]> git.saurik.com Git - wxWidgets.git/blame_incremental - src/tiff/tif_lzw.c
no more _ftime instead of ftime
[wxWidgets.git] / src / tiff / tif_lzw.c
... / ...
CommitLineData
1/* $Header$ */
2
3/*
4 * Copyright (c) 1988-1997 Sam Leffler
5 * Copyright (c) 1991-1997 Silicon Graphics, Inc.
6 *
7 * Permission to use, copy, modify, distribute, and sell this software and
8 * its documentation for any purpose is hereby granted without fee, provided
9 * that (i) the above copyright notices and this permission notice appear in
10 * all copies of the software and related documentation, and (ii) the names of
11 * Sam Leffler and Silicon Graphics may not be used in any advertising or
12 * publicity relating to the software without the specific, prior written
13 * permission of Sam Leffler and Silicon Graphics.
14 *
15 * THE SOFTWARE IS PROVIDED "AS-IS" AND WITHOUT WARRANTY OF ANY KIND,
16 * EXPRESS, IMPLIED OR OTHERWISE, INCLUDING WITHOUT LIMITATION, ANY
17 * WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
18 *
19 * IN NO EVENT SHALL SAM LEFFLER OR SILICON GRAPHICS BE LIABLE FOR
20 * ANY SPECIAL, INCIDENTAL, INDIRECT OR CONSEQUENTIAL DAMAGES OF ANY KIND,
21 * OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS,
22 * WHETHER OR NOT ADVISED OF THE POSSIBILITY OF DAMAGE, AND ON ANY THEORY OF
23 * LIABILITY, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
24 * OF THIS SOFTWARE.
25 */
26
27#include "tiffiop.h"
28#ifdef LZW_SUPPORT
29/*
30 * TIFF Library.
31 * Rev 5.0 Lempel-Ziv & Welch Compression Support
32 *
33 * This code is derived from the compress program whose code is
34 * derived from software contributed to Berkeley by James A. Woods,
35 * derived from original work by Spencer Thomas and Joseph Orost.
36 *
37 * The original Berkeley copyright notice appears below in its entirety.
38 */
39#include "tif_predict.h"
40
41#include <assert.h>
42#include <stdio.h>
43
44/*
45 * NB: The 5.0 spec describes a different algorithm than Aldus
46 * implements. Specifically, Aldus does code length transitions
47 * one code earlier than should be done (for real LZW).
48 * Earlier versions of this library implemented the correct
49 * LZW algorithm, but emitted codes in a bit order opposite
50 * to the TIFF spec. Thus, to maintain compatibility w/ Aldus
51 * we interpret MSB-LSB ordered codes to be images written w/
52 * old versions of this library, but otherwise adhere to the
53 * Aldus "off by one" algorithm.
54 *
55 * Future revisions to the TIFF spec are expected to "clarify this issue".
56 */
57#define LZW_COMPAT /* include backwards compatibility code */
58/*
59 * Each strip of data is supposed to be terminated by a CODE_EOI.
60 * If the following #define is included, the decoder will also
61 * check for end-of-strip w/o seeing this code. This makes the
62 * library more robust, but also slower.
63 */
64#define LZW_CHECKEOS /* include checks for strips w/o EOI code */
65
66#define MAXCODE(n) ((1L<<(n))-1)
67/*
68 * The TIFF spec specifies that encoded bit
69 * strings range from 9 to 12 bits.
70 */
71#define BITS_MIN 9 /* start with 9 bits */
72#define BITS_MAX 12 /* max of 12 bit strings */
73/* predefined codes */
74#define CODE_CLEAR 256 /* code to clear string table */
75#define CODE_EOI 257 /* end-of-information code */
76#define CODE_FIRST 258 /* first free code entry */
77#define CODE_MAX MAXCODE(BITS_MAX)
78#define HSIZE 9001L /* 91% occupancy */
79#define HSHIFT (13-8)
80#ifdef LZW_COMPAT
81/* NB: +1024 is for compatibility with old files */
82#define CSIZE (MAXCODE(BITS_MAX)+1024L)
83#else
84#define CSIZE (MAXCODE(BITS_MAX)+1L)
85#endif
86
87/*
88 * State block for each open TIFF file using LZW
89 * compression/decompression. Note that the predictor
90 * state block must be first in this data structure.
91 */
92typedef struct {
93 TIFFPredictorState predict; /* predictor super class */
94
95 u_short nbits; /* # of bits/code */
96 u_short maxcode; /* maximum code for lzw_nbits */
97 u_short free_ent; /* next free entry in hash table */
98 long nextdata; /* next bits of i/o */
99 long nextbits; /* # of valid bits in lzw_nextdata */
100} LZWBaseState;
101
102#define lzw_nbits base.nbits
103#define lzw_maxcode base.maxcode
104#define lzw_free_ent base.free_ent
105#define lzw_nextdata base.nextdata
106#define lzw_nextbits base.nextbits
107
108/*
109 * Decoding-specific state.
110 */
111typedef struct code_ent {
112 struct code_ent *next;
113 u_short length; /* string len, including this token */
114 u_char value; /* data value */
115 u_char firstchar; /* first token of string */
116} code_t;
117
118typedef int (LINKAGEMODE *decodeFunc)(TIFF*, tidata_t, tsize_t, tsample_t);
119
120typedef struct {
121 LZWBaseState base;
122 long dec_nbitsmask; /* lzw_nbits 1 bits, right adjusted */
123 long dec_restart; /* restart count */
124#ifdef LZW_CHECKEOS
125 long dec_bitsleft; /* available bits in raw data */
126#endif
127 decodeFunc dec_decode; /* regular or backwards compatible */
128 code_t* dec_codep; /* current recognized code */
129 code_t* dec_oldcodep; /* previously recognized code */
130 code_t* dec_free_entp; /* next free entry */
131 code_t* dec_maxcodep; /* max available entry */
132 code_t* dec_codetab; /* kept separate for small machines */
133} LZWDecodeState;
134
135/*
136 * Encoding-specific state.
137 */
138typedef uint16 hcode_t; /* codes fit in 16 bits */
139typedef struct {
140 long hash;
141 hcode_t code;
142} hash_t;
143
144typedef struct {
145 LZWBaseState base;
146 int enc_oldcode; /* last code encountered */
147 long enc_checkpoint; /* point at which to clear table */
148#define CHECK_GAP 10000 /* enc_ratio check interval */
149 long enc_ratio; /* current compression ratio */
150 long enc_incount; /* (input) data bytes encoded */
151 long enc_outcount; /* encoded (output) bytes */
152 tidata_t enc_rawlimit; /* bound on tif_rawdata buffer */
153 hash_t* enc_hashtab; /* kept separate for small machines */
154} LZWEncodeState;
155
156#define LZWState(tif) ((LZWBaseState*) (tif)->tif_data)
157#define DecoderState(tif) ((LZWDecodeState*) LZWState(tif))
158#define EncoderState(tif) ((LZWEncodeState*) LZWState(tif))
159
160static int LINKAGEMODE LZWDecode(TIFF*, tidata_t, tsize_t, tsample_t);
161#ifdef LZW_COMPAT
162static int LINKAGEMODE LZWDecodeCompat(TIFF*, tidata_t, tsize_t, tsample_t);
163#endif
164static void cl_hash(LZWEncodeState*);
165
166/*
167 * LZW Decoder.
168 */
169
170#ifdef LZW_CHECKEOS
171/*
172 * This check shouldn't be necessary because each
173 * strip is suppose to be terminated with CODE_EOI.
174 */
175#define NextCode(_tif, _sp, _bp, _code, _get) { \
176 if ((_sp)->dec_bitsleft < nbits) { \
177 TIFFWarning(_tif->tif_name, \
178 "LZWDecode: Strip %d not terminated with EOI code", \
179 _tif->tif_curstrip); \
180 _code = CODE_EOI; \
181 } else { \
182 _get(_sp,_bp,_code); \
183 (_sp)->dec_bitsleft -= nbits; \
184 } \
185}
186#else
187#define NextCode(tif, sp, bp, code, get) get(sp, bp, code)
188#endif
189
190static int
191LZWSetupDecode(TIFF* tif)
192{
193 LZWDecodeState* sp = DecoderState(tif);
194 static const char module[] = " LZWSetupDecode";
195 int code;
196
197 assert(sp != NULL);
198 if (sp->dec_codetab == NULL) {
199 sp->dec_codetab = (code_t*)_TIFFmalloc(CSIZE*sizeof (code_t));
200 if (sp->dec_codetab == NULL) {
201 TIFFError(module, "No space for LZW code table");
202 return (0);
203 }
204 /*
205 * Pre-load the table.
206 */
207 for (code = 255; code >= 0; code--) {
208 sp->dec_codetab[code].value = code;
209 sp->dec_codetab[code].firstchar = code;
210 sp->dec_codetab[code].length = 1;
211 sp->dec_codetab[code].next = NULL;
212 }
213 }
214 return (1);
215}
216
217/*
218 * Setup state for decoding a strip.
219 */
220static int
221LZWPreDecode(TIFF* tif, tsample_t s)
222{
223 LZWDecodeState *sp = DecoderState(tif);
224
225 (void) s;
226 assert(sp != NULL);
227 /*
228 * Check for old bit-reversed codes.
229 */
230 if (tif->tif_rawdata[0] == 0 && (tif->tif_rawdata[1] & 0x1)) {
231#ifdef LZW_COMPAT
232 if (!sp->dec_decode) {
233 TIFFWarning(tif->tif_name,
234 "Old-style LZW codes, convert file");
235 /*
236 * Override default decoding methods with
237 * ones that deal with the old coding.
238 * Otherwise the predictor versions set
239 * above will call the compatibility routines
240 * through the dec_decode method.
241 */
242 tif->tif_decoderow = LZWDecodeCompat;
243 tif->tif_decodestrip = LZWDecodeCompat;
244 tif->tif_decodetile = LZWDecodeCompat;
245 /*
246 * If doing horizontal differencing, must
247 * re-setup the predictor logic since we
248 * switched the basic decoder methods...
249 */
250 (*tif->tif_setupdecode)(tif);
251 sp->dec_decode = LZWDecodeCompat;
252 }
253 sp->lzw_maxcode = MAXCODE(BITS_MIN);
254#else /* !LZW_COMPAT */
255 if (!sp->dec_decode) {
256 TIFFError(tif->tif_name,
257 "Old-style LZW codes not supported");
258 sp->dec_decode = LZWDecode;
259 }
260 return (0);
261#endif/* !LZW_COMPAT */
262 } else {
263 sp->lzw_maxcode = MAXCODE(BITS_MIN)-1;
264 sp->dec_decode = LZWDecode;
265 }
266 sp->lzw_nbits = BITS_MIN;
267 sp->lzw_nextbits = 0;
268 sp->lzw_nextdata = 0;
269
270 sp->dec_restart = 0;
271 sp->dec_nbitsmask = MAXCODE(BITS_MIN);
272#ifdef LZW_CHECKEOS
273 sp->dec_bitsleft = tif->tif_rawcc << 3;
274#endif
275 sp->dec_free_entp = sp->dec_codetab + CODE_FIRST;
276 /*
277 * Zero entries that are not yet filled in. We do
278 * this to guard against bogus input data that causes
279 * us to index into undefined entries. If you can
280 * come up with a way to safely bounds-check input codes
281 * while decoding then you can remove this operation.
282 */
283 _TIFFmemset(sp->dec_free_entp, 0, (CSIZE-CODE_FIRST)*sizeof (code_t));
284 sp->dec_oldcodep = &sp->dec_codetab[-1];
285 sp->dec_maxcodep = &sp->dec_codetab[sp->dec_nbitsmask-1];
286 return (1);
287}
288
289/*
290 * Decode a "hunk of data".
291 */
292#define GetNextCode(sp, bp, code) { \
293 nextdata = (nextdata<<8) | *(bp)++; \
294 nextbits += 8; \
295 if (nextbits < nbits) { \
296 nextdata = (nextdata<<8) | *(bp)++; \
297 nextbits += 8; \
298 } \
299 code = (hcode_t)((nextdata >> (nextbits-nbits)) & nbitsmask); \
300 nextbits -= nbits; \
301}
302
303static void
304codeLoop(TIFF* tif)
305{
306 TIFFError(tif->tif_name,
307 "LZWDecode: Bogus encoding, loop in the code table; scanline %d",
308 tif->tif_row);
309}
310
311static int
312LZWDecode(TIFF* tif, tidata_t op0, tsize_t occ0, tsample_t s)
313{
314 LZWDecodeState *sp = DecoderState(tif);
315 char *op = (char*) op0;
316 long occ = (long) occ0;
317 char *tp;
318 u_char *bp;
319 hcode_t code;
320 int len;
321 long nbits, nextbits, nextdata, nbitsmask;
322 code_t *codep, *free_entp, *maxcodep, *oldcodep;
323
324 (void) s;
325 assert(sp != NULL);
326 /*
327 * Restart interrupted output operation.
328 */
329 if (sp->dec_restart) {
330 long residue;
331
332 codep = sp->dec_codep;
333 residue = codep->length - sp->dec_restart;
334 if (residue > occ) {
335 /*
336 * Residue from previous decode is sufficient
337 * to satisfy decode request. Skip to the
338 * start of the decoded string, place decoded
339 * values in the output buffer, and return.
340 */
341 sp->dec_restart += occ;
342 do {
343 codep = codep->next;
344 } while (--residue > occ && codep);
345 if (codep) {
346 tp = op + occ;
347 do {
348 *--tp = codep->value;
349 codep = codep->next;
350 } while (--occ && codep);
351 }
352 return (1);
353 }
354 /*
355 * Residue satisfies only part of the decode request.
356 */
357 op += residue, occ -= residue;
358 tp = op;
359 do {
360 int t;
361 --tp;
362 t = codep->value;
363 codep = codep->next;
364 *tp = t;
365 } while (--residue && codep);
366 sp->dec_restart = 0;
367 }
368
369 bp = (u_char *)tif->tif_rawcp;
370 nbits = sp->lzw_nbits;
371 nextdata = sp->lzw_nextdata;
372 nextbits = sp->lzw_nextbits;
373 nbitsmask = sp->dec_nbitsmask;
374 oldcodep = sp->dec_oldcodep;
375 free_entp = sp->dec_free_entp;
376 maxcodep = sp->dec_maxcodep;
377
378 while (occ > 0) {
379 NextCode(tif, sp, bp, code, GetNextCode);
380 if (code == CODE_EOI)
381 break;
382 if (code == CODE_CLEAR) {
383 free_entp = sp->dec_codetab + CODE_FIRST;
384 nbits = BITS_MIN;
385 nbitsmask = MAXCODE(BITS_MIN);
386 maxcodep = sp->dec_codetab + nbitsmask-1;
387 NextCode(tif, sp, bp, code, GetNextCode);
388 if (code == CODE_EOI)
389 break;
390 *op++ = code, occ--;
391 oldcodep = sp->dec_codetab + code;
392 continue;
393 }
394 codep = sp->dec_codetab + code;
395
396 /*
397 * Add the new entry to the code table.
398 */
399 assert(&sp->dec_codetab[0] <= free_entp && free_entp < &sp->dec_codetab[CSIZE]);
400 free_entp->next = oldcodep;
401 free_entp->firstchar = free_entp->next->firstchar;
402 free_entp->length = free_entp->next->length+1;
403 free_entp->value = (codep < free_entp) ?
404 codep->firstchar : free_entp->firstchar;
405 if (++free_entp > maxcodep) {
406 if (++nbits > BITS_MAX) /* should not happen */
407 nbits = BITS_MAX;
408 nbitsmask = MAXCODE(nbits);
409 maxcodep = sp->dec_codetab + nbitsmask-1;
410 }
411 oldcodep = codep;
412 if (code >= 256) {
413 /*
414 * Code maps to a string, copy string
415 * value to output (written in reverse).
416 */
417 if (codep->length > occ) {
418 /*
419 * String is too long for decode buffer,
420 * locate portion that will fit, copy to
421 * the decode buffer, and setup restart
422 * logic for the next decoding call.
423 */
424 sp->dec_codep = codep;
425 do {
426 codep = codep->next;
427 } while (codep && codep->length > occ);
428 if (codep) {
429 sp->dec_restart = occ;
430 tp = op + occ;
431 do {
432 *--tp = codep->value;
433 codep = codep->next;
434 } while (--occ && codep);
435 if (codep)
436 codeLoop(tif);
437 }
438 break;
439 }
440 len = codep->length;
441 tp = op + len;
442 do {
443 int t;
444 --tp;
445 t = codep->value;
446 codep = codep->next;
447 *tp = t;
448 } while (codep && tp > op);
449 if (codep) {
450 codeLoop(tif);
451 break;
452 }
453 op += len, occ -= len;
454 } else
455 *op++ = code, occ--;
456 }
457
458 tif->tif_rawcp = (tidata_t) bp;
459 sp->lzw_nbits = (u_short) nbits;
460 sp->lzw_nextdata = nextdata;
461 sp->lzw_nextbits = nextbits;
462 sp->dec_nbitsmask = nbitsmask;
463 sp->dec_oldcodep = oldcodep;
464 sp->dec_free_entp = free_entp;
465 sp->dec_maxcodep = maxcodep;
466
467 if (occ > 0) {
468 TIFFError(tif->tif_name,
469 "LZWDecode: Not enough data at scanline %d (short %d bytes)",
470 tif->tif_row, occ);
471 return (0);
472 }
473 return (1);
474}
475
476#ifdef LZW_COMPAT
477/*
478 * Decode a "hunk of data" for old images.
479 */
480#define GetNextCodeCompat(sp, bp, code) { \
481 nextdata |= (u_long) *(bp)++ << nextbits; \
482 nextbits += 8; \
483 if (nextbits < nbits) { \
484 nextdata |= (u_long) *(bp)++ << nextbits; \
485 nextbits += 8; \
486 } \
487 code = (hcode_t)(nextdata & nbitsmask); \
488 nextdata >>= nbits; \
489 nextbits -= nbits; \
490}
491
492static int LINKAGEMODE
493LZWDecodeCompat(TIFF* tif, tidata_t op0, tsize_t occ0, tsample_t s)
494{
495 LZWDecodeState *sp = DecoderState(tif);
496 char *op = (char*) op0;
497 long occ = (long) occ0;
498 char *tp;
499 u_char *bp;
500 int code, nbits;
501 long nextbits, nextdata, nbitsmask;
502 code_t *codep, *free_entp, *maxcodep, *oldcodep;
503
504 (void) s;
505 assert(sp != NULL);
506 /*
507 * Restart interrupted output operation.
508 */
509 if (sp->dec_restart) {
510 long residue;
511
512 codep = sp->dec_codep;
513 residue = codep->length - sp->dec_restart;
514 if (residue > occ) {
515 /*
516 * Residue from previous decode is sufficient
517 * to satisfy decode request. Skip to the
518 * start of the decoded string, place decoded
519 * values in the output buffer, and return.
520 */
521 sp->dec_restart += occ;
522 do {
523 codep = codep->next;
524 } while (--residue > occ);
525 tp = op + occ;
526 do {
527 *--tp = codep->value;
528 codep = codep->next;
529 } while (--occ);
530 return (1);
531 }
532 /*
533 * Residue satisfies only part of the decode request.
534 */
535 op += residue, occ -= residue;
536 tp = op;
537 do {
538 *--tp = codep->value;
539 codep = codep->next;
540 } while (--residue);
541 sp->dec_restart = 0;
542 }
543
544 bp = (u_char *)tif->tif_rawcp;
545 nbits = sp->lzw_nbits;
546 nextdata = sp->lzw_nextdata;
547 nextbits = sp->lzw_nextbits;
548 nbitsmask = sp->dec_nbitsmask;
549 oldcodep = sp->dec_oldcodep;
550 free_entp = sp->dec_free_entp;
551 maxcodep = sp->dec_maxcodep;
552
553 while (occ > 0) {
554 NextCode(tif, sp, bp, code, GetNextCodeCompat);
555 if (code == CODE_EOI)
556 break;
557 if (code == CODE_CLEAR) {
558 free_entp = sp->dec_codetab + CODE_FIRST;
559 nbits = BITS_MIN;
560 nbitsmask = MAXCODE(BITS_MIN);
561 maxcodep = sp->dec_codetab + nbitsmask;
562 NextCode(tif, sp, bp, code, GetNextCodeCompat);
563 if (code == CODE_EOI)
564 break;
565 *op++ = code, occ--;
566 oldcodep = sp->dec_codetab + code;
567 continue;
568 }
569 codep = sp->dec_codetab + code;
570
571 /*
572 * Add the new entry to the code table.
573 */
574 assert(&sp->dec_codetab[0] <= free_entp && free_entp < &sp->dec_codetab[CSIZE]);
575 free_entp->next = oldcodep;
576 free_entp->firstchar = free_entp->next->firstchar;
577 free_entp->length = free_entp->next->length+1;
578 free_entp->value = (codep < free_entp) ?
579 codep->firstchar : free_entp->firstchar;
580 if (++free_entp > maxcodep) {
581 if (++nbits > BITS_MAX) /* should not happen */
582 nbits = BITS_MAX;
583 nbitsmask = MAXCODE(nbits);
584 maxcodep = sp->dec_codetab + nbitsmask;
585 }
586 oldcodep = codep;
587 if (code >= 256) {
588 /*
589 * Code maps to a string, copy string
590 * value to output (written in reverse).
591 */
592 if (codep->length > occ) {
593 /*
594 * String is too long for decode buffer,
595 * locate portion that will fit, copy to
596 * the decode buffer, and setup restart
597 * logic for the next decoding call.
598 */
599 sp->dec_codep = codep;
600 do {
601 codep = codep->next;
602 } while (codep->length > occ);
603 sp->dec_restart = occ;
604 tp = op + occ;
605 do {
606 *--tp = codep->value;
607 codep = codep->next;
608 } while (--occ);
609 break;
610 }
611 op += codep->length, occ -= codep->length;
612 tp = op;
613 do {
614 *--tp = codep->value;
615 } while( (codep = codep->next) != NULL);
616 } else
617 *op++ = code, occ--;
618 }
619
620 tif->tif_rawcp = (tidata_t) bp;
621 sp->lzw_nbits = nbits;
622 sp->lzw_nextdata = nextdata;
623 sp->lzw_nextbits = nextbits;
624 sp->dec_nbitsmask = nbitsmask;
625 sp->dec_oldcodep = oldcodep;
626 sp->dec_free_entp = free_entp;
627 sp->dec_maxcodep = maxcodep;
628
629 if (occ > 0) {
630 TIFFError(tif->tif_name,
631 "LZWDecodeCompat: Not enough data at scanline %d (short %d bytes)",
632 tif->tif_row, occ);
633 return (0);
634 }
635 return (1);
636}
637#endif /* LZW_COMPAT */
638
639/*
640 * LZW Encoding.
641 */
642
643static int
644LZWSetupEncode(TIFF* tif)
645{
646 LZWEncodeState* sp = EncoderState(tif);
647 static const char module[] = "LZWSetupEncode";
648
649 assert(sp != NULL);
650 sp->enc_hashtab = (hash_t*) _TIFFmalloc(HSIZE*sizeof (hash_t));
651 if (sp->enc_hashtab == NULL) {
652 TIFFError(module, "No space for LZW hash table");
653 return (0);
654 }
655 return (1);
656}
657
658/*
659 * Reset encoding state at the start of a strip.
660 */
661static int
662LZWPreEncode(TIFF* tif, tsample_t s)
663{
664 LZWEncodeState *sp = EncoderState(tif);
665
666 (void) s;
667 assert(sp != NULL);
668 sp->lzw_nbits = BITS_MIN;
669 sp->lzw_maxcode = MAXCODE(BITS_MIN);
670 sp->lzw_free_ent = CODE_FIRST;
671 sp->lzw_nextbits = 0;
672 sp->lzw_nextdata = 0;
673 sp->enc_checkpoint = CHECK_GAP;
674 sp->enc_ratio = 0;
675 sp->enc_incount = 0;
676 sp->enc_outcount = 0;
677 /*
678 * The 4 here insures there is space for 2 max-sized
679 * codes in LZWEncode and LZWPostDecode.
680 */
681 sp->enc_rawlimit = tif->tif_rawdata + tif->tif_rawdatasize-1 - 4;
682 cl_hash(sp); /* clear hash table */
683 sp->enc_oldcode = (hcode_t) -1; /* generates CODE_CLEAR in LZWEncode */
684 return (1);
685}
686
687#define CALCRATIO(sp, rat) { \
688 if (incount > 0x007fffff) { /* NB: shift will overflow */\
689 rat = outcount >> 8; \
690 rat = (rat == 0 ? 0x7fffffff : incount/rat); \
691 } else \
692 rat = (incount<<8) / outcount; \
693}
694#define PutNextCode(op, c) { \
695 nextdata = (nextdata << nbits) | c; \
696 nextbits += nbits; \
697 *op++ = (u_char)(nextdata >> (nextbits-8)); \
698 nextbits -= 8; \
699 if (nextbits >= 8) { \
700 *op++ = (u_char)(nextdata >> (nextbits-8)); \
701 nextbits -= 8; \
702 } \
703 outcount += nbits; \
704}
705
706/*
707 * Encode a chunk of pixels.
708 *
709 * Uses an open addressing double hashing (no chaining) on the
710 * prefix code/next character combination. We do a variant of
711 * Knuth's algorithm D (vol. 3, sec. 6.4) along with G. Knott's
712 * relatively-prime secondary probe. Here, the modular division
713 * first probe is gives way to a faster exclusive-or manipulation.
714 * Also do block compression with an adaptive reset, whereby the
715 * code table is cleared when the compression ratio decreases,
716 * but after the table fills. The variable-length output codes
717 * are re-sized at this point, and a CODE_CLEAR is generated
718 * for the decoder.
719 */
720static int LINKAGEMODE
721LZWEncode(TIFF* tif, tidata_t bp, tsize_t cc, tsample_t s)
722{
723 register LZWEncodeState *sp = EncoderState(tif);
724 register long fcode;
725 register hash_t *hp;
726 register int h, c;
727 hcode_t ent;
728 long disp;
729 long incount, outcount, checkpoint;
730 long nextdata, nextbits;
731 int free_ent, maxcode, nbits;
732 tidata_t op, limit;
733
734 (void) s;
735 if (sp == NULL)
736 return (0);
737 /*
738 * Load local state.
739 */
740 incount = sp->enc_incount;
741 outcount = sp->enc_outcount;
742 checkpoint = sp->enc_checkpoint;
743 nextdata = sp->lzw_nextdata;
744 nextbits = sp->lzw_nextbits;
745 free_ent = sp->lzw_free_ent;
746 maxcode = sp->lzw_maxcode;
747 nbits = sp->lzw_nbits;
748 op = tif->tif_rawcp;
749 limit = sp->enc_rawlimit;
750 ent = sp->enc_oldcode;
751
752 if (ent == (hcode_t) -1 && cc > 0) {
753 /*
754 * NB: This is safe because it can only happen
755 * at the start of a strip where we know there
756 * is space in the data buffer.
757 */
758 PutNextCode(op, CODE_CLEAR);
759 ent = *bp++; cc--; incount++;
760 }
761 while (cc > 0) {
762 c = *bp++; cc--; incount++;
763 fcode = ((long)c << BITS_MAX) + ent;
764 h = (c << HSHIFT) ^ ent; /* xor hashing */
765#ifdef _WINDOWS
766 /*
767 * Check hash index for an overflow.
768 */
769 if (h >= HSIZE)
770 h -= HSIZE;
771#endif
772 hp = &sp->enc_hashtab[h];
773 if (hp->hash == fcode) {
774 ent = hp->code;
775 continue;
776 }
777 if (hp->hash >= 0) {
778 /*
779 * Primary hash failed, check secondary hash.
780 */
781 disp = HSIZE - h;
782 if (h == 0)
783 disp = 1;
784 do {
785 /*
786 * Avoid pointer arithmetic 'cuz of
787 * wraparound problems with segments.
788 */
789 if ((h -= disp) < 0)
790 h += HSIZE;
791 hp = &sp->enc_hashtab[h];
792 if (hp->hash == fcode) {
793 ent = hp->code;
794 goto hit;
795 }
796 } while (hp->hash >= 0);
797 }
798 /*
799 * New entry, emit code and add to table.
800 */
801 /*
802 * Verify there is space in the buffer for the code
803 * and any potential Clear code that might be emitted
804 * below. The value of limit is setup so that there
805 * are at least 4 bytes free--room for 2 codes.
806 */
807 if (op > limit) {
808 tif->tif_rawcc = (tsize_t)(op - tif->tif_rawdata);
809 TIFFFlushData1(tif);
810 op = tif->tif_rawdata;
811 }
812 PutNextCode(op, ent);
813 ent = c;
814 hp->code = free_ent++;
815 hp->hash = fcode;
816 if (free_ent == CODE_MAX-1) {
817 /* table is full, emit clear code and reset */
818 cl_hash(sp);
819 sp->enc_ratio = 0;
820 incount = 0;
821 outcount = 0;
822 free_ent = CODE_FIRST;
823 PutNextCode(op, CODE_CLEAR);
824 nbits = BITS_MIN;
825 maxcode = MAXCODE(BITS_MIN);
826 } else {
827 /*
828 * If the next entry is going to be too big for
829 * the code size, then increase it, if possible.
830 */
831 if (free_ent > maxcode) {
832 nbits++;
833 assert(nbits <= BITS_MAX);
834 maxcode = (int) MAXCODE(nbits);
835 } else if (incount >= checkpoint) {
836 long rat;
837 /*
838 * Check compression ratio and, if things seem
839 * to be slipping, clear the hash table and
840 * reset state. The compression ratio is a
841 * 24+8-bit fractional number.
842 */
843 checkpoint = incount+CHECK_GAP;
844 CALCRATIO(sp, rat);
845 if (rat <= sp->enc_ratio) {
846 cl_hash(sp);
847 sp->enc_ratio = 0;
848 incount = 0;
849 outcount = 0;
850 free_ent = CODE_FIRST;
851 PutNextCode(op, CODE_CLEAR);
852 nbits = BITS_MIN;
853 maxcode = MAXCODE(BITS_MIN);
854 } else
855 sp->enc_ratio = rat;
856 }
857 }
858 hit:
859 ;
860 }
861
862 /*
863 * Restore global state.
864 */
865 sp->enc_incount = incount;
866 sp->enc_outcount = outcount;
867 sp->enc_checkpoint = checkpoint;
868 sp->enc_oldcode = ent;
869 sp->lzw_nextdata = nextdata;
870 sp->lzw_nextbits = nextbits;
871 sp->lzw_free_ent = free_ent;
872 sp->lzw_maxcode = maxcode;
873 sp->lzw_nbits = nbits;
874 tif->tif_rawcp = op;
875 return (1);
876}
877
878/*
879 * Finish off an encoded strip by flushing the last
880 * string and tacking on an End Of Information code.
881 */
882static int
883LZWPostEncode(TIFF* tif)
884{
885 register LZWEncodeState *sp = EncoderState(tif);
886 tidata_t op = tif->tif_rawcp;
887 long nextbits = sp->lzw_nextbits;
888 long nextdata = sp->lzw_nextdata;
889 long outcount = sp->enc_outcount;
890 int nbits = sp->lzw_nbits;
891
892 if (op > sp->enc_rawlimit) {
893 tif->tif_rawcc = (tsize_t)(op - tif->tif_rawdata);
894 TIFFFlushData1(tif);
895 op = tif->tif_rawdata;
896 }
897 if (sp->enc_oldcode != (hcode_t) -1) {
898 PutNextCode(op, sp->enc_oldcode);
899 sp->enc_oldcode = (hcode_t) -1;
900 }
901 PutNextCode(op, CODE_EOI);
902 if (nextbits > 0)
903 *op++ = (u_char)(nextdata << (8-nextbits));
904 tif->tif_rawcc = (tsize_t)(op - tif->tif_rawdata);
905 return (1);
906}
907
908/*
909 * Reset encoding hash table.
910 */
911static void
912cl_hash(LZWEncodeState* sp)
913{
914 register hash_t *hp = &sp->enc_hashtab[HSIZE-1];
915 register long i = HSIZE-8;
916
917 do {
918 i -= 8;
919 hp[-7].hash = -1;
920 hp[-6].hash = -1;
921 hp[-5].hash = -1;
922 hp[-4].hash = -1;
923 hp[-3].hash = -1;
924 hp[-2].hash = -1;
925 hp[-1].hash = -1;
926 hp[ 0].hash = -1;
927 hp -= 8;
928 } while (i >= 0);
929 for (i += 8; i > 0; i--, hp--)
930 hp->hash = -1;
931}
932
933static void
934LZWCleanup(TIFF* tif)
935{
936 if (tif->tif_data) {
937 if (tif->tif_mode == O_RDONLY) {
938 if (DecoderState(tif)->dec_codetab)
939 _TIFFfree(DecoderState(tif)->dec_codetab);
940 } else {
941 if (EncoderState(tif)->enc_hashtab)
942 _TIFFfree(EncoderState(tif)->enc_hashtab);
943 }
944 _TIFFfree(tif->tif_data);
945 tif->tif_data = NULL;
946 }
947}
948
949int
950TIFFInitLZW(TIFF* tif, int scheme)
951{
952 assert(scheme == COMPRESSION_LZW);
953 /*
954 * Allocate state block so tag methods have storage to record values.
955 */
956 if (tif->tif_mode == O_RDONLY) {
957 tif->tif_data = (tidata_t) _TIFFmalloc(sizeof (LZWDecodeState));
958 if (tif->tif_data == NULL)
959 goto bad;
960 DecoderState(tif)->dec_codetab = NULL;
961 DecoderState(tif)->dec_decode = NULL;
962 } else {
963 tif->tif_data = (tidata_t) _TIFFmalloc(sizeof (LZWEncodeState));
964 if (tif->tif_data == NULL)
965 goto bad;
966 EncoderState(tif)->enc_hashtab = NULL;
967 }
968 /*
969 * Install codec methods.
970 */
971 tif->tif_setupdecode = LZWSetupDecode;
972 tif->tif_predecode = LZWPreDecode;
973 tif->tif_decoderow = LZWDecode;
974 tif->tif_decodestrip = LZWDecode;
975 tif->tif_decodetile = LZWDecode;
976 tif->tif_setupencode = LZWSetupEncode;
977 tif->tif_preencode = LZWPreEncode;
978 tif->tif_postencode = LZWPostEncode;
979 tif->tif_encoderow = LZWEncode;
980 tif->tif_encodestrip = LZWEncode;
981 tif->tif_encodetile = LZWEncode;
982 tif->tif_cleanup = LZWCleanup;
983 /*
984 * Setup predictor setup.
985 */
986 (void) TIFFPredictorInit(tif);
987 return (1);
988bad:
989 TIFFError("TIFFInitLZW", "No space for LZW state block");
990 return (0);
991}
992
993/*
994 * Copyright (c) 1985, 1986 The Regents of the University of California.
995 * All rights reserved.
996 *
997 * This code is derived from software contributed to Berkeley by
998 * James A. Woods, derived from original work by Spencer Thomas
999 * and Joseph Orost.
1000 *
1001 * Redistribution and use in source and binary forms are permitted
1002 * provided that the above copyright notice and this paragraph are
1003 * duplicated in all such forms and that any documentation,
1004 * advertising materials, and other materials related to such
1005 * distribution and use acknowledge that the software was developed
1006 * by the University of California, Berkeley. The name of the
1007 * University may not be used to endorse or promote products derived
1008 * from this software without specific prior written permission.
1009 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR
1010 * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED
1011 * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.
1012 */
1013#endif /* LZW_SUPPORT */