summaryrefslogtreecommitdiffstats
path: root/parts/documentation/protocols/chm/decompress.cpp
blob: b8f45ae286ff5d66ceea8034fe5ed9a5bf455a28 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
/*  Most of this file is taken from:
    cabextract 0.5 - a program to extract Microsoft Cabinet files
    (C) 2000-2001 Stuart Caie <kyzer@4u.net>

    This library is free software; you can redistribute it and/or
    modify it under the terms of the GNU Library General Public
    License as published by the Free Software Foundation; either
    version 2 of the License, or (at your option) any later version.

    This library is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
    Library General Public License for more details.

    You should have received a copy of the GNU Library General Public License
    along with this library; see the file COPYING.LIB.  If not, write to
    the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
    Boston, MA 02110-1301, USA.
*/

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "decompress.h"

int make_decode_table(ULONG nsyms, ULONG nbits, UBYTE *length, UWORD *table);
int lzx_read_lens(UBYTE *lens, ULONG first, ULONG last, lzx_bits *lb);


/*--------------------------------------------------------------------------*/
/* our archiver information / state */

/* LZX stuff */

/* some constants defined by the LZX specification */
#define LZX_MIN_MATCH                (2)
#define LZX_MAX_MATCH                (257)
#define LZX_NUM_CHARS                (256)
#define LZX_BLOCKTYPE_INVALID        (0)   /* also blocktypes 4-7 invalid */
#define LZX_BLOCKTYPE_VERBATIM       (1)
#define LZX_BLOCKTYPE_ALIGNED        (2)
#define LZX_BLOCKTYPE_UNCOMPRESSED   (3)
#define LZX_PRETREE_NUM_ELEMENTS     (20)
#define LZX_ALIGNED_NUM_ELEMENTS     (8)   /* aligned offset tree #elements */
#define LZX_NUM_PRIMARY_LENGTHS      (7)   /* this one missing from spec! */
#define LZX_NUM_SECONDARY_LENGTHS    (249) /* length tree #elements */

/* LZX huffman defines: tweak tablebits as desired */
#define LZX_PRETREE_MAXSYMBOLS  (LZX_PRETREE_NUM_ELEMENTS)
#define LZX_PRETREE_TABLEBITS   (6)
#define LZX_MAINTREE_MAXSYMBOLS (LZX_NUM_CHARS + 50*8)
#define LZX_MAINTREE_TABLEBITS  (12)
#define LZX_LENGTH_MAXSYMBOLS   (LZX_NUM_SECONDARY_LENGTHS+1)
#define LZX_LENGTH_TABLEBITS    (12)
#define LZX_ALIGNED_MAXSYMBOLS  (LZX_ALIGNED_NUM_ELEMENTS)
#define LZX_ALIGNED_TABLEBITS   (7)

#define LZX_LENTABLE_SAFETY (64) /* we allow length table decoding overruns */

#define LZX_DECLARE_TABLE(tbl) \
  UWORD tbl##_table[(1<<LZX_##tbl##_TABLEBITS) + (LZX_##tbl##_MAXSYMBOLS<<1)];\
  UBYTE tbl##_len  [LZX_##tbl##_MAXSYMBOLS + LZX_LENTABLE_SAFETY]

struct LZXstate
{
	UBYTE *window;		   /* the actual decoding window              */
	ULONG window_size;	   /* window size (32Kb through 2Mb)          */
	ULONG actual_size;	   /* window size when it was first allocated */
	ULONG window_posn;	   /* current offset within the window        */
	ULONG R0, R1, R2;	   /* for the LRU offset system               */
	UWORD main_elements;   /* number of main tree elements            */
	int   header_read;	   /* have we started decoding at all yet?    */
	UWORD block_type;	   /* type of this block                      */
	ULONG block_length;	   /* uncompressed length of this block       */
	ULONG block_remaining; /* uncompressed bytes still left to decode */
	ULONG frames_read;	   /* the number of CFDATA blocks processed   */
	LONG  intel_filesize;  /* magic header value used for transform   */
	LONG  intel_curpos;	   /* current offset in transform space       */
	int   intel_started;   /* have we seen any translatable data yet? */

	LZX_DECLARE_TABLE(PRETREE);
	LZX_DECLARE_TABLE(MAINTREE);
	LZX_DECLARE_TABLE(LENGTH);
	LZX_DECLARE_TABLE(ALIGNED);
};


/* generic stuff */
#define CAB(x) (decomp_state.x)
#define LZX(x) (decomp_state.lzx.x)
#define DECR_OK           (0)
#define DECR_DATAFORMAT   (1)
#define DECR_ILLEGALDATA  (2)
#define DECR_NOMEMORY     (3)
#define DECR_CHECKSUM     (4)
#define DECR_INPUT        (5)
#define DECR_OUTPUT       (6)

struct
{
    struct LZXstate lzx;
} decomp_state;


/* LZX decruncher */

/* Microsoft's LZX document and their implementation of the
 * com.ms.util.cab Java package do not concur.
 *
 * In the LZX document, there is a table showing the correlation between
 * window size and the number of position slots. It states that the 1MB
 * window = 40 slots and the 2MB window = 42 slots. In the implementation,
 * 1MB = 42 slots, 2MB = 50 slots. The actual calculation is 'find the
 * first slot whose position base is equal to or more than the required
 * window size'. This would explain why other tables in the document refer
 * to 50 slots rather than 42.
 *
 * The constant NUM_PRIMARY_LENGTHS used in the decompression pseudocode
 * is not defined in the specification.
 *
 * The LZX document does not state the uncompressed block has an
 * uncompressed length field. Where does this length field come from, so
 * we can know how large the block is? The implementation has it as the 24
 * bits following after the 3 blocktype bits, before the alignment
 * padding.
 *
 * The LZX document states that aligned offset blocks have their aligned
 * offset huffman tree AFTER the main and length trees. The implementation
 * suggests that the aligned offset tree is BEFORE the main and length
 * trees.
 *
 * The LZX document decoding algorithm states that, in an aligned offset
 * block, if an extra_bits value is 1, 2 or 3, then that number of bits
 * should be read and the result added to the match offset. This is
 * correct for 1 and 2, but not 3, where just a huffman symbol (using the
 * aligned tree) should be read.
 *
 * Regarding the E8 preprocessing, the LZX document states 'No translation
 * may be performed on the last 6 bytes of the input block'. This is
 * correct.  However, the pseudocode provided checks for the *E8 leader*
 * up to the last 6 bytes. If the leader appears between -10 and -7 bytes
 * from the end, this would cause the next four bytes to be modified, at
 * least one of which would be in the last 6 bytes, which is not allowed
 * according to the spec.
 *
 * The specification states that the huffman trees must always contain at
 * least one element. However, many CAB files contain blocks where the
 * length tree is completely empty (because there are no matches), and
 * this is expected to succeed.
 */


/* LZX uses what it calls 'position slots' to represent match offsets.
 * What this means is that a small 'position slot' number and a small
 * offset from that slot are encoded instead of one large offset for
 * every match.
 * - position_base is an index to the position slot bases
 * - extra_bits states how many bits of offset-from-base data is needed.
 */
static ULONG position_base[51];
static UBYTE extra_bits[51];


int LZXinit(int window) {
	ULONG wndsize = 1 << window;
	int i, j, posn_slots;

	/* LZX supports window sizes of 2^15 (32Kb) through 2^21 (2Mb) */
	/* if a previously allocated window is big enough, keep it     */
	if (window < 15 || window > 21)	return DECR_DATAFORMAT;
	if (LZX(actual_size) < wndsize)
	{
		if (LZX(window)) free(LZX(window));
		LZX(window) = NULL;
	}
	if (!LZX(window))
	{
		if (!(LZX(window) = (UBYTE*)malloc(wndsize))) return DECR_NOMEMORY;
		LZX(actual_size) = wndsize;
	}
	LZX(window_size) = wndsize;

	/* initialise static tables */
	for (i=0, j=0; i <= 49; i += 2)
	{
		extra_bits[i] = extra_bits[i+1] = j; /* 0,0,0,0,1,1,2,2,3,3... */
		if ((i != 0) && (j < 17)) j++; /* 0,0,1,2,3,4...15,16,17,17,17,17... */
	}
	for (i=0, j=0; i <= 50; i++)
	{
		position_base[i] = j; /* 0,1,2,3,4,6,8,12,16,24,32,... */
		j += 1 << extra_bits[i]; /* 1,1,1,1,2,2,4,4,8,8,16,16,32,32,... */
	}

	/* calculate required position slots */
	if (window == 20) posn_slots = 42;
	else if (window == 21) posn_slots = 50;
	else posn_slots	= window << 1;

	/*posn_slots=i=0; while (i < wndsize) i += 1 << extra_bits[posn_slots++]; */


	LZX(R0)  =  LZX(R1)  = LZX(R2) = 1;
	LZX(main_elements)   = LZX_NUM_CHARS + (posn_slots << 3);
	LZX(header_read)     = 0;
	LZX(frames_read)     = 0;
	LZX(block_remaining) = 0;
	LZX(block_type)      = LZX_BLOCKTYPE_INVALID;
	LZX(intel_curpos)    = 0;
	LZX(intel_started)   = 0;
	LZX(window_posn)     = 0;

	/* initialise tables to 0 (because deltas will be applied to them) */
	for (i = 0; i < LZX_MAINTREE_MAXSYMBOLS; i++) LZX(MAINTREE_len)[i] = 0;
	for (i = 0; i < LZX_LENGTH_MAXSYMBOLS; i++)	  LZX(LENGTH_len)[i]   = 0;

	return DECR_OK;
}


/* Bitstream reading macros:
 *
 * INIT_BITSTREAM    should be used first to set up the system
 * READ_BITS(var,n)  takes N bits from the buffer and puts them in var
 *
 * ENSURE_BITS(n)    ensures there are at least N bits in the bit buffer
 * PEEK_BITS(n)      extracts (without removing) N bits from the bit buffer
 * REMOVE_BITS(n)    removes N bits from the bit buffer
 *
 * These bit access routines work by using the area beyond the MSB and the
 * LSB as a free source of zeroes. This avoids having to mask any bits.
 * So we have to know the bit width of the bitbuffer variable. This is
 * sizeof(ULONG) * 8, also defined as ULONG_BITS
 */

/* number of bits in ULONG. Note: This must be at multiple of 16, and at
 * least 32 for the bitbuffer code to work (ie, it must be able to ensure
 * up to 17 bits - that's adding 16 bits when there's one bit left, or
 * adding 32 bits when there are no bits left. The code should work fine
 * for machines where ULONG >= 32 bits.
 */
#define ULONG_BITS (sizeof(ULONG)<<3)

#define INIT_BITSTREAM do { bitsleft = 0; bitbuf = 0; } while (0)

#define ENSURE_BITS(n)							\
  while (bitsleft < (n)) {						\
    bitbuf |= ((inpos[1]<<8)|inpos[0]) << (ULONG_BITS-16 - bitsleft);	\
    bitsleft += 16; inpos+=2;						\
  }

#define PEEK_BITS(n)   (bitbuf >> (ULONG_BITS - (n)))
#define REMOVE_BITS(n) ((bitbuf <<= (n)), (bitsleft -= (n)))

#define READ_BITS(v,n) do {						\
  ENSURE_BITS(n);							\
  (v) = PEEK_BITS(n);							\
  REMOVE_BITS(n);							\
} while (0)


/* Huffman macros */

#define TABLEBITS(tbl)   (LZX_##tbl##_TABLEBITS)
#define MAXSYMBOLS(tbl)  (LZX_##tbl##_MAXSYMBOLS)
#define SYMTABLE(tbl)    (LZX(tbl##_table))
#define LENTABLE(tbl)    (LZX(tbl##_len))

/* BUILD_TABLE(tablename) builds a huffman lookup table from code lengths.
 * In reality, it just calls make_decode_table() with the appropriate
 * values - they're all fixed by some #defines anyway, so there's no point
 * writing each call out in full by hand.
 */
#define BUILD_TABLE(tbl)						\
  if (make_decode_table(						\
    MAXSYMBOLS(tbl), TABLEBITS(tbl), LENTABLE(tbl), SYMTABLE(tbl)	\
  )) { return DECR_ILLEGALDATA; }


/* READ_HUFFSYM(tablename, var) decodes one huffman symbol from the
 * bitstream using the stated table and puts it in var.
 */
#define READ_HUFFSYM(tbl,var) do {					\
  ENSURE_BITS(16);							\
  hufftbl = SYMTABLE(tbl);						\
  if ((i = hufftbl[PEEK_BITS(TABLEBITS(tbl))]) >= MAXSYMBOLS(tbl)) {	\
    j = 1 << (ULONG_BITS - TABLEBITS(tbl));				\
    do {								\
      j >>= 1; i <<= 1; i |= (bitbuf & j) ? 1 : 0;			\
      if (!j) { return DECR_ILLEGALDATA; }	                        \
    } while ((i = hufftbl[i]) >= MAXSYMBOLS(tbl));			\
  }									\
  j = LENTABLE(tbl)[(var) = i];						\
  REMOVE_BITS(j);							\
} while (0)


/* READ_LENGTHS(tablename, first, last) reads in code lengths for symbols
 * first to last in the given table. The code lengths are stored in their
 * own special LZX way.
 */
#define READ_LENGTHS(tbl,first,last) do { \
  lb.bb = bitbuf; lb.bl = bitsleft; lb.ip = inpos; \
  if (lzx_read_lens(LENTABLE(tbl),(first),(last),&lb)) { \
    return DECR_ILLEGALDATA; \
  } \
  bitbuf = lb.bb; bitsleft = lb.bl; inpos = lb.ip; \
} while (0)


/* make_decode_table(nsyms, nbits, length[], table[])
 *
 * This function was coded by David Tritscher. It builds a fast huffman
 * decoding table out of just a canonical huffman code lengths table.
 *
 * nsyms  = total number of symbols in this huffman tree.
 * nbits  = any symbols with a code length of nbits or less can be decoded
 *          in one lookup of the table.
 * length = A table to get code lengths from [0 to syms-1]
 * table  = The table to fill up with decoded symbols and pointers.
 *
 * Returns 0 for OK or 1 for error
 */

int make_decode_table(ULONG nsyms, ULONG nbits, UBYTE *length, UWORD *table) {
	register UWORD sym;
	register ULONG leaf;
	register UBYTE bit_num = 1;
	ULONG fill;
	ULONG pos         = 0; /* the current position in the decode table */
	ULONG table_mask  = 1 << nbits;
	ULONG bit_mask    = table_mask >> 1; /* don't do 0 length codes */
	ULONG next_symbol = bit_mask; /* base of allocation for long codes */

	/* fill entries for codes short enough for a direct mapping */
	while (bit_num <= nbits)
	{
		for (sym = 0; sym < nsyms; sym++)
		{
			if (length[sym] == bit_num)
			{
				leaf = pos;

				if ((pos += bit_mask) > table_mask)	return 1; /* table overrun */

				/* fill all possible lookups of this symbol with the symbol itself */
				fill = bit_mask;
				while (fill-- > 0) table[leaf++] = sym;
			}
		}
		bit_mask >>= 1;
		bit_num++;
	}

	/* if there are any codes longer than nbits */
	if (pos != table_mask)
	{
		/* clear the remainder of the table */
		for (sym = pos; sym < table_mask; sym++) table[sym] = 0;

		/* give ourselves room for codes to grow by up to 16 more bits */
		pos <<= 16;
		table_mask <<= 16;
		bit_mask = 1 << 15;

		while (bit_num <= 16)
		{
			for (sym = 0; sym < nsyms; sym++)
			{
				if (length[sym] == bit_num)
				{
					leaf = pos >> 16;
					for (fill = 0; fill < bit_num - nbits; fill++)
					{
						/* if this path hasn't been taken yet, 'allocate' two entries */
						if (table[leaf] == 0)
						{
							table[(next_symbol << 1)] = 0;
							table[(next_symbol << 1) + 1] = 0;
							table[leaf] = next_symbol++;
						}
						/* follow the path and select either left or right for next bit */
						leaf = table[leaf] << 1;
						if ((pos >> (15-fill)) & 1)	leaf++;
					}
					table[leaf] = sym;

					if ((pos += bit_mask) > table_mask)	return 1; /* table overflow */
				}
			}
			bit_mask >>= 1;
			bit_num++;
		}
	}

	/* full table? */
	if (pos == table_mask) return 0;

	/* either erroneous table, or all elements are 0 - let's find out. */
	for (sym = 0; sym < nsyms; sym++) if (length[sym]) return 1;
	return 0;
}

int lzx_read_lens(UBYTE *lens, ULONG first, ULONG last, lzx_bits *lb) {
	ULONG i,j, x,y;
	int z;

	register ULONG bitbuf = lb->bb;
	register int bitsleft = lb->bl;
	UBYTE *inpos = lb->ip;
	UWORD *hufftbl;

	for (x = 0; x < 20; x++)
	{
		READ_BITS(y, 4);
		LENTABLE(PRETREE)[x] = y;
	}
	BUILD_TABLE(PRETREE);

	for (x = first; x < last;)
	{
		READ_HUFFSYM(PRETREE, z);
		if (z == 17)
		{
			READ_BITS(y, 4); y += 4;
			while (y--)	lens[x++] = 0;
		}
		else if (z == 18)
		{
			READ_BITS(y, 5); y += 20;
			while (y--)	lens[x++] = 0;
		}
		else if (z == 19)
		{
			READ_BITS(y, 1); y += 4;
			READ_HUFFSYM(PRETREE, z);
			z = lens[x] - z; if (z < 0)	z += 17;
			while (y--)	lens[x++] = z;
		}
		else
		{
			z = lens[x] - z; if (z < 0)	z += 17;
			lens[x++] = z;
		}
	}

	lb->bb = bitbuf;
	lb->bl = bitsleft;
	lb->ip = inpos;
	return 0;
}

int LZXdecompress(UBYTE* inpos, int inlen, UBYTE* outpos, int outlen) {
	UBYTE *endinp = inpos + inlen;
	UBYTE *window = LZX(window);
	UBYTE *runsrc, *rundest;
	UWORD *hufftbl;	/* used in READ_HUFFSYM macro as chosen decoding table */

	ULONG window_posn = LZX(window_posn);
	ULONG window_size = LZX(window_size);
	ULONG R0 = LZX(R0);
	ULONG R1 = LZX(R1);
	ULONG R2 = LZX(R2);

	register ULONG bitbuf;
	register int bitsleft;
	ULONG match_offset, i,j,k; /* ijk used in READ_HUFFSYM macro */
	lzx_bits lb;	/* used in READ_LENGTHS macro */

	int togo = outlen, this_run, main_element, aligned_bits;
	int match_length, length_footer, extra, verbatim_bits;

	INIT_BITSTREAM;

	/* read header if necessary */
	if (!LZX(header_read))
	{
		i = j = 0;
		READ_BITS(k, 1); if (k)
		{
			READ_BITS(i,16); READ_BITS(j,16);
		}
		LZX(intel_filesize) = (i << 16) | j; /* or 0 if not encoded */
		LZX(header_read) = 1;
	}

	/* main decoding loop */
	while (togo > 0)
	{
		/* last block finished, new block expected */
		if (LZX(block_remaining) == 0)
		{
			if (LZX(block_type) == LZX_BLOCKTYPE_UNCOMPRESSED)
			{
				if (LZX(block_length) & 1) inpos++;	/* realign bitstream to word */
				INIT_BITSTREAM;
			}

			READ_BITS(LZX(block_type), 3);
			READ_BITS(i, 16);
			READ_BITS(j, 8);
			LZX(block_remaining) = LZX(block_length) = (i << 8) | j;

			switch (LZX(block_type))
			{
				case LZX_BLOCKTYPE_ALIGNED:
					for (i = 0; i < 8; i++)
					{
						READ_BITS(j, 3); LENTABLE(ALIGNED)[i] = j;
					}
					BUILD_TABLE(ALIGNED);
					/* rest of aligned header is same as verbatim */

				case LZX_BLOCKTYPE_VERBATIM:
					READ_LENGTHS(MAINTREE, 0, 256);
					READ_LENGTHS(MAINTREE, 256, LZX(main_elements));
					BUILD_TABLE(MAINTREE);
					if (LENTABLE(MAINTREE)[0xE8] != 0) LZX(intel_started) = 1;

					READ_LENGTHS(LENGTH, 0, LZX_NUM_SECONDARY_LENGTHS);
					BUILD_TABLE(LENGTH);
					break;

				case LZX_BLOCKTYPE_UNCOMPRESSED:
					LZX(intel_started) = 1;	/* because we can't assume otherwise */
					ENSURE_BITS(16); /* get up to 16 pad bits into the buffer */
					if (bitsleft > 16) inpos -= 2; /* and align the bitstream! */
					R0 = inpos[0]|(inpos[1]<<8)|(inpos[2]<<16)|(inpos[3]<<24);inpos+=4;
					R1 = inpos[0]|(inpos[1]<<8)|(inpos[2]<<16)|(inpos[3]<<24);inpos+=4;
					R2 = inpos[0]|(inpos[1]<<8)|(inpos[2]<<16)|(inpos[3]<<24);inpos+=4;
					break;

				default:
					return DECR_ILLEGALDATA;
			}
		}

		/* buffer exhaustion check */
		if (inpos > endinp)
		{
			/* it's possible to have a file where the next run is less than
			 * 16 bits in size. In this case, the READ_HUFFSYM() macro used
			 * in building the tables will exhaust the buffer, so we should
			 * allow for this, but not allow those accidentally read bits to
			 * be used (so we check that there are at least 16 bits
			 * remaining - in this boundary case they aren't really part of
			 * the compressed data)
			 */
			if (inpos > (endinp+2) || bitsleft < 16) return DECR_ILLEGALDATA;
		}

		while ((this_run = LZX(block_remaining)) > 0 && togo > 0)
		{
			if (this_run > togo) this_run = togo;
			togo -= this_run;
			LZX(block_remaining) -= this_run;

			/* apply 2^x-1 mask */
			window_posn &= window_size - 1;
			/* runs can't straddle the window wraparound */
			if ((window_posn + this_run) > window_size)
				return DECR_DATAFORMAT;

			switch (LZX(block_type))
			{
				
				case LZX_BLOCKTYPE_VERBATIM:
					while (this_run > 0)
					{
						READ_HUFFSYM(MAINTREE, main_element);

						if (main_element < LZX_NUM_CHARS)
						{
							/* literal: 0 to LZX_NUM_CHARS-1 */
							window[window_posn++] = main_element;
							this_run--;
						}
						else
						{
							/* match: LZX_NUM_CHARS + ((slot<<3) | length_header (3 bits)) */
							main_element -= LZX_NUM_CHARS;

							match_length = main_element & LZX_NUM_PRIMARY_LENGTHS;
							if (match_length == LZX_NUM_PRIMARY_LENGTHS)
							{
								READ_HUFFSYM(LENGTH, length_footer);
								match_length += length_footer;
							}
							match_length += LZX_MIN_MATCH;

							match_offset = main_element >> 3;

							if (match_offset > 2)
							{
								/* not repeated offset */
								if (match_offset != 3)
								{
									extra = extra_bits[match_offset];
									READ_BITS(verbatim_bits, extra);
									match_offset = position_base[match_offset] - 2 + verbatim_bits;
								}
								else
								{
									match_offset = 1;
								}

								/* update repeated offset LRU queue */
								R2 = R1; R1 = R0; R0 = match_offset;
							}
							else if (match_offset == 0)
							{
								match_offset = R0;
							}
							else if (match_offset == 1)
							{
								match_offset = R1;
								R1 = R0; R0 = match_offset;
							}
							else /* match_offset == 2 */
							{
								match_offset = R2;
								R2 = R0; R0 = match_offset;
							}

							rundest = window + window_posn;
							runsrc  = rundest - match_offset;
							window_posn += match_length;
							this_run -= match_length;

							/* copy any wrapped around source data */
							while ((runsrc < window) && (match_length-- > 0))
							{
								*rundest++ = *(runsrc + window_size); runsrc++;
							}
							/* copy match data - no worries about destination wraps */
							while (match_length-- > 0) *rundest++ = *runsrc++;

						}
					}
					break;

				case LZX_BLOCKTYPE_ALIGNED:
					while (this_run > 0)
					{
						READ_HUFFSYM(MAINTREE, main_element);

						if (main_element < LZX_NUM_CHARS)
						{
							/* literal: 0 to LZX_NUM_CHARS-1 */
							window[window_posn++] = main_element;
							this_run--;
						}
						else
						{
							/* match: LZX_NUM_CHARS + ((slot<<3) | length_header (3 bits)) */
							main_element -= LZX_NUM_CHARS;

							match_length = main_element & LZX_NUM_PRIMARY_LENGTHS;
							if (match_length == LZX_NUM_PRIMARY_LENGTHS)
							{
								READ_HUFFSYM(LENGTH, length_footer);
								match_length += length_footer;
							}
							match_length += LZX_MIN_MATCH;

							match_offset = main_element >> 3;

							if (match_offset > 2)
							{
								/* not repeated offset */
								extra = extra_bits[match_offset];
								match_offset = position_base[match_offset] - 2;
								if (extra > 3)
								{
									/* verbatim and aligned bits */
									extra -= 3;
									READ_BITS(verbatim_bits, extra);
									match_offset += (verbatim_bits << 3);
									READ_HUFFSYM(ALIGNED, aligned_bits);
									match_offset += aligned_bits;
								}
								else if (extra == 3)
								{
									/* aligned bits only */
									READ_HUFFSYM(ALIGNED, aligned_bits);
									match_offset += aligned_bits;
								}
								else if (extra > 0)
								{ /* extra==1, extra==2 */
									/* verbatim bits only */
									READ_BITS(verbatim_bits, extra);
									match_offset += verbatim_bits;
								}
								else /* extra == 0 */
								{
									/* ??? */
									match_offset = 1;
								}

								/* update repeated offset LRU queue */
								R2 = R1; R1 = R0; R0 = match_offset;
							}
							else if (match_offset == 0)
							{
								match_offset = R0;
							}
							else if (match_offset == 1)
							{
								match_offset = R1;
								R1 = R0; R0 = match_offset;
							}
							else /* match_offset == 2 */
							{
								match_offset = R2;
								R2 = R0; R0 = match_offset;
							}

							rundest = window + window_posn;
							runsrc  = rundest - match_offset;
							window_posn += match_length;
							this_run -= match_length;

							/* copy any wrapped around source data */
							while ((runsrc < window) && (match_length-- > 0))
							{
								*rundest++ = *(runsrc + window_size); runsrc++;
							}
							/* copy match data - no worries about destination wraps */
							while (match_length-- > 0) *rundest++ = *runsrc++;

						}
					}
					break;

				case LZX_BLOCKTYPE_UNCOMPRESSED:
					if ((inpos + this_run) > endinp) return DECR_ILLEGALDATA;
					memcpy(window + window_posn, inpos, (size_t) this_run);
					inpos += this_run; window_posn += this_run;
					break;

				default:
					return DECR_ILLEGALDATA; /* might as well */
			}

		}
	}

	if (togo != 0) return DECR_ILLEGALDATA;
	memcpy(outpos, window + ((!window_posn) ? window_size : window_posn) -
		   outlen, (size_t) outlen);

	LZX(window_posn) = window_posn;
	LZX(R0) = R0;
	LZX(R1) = R1;
	LZX(R2) = R2;

	/* intel E8 decoding */
	if ((LZX(frames_read)++ < 32768) && LZX(intel_filesize) != 0)
	{
		if (outlen <= 6 || !LZX(intel_started))
		{
			LZX(intel_curpos) += outlen;
		}
		else
		{
			UBYTE *data    = outpos;
			UBYTE *dataend = data + outlen - 10;
			LONG curpos    = LZX(intel_curpos);
			LONG filesize  = LZX(intel_filesize);
			LONG abs_off, rel_off;

			LZX(intel_curpos) = curpos + outlen;

			while (data < dataend)
			{
				if (*data++ != 0xE8)
				{
					curpos++; continue;
				}
				abs_off = data[0] | (data[1]<<8) | (data[2]<<16) | (data[3]<<24);
				if ((abs_off >= -curpos) && (abs_off < filesize))
				{
					rel_off = (abs_off >= 0) ? abs_off - curpos : abs_off + filesize;
					data[0] = (UBYTE) rel_off;
					data[1] = (UBYTE) (rel_off >> 8);
					data[2] = (UBYTE) (rel_off >> 16);
					data[3] = (UBYTE) (rel_off >> 24);
				}
				data += 4;
				curpos += 5;
			}
		}
	}
	return DECR_OK;
}