1
|
|
|
<?php |
2
|
|
|
|
3
|
|
|
namespace CHMLib\LZX; |
4
|
|
|
|
5
|
|
|
use Exception; |
6
|
|
|
use CHMLib\Reader\BitReader; |
7
|
|
|
|
8
|
|
|
/** |
9
|
|
|
* Decompressor of LXZ-compressed data. |
10
|
|
|
*/ |
11
|
|
|
class Inflater |
12
|
|
|
{ |
13
|
|
|
/** |
14
|
|
|
* The smallest allowable match length. |
15
|
|
|
* |
16
|
|
|
* @var int |
17
|
|
|
*/ |
18
|
|
|
const MIN_MATCH = 2; |
19
|
|
|
|
20
|
|
|
/** |
21
|
|
|
* The number of uncompressed character types. |
22
|
|
|
* |
23
|
|
|
* @var int |
24
|
|
|
*/ |
25
|
|
|
const NUM_CHARS = 256; |
26
|
|
|
|
27
|
|
|
/** |
28
|
|
|
* Block type: verbatim. |
29
|
|
|
* |
30
|
|
|
* @var int |
31
|
|
|
*/ |
32
|
|
|
const BLOCKTYPE_VERBATIM = 1; |
33
|
|
|
|
34
|
|
|
/** |
35
|
|
|
* Block type: aligned offset. |
36
|
|
|
* |
37
|
|
|
* @var int |
38
|
|
|
*/ |
39
|
|
|
const BLOCKTYPE_ALIGNED = 2; |
40
|
|
|
|
41
|
|
|
/** |
42
|
|
|
* Block type: uncompressed. |
43
|
|
|
* |
44
|
|
|
* @var int |
45
|
|
|
*/ |
46
|
|
|
const BLOCKTYPE_UNCOMPRESSED = 3; |
47
|
|
|
|
48
|
|
|
/** |
49
|
|
|
* The number of elements in the aligned offset tree. |
50
|
|
|
* |
51
|
|
|
* @var int |
52
|
|
|
*/ |
53
|
|
|
const ALIGNED_NUM_ELEMENTS = 8; |
54
|
|
|
|
55
|
|
|
/** |
56
|
|
|
* Unknown. |
57
|
|
|
* |
58
|
|
|
* @var int |
59
|
|
|
*/ |
60
|
|
|
const NUM_PRIMARY_LENGTHS = 7; |
61
|
|
|
|
62
|
|
|
/** |
63
|
|
|
* The number of elements in the length tree. |
64
|
|
|
* |
65
|
|
|
* @var int |
66
|
|
|
*/ |
67
|
|
|
const NUM_SECONDARY_LENGTHS = 249; |
68
|
|
|
|
69
|
|
|
/** |
70
|
|
|
* The index matrix of the position slot bases. |
71
|
|
|
* |
72
|
|
|
* @var int[] |
73
|
|
|
*/ |
74
|
|
|
protected static $POSITION_BASE = array( |
75
|
|
|
0, 1, 2, 3, 4, 6, 8, 12, 16, 24, 32, 48, 64, 96, 128, 192, |
76
|
|
|
256, 384, 512, 768, 1024, 1536, 2048, 3072, 4096, 6144, 8192, 12288, 16384, 24576, 32768, 49152, |
77
|
|
|
65536, 98304, 131072, 196608, 262144, 393216, 524288, 655360, 786432, 917504, 1048576, 1179648, 1310720, 1441792, 1572864, 1703936, |
78
|
|
|
1835008, 1966080, 2097152, |
79
|
|
|
); |
80
|
|
|
|
81
|
|
|
/** |
82
|
|
|
* Number of needed bits for offset-from-base data. |
83
|
|
|
* |
84
|
|
|
* @var int |
85
|
|
|
*/ |
86
|
|
|
protected static $EXTRA_BITS = array( |
87
|
|
|
0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, |
88
|
|
|
7, 7, 8, 8, 9, 9, 10, 10, 11, 11, 12, 12, 13, 13, 14, 14, |
89
|
|
|
15, 15, 16, 16, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, |
90
|
|
|
17, 17, 17, |
91
|
|
|
); |
92
|
|
|
|
93
|
|
|
/** |
94
|
|
|
* The size of the actual decoding window. |
95
|
|
|
* |
96
|
|
|
* @var int |
97
|
|
|
*/ |
98
|
|
|
protected $windowSize; |
99
|
|
|
|
100
|
|
|
/** |
101
|
|
|
* The actual decoding window. |
102
|
|
|
* |
103
|
|
|
* @var int[] |
104
|
|
|
*/ |
105
|
|
|
protected $window; |
106
|
|
|
|
107
|
|
|
/** |
108
|
|
|
* The current offset within the window. |
109
|
|
|
* |
110
|
|
|
* @var int |
111
|
|
|
*/ |
112
|
|
|
protected $windowPosition; |
113
|
|
|
|
114
|
|
|
/** |
115
|
|
|
* The most recent non-repeated offset (LRU offset system). |
116
|
|
|
* |
117
|
|
|
* @var int |
118
|
|
|
*/ |
119
|
|
|
protected $r0; |
120
|
|
|
|
121
|
|
|
/** |
122
|
|
|
* The second most recent non-repeated offset (LRU offset system). |
123
|
|
|
* |
124
|
|
|
* @var int |
125
|
|
|
*/ |
126
|
|
|
protected $r1; |
127
|
|
|
|
128
|
|
|
/** |
129
|
|
|
* The third most recent non-repeated offset (LRU offset system). |
130
|
|
|
* |
131
|
|
|
* @var int |
132
|
|
|
*/ |
133
|
|
|
protected $r2; |
134
|
|
|
|
135
|
|
|
/** |
136
|
|
|
* The number of main tree elements. |
137
|
|
|
* |
138
|
|
|
* @var int |
139
|
|
|
*/ |
140
|
|
|
protected $mainElements; |
141
|
|
|
|
142
|
|
|
/** |
143
|
|
|
* Decoding has already been started? |
144
|
|
|
* |
145
|
|
|
* @var bool |
146
|
|
|
*/ |
147
|
|
|
protected $headerRead; |
148
|
|
|
|
149
|
|
|
/** |
150
|
|
|
* The type of the current block. |
151
|
|
|
* |
152
|
|
|
* @var int |
153
|
|
|
*/ |
154
|
|
|
protected $blockType; |
155
|
|
|
|
156
|
|
|
/** |
157
|
|
|
* The uncompressed length of the current block. |
158
|
|
|
* |
159
|
|
|
* @var int |
160
|
|
|
*/ |
161
|
|
|
protected $blockLength; |
162
|
|
|
|
163
|
|
|
/** |
164
|
|
|
* The number of ncompressed bytes still left to decode in the current block. |
165
|
|
|
* |
166
|
|
|
* @var int |
167
|
|
|
*/ |
168
|
|
|
protected $remainingInBlock; |
169
|
|
|
|
170
|
|
|
/** |
171
|
|
|
* The number of CFDATA blocks processed. |
172
|
|
|
* |
173
|
|
|
* @var int |
174
|
|
|
*/ |
175
|
|
|
protected $framesRead; |
176
|
|
|
|
177
|
|
|
/** |
178
|
|
|
* The magic header value used for transform (0 if not encoded). |
179
|
|
|
* |
180
|
|
|
* @var int |
181
|
|
|
*/ |
182
|
|
|
protected $intelFilesize; |
183
|
|
|
|
184
|
|
|
/** |
185
|
|
|
* The current offset in transform space. |
186
|
|
|
* |
187
|
|
|
* @var int |
188
|
|
|
*/ |
189
|
|
|
protected $intelCurrentPosition; |
190
|
|
|
|
191
|
|
|
/** |
192
|
|
|
* Have we seen any translatable data yet? |
193
|
|
|
* |
194
|
|
|
* @var bool |
195
|
|
|
*/ |
196
|
|
|
protected $intelStarted; |
197
|
|
|
|
198
|
|
|
/** |
199
|
|
|
* The main Huffman tree. |
200
|
|
|
* |
201
|
|
|
* @var \CHMLib\LZX\Tree |
202
|
|
|
*/ |
203
|
|
|
protected $mainTree; |
204
|
|
|
|
205
|
|
|
/** |
206
|
|
|
* The length Huffman tree. |
207
|
|
|
* |
208
|
|
|
* @var \CHMLib\LZX\Tree |
209
|
|
|
*/ |
210
|
|
|
protected $lengthTree; |
211
|
|
|
|
212
|
|
|
/** |
213
|
|
|
* The aligned offset Huffman tree. |
214
|
|
|
* |
215
|
|
|
* @var \CHMLib\LZX\Tree |
216
|
|
|
*/ |
217
|
|
|
protected $alignedTree; |
218
|
|
|
|
219
|
|
|
/** |
220
|
|
|
* Initializes the instance. |
221
|
|
|
* |
222
|
|
|
* @param int $windowSize The window size (determines the number of window subdivisions, or "position slots"), |
223
|
|
|
* |
224
|
|
|
* @throws \Exception Throws an Exception in case of errors. |
225
|
|
|
*/ |
226
|
483 |
|
public function __construct($windowSize) |
227
|
|
|
{ |
228
|
483 |
|
$this->windowSize = (int) $windowSize; |
229
|
483 |
|
if ($this->windowSize < 0x8000 || $this->windowSize > 0x200000) { |
230
|
|
|
throw new Exception("Unsupported window size: $windowSize"); |
231
|
|
|
} |
232
|
483 |
|
$this->mainTree = new Tree(12, static::NUM_CHARS + 50 * 8); |
233
|
483 |
|
$this->lengthTree = new Tree(12, static::NUM_SECONDARY_LENGTHS + 1); |
234
|
483 |
|
$this->alignedTree = new Tree(7, static::ALIGNED_NUM_ELEMENTS); |
235
|
483 |
|
$this->window = array_fill(0, $this->windowSize, 0); |
236
|
483 |
|
$this->intelFilesize = 0; |
237
|
483 |
|
$numPositionSlots = 0; |
238
|
483 |
|
$windowSize = $this->windowSize; |
239
|
483 |
|
while ($windowSize > 1) { |
240
|
483 |
|
$windowSize >>= 1; |
241
|
483 |
|
$numPositionSlots += 2; |
242
|
|
|
} |
243
|
|
|
switch ($numPositionSlots) { |
244
|
483 |
|
case 40: |
245
|
|
|
$numPositionSlots = 42; |
246
|
|
|
break; |
247
|
483 |
|
case 42: |
248
|
|
|
$numPositionSlots = 50; |
249
|
|
|
break; |
250
|
|
|
} |
251
|
483 |
|
$this->mainElements = static::NUM_CHARS + ($numPositionSlots << 3); |
252
|
483 |
|
} |
253
|
|
|
|
254
|
|
|
/** |
255
|
|
|
* Uncompress bytes. |
256
|
|
|
* |
257
|
|
|
* @param bool $reset Reset the LXZ state? |
258
|
|
|
* @param \CHMLib\Reader\BitReader $reader The reader that provides the data. |
259
|
|
|
* @param int $numberOfBytes The number of decompressed bytes to retrieve. |
260
|
|
|
* |
261
|
|
|
* @throws \Exception Throws an Exception in case of errors. |
262
|
|
|
* |
263
|
|
|
* @return string |
264
|
|
|
*/ |
265
|
26 |
|
public function inflate($reset, BitReader $reader, $numberOfBytes) |
266
|
|
|
{ |
267
|
26 |
|
if ($reset) { |
268
|
26 |
|
$this->r2 = $this->r1 = $this->r0 = 1; |
269
|
26 |
|
$this->headerRead = false; |
270
|
26 |
|
$this->framesRead = 0; |
271
|
26 |
|
$this->remainingInBlock = 0; |
272
|
26 |
|
$this->blockType = null; |
273
|
26 |
|
$this->intelCurrentPosition = 0; |
274
|
26 |
|
$this->intelStarted = false; |
275
|
26 |
|
$this->windowPosition = 0; |
276
|
26 |
|
$this->mainTree->clear(); |
277
|
26 |
|
$this->lengthTree->clear(); |
278
|
|
|
} |
279
|
|
|
|
280
|
26 |
|
if ($this->headerRead === false) { |
281
|
26 |
|
if ($reader->readLE(1) > 0) { |
282
|
|
|
$this->intelFilesize = ($reader->readLE(16) << 16) | $reader->readLE(16); |
283
|
|
|
} |
284
|
26 |
|
$this->headerRead = true; |
285
|
|
|
} |
286
|
|
|
|
287
|
26 |
|
$togo = $numberOfBytes; |
288
|
26 |
|
while ($togo > 0) { |
289
|
26 |
|
if ($this->remainingInBlock === 0) { |
290
|
26 |
|
if ($this->blockType === static::BLOCKTYPE_UNCOMPRESSED) { |
291
|
|
|
if (($this->blockLength & 1) !== 0) { |
292
|
|
|
$reader->skip(1); |
293
|
|
|
} |
294
|
|
|
} |
295
|
26 |
|
$this->blockType = $reader->readLE(3); |
296
|
26 |
|
$this->remainingInBlock = $this->blockLength = $reader->readLE(16) << 8 | $reader->readLE(8); |
297
|
26 |
|
switch ($this->blockType) { |
298
|
26 |
|
case static::BLOCKTYPE_ALIGNED: |
299
|
6 |
|
$this->alignedTree->readAlignLengthTable($reader); |
300
|
6 |
|
$this->alignedTree->makeSymbolTable(); |
301
|
|
|
/* @noinspection PhpMissingBreakStatementInspection */ |
302
|
22 |
|
case static::BLOCKTYPE_VERBATIM: |
303
|
26 |
|
$this->mainTree->readLengthTable($reader, 0, static::NUM_CHARS); |
304
|
26 |
|
$this->mainTree->readLengthTable($reader, static::NUM_CHARS, $this->mainElements); |
305
|
26 |
|
$this->mainTree->makeSymbolTable(); |
306
|
26 |
|
if ($this->mainTree->isIntel()) { |
307
|
6 |
|
$this->intelStarted = true; |
308
|
|
|
} |
309
|
26 |
|
$this->lengthTree->readLengthTable($reader, 0, static::NUM_SECONDARY_LENGTHS); |
310
|
26 |
|
$this->lengthTree->makeSymbolTable(); |
311
|
26 |
|
break; |
312
|
|
|
case static::BLOCKTYPE_UNCOMPRESSED: |
313
|
|
|
$this->intelStarted = true; |
314
|
|
|
if ($reader->ensure(16) > 16) { |
315
|
|
|
$reader->skip(-2); |
316
|
|
|
} |
317
|
|
|
$this->r0 = $reader->readUInt32(); |
318
|
|
|
$this->r1 = $reader->readUInt32(); |
319
|
|
|
$this->r2 = $reader->readUInt32(); |
320
|
|
|
break; |
321
|
|
|
default: |
322
|
|
|
throw new Exception('Unexpected block type '.$this->blockType); |
323
|
|
|
} |
324
|
|
|
} |
325
|
26 |
|
while (($thisRun = $this->remainingInBlock) > 0 && $togo > 0) { |
326
|
26 |
|
if ($thisRun > $togo) { |
327
|
23 |
|
$thisRun = $togo; |
328
|
|
|
} |
329
|
26 |
|
$togo -= $thisRun; |
330
|
26 |
|
$this->remainingInBlock -= $thisRun; |
331
|
26 |
|
$this->windowPosition %= $this->windowSize; |
332
|
26 |
|
if ($this->windowPosition + $thisRun > $this->windowSize) { |
333
|
|
|
throw new Exception('Trying to read more that window size bytes'); |
334
|
|
|
} |
335
|
26 |
|
if ($this->blockType === static::BLOCKTYPE_UNCOMPRESSED) { |
336
|
|
|
$this->window = $reader->readFully($this->window, $this->windowPosition, $thisRun); |
337
|
|
|
$this->windowPosition += $thisRun; |
338
|
|
|
} else { |
339
|
26 |
|
while ($thisRun > 0) { |
340
|
26 |
|
$mainElement = $this->mainTree->readHuffmanSymbol($reader); |
341
|
26 |
|
if ($mainElement < static::NUM_CHARS) { |
342
|
26 |
|
$this->window[$this->windowPosition++] = $mainElement; |
343
|
26 |
|
--$thisRun; |
344
|
|
|
} else { |
345
|
26 |
|
$mainElement -= static::NUM_CHARS; |
346
|
26 |
|
$matchLength = $mainElement & static::NUM_PRIMARY_LENGTHS; |
347
|
26 |
|
if ($matchLength === static::NUM_PRIMARY_LENGTHS) { |
348
|
26 |
|
$matchLength += $this->lengthTree->readHuffmanSymbol($reader); |
349
|
|
|
} |
350
|
26 |
|
$matchLength += static::MIN_MATCH; |
351
|
26 |
|
$matchOffset = $mainElement >> 3; |
352
|
|
|
switch ($matchOffset) { |
353
|
26 |
|
case 0: |
354
|
26 |
|
$matchOffset = $this->r0; |
355
|
26 |
|
break; |
356
|
26 |
|
case 1: |
357
|
26 |
|
$matchOffset = $this->r1; |
358
|
26 |
|
$this->r1 = $this->r0; |
359
|
26 |
|
$this->r0 = $matchOffset; |
360
|
26 |
|
break; |
361
|
26 |
|
case 2: |
362
|
26 |
|
$matchOffset = $this->r2; |
363
|
26 |
|
$this->r2 = $this->r0; |
364
|
26 |
|
$this->r0 = $matchOffset; |
365
|
26 |
|
break; |
366
|
|
|
default: |
367
|
26 |
|
switch ($this->blockType) { |
368
|
26 |
|
case static::BLOCKTYPE_VERBATIM: |
369
|
22 |
|
if ($matchOffset !== 3) { |
370
|
22 |
|
$extra = static::$EXTRA_BITS[$matchOffset]; |
371
|
22 |
|
$matchOffset = static::$POSITION_BASE[$matchOffset] - 2 + $reader->readLE($extra); |
372
|
|
|
} else { |
373
|
21 |
|
$matchOffset = 1; |
374
|
|
|
} |
375
|
22 |
|
break; |
376
|
6 |
|
case static::BLOCKTYPE_ALIGNED: |
377
|
6 |
|
$extra = static::$EXTRA_BITS[$matchOffset]; |
378
|
6 |
|
$matchOffset = static::$POSITION_BASE[$matchOffset] - 2; |
379
|
6 |
|
switch ($extra) { |
380
|
6 |
|
case 0: |
381
|
6 |
|
$matchOffset = 1; |
382
|
6 |
|
break; |
383
|
6 |
|
case 1: |
384
|
6 |
|
case 2: |
385
|
|
|
// verbatim bits only |
386
|
6 |
|
$matchOffset += $reader->readLE($extra); |
387
|
6 |
|
break; |
388
|
6 |
|
case 3: |
389
|
|
|
// aligned bits only |
390
|
6 |
|
$matchOffset += $this->alignedTree->readHuffmanSymbol($reader); |
391
|
6 |
|
break; |
392
|
|
|
default: |
393
|
|
|
// verbatim and aligned bits |
394
|
6 |
|
$extra -= 3; |
395
|
6 |
|
$matchOffset += ($reader->readLE($extra) << 3); |
396
|
6 |
|
$matchOffset += $this->alignedTree->readHuffmanSymbol($reader); |
397
|
6 |
|
break; |
398
|
|
|
} |
399
|
6 |
|
break; |
400
|
|
|
default: |
401
|
|
|
throw new Exception('Unexpected block type ' + $this->blockType); |
402
|
|
|
} |
403
|
26 |
|
$this->r2 = $this->r1; |
404
|
26 |
|
$this->r1 = $this->r0; |
405
|
26 |
|
$this->r0 = $matchOffset; |
406
|
26 |
|
break; |
407
|
|
|
} |
408
|
26 |
|
$runSrc = 0; |
|
|
|
|
409
|
26 |
|
$runDest = $this->windowPosition; |
410
|
26 |
|
$thisRun -= $matchLength; |
411
|
|
|
// copy any wrapped around source data |
412
|
26 |
|
if ($this->windowPosition >= $matchOffset) { |
413
|
|
|
// no wrap |
414
|
26 |
|
$runSrc = $runDest - $matchOffset; |
415
|
|
|
} else { |
416
|
|
|
// wrap around |
417
|
|
|
$runSrc = $runDest + ($this->windowSize - $matchOffset); |
418
|
|
|
$copyLength = $matchOffset - $this->windowPosition; |
419
|
|
|
if ($copyLength < $matchLength) { |
420
|
|
|
$matchLength -= $copyLength; |
421
|
|
|
$this->windowPosition += $copyLength; |
422
|
|
|
while ($copyLength-- > 0) { |
423
|
|
|
$this->window[$runDest++] = $this->window[$runSrc++]; |
424
|
|
|
} |
425
|
|
|
$runSrc = 0; |
426
|
|
|
} |
427
|
|
|
} |
428
|
26 |
|
$this->windowPosition += $matchLength; |
429
|
|
|
// copy match data - no worries about destination wraps |
430
|
26 |
|
while ($matchLength-- > 0) { |
431
|
26 |
|
$this->window[$runDest++] = $this->window[$runSrc++]; |
432
|
|
|
} |
433
|
|
|
} |
434
|
|
|
} |
435
|
|
|
} |
436
|
|
|
} |
437
|
|
|
} |
438
|
|
|
|
439
|
26 |
|
if ($togo !== 0) { |
440
|
|
|
throw new Exception('should never happens'); |
441
|
|
|
} |
442
|
|
|
|
443
|
26 |
|
$result = array_slice($this->window, ($this->windowPosition === 0 ? $this->windowSize : $this->windowPosition) - $numberOfBytes, $numberOfBytes); |
444
|
|
|
|
445
|
|
|
// Intel E8 decoding |
446
|
26 |
|
if (($this->intelFilesize !== 0)) { |
447
|
|
|
if ($this->framesRead++ < 32768) { |
448
|
|
|
if ($numberOfBytes <= 6 || $this->intelStarted === false) { |
449
|
|
|
$this->intelCurrentPosition += $numberOfBytes; |
450
|
|
|
} else { |
451
|
|
|
$currentPosition = $this->intelCurrentPosition; |
452
|
|
|
$this->intelCurrentPosition += $numberOfBytes; |
453
|
|
|
for ($i = 0; $i < $numberOfBytes - 10;) { |
454
|
|
|
if ($result[$i++] !== 0xe8) { |
455
|
|
|
++$currentPosition; |
456
|
|
|
} else { |
457
|
|
|
$absoluteOffset = ($result[$i] & 0xff) | (($result[$i + 1] & 0xff) << 8) | (($result[$i + 2] & 0xff) << 16) | (($result[$i + 3] & 0xff) << 24); |
458
|
|
|
if (($absoluteOffset >= -$currentPosition) && ($absoluteOffset < $this->intelFilesize)) { |
459
|
|
|
$referenceOffset = ($absoluteOffset >= 0) ? $absoluteOffset - $currentPosition : $absoluteOffset + $this->intelFilesize; |
460
|
|
|
$result[$i] = $referenceOffset; |
461
|
|
|
$result[$i + 1] = $referenceOffset >> 8; |
462
|
|
|
$result[$i + 2] = $referenceOffset >> 16; |
463
|
|
|
$result[$i + 3] = $referenceOffset >> 24; |
464
|
|
|
} |
465
|
|
|
$i += 4; |
466
|
|
|
$currentPosition += 5; |
467
|
|
|
} |
468
|
|
|
} |
469
|
|
|
} |
470
|
|
|
} |
471
|
|
|
} |
472
|
|
|
|
473
|
26 |
|
return implode('', array_map('chr', $result)); |
474
|
|
|
} |
475
|
|
|
} |
476
|
|
|
|
This check looks for variable assignements that are either overwritten by other assignments or where the variable is not used subsequently.
Both the
$myVar
assignment in line 1 and the$higher
assignment in line 2 are dead. The first because$myVar
is never used and the second because$higher
is always overwritten for every possible time line.