (U32)(hc4->end - hc4->inputBuffer) - 64 KB;
|
384
|
+
distance = (distance >> 16) << 16; // Must be a multiple of 64 KB
|
385
|
+
LZ4HC_Insert(hc4, hc4->end - MINMATCH);
|
386
|
+
memcpy((void*)(hc4->end - 64 KB - distance), (const void*)(hc4->end - 64 KB), 64 KB);
|
387
|
+
hc4->nextToUpdate -= distance;
|
388
|
+
hc4->base -= distance;
|
389
|
+
if ((U32)(hc4->inputBuffer - hc4->base) > 1 GB + 64 KB) // Avoid overflow
|
390
|
+
{
|
391
|
+
int i;
|
392
|
+
hc4->base += 1 GB;
|
393
|
+
for (i=0; i<HASHTABLESIZE; i++) hc4->hashTable[i] -= 1 GB;
|
394
|
+
}
|
395
|
+
hc4->end -= distance;
|
396
|
+
return (char*)(hc4->end);
|
397
|
+
}
|
398
|
+
|
399
|
+
|
400
|
+
FORCE_INLINE size_t LZ4HC_CommonLength (const BYTE* p1, const BYTE* p2, const BYTE* const matchlimit)
|
401
|
+
{
|
402
|
+
const BYTE* p1t = p1;
|
403
|
+
|
404
|
+
while (p1t<matchlimit-(STEPSIZE-1))
|
405
|
+
{
|
406
|
+
UARCH diff = AARCH(p2) ^ AARCH(p1t);
|
407
|
+
if (!diff) { p1t+=STEPSIZE; p2+=STEPSIZE; continue; }
|
408
|
+
p1t += LZ4_NbCommonBytes(diff);
|
409
|
+
return (p1t - p1);
|
410
|
+
}
|
411
|
+
if (LZ4_ARCH64) if ((p1t<(matchlimit-3)) && (A32(p2) == A32(p1t))) { p1t+=4; p2+=4; }
|
412
|
+
if ((p1t<(matchlimit-1)) && (A16(p2) == A16(p1t))) { p1t+=2; p2+=2; }
|
413
|
+
if ((p1t<matchlimit) && (*p2 == *p1t)) p1t++;
|
414
|
+
return (p1t - p1);
|
415
|
+
}
|
416
|
+
|
417
|
+
|
418
|
+
FORCE_INLINE int LZ4HC_InsertAndFindBestMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* const matchlimit, const BYTE** matchpos)
|
419
|
+
{
|
420
|
+
U16* const chainTable = hc4->chainTable;
|
421
|
+
HTYPE* const HashTable = hc4->hashTable;
|
422
|
+
const BYTE* ref;
|
423
|
+
INITBASE(base,hc4->base);
|
424
|
+
int nbAttempts=MAX_NB_ATTEMPTS;
|
425
|
+
size_t repl=0, ml=0;
|
426
|
+
U16 delta=0; // useless assignment, to remove an uninitialization warning
|
427
|
+
|
428
|
+
// HC4 match finder
|
429
|
+
LZ4HC_Insert(hc4, ip);
|
430
|
+
ref = HASH_POINTER(ip);
|
431
|
+
|
432
|
+
#define REPEAT_OPTIMIZATION
|
433
|
+
#ifdef REPEAT_OPTIMIZATION
|
434
|
+
// Detect repetitive sequences of length <= 4
|
435
|
+
if ((U32)(ip-ref) <= 4) // potential repetition
|
436
|
+
{
|
437
|
+
if (A32(ref) == A32(ip)) // confirmed
|
438
|
+
{
|
439
|
+
delta = (U16)(ip-ref);
|
440
|
+
repl = ml = LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit) + MINMATCH;
|
441
|
+
*matchpos = ref;
|
442
|
+
}
|
443
|
+
ref = GETNEXT(ref);
|
444
|
+
}
|
445
|
+
#endif
|
446
|
+
|
447
|
+
while (((U32)(ip-ref) <= MAX_DISTANCE) && (nbAttempts))
|
448
|
+
{
|
449
|
+
nbAttempts--;
|
450
|
+
if (*(ref+ml) == *(ip+ml))
|
451
|
+
if (A32(ref) == A32(ip))
|
452
|
+
{
|
453
|
+
size_t mlt = LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit) + MINMATCH;
|
454
|
+
if (mlt > ml) { ml = mlt; *matchpos = ref; }
|
455
|
+
}
|
456
|
+
ref = GETNEXT(ref);
|
457
|
+
}
|
458
|
+
|
459
|
+
#ifdef REPEAT_OPTIMIZATION
|
460
|
+
// Complete table
|
461
|
+
if (repl)
|
462
|
+
{
|
463
|
+
const BYTE* ptr = ip;
|
464
|
+
const BYTE* end;
|
465
|
+
|
466
|
+
end = ip + repl - (MINMATCH-1);
|
467
|
+
while(ptr < end-delta)
|
468
|
+
{
|
469
|
+
DELTANEXT(ptr) = delta; // Pre-Load
|
470
|
+
ptr++;
|
471
|
+
}
|
472
|
+
do
|
473
|
+
{
|
474
|
+
DELTANEXT(ptr) = delta;
|
475
|
+
HashTable[HASH_VALUE(ptr)] = (HTYPE)((ptr) - base); // Head of chain
|
476
|
+
ptr++;
|
477
|
+
} while(ptr < end);
|
478
|
+
hc4->nextToUpdate = end;
|
479
|
+
}
|
480
|
+
#endif
|
481
|
+
|
482
|
+
return (int)ml;
|
483
|
+
}
|
484
|
+
|
485
|
+
|
486
|
+
FORCE_INLINE int LZ4HC_InsertAndGetWiderMatch (LZ4HC_Data_Structure* hc4, const BYTE* ip, const BYTE* startLimit, const BYTE* matchlimit, int longest, const BYTE** matchpos, const BYTE** startpos)
|
487
|
+
{
|
488
|
+
U16* const chainTable = hc4->chainTable;
|
489
|
+
HTYPE* const HashTable = hc4->hashTable;
|
490
|
+
INITBASE(base,hc4->base);
|
491
|
+
const BYTE* ref;
|
492
|
+
int nbAttempts = MAX_NB_ATTEMPTS;
|
493
|
+
int delta = (int)(ip-startLimit);
|
494
|
+
|
495
|
+
// First Match
|
496
|
+
LZ4HC_Insert(hc4, ip);
|
497
|
+
ref = HASH_POINTER(ip);
|
498
|
+
|
499
|
+
while (((U32)(ip-ref) <= MAX_DISTANCE) && (nbAttempts))
|
500
|
+
{
|
501
|
+
nbAttempts--;
|
502
|
+
if (*(startLimit + longest) == *(ref - delta + longest))
|
503
|
+
if (A32(ref) == A32(ip))
|
504
|
+
{
|
505
|
+
#if 1
|
506
|
+
const BYTE* reft = ref+MINMATCH;
|
507
|
+
const BYTE* ipt = ip+MINMATCH;
|
508
|
+
const BYTE* startt = ip;
|
509
|
+
|
510
|
+
while (ipt<matchlimit-(STEPSIZE-1))
|
511
|
+
{
|
512
|
+
UARCH diff = AARCH(reft) ^ AARCH(ipt);
|
513
|
+
if (!diff) { ipt+=STEPSIZE; reft+=STEPSIZE; continue; }
|
514
|
+
ipt += LZ4_NbCommonBytes(diff);
|
515
|
+
goto _endCount;
|
516
|
+
}
|
517
|
+
if (LZ4_ARCH64) if ((ipt<(matchlimit-3)) && (A32(reft) == A32(ipt))) { ipt+=4; reft+=4; }
|
518
|
+
if ((ipt<(matchlimit-1)) && (A16(reft) == A16(ipt))) { ipt+=2; reft+=2; }
|
519
|
+
if ((ipt<matchlimit) && (*reft == *ipt)) ipt++;
|
520
|
+
_endCount:
|
521
|
+
reft = ref;
|
522
|
+
#else
|
523
|
+
// Easier for code maintenance, but unfortunately slower too
|
524
|
+
const BYTE* startt = ip;
|
525
|
+
const BYTE* reft = ref;
|
526
|
+
const BYTE* ipt = ip + MINMATCH + LZ4HC_CommonLength(ip+MINMATCH, ref+MINMATCH, matchlimit);
|
527
|
+
#endif
|
528
|
+
|
529
|
+
while ((startt>startLimit) && (reft > hc4->inputBuffer) && (startt[-1] == reft[-1])) {startt--; reft--;}
|
530
|
+
|
531
|
+
if ((ipt-startt) > longest)
|
532
|
+
{
|
533
|
+
longest = (int)(ipt-startt);
|
534
|
+
*matchpos = reft;
|
535
|
+
*startpos = startt;
|
536
|
+
}
|
537
|
+
}
|
538
|
+
ref = GETNEXT(ref);
|
539
|
+
}
|
540
|
+
|
541
|
+
return longest;
|
542
|
+
}
|
543
|
+
|
544
|
+
|
545
|
+
typedef enum { noLimit = 0, limitedOutput = 1 } limitedOutput_directive;
|
546
|
+
|
547
|
+
FORCE_INLINE int LZ4HC_encodeSequence (
|
548
|
+
const BYTE** ip,
|
549
|
+
BYTE** op,
|
550
|
+
const BYTE** anchor,
|
551
|
+
int matchLength,
|
552
|
+
const BYTE* ref,
|
553
|
+
limitedOutput_directive limitedOutputBuffer,
|
554
|
+
BYTE* oend)
|
555
|
+
{
|
556
|
+
int length;
|
557
|
+
BYTE* token;
|
558
|
+
|
559
|
+
// Encode Literal length
|
560
|
+
length = (int)(*ip - *anchor);
|
561
|
+
token = (*op)++;
|
562
|
+
if ((limitedOutputBuffer) && ((*op + length + (2 + 1 + LASTLITERALS) + (length>>8)) > oend)) return 1; // Check output limit
|
563
|
+
if (length>=(int)RUN_MASK) { int len; *token=(RUN_MASK<<ML_BITS); len = length-RUN_MASK; for(; len > 254 ; len-=255) *(*op)++ = 255; *(*op)++ = (BYTE)len; }
|
564
|
+
else *token = (BYTE)(length<<ML_BITS);
|
565
|
+
|
566
|
+
// Copy Literals
|
567
|
+
LZ4_BLINDCOPY(*anchor, *op, length);
|
568
|
+
|
569
|
+
// Encode Offset
|
570
|
+
LZ4_WRITE_LITTLEENDIAN_16(*op,(U16)(*ip-ref));
|
571
|
+
|
572
|
+
// Encode MatchLength
|
573
|
+
length = (int)(matchLength-MINMATCH);
|
574
|
+
if ((limitedOutputBuffer) && (*op + (1 + LASTLITERALS) + (length>>8) > oend)) return 1; // Check output limit
|
575
|
+
if (length>=(int)ML_MASK) { *token+=ML_MASK; length-=ML_MASK; for(; length > 509 ; length-=510) { *(*op)++ = 255; *(*op)++ = 255; } if (length > 254) { length-=255; *(*op)++ = 255; } *(*op)++ = (BYTE)length; }
|
576
|
+
else *token += (BYTE)(length);
|
577
|
+
|
578
|
+
// Prepare next loop
|
579
|
+
*ip += matchLength;
|
580
|
+
*anchor = *ip;
|
581
|
+
|
582
|
+
return 0;
|
583
|
+
}
|
584
|
+
|
585
|
+
|
586
|
+
static int LZ4HC_compress_generic (
|
587
|
+
void* ctxvoid,
|
588
|
+
const char* source,
|
589
|
+
char* dest,
|
590
|
+
int inputSize,
|
591
|
+
int maxOutputSize,
|
592
|
+
limitedOutput_directive limit
|
593
|
+
)
|
594
|
+
{
|
595
|
+
LZ4HC_Data_Structure* ctx = (LZ4HC_Data_Structure*) ctxvoid;
|
596
|
+
const BYTE* ip = (const BYTE*) source;
|
597
|
+
const BYTE* anchor = ip;
|
598
|
+
const BYTE* const iend = ip + inputSize;
|
599
|
+
const BYTE* const mflimit = iend - MFLIMIT;
|
600
|
+
const BYTE* const matchlimit = (iend - LASTLITERALS);
|
601
|
+
|
602
|
+
BYTE* op = (BYTE*) dest;
|
603
|
+
BYTE* const oend = op + maxOutputSize;
|
604
|
+
|
605
|
+
int ml, ml2, ml3, ml0;
|
606
|
+
const BYTE* ref=NULL;
|
607
|
+
const BYTE* start2=NULL;
|
608
|
+
const BYTE* ref2=NULL;
|
609
|
+
const BYTE* start3=NULL;
|
610
|
+
const BYTE* ref3=NULL;
|
611
|
+
const BYTE* start0;
|
612
|
+
const BYTE* ref0;
|
613
|
+
|
614
|
+
|
615
|
+
// Ensure blocks follow each other
|
616
|
+
if (ip != ctx->end) return 0;
|
617
|
+
ctx->end += inputSize;
|
618
|
+
|
619
|
+
ip++;
|
620
|
+
|
621
|
+
// Main Loop
|
622
|
+
while (ip < mflimit)
|
623
|
+
{
|
624
|
+
ml = LZ4HC_InsertAndFindBestMatch (ctx, ip, matchlimit, (&ref));
|
625
|
+
if (!ml) { ip++; continue; }
|
626
|
+
|
627
|
+
// saved, in case we would skip too much
|
628
|
+
start0 = ip;
|
629
|
+
ref0 = ref;
|
630
|
+
ml0 = ml;
|
631
|
+
|
632
|
+
_Search2:
|
633
|
+
if (ip+ml < mflimit)
|
634
|
+
ml2 = LZ4HC_InsertAndGetWiderMatch(ctx, ip + ml - 2, ip + 1, matchlimit, ml, &ref2, &start2);
|
635
|
+
else ml2 = ml;
|
636
|
+
|
637
|
+
if (ml2 == ml) // No better match
|
638
|
+
{
|
639
|
+
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0;
|
640
|
+
continue;
|
641
|
+
}
|
642
|
+
|
643
|
+
if (start0 < ip)
|
644
|
+
{
|
645
|
+
if (start2 < ip + ml0) // empirical
|
646
|
+
{
|
647
|
+
ip = start0;
|
648
|
+
ref = ref0;
|
649
|
+
ml = ml0;
|
650
|
+
}
|
651
|
+
}
|
652
|
+
|
653
|
+
// Here, start0==ip
|
654
|
+
if ((start2 - ip) < 3) // First Match too small : removed
|
655
|
+
{
|
656
|
+
ml = ml2;
|
657
|
+
ip = start2;
|
658
|
+
ref =ref2;
|
659
|
+
goto _Search2;
|
660
|
+
}
|
661
|
+
|
662
|
+
_Search3:
|
663
|
+
// Currently we have :
|
664
|
+
// ml2 > ml1, and
|
665
|
+
// ip1+3 <= ip2 (usually < ip1+ml1)
|
666
|
+
if ((start2 - ip) < OPTIMAL_ML)
|
667
|
+
{
|
668
|
+
int correction;
|
669
|
+
int new_ml = ml;
|
670
|
+
if (new_ml > OPTIMAL_ML) new_ml = OPTIMAL_ML;
|
671
|
+
if (ip+new_ml > start2 + ml2 - MINMATCH) new_ml = (int)(start2 - ip) + ml2 - MINMATCH;
|
672
|
+
correction = new_ml - (int)(start2 - ip);
|
673
|
+
if (correction > 0)
|
674
|
+
{
|
675
|
+
start2 += correction;
|
676
|
+
ref2 += correction;
|
677
|
+
ml2 -= correction;
|
678
|
+
}
|
679
|
+
}
|
680
|
+
// Now, we have start2 = ip+new_ml, with new_ml = min(ml, OPTIMAL_ML=18)
|
681
|
+
|
682
|
+
if (start2 + ml2 < mflimit)
|
683
|
+
ml3 = LZ4HC_InsertAndGetWiderMatch(ctx, start2 + ml2 - 3, start2, matchlimit, ml2, &ref3, &start3);
|
684
|
+
else ml3 = ml2;
|
685
|
+
|
686
|
+
if (ml3 == ml2) // No better match : 2 sequences to encode
|
687
|
+
{
|
688
|
+
// ip & ref are known; Now for ml
|
689
|
+
if (start2 < ip+ml) ml = (int)(start2 - ip);
|
690
|
+
// Now, encode 2 sequences
|
691
|
+
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0;
|
692
|
+
ip = start2;
|
693
|
+
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml2, ref2, limit, oend)) return 0;
|
694
|
+
continue;
|
695
|
+
}
|
696
|
+
|
697
|
+
if (start3 < ip+ml+3) // Not enough space for match 2 : remove it
|
698
|
+
{
|
699
|
+
if (start3 >= (ip+ml)) // can write Seq1 immediately ==> Seq2 is removed, so Seq3 becomes Seq1
|
700
|
+
{
|
701
|
+
if (start2 < ip+ml)
|
702
|
+
{
|
703
|
+
int correction = (int)(ip+ml - start2);
|
704
|
+
start2 += correction;
|
705
|
+
ref2 += correction;
|
706
|
+
ml2 -= correction;
|
707
|
+
if (ml2 < MINMATCH)
|
708
|
+
{
|
709
|
+
start2 = start3;
|
710
|
+
ref2 = ref3;
|
711
|
+
ml2 = ml3;
|
712
|
+
}
|
713
|
+
}
|
714
|
+
|
715
|
+
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0;
|
716
|
+
ip = start3;
|
717
|
+
ref = ref3;
|
718
|
+
ml = ml3;
|
719
|
+
|
720
|
+
start0 = start2;
|
721
|
+
ref0 = ref2;
|
722
|
+
ml0 = ml2;
|
723
|
+
goto _Search2;
|
724
|
+
}
|
725
|
+
|
726
|
+
start2 = start3;
|
727
|
+
ref2 = ref3;
|
728
|
+
ml2 = ml3;
|
729
|
+
goto _Search3;
|
730
|
+
}
|
731
|
+
|
732
|
+
// OK, now we have 3 ascending matches; let's write at least the first one
|
733
|
+
// ip & ref are known; Now for ml
|
734
|
+
if (start2 < ip+ml)
|
735
|
+
{
|
736
|
+
if ((start2 - ip) < (int)ML_MASK)
|
737
|
+
{
|
738
|
+
int correction;
|
739
|
+
if (ml > OPTIMAL_ML) ml = OPTIMAL_ML;
|
740
|
+
if (ip + ml > start2 + ml2 - MINMATCH) ml = (int)(start2 - ip) + ml2 - MINMATCH;
|
741
|
+
correction = ml - (int)(start2 - ip);
|
742
|
+
if (correction > 0)
|
743
|
+
{
|
744
|
+
start2 += correction;
|
745
|
+
ref2 += correction;
|
746
|
+
ml2 -= correction;
|
747
|
+
}
|
748
|
+
}
|
749
|
+
else
|
750
|
+
{
|
751
|
+
ml = (int)(start2 - ip);
|
752
|
+
}
|
753
|
+
}
|
754
|
+
if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) return 0;
|
755
|
+
|
756
|
+
ip = start2;
|
757
|
+
ref = ref2;
|
758
|
+
ml = ml2;
|
759
|
+
|
760
|
+
start2 = start3;
|
761
|
+
ref2 = ref3;
|
762
|
+
ml2 = ml3;
|
763
|
+
|
764
|
+
goto _Search3;
|
765
|
+
|
766
|
+
}
|
767
|
+
|
768
|
+
// Encode Last Literals
|
769
|
+
{
|
770
|
+
int lastRun = (int)(iend - anchor);
|
771
|
+
if ((limit) && (((char*)op - dest) + lastRun + 1 + ((lastRun+255-RUN_MASK)/255) > (U32)maxOutputSize)) return 0; // Check output limit
|
772
|
+
if (lastRun>=(int)RUN_MASK) { *op++=(RUN_MASK<<ML_BITS); lastRun-=RUN_MASK; for(; lastRun > 254 ; lastRun-=255) *op++ = 255; *op++ = (BYTE) lastRun; }
|
773
|
+
else *op++ = (BYTE)(lastRun<<ML_BITS);
|
774
|
+
memcpy(op, anchor, iend - anchor);
|
775
|
+
op += iend-anchor;
|
776
|
+
}
|
777
|
+
|
778
|
+
// End
|
779
|
+
return (int) (((char*)op)-dest);
|
780
|
+
}
|
781
|
+
|
782
|
+
|
783
|
+
int LZ4_compressHC(const char* source, char* dest, int inputSize)
|
784
|
+
{
|
785
|
+
void* ctx = LZ4_createHC(source);
|
786
|
+
int result;
|
787
|
+
if (ctx==NULL) return 0;
|
788
|
+
|
789
|
+
result = LZ4HC_compress_generic (ctx, source, dest, inputSize, 0, noLimit);
|
790
|
+
|
791
|
+
LZ4_freeHC(ctx);
|
792
|
+
return result;
|
793
|
+
}
|
794
|
+
|
795
|
+
int LZ4_compressHC_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize)
|
796
|
+
{
|
797
|
+
return LZ4HC_compress_generic (LZ4HC_Data, source, dest, inputSize, 0, noLimit);
|
798
|
+
}
|
799
|
+
|
800
|
+
|
801
|
+
int LZ4_compressHC_limitedOutput(const char* source, char* dest, int inputSize, int maxOutputSize)
|
802
|
+
{
|
803
|
+
void* ctx = LZ4_createHC(source);
|
804
|
+
int result;
|
805
|
+
if (ctx==NULL) return 0;
|
806
|
+
|
807
|
+
result = LZ4HC_compress_generic (ctx, source, dest, inputSize, maxOutputSize, limitedOutput);
|
808
|
+
|
809
|
+
LZ4_freeHC(ctx);
|
810
|
+
return result;
|
811
|
+
}
|
812
|
+
|
813
|
+
int LZ4_compressHC_limitedOutput_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize, int maxOutputSize)
|
814
|
+
{
|
815
|
+
return LZ4HC_compress_generic (LZ4HC_Data, source, dest, inputSize, maxOutputSize, limitedOutput);
|
816
|
+
}
|
817
|
+
|