zopfli-bin 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.document +5 -0
- data/.gitmodules +3 -0
- data/.testguardrc +1 -0
- data/Gemfile +17 -0
- data/Gemfile.lock +111 -0
- data/LICENSE.txt +20 -0
- data/README.rdoc +19 -0
- data/Rakefile +39 -0
- data/VERSION +1 -0
- data/ext/Makefile +8 -0
- data/ext/extconf.rb +4 -0
- data/lib/zopfli-bin.rb +5 -0
- data/lib/zopfli/bin.rb +34 -0
- data/test/helper.rb +19 -0
- data/test/test_zopfli-bin.rb +33 -0
- data/vendor/zopfli/CONTRIBUTORS +7 -0
- data/vendor/zopfli/COPYING +201 -0
- data/vendor/zopfli/Makefile +37 -0
- data/vendor/zopfli/README +32 -0
- data/vendor/zopfli/README.zopflipng +35 -0
- data/vendor/zopfli/src/zopfli/blocksplitter.c +342 -0
- data/vendor/zopfli/src/zopfli/blocksplitter.h +77 -0
- data/vendor/zopfli/src/zopfli/cache.c +119 -0
- data/vendor/zopfli/src/zopfli/cache.h +66 -0
- data/vendor/zopfli/src/zopfli/deflate.c +866 -0
- data/vendor/zopfli/src/zopfli/deflate.h +86 -0
- data/vendor/zopfli/src/zopfli/gzip_container.c +117 -0
- data/vendor/zopfli/src/zopfli/gzip_container.h +50 -0
- data/vendor/zopfli/src/zopfli/hash.c +135 -0
- data/vendor/zopfli/src/zopfli/hash.h +70 -0
- data/vendor/zopfli/src/zopfli/katajainen.c +251 -0
- data/vendor/zopfli/src/zopfli/katajainen.h +42 -0
- data/vendor/zopfli/src/zopfli/lz77.c +482 -0
- data/vendor/zopfli/src/zopfli/lz77.h +129 -0
- data/vendor/zopfli/src/zopfli/squeeze.c +546 -0
- data/vendor/zopfli/src/zopfli/squeeze.h +60 -0
- data/vendor/zopfli/src/zopfli/tree.c +101 -0
- data/vendor/zopfli/src/zopfli/tree.h +51 -0
- data/vendor/zopfli/src/zopfli/util.c +213 -0
- data/vendor/zopfli/src/zopfli/util.h +175 -0
- data/vendor/zopfli/src/zopfli/zlib_container.c +79 -0
- data/vendor/zopfli/src/zopfli/zlib_container.h +50 -0
- data/vendor/zopfli/src/zopfli/zopfli.h +97 -0
- data/vendor/zopfli/src/zopfli/zopfli_bin.c +203 -0
- data/vendor/zopfli/src/zopfli/zopfli_lib.c +42 -0
- data/vendor/zopfli/src/zopflipng/lodepng/lodepng.cpp +6260 -0
- data/vendor/zopfli/src/zopflipng/lodepng/lodepng.h +1716 -0
- data/vendor/zopfli/src/zopflipng/lodepng/lodepng_util.cpp +656 -0
- data/vendor/zopfli/src/zopflipng/lodepng/lodepng_util.h +151 -0
- data/vendor/zopfli/src/zopflipng/zopflipng_bin.cc +407 -0
- data/vendor/zopfli/src/zopflipng/zopflipng_lib.cc +425 -0
- data/vendor/zopfli/src/zopflipng/zopflipng_lib.h +79 -0
- data/zopfli-bin.gemspec +119 -0
- metadata +225 -0
@@ -0,0 +1,77 @@
|
|
1
|
+
/*
|
2
|
+
Copyright 2011 Google Inc. All Rights Reserved.
|
3
|
+
|
4
|
+
Licensed under the Apache License, Version 2.0 (the "License");
|
5
|
+
you may not use this file except in compliance with the License.
|
6
|
+
You may obtain a copy of the License at
|
7
|
+
|
8
|
+
http://www.apache.org/licenses/LICENSE-2.0
|
9
|
+
|
10
|
+
Unless required by applicable law or agreed to in writing, software
|
11
|
+
distributed under the License is distributed on an "AS IS" BASIS,
|
12
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13
|
+
See the License for the specific language governing permissions and
|
14
|
+
limitations under the License.
|
15
|
+
|
16
|
+
Author: lode.vandevenne@gmail.com (Lode Vandevenne)
|
17
|
+
Author: jyrki.alakuijala@gmail.com (Jyrki Alakuijala)
|
18
|
+
*/
|
19
|
+
|
20
|
+
/*
|
21
|
+
Functions to choose good boundaries for block splitting. Deflate allows encoding
|
22
|
+
the data in multiple blocks, with a separate Huffman tree for each block. The
|
23
|
+
Huffman tree itself requires some bytes to encode, so by choosing certain
|
24
|
+
blocks, you can either hurt, or enhance compression. These functions choose good
|
25
|
+
ones that enhance it.
|
26
|
+
*/
|
27
|
+
|
28
|
+
#ifndef ZOPFLI_BLOCKSPLITTER_H_
|
29
|
+
#define ZOPFLI_BLOCKSPLITTER_H_
|
30
|
+
|
31
|
+
#include <stdlib.h>
|
32
|
+
|
33
|
+
#include "zopfli.h"
|
34
|
+
|
35
|
+
|
36
|
+
/*
|
37
|
+
Does blocksplitting on LZ77 data.
|
38
|
+
The output splitpoints are indices in the LZ77 data.
|
39
|
+
litlens: lz77 lit/lengths
|
40
|
+
dists: lz77 distances
|
41
|
+
llsize: size of litlens and dists
|
42
|
+
maxblocks: set a limit to the amount of blocks. Set to 0 to mean no limit.
|
43
|
+
*/
|
44
|
+
void ZopfliBlockSplitLZ77(const ZopfliOptions* options,
|
45
|
+
const unsigned short* litlens,
|
46
|
+
const unsigned short* dists,
|
47
|
+
size_t llsize, size_t maxblocks,
|
48
|
+
size_t** splitpoints, size_t* npoints);
|
49
|
+
|
50
|
+
/*
|
51
|
+
Does blocksplitting on uncompressed data.
|
52
|
+
The output splitpoints are indices in the uncompressed bytes.
|
53
|
+
|
54
|
+
options: general program options.
|
55
|
+
in: uncompressed input data
|
56
|
+
instart: where to start splitting
|
57
|
+
inend: where to end splitting (not inclusive)
|
58
|
+
maxblocks: maximum amount of blocks to split into, or 0 for no limit
|
59
|
+
splitpoints: dynamic array to put the resulting split point coordinates into.
|
60
|
+
The coordinates are indices in the input array.
|
61
|
+
npoints: pointer to amount of splitpoints, for the dynamic array. The amount of
|
62
|
+
blocks is the amount of splitpoitns + 1.
|
63
|
+
*/
|
64
|
+
void ZopfliBlockSplit(const ZopfliOptions* options,
|
65
|
+
const unsigned char* in, size_t instart, size_t inend,
|
66
|
+
size_t maxblocks, size_t** splitpoints, size_t* npoints);
|
67
|
+
|
68
|
+
/*
|
69
|
+
Divides the input into equal blocks, does not even take LZ77 lengths into
|
70
|
+
account.
|
71
|
+
*/
|
72
|
+
void ZopfliBlockSplitSimple(const unsigned char* in,
|
73
|
+
size_t instart, size_t inend,
|
74
|
+
size_t blocksize,
|
75
|
+
size_t** splitpoints, size_t* npoints);
|
76
|
+
|
77
|
+
#endif /* ZOPFLI_BLOCKSPLITTER_H_ */
|
@@ -0,0 +1,119 @@
|
|
1
|
+
/*
|
2
|
+
Copyright 2011 Google Inc. All Rights Reserved.
|
3
|
+
|
4
|
+
Licensed under the Apache License, Version 2.0 (the "License");
|
5
|
+
you may not use this file except in compliance with the License.
|
6
|
+
You may obtain a copy of the License at
|
7
|
+
|
8
|
+
http://www.apache.org/licenses/LICENSE-2.0
|
9
|
+
|
10
|
+
Unless required by applicable law or agreed to in writing, software
|
11
|
+
distributed under the License is distributed on an "AS IS" BASIS,
|
12
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13
|
+
See the License for the specific language governing permissions and
|
14
|
+
limitations under the License.
|
15
|
+
|
16
|
+
Author: lode.vandevenne@gmail.com (Lode Vandevenne)
|
17
|
+
Author: jyrki.alakuijala@gmail.com (Jyrki Alakuijala)
|
18
|
+
*/
|
19
|
+
|
20
|
+
#include "cache.h"
|
21
|
+
|
22
|
+
#include <assert.h>
|
23
|
+
#include <stdio.h>
|
24
|
+
#include <stdlib.h>
|
25
|
+
|
26
|
+
#ifdef ZOPFLI_LONGEST_MATCH_CACHE
|
27
|
+
|
28
|
+
void ZopfliInitCache(size_t blocksize, ZopfliLongestMatchCache* lmc) {
|
29
|
+
size_t i;
|
30
|
+
lmc->length = (unsigned short*)malloc(sizeof(unsigned short) * blocksize);
|
31
|
+
lmc->dist = (unsigned short*)malloc(sizeof(unsigned short) * blocksize);
|
32
|
+
/* Rather large amount of memory. */
|
33
|
+
lmc->sublen = (unsigned char*)malloc(ZOPFLI_CACHE_LENGTH * 3 * blocksize);
|
34
|
+
|
35
|
+
/* length > 0 and dist 0 is invalid combination, which indicates on purpose
|
36
|
+
that this cache value is not filled in yet. */
|
37
|
+
for (i = 0; i < blocksize; i++) lmc->length[i] = 1;
|
38
|
+
for (i = 0; i < blocksize; i++) lmc->dist[i] = 0;
|
39
|
+
for (i = 0; i < ZOPFLI_CACHE_LENGTH * blocksize * 3; i++) lmc->sublen[i] = 0;
|
40
|
+
}
|
41
|
+
|
42
|
+
void ZopfliCleanCache(ZopfliLongestMatchCache* lmc) {
|
43
|
+
free(lmc->length);
|
44
|
+
free(lmc->dist);
|
45
|
+
free(lmc->sublen);
|
46
|
+
}
|
47
|
+
|
48
|
+
void ZopfliSublenToCache(const unsigned short* sublen,
|
49
|
+
size_t pos, size_t length,
|
50
|
+
ZopfliLongestMatchCache* lmc) {
|
51
|
+
size_t i;
|
52
|
+
size_t j = 0;
|
53
|
+
unsigned bestlength = 0;
|
54
|
+
unsigned char* cache;
|
55
|
+
|
56
|
+
#if ZOPFLI_CACHE_LENGTH == 0
|
57
|
+
return;
|
58
|
+
#endif
|
59
|
+
|
60
|
+
cache = &lmc->sublen[ZOPFLI_CACHE_LENGTH * pos * 3];
|
61
|
+
if (length < 3) return;
|
62
|
+
for (i = 3; i <= length; i++) {
|
63
|
+
if (i == length || sublen[i] != sublen[i + 1]) {
|
64
|
+
cache[j * 3] = i - 3;
|
65
|
+
cache[j * 3 + 1] = sublen[i] % 256;
|
66
|
+
cache[j * 3 + 2] = (sublen[i] >> 8) % 256;
|
67
|
+
bestlength = i;
|
68
|
+
j++;
|
69
|
+
if (j >= ZOPFLI_CACHE_LENGTH) break;
|
70
|
+
}
|
71
|
+
}
|
72
|
+
if (j < ZOPFLI_CACHE_LENGTH) {
|
73
|
+
assert(bestlength == length);
|
74
|
+
cache[(ZOPFLI_CACHE_LENGTH - 1) * 3] = bestlength - 3;
|
75
|
+
} else {
|
76
|
+
assert(bestlength <= length);
|
77
|
+
}
|
78
|
+
assert(bestlength == ZopfliMaxCachedSublen(lmc, pos, length));
|
79
|
+
}
|
80
|
+
|
81
|
+
void ZopfliCacheToSublen(const ZopfliLongestMatchCache* lmc,
|
82
|
+
size_t pos, size_t length,
|
83
|
+
unsigned short* sublen) {
|
84
|
+
size_t i, j;
|
85
|
+
unsigned maxlength = ZopfliMaxCachedSublen(lmc, pos, length);
|
86
|
+
unsigned prevlength = 0;
|
87
|
+
unsigned char* cache;
|
88
|
+
#if ZOPFLI_CACHE_LENGTH == 0
|
89
|
+
return;
|
90
|
+
#endif
|
91
|
+
if (length < 3) return;
|
92
|
+
cache = &lmc->sublen[ZOPFLI_CACHE_LENGTH * pos * 3];
|
93
|
+
for (j = 0; j < ZOPFLI_CACHE_LENGTH; j++) {
|
94
|
+
unsigned length = cache[j * 3] + 3;
|
95
|
+
unsigned dist = cache[j * 3 + 1] + 256 * cache[j * 3 + 2];
|
96
|
+
for (i = prevlength; i <= length; i++) {
|
97
|
+
sublen[i] = dist;
|
98
|
+
}
|
99
|
+
if (length == maxlength) break;
|
100
|
+
prevlength = length + 1;
|
101
|
+
}
|
102
|
+
}
|
103
|
+
|
104
|
+
/*
|
105
|
+
Returns the length up to which could be stored in the cache.
|
106
|
+
*/
|
107
|
+
unsigned ZopfliMaxCachedSublen(const ZopfliLongestMatchCache* lmc,
|
108
|
+
size_t pos, size_t length) {
|
109
|
+
unsigned char* cache;
|
110
|
+
#if ZOPFLI_CACHE_LENGTH == 0
|
111
|
+
return 0;
|
112
|
+
#endif
|
113
|
+
cache = &lmc->sublen[ZOPFLI_CACHE_LENGTH * pos * 3];
|
114
|
+
(void)length;
|
115
|
+
if (cache[1] == 0 && cache[2] == 0) return 0; /* No sublen cached. */
|
116
|
+
return cache[(ZOPFLI_CACHE_LENGTH - 1) * 3] + 3;
|
117
|
+
}
|
118
|
+
|
119
|
+
#endif /* ZOPFLI_LONGEST_MATCH_CACHE */
|
@@ -0,0 +1,66 @@
|
|
1
|
+
/*
|
2
|
+
Copyright 2011 Google Inc. All Rights Reserved.
|
3
|
+
|
4
|
+
Licensed under the Apache License, Version 2.0 (the "License");
|
5
|
+
you may not use this file except in compliance with the License.
|
6
|
+
You may obtain a copy of the License at
|
7
|
+
|
8
|
+
http://www.apache.org/licenses/LICENSE-2.0
|
9
|
+
|
10
|
+
Unless required by applicable law or agreed to in writing, software
|
11
|
+
distributed under the License is distributed on an "AS IS" BASIS,
|
12
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13
|
+
See the License for the specific language governing permissions and
|
14
|
+
limitations under the License.
|
15
|
+
|
16
|
+
Author: lode.vandevenne@gmail.com (Lode Vandevenne)
|
17
|
+
Author: jyrki.alakuijala@gmail.com (Jyrki Alakuijala)
|
18
|
+
*/
|
19
|
+
|
20
|
+
/*
|
21
|
+
The cache that speeds up ZopfliFindLongestMatch of lz77.c.
|
22
|
+
*/
|
23
|
+
|
24
|
+
#ifndef ZOPFLI_CACHE_H_
|
25
|
+
#define ZOPFLI_CACHE_H_
|
26
|
+
|
27
|
+
#include "util.h"
|
28
|
+
|
29
|
+
#ifdef ZOPFLI_LONGEST_MATCH_CACHE
|
30
|
+
|
31
|
+
/*
|
32
|
+
Cache used by ZopfliFindLongestMatch to remember previously found length/dist
|
33
|
+
values.
|
34
|
+
This is needed because the squeeze runs will ask these values multiple times for
|
35
|
+
the same position.
|
36
|
+
Uses large amounts of memory, since it has to remember the distance belonging
|
37
|
+
to every possible shorter-than-the-best length (the so called "sublen" array).
|
38
|
+
*/
|
39
|
+
typedef struct ZopfliLongestMatchCache {
|
40
|
+
unsigned short* length;
|
41
|
+
unsigned short* dist;
|
42
|
+
unsigned char* sublen;
|
43
|
+
} ZopfliLongestMatchCache;
|
44
|
+
|
45
|
+
/* Initializes the ZopfliLongestMatchCache. */
|
46
|
+
void ZopfliInitCache(size_t blocksize, ZopfliLongestMatchCache* lmc);
|
47
|
+
|
48
|
+
/* Frees up the memory of the ZopfliLongestMatchCache. */
|
49
|
+
void ZopfliCleanCache(ZopfliLongestMatchCache* lmc);
|
50
|
+
|
51
|
+
/* Stores sublen array in the cache. */
|
52
|
+
void ZopfliSublenToCache(const unsigned short* sublen,
|
53
|
+
size_t pos, size_t length,
|
54
|
+
ZopfliLongestMatchCache* lmc);
|
55
|
+
|
56
|
+
/* Extracts sublen array from the cache. */
|
57
|
+
void ZopfliCacheToSublen(const ZopfliLongestMatchCache* lmc,
|
58
|
+
size_t pos, size_t length,
|
59
|
+
unsigned short* sublen);
|
60
|
+
/* Returns the length up to which could be stored in the cache. */
|
61
|
+
unsigned ZopfliMaxCachedSublen(const ZopfliLongestMatchCache* lmc,
|
62
|
+
size_t pos, size_t length);
|
63
|
+
|
64
|
+
#endif /* ZOPFLI_LONGEST_MATCH_CACHE */
|
65
|
+
|
66
|
+
#endif /* ZOPFLI_CACHE_H_ */
|
@@ -0,0 +1,866 @@
|
|
1
|
+
/*
|
2
|
+
Copyright 2011 Google Inc. All Rights Reserved.
|
3
|
+
|
4
|
+
Licensed under the Apache License, Version 2.0 (the "License");
|
5
|
+
you may not use this file except in compliance with the License.
|
6
|
+
You may obtain a copy of the License at
|
7
|
+
|
8
|
+
http://www.apache.org/licenses/LICENSE-2.0
|
9
|
+
|
10
|
+
Unless required by applicable law or agreed to in writing, software
|
11
|
+
distributed under the License is distributed on an "AS IS" BASIS,
|
12
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13
|
+
See the License for the specific language governing permissions and
|
14
|
+
limitations under the License.
|
15
|
+
|
16
|
+
Author: lode.vandevenne@gmail.com (Lode Vandevenne)
|
17
|
+
Author: jyrki.alakuijala@gmail.com (Jyrki Alakuijala)
|
18
|
+
*/
|
19
|
+
|
20
|
+
#include "deflate.h"
|
21
|
+
|
22
|
+
#include <assert.h>
|
23
|
+
#include <stdio.h>
|
24
|
+
#include <stdlib.h>
|
25
|
+
|
26
|
+
#include "blocksplitter.h"
|
27
|
+
#include "lz77.h"
|
28
|
+
#include "squeeze.h"
|
29
|
+
#include "tree.h"
|
30
|
+
|
31
|
+
/*
|
32
|
+
bp = bitpointer, always in range [0, 7].
|
33
|
+
The outsize is number of necessary bytes to encode the bits.
|
34
|
+
Given the value of bp and the amount of bytes, the amount of bits represented
|
35
|
+
is not simply bytesize * 8 + bp because even representing one bit requires a
|
36
|
+
whole byte. It is: (bp == 0) ? (bytesize * 8) : ((bytesize - 1) * 8 + bp)
|
37
|
+
*/
|
38
|
+
static void AddBit(int bit,
|
39
|
+
unsigned char* bp, unsigned char** out, size_t* outsize) {
|
40
|
+
if (*bp == 0) ZOPFLI_APPEND_DATA(0, out, outsize);
|
41
|
+
(*out)[*outsize - 1] |= bit << *bp;
|
42
|
+
*bp = (*bp + 1) & 7;
|
43
|
+
}
|
44
|
+
|
45
|
+
static void AddBits(unsigned symbol, unsigned length,
|
46
|
+
unsigned char* bp, unsigned char** out, size_t* outsize) {
|
47
|
+
/* TODO(lode): make more efficient (add more bits at once). */
|
48
|
+
unsigned i;
|
49
|
+
for (i = 0; i < length; i++) {
|
50
|
+
unsigned bit = (symbol >> i) & 1;
|
51
|
+
if (*bp == 0) ZOPFLI_APPEND_DATA(0, out, outsize);
|
52
|
+
(*out)[*outsize - 1] |= bit << *bp;
|
53
|
+
*bp = (*bp + 1) & 7;
|
54
|
+
}
|
55
|
+
}
|
56
|
+
|
57
|
+
/*
|
58
|
+
Adds bits, like AddBits, but the order is inverted. The deflate specification
|
59
|
+
uses both orders in one standard.
|
60
|
+
*/
|
61
|
+
static void AddHuffmanBits(unsigned symbol, unsigned length,
|
62
|
+
unsigned char* bp, unsigned char** out,
|
63
|
+
size_t* outsize) {
|
64
|
+
/* TODO(lode): make more efficient (add more bits at once). */
|
65
|
+
unsigned i;
|
66
|
+
for (i = 0; i < length; i++) {
|
67
|
+
unsigned bit = (symbol >> (length - i - 1)) & 1;
|
68
|
+
if (*bp == 0) ZOPFLI_APPEND_DATA(0, out, outsize);
|
69
|
+
(*out)[*outsize - 1] |= bit << *bp;
|
70
|
+
*bp = (*bp + 1) & 7;
|
71
|
+
}
|
72
|
+
}
|
73
|
+
|
74
|
+
/*
|
75
|
+
Ensures there are at least 2 distance codes to support buggy decoders.
|
76
|
+
Zlib 1.2.1 and below have a bug where it fails if there isn't at least 1
|
77
|
+
distance code (with length > 0), even though it's valid according to the
|
78
|
+
deflate spec to have 0 distance codes. On top of that, some mobile phones
|
79
|
+
require at least two distance codes. To support these decoders too (but
|
80
|
+
potentially at the cost of a few bytes), add dummy code lengths of 1.
|
81
|
+
References to this bug can be found in the changelog of
|
82
|
+
Zlib 1.2.2 and here: http://www.jonof.id.au/forum/index.php?topic=515.0.
|
83
|
+
|
84
|
+
d_lengths: the 32 lengths of the distance codes.
|
85
|
+
*/
|
86
|
+
static void PatchDistanceCodesForBuggyDecoders(unsigned* d_lengths) {
|
87
|
+
int num_dist_codes = 0; /* Amount of non-zero distance codes */
|
88
|
+
int i;
|
89
|
+
for (i = 0; i < 30 /* Ignore the two unused codes from the spec */; i++) {
|
90
|
+
if (d_lengths[i]) num_dist_codes++;
|
91
|
+
if (num_dist_codes >= 2) return; /* Two or more codes is fine. */
|
92
|
+
}
|
93
|
+
|
94
|
+
if (num_dist_codes == 0) {
|
95
|
+
d_lengths[0] = d_lengths[1] = 1;
|
96
|
+
} else if (num_dist_codes == 1) {
|
97
|
+
d_lengths[d_lengths[0] ? 1 : 0] = 1;
|
98
|
+
}
|
99
|
+
}
|
100
|
+
|
101
|
+
/*
|
102
|
+
Encodes the Huffman tree and returns how many bits its encoding takes. If out
|
103
|
+
is a null pointer, only returns the size and runs faster.
|
104
|
+
*/
|
105
|
+
static size_t EncodeTree(const unsigned* ll_lengths,
|
106
|
+
const unsigned* d_lengths,
|
107
|
+
int use_16, int use_17, int use_18,
|
108
|
+
unsigned char* bp,
|
109
|
+
unsigned char** out, size_t* outsize) {
|
110
|
+
unsigned lld_total; /* Total amount of literal, length, distance codes. */
|
111
|
+
/* Runlength encoded version of lengths of litlen and dist trees. */
|
112
|
+
unsigned* rle = 0;
|
113
|
+
unsigned* rle_bits = 0; /* Extra bits for rle values 16, 17 and 18. */
|
114
|
+
size_t rle_size = 0; /* Size of rle array. */
|
115
|
+
size_t rle_bits_size = 0; /* Should have same value as rle_size. */
|
116
|
+
unsigned hlit = 29; /* 286 - 257 */
|
117
|
+
unsigned hdist = 29; /* 32 - 1, but gzip does not like hdist > 29.*/
|
118
|
+
unsigned hclen;
|
119
|
+
unsigned hlit2;
|
120
|
+
size_t i, j;
|
121
|
+
size_t clcounts[19];
|
122
|
+
unsigned clcl[19]; /* Code length code lengths. */
|
123
|
+
unsigned clsymbols[19];
|
124
|
+
/* The order in which code length code lengths are encoded as per deflate. */
|
125
|
+
static const unsigned order[19] = {
|
126
|
+
16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15
|
127
|
+
};
|
128
|
+
int size_only = !out;
|
129
|
+
size_t result_size = 0;
|
130
|
+
|
131
|
+
for(i = 0; i < 19; i++) clcounts[i] = 0;
|
132
|
+
|
133
|
+
/* Trim zeros. */
|
134
|
+
while (hlit > 0 && ll_lengths[257 + hlit - 1] == 0) hlit--;
|
135
|
+
while (hdist > 0 && d_lengths[1 + hdist - 1] == 0) hdist--;
|
136
|
+
hlit2 = hlit + 257;
|
137
|
+
|
138
|
+
lld_total = hlit2 + hdist + 1;
|
139
|
+
|
140
|
+
for (i = 0; i < lld_total; i++) {
|
141
|
+
/* This is an encoding of a huffman tree, so now the length is a symbol */
|
142
|
+
unsigned char symbol = i < hlit2 ? ll_lengths[i] : d_lengths[i - hlit2];
|
143
|
+
unsigned count = 1;
|
144
|
+
if(use_16 || (symbol == 0 && (use_17 || use_18))) {
|
145
|
+
for (j = i + 1; j < lld_total && symbol ==
|
146
|
+
(j < hlit2 ? ll_lengths[j] : d_lengths[j - hlit2]); j++) {
|
147
|
+
count++;
|
148
|
+
}
|
149
|
+
}
|
150
|
+
i += count - 1;
|
151
|
+
|
152
|
+
/* Repetitions of zeroes */
|
153
|
+
if (symbol == 0 && count >= 3) {
|
154
|
+
if (use_18) {
|
155
|
+
while (count >= 11) {
|
156
|
+
unsigned count2 = count > 138 ? 138 : count;
|
157
|
+
if (!size_only) {
|
158
|
+
ZOPFLI_APPEND_DATA(18, &rle, &rle_size);
|
159
|
+
ZOPFLI_APPEND_DATA(count2 - 11, &rle_bits, &rle_bits_size);
|
160
|
+
}
|
161
|
+
clcounts[18]++;
|
162
|
+
count -= count2;
|
163
|
+
}
|
164
|
+
}
|
165
|
+
if (use_17) {
|
166
|
+
while (count >= 3) {
|
167
|
+
unsigned count2 = count > 10 ? 10 : count;
|
168
|
+
if (!size_only) {
|
169
|
+
ZOPFLI_APPEND_DATA(17, &rle, &rle_size);
|
170
|
+
ZOPFLI_APPEND_DATA(count2 - 3, &rle_bits, &rle_bits_size);
|
171
|
+
}
|
172
|
+
clcounts[17]++;
|
173
|
+
count -= count2;
|
174
|
+
}
|
175
|
+
}
|
176
|
+
}
|
177
|
+
|
178
|
+
/* Repetitions of any symbol */
|
179
|
+
if (use_16 && count >= 4) {
|
180
|
+
count--; /* Since the first one is hardcoded. */
|
181
|
+
clcounts[symbol]++;
|
182
|
+
if (!size_only) {
|
183
|
+
ZOPFLI_APPEND_DATA(symbol, &rle, &rle_size);
|
184
|
+
ZOPFLI_APPEND_DATA(0, &rle_bits, &rle_bits_size);
|
185
|
+
}
|
186
|
+
while (count >= 3) {
|
187
|
+
unsigned count2 = count > 6 ? 6 : count;
|
188
|
+
if (!size_only) {
|
189
|
+
ZOPFLI_APPEND_DATA(16, &rle, &rle_size);
|
190
|
+
ZOPFLI_APPEND_DATA(count2 - 3, &rle_bits, &rle_bits_size);
|
191
|
+
}
|
192
|
+
clcounts[16]++;
|
193
|
+
count -= count2;
|
194
|
+
}
|
195
|
+
}
|
196
|
+
|
197
|
+
/* No or insufficient repetition */
|
198
|
+
clcounts[symbol] += count;
|
199
|
+
while (count > 0) {
|
200
|
+
if (!size_only) {
|
201
|
+
ZOPFLI_APPEND_DATA(symbol, &rle, &rle_size);
|
202
|
+
ZOPFLI_APPEND_DATA(0, &rle_bits, &rle_bits_size);
|
203
|
+
}
|
204
|
+
count--;
|
205
|
+
}
|
206
|
+
}
|
207
|
+
|
208
|
+
ZopfliCalculateBitLengths(clcounts, 19, 7, clcl);
|
209
|
+
if (!size_only) ZopfliLengthsToSymbols(clcl, 19, 7, clsymbols);
|
210
|
+
|
211
|
+
hclen = 15;
|
212
|
+
/* Trim zeros. */
|
213
|
+
while (hclen > 0 && clcounts[order[hclen + 4 - 1]] == 0) hclen--;
|
214
|
+
|
215
|
+
if (!size_only) {
|
216
|
+
AddBits(hlit, 5, bp, out, outsize);
|
217
|
+
AddBits(hdist, 5, bp, out, outsize);
|
218
|
+
AddBits(hclen, 4, bp, out, outsize);
|
219
|
+
|
220
|
+
for (i = 0; i < hclen + 4; i++) {
|
221
|
+
AddBits(clcl[order[i]], 3, bp, out, outsize);
|
222
|
+
}
|
223
|
+
|
224
|
+
for (i = 0; i < rle_size; i++) {
|
225
|
+
unsigned symbol = clsymbols[rle[i]];
|
226
|
+
AddHuffmanBits(symbol, clcl[rle[i]], bp, out, outsize);
|
227
|
+
/* Extra bits. */
|
228
|
+
if (rle[i] == 16) AddBits(rle_bits[i], 2, bp, out, outsize);
|
229
|
+
else if (rle[i] == 17) AddBits(rle_bits[i], 3, bp, out, outsize);
|
230
|
+
else if (rle[i] == 18) AddBits(rle_bits[i], 7, bp, out, outsize);
|
231
|
+
}
|
232
|
+
}
|
233
|
+
|
234
|
+
result_size += 14; /* hlit, hdist, hclen bits */
|
235
|
+
result_size += (hclen + 4) * 3; /* clcl bits */
|
236
|
+
for(i = 0; i < 19; i++) {
|
237
|
+
result_size += clcl[i] * clcounts[i];
|
238
|
+
}
|
239
|
+
/* Extra bits. */
|
240
|
+
result_size += clcounts[16] * 2;
|
241
|
+
result_size += clcounts[17] * 3;
|
242
|
+
result_size += clcounts[18] * 7;
|
243
|
+
|
244
|
+
/* Note: in case of "size_only" these are null pointers so no effect. */
|
245
|
+
free(rle);
|
246
|
+
free(rle_bits);
|
247
|
+
|
248
|
+
return result_size;
|
249
|
+
}
|
250
|
+
|
251
|
+
static void AddDynamicTree(const unsigned* ll_lengths,
|
252
|
+
const unsigned* d_lengths,
|
253
|
+
unsigned char* bp,
|
254
|
+
unsigned char** out, size_t* outsize) {
|
255
|
+
int i;
|
256
|
+
int best = 0;
|
257
|
+
size_t bestsize = 0;
|
258
|
+
|
259
|
+
for(i = 0; i < 8; i++) {
|
260
|
+
size_t size = EncodeTree(ll_lengths, d_lengths,
|
261
|
+
i & 1, i & 2, i & 4,
|
262
|
+
0, 0, 0);
|
263
|
+
if (bestsize == 0 || size < bestsize) {
|
264
|
+
bestsize = size;
|
265
|
+
best = i;
|
266
|
+
}
|
267
|
+
}
|
268
|
+
|
269
|
+
EncodeTree(ll_lengths, d_lengths,
|
270
|
+
best & 1, best & 2, best & 4,
|
271
|
+
bp, out, outsize);
|
272
|
+
}
|
273
|
+
|
274
|
+
/*
|
275
|
+
Gives the exact size of the tree, in bits, as it will be encoded in DEFLATE.
|
276
|
+
*/
|
277
|
+
static size_t CalculateTreeSize(const unsigned* ll_lengths,
|
278
|
+
const unsigned* d_lengths) {
|
279
|
+
size_t result = 0;
|
280
|
+
int i;
|
281
|
+
|
282
|
+
for(i = 0; i < 8; i++) {
|
283
|
+
size_t size = EncodeTree(ll_lengths, d_lengths,
|
284
|
+
i & 1, i & 2, i & 4,
|
285
|
+
0, 0, 0);
|
286
|
+
if (result == 0 || size < result) result = size;
|
287
|
+
}
|
288
|
+
|
289
|
+
return result;
|
290
|
+
}
|
291
|
+
|
292
|
+
/*
|
293
|
+
Adds all lit/len and dist codes from the lists as huffman symbols. Does not add
|
294
|
+
end code 256. expected_data_size is the uncompressed block size, used for
|
295
|
+
assert, but you can set it to 0 to not do the assertion.
|
296
|
+
*/
|
297
|
+
static void AddLZ77Data(const unsigned short* litlens,
|
298
|
+
const unsigned short* dists,
|
299
|
+
size_t lstart, size_t lend,
|
300
|
+
size_t expected_data_size,
|
301
|
+
const unsigned* ll_symbols, const unsigned* ll_lengths,
|
302
|
+
const unsigned* d_symbols, const unsigned* d_lengths,
|
303
|
+
unsigned char* bp,
|
304
|
+
unsigned char** out, size_t* outsize) {
|
305
|
+
size_t testlength = 0;
|
306
|
+
size_t i;
|
307
|
+
|
308
|
+
for (i = lstart; i < lend; i++) {
|
309
|
+
unsigned dist = dists[i];
|
310
|
+
unsigned litlen = litlens[i];
|
311
|
+
if (dist == 0) {
|
312
|
+
assert(litlen < 256);
|
313
|
+
assert(ll_lengths[litlen] > 0);
|
314
|
+
AddHuffmanBits(ll_symbols[litlen], ll_lengths[litlen], bp, out, outsize);
|
315
|
+
testlength++;
|
316
|
+
} else {
|
317
|
+
unsigned lls = ZopfliGetLengthSymbol(litlen);
|
318
|
+
unsigned ds = ZopfliGetDistSymbol(dist);
|
319
|
+
assert(litlen >= 3 && litlen <= 288);
|
320
|
+
assert(ll_lengths[lls] > 0);
|
321
|
+
assert(d_lengths[ds] > 0);
|
322
|
+
AddHuffmanBits(ll_symbols[lls], ll_lengths[lls], bp, out, outsize);
|
323
|
+
AddBits(ZopfliGetLengthExtraBitsValue(litlen),
|
324
|
+
ZopfliGetLengthExtraBits(litlen),
|
325
|
+
bp, out, outsize);
|
326
|
+
AddHuffmanBits(d_symbols[ds], d_lengths[ds], bp, out, outsize);
|
327
|
+
AddBits(ZopfliGetDistExtraBitsValue(dist),
|
328
|
+
ZopfliGetDistExtraBits(dist),
|
329
|
+
bp, out, outsize);
|
330
|
+
testlength += litlen;
|
331
|
+
}
|
332
|
+
}
|
333
|
+
assert(expected_data_size == 0 || testlength == expected_data_size);
|
334
|
+
}
|
335
|
+
|
336
|
+
static void GetFixedTree(unsigned* ll_lengths, unsigned* d_lengths) {
|
337
|
+
size_t i;
|
338
|
+
for (i = 0; i < 144; i++) ll_lengths[i] = 8;
|
339
|
+
for (i = 144; i < 256; i++) ll_lengths[i] = 9;
|
340
|
+
for (i = 256; i < 280; i++) ll_lengths[i] = 7;
|
341
|
+
for (i = 280; i < 288; i++) ll_lengths[i] = 8;
|
342
|
+
for (i = 0; i < 32; i++) d_lengths[i] = 5;
|
343
|
+
}
|
344
|
+
|
345
|
+
/*
|
346
|
+
Calculates size of the part after the header and tree of an LZ77 block, in bits.
|
347
|
+
*/
|
348
|
+
static size_t CalculateBlockSymbolSize(const unsigned* ll_lengths,
|
349
|
+
const unsigned* d_lengths,
|
350
|
+
const unsigned short* litlens,
|
351
|
+
const unsigned short* dists,
|
352
|
+
size_t lstart, size_t lend) {
|
353
|
+
size_t result = 0;
|
354
|
+
size_t i;
|
355
|
+
for (i = lstart; i < lend; i++) {
|
356
|
+
if (dists[i] == 0) {
|
357
|
+
result += ll_lengths[litlens[i]];
|
358
|
+
} else {
|
359
|
+
result += ll_lengths[ZopfliGetLengthSymbol(litlens[i])];
|
360
|
+
result += d_lengths[ZopfliGetDistSymbol(dists[i])];
|
361
|
+
result += ZopfliGetLengthExtraBits(litlens[i]);
|
362
|
+
result += ZopfliGetDistExtraBits(dists[i]);
|
363
|
+
}
|
364
|
+
}
|
365
|
+
result += ll_lengths[256]; /*end symbol*/
|
366
|
+
return result;
|
367
|
+
}
|
368
|
+
|
369
|
+
static size_t AbsDiff(size_t x, size_t y) {
|
370
|
+
if (x > y)
|
371
|
+
return x - y;
|
372
|
+
else
|
373
|
+
return y - x;
|
374
|
+
}
|
375
|
+
|
376
|
+
/*
|
377
|
+
Change the population counts in a way that the consequent Hufmann tree
|
378
|
+
compression, especially its rle-part will be more likely to compress this data
|
379
|
+
more efficiently. length containts the size of the histogram.
|
380
|
+
*/
|
381
|
+
void OptimizeHuffmanForRle(int length, size_t* counts) {
|
382
|
+
int i, k, stride;
|
383
|
+
size_t symbol, sum, limit;
|
384
|
+
int* good_for_rle;
|
385
|
+
|
386
|
+
/* 1) We don't want to touch the trailing zeros. We may break the
|
387
|
+
rules of the format by adding more data in the distance codes. */
|
388
|
+
for (; length >= 0; --length) {
|
389
|
+
if (length == 0) {
|
390
|
+
return;
|
391
|
+
}
|
392
|
+
if (counts[length - 1] != 0) {
|
393
|
+
/* Now counts[0..length - 1] does not have trailing zeros. */
|
394
|
+
break;
|
395
|
+
}
|
396
|
+
}
|
397
|
+
/* 2) Let's mark all population counts that already can be encoded
|
398
|
+
with an rle code.*/
|
399
|
+
good_for_rle = (int*)malloc(length * sizeof(int));
|
400
|
+
for (i = 0; i < length; ++i) good_for_rle[i] = 0;
|
401
|
+
|
402
|
+
/* Let's not spoil any of the existing good rle codes.
|
403
|
+
Mark any seq of 0's that is longer than 5 as a good_for_rle.
|
404
|
+
Mark any seq of non-0's that is longer than 7 as a good_for_rle.*/
|
405
|
+
symbol = counts[0];
|
406
|
+
stride = 0;
|
407
|
+
for (i = 0; i < length + 1; ++i) {
|
408
|
+
if (i == length || counts[i] != symbol) {
|
409
|
+
if ((symbol == 0 && stride >= 5) || (symbol != 0 && stride >= 7)) {
|
410
|
+
for (k = 0; k < stride; ++k) {
|
411
|
+
good_for_rle[i - k - 1] = 1;
|
412
|
+
}
|
413
|
+
}
|
414
|
+
stride = 1;
|
415
|
+
if (i != length) {
|
416
|
+
symbol = counts[i];
|
417
|
+
}
|
418
|
+
} else {
|
419
|
+
++stride;
|
420
|
+
}
|
421
|
+
}
|
422
|
+
|
423
|
+
/* 3) Let's replace those population counts that lead to more rle codes. */
|
424
|
+
stride = 0;
|
425
|
+
limit = counts[0];
|
426
|
+
sum = 0;
|
427
|
+
for (i = 0; i < length + 1; ++i) {
|
428
|
+
if (i == length || good_for_rle[i]
|
429
|
+
/* Heuristic for selecting the stride ranges to collapse. */
|
430
|
+
|| AbsDiff(counts[i], limit) >= 4) {
|
431
|
+
if (stride >= 4 || (stride >= 3 && sum == 0)) {
|
432
|
+
/* The stride must end, collapse what we have, if we have enough (4). */
|
433
|
+
int count = (sum + stride / 2) / stride;
|
434
|
+
if (count < 1) count = 1;
|
435
|
+
if (sum == 0) {
|
436
|
+
/* Don't make an all zeros stride to be upgraded to ones. */
|
437
|
+
count = 0;
|
438
|
+
}
|
439
|
+
for (k = 0; k < stride; ++k) {
|
440
|
+
/* We don't want to change value at counts[i],
|
441
|
+
that is already belonging to the next stride. Thus - 1. */
|
442
|
+
counts[i - k - 1] = count;
|
443
|
+
}
|
444
|
+
}
|
445
|
+
stride = 0;
|
446
|
+
sum = 0;
|
447
|
+
if (i < length - 3) {
|
448
|
+
/* All interesting strides have a count of at least 4,
|
449
|
+
at least when non-zeros. */
|
450
|
+
limit = (counts[i] + counts[i + 1] +
|
451
|
+
counts[i + 2] + counts[i + 3] + 2) / 4;
|
452
|
+
} else if (i < length) {
|
453
|
+
limit = counts[i];
|
454
|
+
} else {
|
455
|
+
limit = 0;
|
456
|
+
}
|
457
|
+
}
|
458
|
+
++stride;
|
459
|
+
if (i != length) {
|
460
|
+
sum += counts[i];
|
461
|
+
}
|
462
|
+
}
|
463
|
+
|
464
|
+
free(good_for_rle);
|
465
|
+
}
|
466
|
+
|
467
|
+
/*
|
468
|
+
Calculates the bit lengths for the symbols for dynamic blocks. Chooses bit
|
469
|
+
lengths that give the smallest size of tree encoding + encoding of all the
|
470
|
+
symbols to have smallest output size. This are not necessarily the ideal Huffman
|
471
|
+
bit lengths.
|
472
|
+
*/
|
473
|
+
static void GetDynamicLengths(const unsigned short* litlens,
|
474
|
+
const unsigned short* dists,
|
475
|
+
size_t lstart, size_t lend,
|
476
|
+
unsigned* ll_lengths, unsigned* d_lengths) {
|
477
|
+
size_t ll_counts[288];
|
478
|
+
size_t d_counts[32];
|
479
|
+
|
480
|
+
ZopfliLZ77Counts(litlens, dists, lstart, lend, ll_counts, d_counts);
|
481
|
+
OptimizeHuffmanForRle(288, ll_counts);
|
482
|
+
OptimizeHuffmanForRle(32, d_counts);
|
483
|
+
ZopfliCalculateBitLengths(ll_counts, 288, 15, ll_lengths);
|
484
|
+
ZopfliCalculateBitLengths(d_counts, 32, 15, d_lengths);
|
485
|
+
PatchDistanceCodesForBuggyDecoders(d_lengths);
|
486
|
+
}
|
487
|
+
|
488
|
+
double ZopfliCalculateBlockSize(const unsigned short* litlens,
|
489
|
+
const unsigned short* dists,
|
490
|
+
size_t lstart, size_t lend, int btype) {
|
491
|
+
unsigned ll_lengths[288];
|
492
|
+
unsigned d_lengths[32];
|
493
|
+
|
494
|
+
double result = 3; /* bfinal and btype bits */
|
495
|
+
|
496
|
+
assert(btype == 1 || btype == 2); /* This is not for uncompressed blocks. */
|
497
|
+
|
498
|
+
if(btype == 1) {
|
499
|
+
GetFixedTree(ll_lengths, d_lengths);
|
500
|
+
} else {
|
501
|
+
GetDynamicLengths(litlens, dists, lstart, lend, ll_lengths, d_lengths);
|
502
|
+
result += CalculateTreeSize(ll_lengths, d_lengths);
|
503
|
+
}
|
504
|
+
|
505
|
+
result += CalculateBlockSymbolSize(
|
506
|
+
ll_lengths, d_lengths, litlens, dists, lstart, lend);
|
507
|
+
|
508
|
+
return result;
|
509
|
+
}
|
510
|
+
|
511
|
+
/*
|
512
|
+
Adds a deflate block with the given LZ77 data to the output.
|
513
|
+
options: global program options
|
514
|
+
btype: the block type, must be 1 or 2
|
515
|
+
final: whether to set the "final" bit on this block, must be the last block
|
516
|
+
litlens: literal/length array of the LZ77 data, in the same format as in
|
517
|
+
ZopfliLZ77Store.
|
518
|
+
dists: distance array of the LZ77 data, in the same format as in
|
519
|
+
ZopfliLZ77Store.
|
520
|
+
lstart: where to start in the LZ77 data
|
521
|
+
lend: where to end in the LZ77 data (not inclusive)
|
522
|
+
expected_data_size: the uncompressed block size, used for assert, but you can
|
523
|
+
set it to 0 to not do the assertion.
|
524
|
+
bp: output bit pointer
|
525
|
+
out: dynamic output array to append to
|
526
|
+
outsize: dynamic output array size
|
527
|
+
*/
|
528
|
+
static void AddLZ77Block(const ZopfliOptions* options, int btype, int final,
|
529
|
+
const unsigned short* litlens,
|
530
|
+
const unsigned short* dists,
|
531
|
+
size_t lstart, size_t lend,
|
532
|
+
size_t expected_data_size,
|
533
|
+
unsigned char* bp,
|
534
|
+
unsigned char** out, size_t* outsize) {
|
535
|
+
unsigned ll_lengths[288];
|
536
|
+
unsigned d_lengths[32];
|
537
|
+
unsigned ll_symbols[288];
|
538
|
+
unsigned d_symbols[32];
|
539
|
+
size_t detect_block_size = *outsize;
|
540
|
+
size_t compressed_size;
|
541
|
+
size_t uncompressed_size = 0;
|
542
|
+
size_t i;
|
543
|
+
|
544
|
+
AddBit(final, bp, out, outsize);
|
545
|
+
AddBit(btype & 1, bp, out, outsize);
|
546
|
+
AddBit((btype & 2) >> 1, bp, out, outsize);
|
547
|
+
|
548
|
+
if (btype == 1) {
|
549
|
+
/* Fixed block. */
|
550
|
+
GetFixedTree(ll_lengths, d_lengths);
|
551
|
+
} else {
|
552
|
+
/* Dynamic block. */
|
553
|
+
unsigned detect_tree_size;
|
554
|
+
assert(btype == 2);
|
555
|
+
|
556
|
+
GetDynamicLengths(litlens, dists, lstart, lend, ll_lengths, d_lengths);
|
557
|
+
|
558
|
+
detect_tree_size = *outsize;
|
559
|
+
AddDynamicTree(ll_lengths, d_lengths, bp, out, outsize);
|
560
|
+
if (options->verbose) {
|
561
|
+
fprintf(stderr, "treesize: %d\n", (int)(*outsize - detect_tree_size));
|
562
|
+
}
|
563
|
+
}
|
564
|
+
|
565
|
+
ZopfliLengthsToSymbols(ll_lengths, 288, 15, ll_symbols);
|
566
|
+
ZopfliLengthsToSymbols(d_lengths, 32, 15, d_symbols);
|
567
|
+
|
568
|
+
detect_block_size = *outsize;
|
569
|
+
AddLZ77Data(litlens, dists, lstart, lend, expected_data_size,
|
570
|
+
ll_symbols, ll_lengths, d_symbols, d_lengths,
|
571
|
+
bp, out, outsize);
|
572
|
+
/* End symbol. */
|
573
|
+
AddHuffmanBits(ll_symbols[256], ll_lengths[256], bp, out, outsize);
|
574
|
+
|
575
|
+
for (i = lstart; i < lend; i++) {
|
576
|
+
uncompressed_size += dists[i] == 0 ? 1 : litlens[i];
|
577
|
+
}
|
578
|
+
compressed_size = *outsize - detect_block_size;
|
579
|
+
if (options->verbose) {
|
580
|
+
fprintf(stderr, "compressed block size: %d (%dk) (unc: %d)\n",
|
581
|
+
(int)compressed_size, (int)(compressed_size / 1024),
|
582
|
+
(int)(uncompressed_size));
|
583
|
+
}
|
584
|
+
}
|
585
|
+
|
586
|
+
static void DeflateDynamicBlock(const ZopfliOptions* options, int final,
|
587
|
+
const unsigned char* in,
|
588
|
+
size_t instart, size_t inend,
|
589
|
+
unsigned char* bp,
|
590
|
+
unsigned char** out, size_t* outsize) {
|
591
|
+
ZopfliBlockState s;
|
592
|
+
size_t blocksize = inend - instart;
|
593
|
+
ZopfliLZ77Store store;
|
594
|
+
int btype = 2;
|
595
|
+
|
596
|
+
ZopfliInitLZ77Store(&store);
|
597
|
+
|
598
|
+
s.options = options;
|
599
|
+
s.blockstart = instart;
|
600
|
+
s.blockend = inend;
|
601
|
+
#ifdef ZOPFLI_LONGEST_MATCH_CACHE
|
602
|
+
s.lmc = (ZopfliLongestMatchCache*)malloc(sizeof(ZopfliLongestMatchCache));
|
603
|
+
ZopfliInitCache(blocksize, s.lmc);
|
604
|
+
#endif
|
605
|
+
|
606
|
+
ZopfliLZ77Optimal(&s, in, instart, inend, &store);
|
607
|
+
|
608
|
+
/* For small block, encoding with fixed tree can be smaller. For large block,
|
609
|
+
don't bother doing this expensive test, dynamic tree will be better.*/
|
610
|
+
if (store.size < 1000) {
|
611
|
+
double dyncost, fixedcost;
|
612
|
+
ZopfliLZ77Store fixedstore;
|
613
|
+
ZopfliInitLZ77Store(&fixedstore);
|
614
|
+
ZopfliLZ77OptimalFixed(&s, in, instart, inend, &fixedstore);
|
615
|
+
dyncost = ZopfliCalculateBlockSize(store.litlens, store.dists,
|
616
|
+
0, store.size, 2);
|
617
|
+
fixedcost = ZopfliCalculateBlockSize(fixedstore.litlens, fixedstore.dists,
|
618
|
+
0, fixedstore.size, 1);
|
619
|
+
if (fixedcost < dyncost) {
|
620
|
+
btype = 1;
|
621
|
+
ZopfliCleanLZ77Store(&store);
|
622
|
+
store = fixedstore;
|
623
|
+
} else {
|
624
|
+
ZopfliCleanLZ77Store(&fixedstore);
|
625
|
+
}
|
626
|
+
}
|
627
|
+
|
628
|
+
AddLZ77Block(s.options, btype, final,
|
629
|
+
store.litlens, store.dists, 0, store.size,
|
630
|
+
blocksize, bp, out, outsize);
|
631
|
+
|
632
|
+
#ifdef ZOPFLI_LONGEST_MATCH_CACHE
|
633
|
+
ZopfliCleanCache(s.lmc);
|
634
|
+
free(s.lmc);
|
635
|
+
#endif
|
636
|
+
ZopfliCleanLZ77Store(&store);
|
637
|
+
}
|
638
|
+
|
639
|
+
static void DeflateFixedBlock(const ZopfliOptions* options, int final,
|
640
|
+
const unsigned char* in,
|
641
|
+
size_t instart, size_t inend,
|
642
|
+
unsigned char* bp,
|
643
|
+
unsigned char** out, size_t* outsize) {
|
644
|
+
ZopfliBlockState s;
|
645
|
+
size_t blocksize = inend - instart;
|
646
|
+
ZopfliLZ77Store store;
|
647
|
+
|
648
|
+
ZopfliInitLZ77Store(&store);
|
649
|
+
|
650
|
+
s.options = options;
|
651
|
+
s.blockstart = instart;
|
652
|
+
s.blockend = inend;
|
653
|
+
#ifdef ZOPFLI_LONGEST_MATCH_CACHE
|
654
|
+
s.lmc = (ZopfliLongestMatchCache*)malloc(sizeof(ZopfliLongestMatchCache));
|
655
|
+
ZopfliInitCache(blocksize, s.lmc);
|
656
|
+
#endif
|
657
|
+
|
658
|
+
ZopfliLZ77OptimalFixed(&s, in, instart, inend, &store);
|
659
|
+
|
660
|
+
AddLZ77Block(s.options, 1, final, store.litlens, store.dists, 0, store.size,
|
661
|
+
blocksize, bp, out, outsize);
|
662
|
+
|
663
|
+
#ifdef ZOPFLI_LONGEST_MATCH_CACHE
|
664
|
+
ZopfliCleanCache(s.lmc);
|
665
|
+
free(s.lmc);
|
666
|
+
#endif
|
667
|
+
ZopfliCleanLZ77Store(&store);
|
668
|
+
}
|
669
|
+
|
670
|
+
static void DeflateNonCompressedBlock(const ZopfliOptions* options, int final,
|
671
|
+
const unsigned char* in, size_t instart,
|
672
|
+
size_t inend,
|
673
|
+
unsigned char* bp,
|
674
|
+
unsigned char** out, size_t* outsize) {
|
675
|
+
size_t i;
|
676
|
+
size_t blocksize = inend - instart;
|
677
|
+
unsigned short nlen = ~blocksize;
|
678
|
+
|
679
|
+
(void)options;
|
680
|
+
assert(blocksize < 65536); /* Non compressed blocks are max this size. */
|
681
|
+
|
682
|
+
AddBit(final, bp, out, outsize);
|
683
|
+
/* BTYPE 00 */
|
684
|
+
AddBit(0, bp, out, outsize);
|
685
|
+
AddBit(0, bp, out, outsize);
|
686
|
+
|
687
|
+
/* Any bits of input up to the next byte boundary are ignored. */
|
688
|
+
*bp = 0;
|
689
|
+
|
690
|
+
ZOPFLI_APPEND_DATA(blocksize % 256, out, outsize);
|
691
|
+
ZOPFLI_APPEND_DATA((blocksize / 256) % 256, out, outsize);
|
692
|
+
ZOPFLI_APPEND_DATA(nlen % 256, out, outsize);
|
693
|
+
ZOPFLI_APPEND_DATA((nlen / 256) % 256, out, outsize);
|
694
|
+
|
695
|
+
for (i = instart; i < inend; i++) {
|
696
|
+
ZOPFLI_APPEND_DATA(in[i], out, outsize);
|
697
|
+
}
|
698
|
+
}
|
699
|
+
|
700
|
+
static void DeflateBlock(const ZopfliOptions* options,
|
701
|
+
int btype, int final,
|
702
|
+
const unsigned char* in, size_t instart, size_t inend,
|
703
|
+
unsigned char* bp,
|
704
|
+
unsigned char** out, size_t* outsize) {
|
705
|
+
if (btype == 0) {
|
706
|
+
DeflateNonCompressedBlock(
|
707
|
+
options, final, in, instart, inend, bp, out, outsize);
|
708
|
+
} else if (btype == 1) {
|
709
|
+
DeflateFixedBlock(options, final, in, instart, inend, bp, out, outsize);
|
710
|
+
} else {
|
711
|
+
assert (btype == 2);
|
712
|
+
DeflateDynamicBlock(options, final, in, instart, inend, bp, out, outsize);
|
713
|
+
}
|
714
|
+
}
|
715
|
+
|
716
|
+
/*
|
717
|
+
Does squeeze strategy where first block splitting is done, then each block is
|
718
|
+
squeezed.
|
719
|
+
Parameters: see description of the ZopfliDeflate function.
|
720
|
+
*/
|
721
|
+
static void DeflateSplittingFirst(const ZopfliOptions* options,
|
722
|
+
int btype, int final,
|
723
|
+
const unsigned char* in,
|
724
|
+
size_t instart, size_t inend,
|
725
|
+
unsigned char* bp,
|
726
|
+
unsigned char** out, size_t* outsize) {
|
727
|
+
size_t i;
|
728
|
+
size_t* splitpoints = 0;
|
729
|
+
size_t npoints = 0;
|
730
|
+
if (btype == 0) {
|
731
|
+
ZopfliBlockSplitSimple(in, instart, inend, 65535, &splitpoints, &npoints);
|
732
|
+
} else if (btype == 1) {
|
733
|
+
/* If all blocks are fixed tree, splitting into separate blocks only
|
734
|
+
increases the total size. Leave npoints at 0, this represents 1 block. */
|
735
|
+
} else {
|
736
|
+
ZopfliBlockSplit(options, in, instart, inend,
|
737
|
+
options->blocksplittingmax, &splitpoints, &npoints);
|
738
|
+
}
|
739
|
+
|
740
|
+
for (i = 0; i <= npoints; i++) {
|
741
|
+
size_t start = i == 0 ? instart : splitpoints[i - 1];
|
742
|
+
size_t end = i == npoints ? inend : splitpoints[i];
|
743
|
+
DeflateBlock(options, btype, i == npoints && final, in, start, end,
|
744
|
+
bp, out, outsize);
|
745
|
+
}
|
746
|
+
|
747
|
+
free(splitpoints);
|
748
|
+
}
|
749
|
+
|
750
|
+
/*
|
751
|
+
Does squeeze strategy where first the best possible lz77 is done, and then based
|
752
|
+
on that data, block splitting is done.
|
753
|
+
Parameters: see description of the ZopfliDeflate function.
|
754
|
+
*/
|
755
|
+
static void DeflateSplittingLast(const ZopfliOptions* options,
|
756
|
+
int btype, int final,
|
757
|
+
const unsigned char* in,
|
758
|
+
size_t instart, size_t inend,
|
759
|
+
unsigned char* bp,
|
760
|
+
unsigned char** out, size_t* outsize) {
|
761
|
+
size_t i;
|
762
|
+
ZopfliBlockState s;
|
763
|
+
ZopfliLZ77Store store;
|
764
|
+
size_t* splitpoints = 0;
|
765
|
+
size_t npoints = 0;
|
766
|
+
|
767
|
+
if (btype == 0) {
|
768
|
+
/* This function only supports LZ77 compression. DeflateSplittingFirst
|
769
|
+
supports the special case of noncompressed data. Punt it to that one. */
|
770
|
+
DeflateSplittingFirst(options, btype, final,
|
771
|
+
in, instart, inend,
|
772
|
+
bp, out, outsize);
|
773
|
+
}
|
774
|
+
assert(btype == 1 || btype == 2);
|
775
|
+
|
776
|
+
ZopfliInitLZ77Store(&store);
|
777
|
+
|
778
|
+
s.options = options;
|
779
|
+
s.blockstart = instart;
|
780
|
+
s.blockend = inend;
|
781
|
+
#ifdef ZOPFLI_LONGEST_MATCH_CACHE
|
782
|
+
s.lmc = (ZopfliLongestMatchCache*)malloc(sizeof(ZopfliLongestMatchCache));
|
783
|
+
ZopfliInitCache(inend - instart, s.lmc);
|
784
|
+
#endif
|
785
|
+
|
786
|
+
if (btype == 2) {
|
787
|
+
ZopfliLZ77Optimal(&s, in, instart, inend, &store);
|
788
|
+
} else {
|
789
|
+
assert (btype == 1);
|
790
|
+
ZopfliLZ77OptimalFixed(&s, in, instart, inend, &store);
|
791
|
+
}
|
792
|
+
|
793
|
+
if (btype == 1) {
|
794
|
+
/* If all blocks are fixed tree, splitting into separate blocks only
|
795
|
+
increases the total size. Leave npoints at 0, this represents 1 block. */
|
796
|
+
} else {
|
797
|
+
ZopfliBlockSplitLZ77(options, store.litlens, store.dists, store.size,
|
798
|
+
options->blocksplittingmax, &splitpoints, &npoints);
|
799
|
+
}
|
800
|
+
|
801
|
+
for (i = 0; i <= npoints; i++) {
|
802
|
+
size_t start = i == 0 ? 0 : splitpoints[i - 1];
|
803
|
+
size_t end = i == npoints ? store.size : splitpoints[i];
|
804
|
+
AddLZ77Block(options, btype, i == npoints && final,
|
805
|
+
store.litlens, store.dists, start, end, 0,
|
806
|
+
bp, out, outsize);
|
807
|
+
}
|
808
|
+
|
809
|
+
#ifdef ZOPFLI_LONGEST_MATCH_CACHE
|
810
|
+
ZopfliCleanCache(s.lmc);
|
811
|
+
free(s.lmc);
|
812
|
+
#endif
|
813
|
+
|
814
|
+
ZopfliCleanLZ77Store(&store);
|
815
|
+
free(splitpoints);
|
816
|
+
}
|
817
|
+
|
818
|
+
/*
|
819
|
+
Deflate a part, to allow ZopfliDeflate() to use multiple master blocks if
|
820
|
+
needed.
|
821
|
+
It is possible to call this function multiple times in a row, shifting
|
822
|
+
instart and inend to next bytes of the data. If instart is larger than 0, then
|
823
|
+
previous bytes are used as the initial dictionary for LZ77.
|
824
|
+
This function will usually output multiple deflate blocks. If final is 1, then
|
825
|
+
the final bit will be set on the last block.
|
826
|
+
*/
|
827
|
+
void ZopfliDeflatePart(const ZopfliOptions* options, int btype, int final,
|
828
|
+
const unsigned char* in, size_t instart, size_t inend,
|
829
|
+
unsigned char* bp, unsigned char** out,
|
830
|
+
size_t* outsize) {
|
831
|
+
if (options->blocksplitting) {
|
832
|
+
if (options->blocksplittinglast) {
|
833
|
+
DeflateSplittingLast(options, btype, final, in, instart, inend,
|
834
|
+
bp, out, outsize);
|
835
|
+
} else {
|
836
|
+
DeflateSplittingFirst(options, btype, final, in, instart, inend,
|
837
|
+
bp, out, outsize);
|
838
|
+
}
|
839
|
+
} else {
|
840
|
+
DeflateBlock(options, btype, final, in, instart, inend, bp, out, outsize);
|
841
|
+
}
|
842
|
+
}
|
843
|
+
|
844
|
+
void ZopfliDeflate(const ZopfliOptions* options, int btype, int final,
|
845
|
+
const unsigned char* in, size_t insize,
|
846
|
+
unsigned char* bp, unsigned char** out, size_t* outsize) {
|
847
|
+
#if ZOPFLI_MASTER_BLOCK_SIZE == 0
|
848
|
+
ZopfliDeflatePart(options, btype, final, in, 0, insize, bp, out, outsize);
|
849
|
+
#else
|
850
|
+
size_t i = 0;
|
851
|
+
while (i < insize) {
|
852
|
+
int masterfinal = (i + ZOPFLI_MASTER_BLOCK_SIZE >= insize);
|
853
|
+
int final2 = final && masterfinal;
|
854
|
+
size_t size = masterfinal ? insize - i : ZOPFLI_MASTER_BLOCK_SIZE;
|
855
|
+
ZopfliDeflatePart(options, btype, final2,
|
856
|
+
in, i, i + size, bp, out, outsize);
|
857
|
+
i += size;
|
858
|
+
}
|
859
|
+
#endif
|
860
|
+
if (options->verbose) {
|
861
|
+
fprintf(stderr,
|
862
|
+
"Original Size: %d, Deflate: %d, Compression: %f%% Removed\n",
|
863
|
+
(int)insize, (int)*outsize,
|
864
|
+
100.0 * (double)(insize - *outsize) / (double)insize);
|
865
|
+
}
|
866
|
+
}
|