Subversion Repositories shark

Rev

Go to most recent revision | Details | Last modification | View Log | RSS feed

Rev Author Line No. Line
107 pj 1
/* deflate.c -- compress data using the deflation algorithm
2
 * Copyright (C) 1995-2002 Jean-loup Gailly.
3
 * For conditions of distribution and use, see copyright notice in zlib.h
4
 */
5
 
6
/*
7
 *  ALGORITHM
8
 *
9
 *      The "deflation" process depends on being able to identify portions
10
 *      of the input text which are identical to earlier input (within a
11
 *      sliding window trailing behind the input currently being processed).
12
 *
13
 *      The most straightforward technique turns out to be the fastest for
14
 *      most input files: try all possible matches and select the longest.
15
 *      The key feature of this algorithm is that insertions into the string
16
 *      dictionary are very simple and thus fast, and deletions are avoided
17
 *      completely. Insertions are performed at each input character, whereas
18
 *      string matches are performed only when the previous match ends. So it
19
 *      is preferable to spend more time in matches to allow very fast string
20
 *      insertions and avoid deletions. The matching algorithm for small
21
 *      strings is inspired from that of Rabin & Karp. A brute force approach
22
 *      is used to find longer strings when a small match has been found.
23
 *      A similar algorithm is used in comic (by Jan-Mark Wams) and freeze
24
 *      (by Leonid Broukhis).
25
 *         A previous version of this file used a more sophisticated algorithm
26
 *      (by Fiala and Greene) which is guaranteed to run in linear amortized
27
 *      time, but has a larger average cost, uses more memory and is patented.
28
 *      However the F&G algorithm may be faster for some highly redundant
29
 *      files if the parameter max_chain_length (described below) is too large.
30
 *
31
 *  ACKNOWLEDGEMENTS
32
 *
33
 *      The idea of lazy evaluation of matches is due to Jan-Mark Wams, and
34
 *      I found it in 'freeze' written by Leonid Broukhis.
35
 *      Thanks to many people for bug reports and testing.
36
 *
37
 *  REFERENCES
38
 *
39
 *      Deutsch, L.P.,"DEFLATE Compressed Data Format Specification".
40
 *      Available in ftp://ds.internic.net/rfc/rfc1951.txt
41
 *
42
 *      A description of the Rabin and Karp algorithm is given in the book
43
 *         "Algorithms" by R. Sedgewick, Addison-Wesley, p252.
44
 *
45
 *      Fiala,E.R., and Greene,D.H.
46
 *         Data Compression with Finite Windows, Comm.ACM, 32,4 (1989) 490-595
47
 *
48
 */
49
 
50
/* @(#) $Id: deflate.c,v 1.1 2003-03-24 11:13:44 pj Exp $ */
51
 
52
#include "deflate.h"
53
 
54
const char deflate_copyright[] =
55
   " deflate 1.1.4 Copyright 1995-2002 Jean-loup Gailly ";
56
/*
57
  If you use the zlib library in a product, an acknowledgment is welcome
58
  in the documentation of your product. If for some reason you cannot
59
  include such an acknowledgment, I would appreciate that you keep this
60
  copyright string in the executable of your product.
61
 */
62
 
63
/* ===========================================================================
64
 *  Function prototypes.
65
 */
66
typedef enum {
67
    need_more,      /* block not completed, need more input or more output */
68
    block_done,     /* block flush performed */
69
    finish_started, /* finish started, need only more output at next deflate */
70
    finish_done     /* finish done, accept no more input or output */
71
} block_state;
72
 
73
typedef block_state (*compress_func) OF((deflate_state *s, int flush));
74
/* Compression function. Returns the block state after the call. */
75
 
76
local void fill_window    OF((deflate_state *s));
77
local block_state deflate_stored OF((deflate_state *s, int flush));
78
local block_state deflate_fast   OF((deflate_state *s, int flush));
79
local block_state deflate_slow   OF((deflate_state *s, int flush));
80
local void lm_init        OF((deflate_state *s));
81
local void putShortMSB    OF((deflate_state *s, uInt b));
82
local void flush_pending  OF((z_streamp strm));
83
local int read_buf        OF((z_streamp strm, Bytef *buf, unsigned size));
84
#ifdef ASMV
85
      void match_init OF((void)); /* asm code initialization */
86
      uInt longest_match  OF((deflate_state *s, IPos cur_match));
87
#else
88
local uInt longest_match  OF((deflate_state *s, IPos cur_match));
89
#endif
90
 
91
#ifdef DEBUG
92
local  void check_match OF((deflate_state *s, IPos start, IPos match,
93
                            int length));
94
#endif
95
 
96
/* ===========================================================================
97
 * Local data
98
 */
99
 
100
#define NIL 0
101
/* Tail of hash chains */
102
 
103
#ifndef TOO_FAR
104
#  define TOO_FAR 4096
105
#endif
106
/* Matches of length 3 are discarded if their distance exceeds TOO_FAR */
107
 
108
#define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
109
/* Minimum amount of lookahead, except at the end of the input file.
110
 * See deflate.c for comments about the MIN_MATCH+1.
111
 */
112
 
113
/* Values for max_lazy_match, good_match and max_chain_length, depending on
114
 * the desired pack level (0..9). The values given below have been tuned to
115
 * exclude worst case performance for pathological files. Better values may be
116
 * found for specific files.
117
 */
118
typedef struct config_s {
119
   ush good_length; /* reduce lazy search above this match length */
120
   ush max_lazy;    /* do not perform lazy search above this match length */
121
   ush nice_length; /* quit search above this match length */
122
   ush max_chain;
123
   compress_func func;
124
} config;
125
 
126
local const config configuration_table[10] = {
127
/*      good lazy nice chain */
128
/* 0 */ {0,    0,  0,    0, deflate_stored},  /* store only */
129
/* 1 */ {4,    4,  8,    4, deflate_fast}, /* maximum speed, no lazy matches */
130
/* 2 */ {4,    5, 16,    8, deflate_fast},
131
/* 3 */ {4,    6, 32,   32, deflate_fast},
132
 
133
/* 4 */ {4,    4, 16,   16, deflate_slow},  /* lazy matches */
134
/* 5 */ {8,   16, 32,   32, deflate_slow},
135
/* 6 */ {8,   16, 128, 128, deflate_slow},
136
/* 7 */ {8,   32, 128, 256, deflate_slow},
137
/* 8 */ {32, 128, 258, 1024, deflate_slow},
138
/* 9 */ {32, 258, 258, 4096, deflate_slow}}; /* maximum compression */
139
 
140
/* Note: the deflate() code requires max_lazy >= MIN_MATCH and max_chain >= 4
141
 * For deflate_fast() (levels <= 3) good is ignored and lazy has a different
142
 * meaning.
143
 */
144
 
145
#define EQUAL 0
146
/* result of memcmp for equal strings */
147
 
148
struct static_tree_desc_s {int dummy;}; /* for buggy compilers */
149
 
150
/* ===========================================================================
151
 * Update a hash value with the given input byte
152
 * IN  assertion: all calls to to UPDATE_HASH are made with consecutive
153
 *    input characters, so that a running hash key can be computed from the
154
 *    previous key instead of complete recalculation each time.
155
 */
156
#define UPDATE_HASH(s,h,c) (h = (((h)<<s->hash_shift) ^ (c)) & s->hash_mask)
157
 
158
 
159
/* ===========================================================================
160
 * Insert string str in the dictionary and set match_head to the previous head
161
 * of the hash chain (the most recent string with same hash key). Return
162
 * the previous length of the hash chain.
163
 * If this file is compiled with -DFASTEST, the compression level is forced
164
 * to 1, and no hash chains are maintained.
165
 * IN  assertion: all calls to to INSERT_STRING are made with consecutive
166
 *    input characters and the first MIN_MATCH bytes of str are valid
167
 *    (except for the last MIN_MATCH-1 bytes of the input file).
168
 */
169
#ifdef FASTEST
170
#define INSERT_STRING(s, str, match_head) \
171
   (UPDATE_HASH(s, s->ins_h, s->window[(str) + (MIN_MATCH-1)]), \
172
    match_head = s->head[s->ins_h], \
173
    s->head[s->ins_h] = (Pos)(str))
174
#else
175
#define INSERT_STRING(s, str, match_head) \
176
   (UPDATE_HASH(s, s->ins_h, s->window[(str) + (MIN_MATCH-1)]), \
177
    s->prev[(str) & s->w_mask] = match_head = s->head[s->ins_h], \
178
    s->head[s->ins_h] = (Pos)(str))
179
#endif
180
 
181
/* ===========================================================================
182
 * Initialize the hash table (avoiding 64K overflow for 16 bit systems).
183
 * prev[] will be initialized on the fly.
184
 */
185
#define CLEAR_HASH(s) \
186
    s->head[s->hash_size-1] = NIL; \
187
    zmemzero((Bytef *)s->head, (unsigned)(s->hash_size-1)*sizeof(*s->head));
188
 
189
/* ========================================================================= */
190
int ZEXPORT deflateInit_(strm, level, version, stream_size)
191
    z_streamp strm;
192
    int level;
193
    const char *version;
194
    int stream_size;
195
{
196
    return deflateInit2_(strm, level, Z_DEFLATED, MAX_WBITS, DEF_MEM_LEVEL,
197
                         Z_DEFAULT_STRATEGY, version, stream_size);
198
    /* To do: ignore strm->next_in if we use it as window */
199
}
200
 
201
/* ========================================================================= */
202
int ZEXPORT deflateInit2_(strm, level, method, windowBits, memLevel, strategy,
203
                  version, stream_size)
204
    z_streamp strm;
205
    int  level;
206
    int  method;
207
    int  windowBits;
208
    int  memLevel;
209
    int  strategy;
210
    const char *version;
211
    int stream_size;
212
{
213
    deflate_state *s;
214
    int noheader = 0;
215
    static const char* my_version = ZLIB_VERSION;
216
 
217
    ushf *overlay;
218
    /* We overlay pending_buf and d_buf+l_buf. This works since the average
219
     * output size for (length,distance) codes is <= 24 bits.
220
     */
221
 
222
    if (version == Z_NULL || version[0] != my_version[0] ||
223
        stream_size != sizeof(z_stream)) {
224
        return Z_VERSION_ERROR;
225
    }
226
    if (strm == Z_NULL) return Z_STREAM_ERROR;
227
 
228
    strm->msg = Z_NULL;
229
    if (strm->zalloc == Z_NULL) {
230
        strm->zalloc = zcalloc;
231
        strm->opaque = (voidpf)0;
232
    }
233
    if (strm->zfree == Z_NULL) strm->zfree = zcfree;
234
 
235
    if (level == Z_DEFAULT_COMPRESSION) level = 6;
236
#ifdef FASTEST
237
    level = 1;
238
#endif
239
 
240
    if (windowBits < 0) { /* undocumented feature: suppress zlib header */
241
        noheader = 1;
242
        windowBits = -windowBits;
243
    }
244
    if (memLevel < 1 || memLevel > MAX_MEM_LEVEL || method != Z_DEFLATED ||
245
        windowBits < 9 || windowBits > 15 || level < 0 || level > 9 ||
246
        strategy < 0 || strategy > Z_HUFFMAN_ONLY) {
247
        return Z_STREAM_ERROR;
248
    }
249
    s = (deflate_state *) ZALLOC(strm, 1, sizeof(deflate_state));
250
    if (s == Z_NULL) return Z_MEM_ERROR;
251
    strm->state = (struct internal_state FAR *)s;
252
    s->strm = strm;
253
 
254
    s->noheader = noheader;
255
    s->w_bits = windowBits;
256
    s->w_size = 1 << s->w_bits;
257
    s->w_mask = s->w_size - 1;
258
 
259
    s->hash_bits = memLevel + 7;
260
    s->hash_size = 1 << s->hash_bits;
261
    s->hash_mask = s->hash_size - 1;
262
    s->hash_shift =  ((s->hash_bits+MIN_MATCH-1)/MIN_MATCH);
263
 
264
    s->window = (Bytef *) ZALLOC(strm, s->w_size, 2*sizeof(Byte));
265
    s->prev   = (Posf *)  ZALLOC(strm, s->w_size, sizeof(Pos));
266
    s->head   = (Posf *)  ZALLOC(strm, s->hash_size, sizeof(Pos));
267
 
268
    s->lit_bufsize = 1 << (memLevel + 6); /* 16K elements by default */
269
 
270
    overlay = (ushf *) ZALLOC(strm, s->lit_bufsize, sizeof(ush)+2);
271
    s->pending_buf = (uchf *) overlay;
272
    s->pending_buf_size = (ulg)s->lit_bufsize * (sizeof(ush)+2L);
273
 
274
    if (s->window == Z_NULL || s->prev == Z_NULL || s->head == Z_NULL ||
275
        s->pending_buf == Z_NULL) {
276
        strm->msg = (char*)ERR_MSG(Z_MEM_ERROR);
277
        deflateEnd (strm);
278
        return Z_MEM_ERROR;
279
    }
280
    s->d_buf = overlay + s->lit_bufsize/sizeof(ush);
281
    s->l_buf = s->pending_buf + (1+sizeof(ush))*s->lit_bufsize;
282
 
283
    s->level = level;
284
    s->strategy = strategy;
285
    s->method = (Byte)method;
286
 
287
    return deflateReset(strm);
288
}
289
 
290
/* ========================================================================= */
291
int ZEXPORT deflateSetDictionary (strm, dictionary, dictLength)
292
    z_streamp strm;
293
    const Bytef *dictionary;
294
    uInt  dictLength;
295
{
296
    deflate_state *s;
297
    uInt length = dictLength;
298
    uInt n;
299
    IPos hash_head = 0;
300
 
301
    if (strm == Z_NULL || strm->state == Z_NULL || dictionary == Z_NULL ||
302
        strm->state->status != INIT_STATE) return Z_STREAM_ERROR;
303
 
304
    s = strm->state;
305
    strm->adler = adler32(strm->adler, dictionary, dictLength);
306
 
307
    if (length < MIN_MATCH) return Z_OK;
308
    if (length > MAX_DIST(s)) {
309
        length = MAX_DIST(s);
310
#ifndef USE_DICT_HEAD
311
        dictionary += dictLength - length; /* use the tail of the dictionary */
312
#endif
313
    }
314
    zmemcpy(s->window, dictionary, length);
315
    s->strstart = length;
316
    s->block_start = (long)length;
317
 
318
    /* Insert all strings in the hash table (except for the last two bytes).
319
     * s->lookahead stays null, so s->ins_h will be recomputed at the next
320
     * call of fill_window.
321
     */
322
    s->ins_h = s->window[0];
323
    UPDATE_HASH(s, s->ins_h, s->window[1]);
324
    for (n = 0; n <= length - MIN_MATCH; n++) {
325
        INSERT_STRING(s, n, hash_head);
326
    }
327
    if (hash_head) hash_head = 0;  /* to make compiler happy */
328
    return Z_OK;
329
}
330
 
331
/* ========================================================================= */
332
int ZEXPORT deflateReset (strm)
333
    z_streamp strm;
334
{
335
    deflate_state *s;
336
 
337
    if (strm == Z_NULL || strm->state == Z_NULL ||
338
        strm->zalloc == Z_NULL || strm->zfree == Z_NULL) return Z_STREAM_ERROR;
339
 
340
    strm->total_in = strm->total_out = 0;
341
    strm->msg = Z_NULL; /* use zfree if we ever allocate msg dynamically */
342
    strm->data_type = Z_UNKNOWN;
343
 
344
    s = (deflate_state *)strm->state;
345
    s->pending = 0;
346
    s->pending_out = s->pending_buf;
347
 
348
    if (s->noheader < 0) {
349
        s->noheader = 0; /* was set to -1 by deflate(..., Z_FINISH); */
350
    }
351
    s->status = s->noheader ? BUSY_STATE : INIT_STATE;
352
    strm->adler = 1;
353
    s->last_flush = Z_NO_FLUSH;
354
 
355
    _tr_init(s);
356
    lm_init(s);
357
 
358
    return Z_OK;
359
}
360
 
361
/* ========================================================================= */
362
int ZEXPORT deflateParams(strm, level, strategy)
363
    z_streamp strm;
364
    int level;
365
    int strategy;
366
{
367
    deflate_state *s;
368
    compress_func func;
369
    int err = Z_OK;
370
 
371
    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
372
    s = strm->state;
373
 
374
    if (level == Z_DEFAULT_COMPRESSION) {
375
        level = 6;
376
    }
377
    if (level < 0 || level > 9 || strategy < 0 || strategy > Z_HUFFMAN_ONLY) {
378
        return Z_STREAM_ERROR;
379
    }
380
    func = configuration_table[s->level].func;
381
 
382
    if (func != configuration_table[level].func && strm->total_in != 0) {
383
        /* Flush the last buffer: */
384
        err = deflate(strm, Z_PARTIAL_FLUSH);
385
    }
386
    if (s->level != level) {
387
        s->level = level;
388
        s->max_lazy_match   = configuration_table[level].max_lazy;
389
        s->good_match       = configuration_table[level].good_length;
390
        s->nice_match       = configuration_table[level].nice_length;
391
        s->max_chain_length = configuration_table[level].max_chain;
392
    }
393
    s->strategy = strategy;
394
    return err;
395
}
396
 
397
/* =========================================================================
398
 * Put a short in the pending buffer. The 16-bit value is put in MSB order.
399
 * IN assertion: the stream state is correct and there is enough room in
400
 * pending_buf.
401
 */
402
local void putShortMSB (s, b)
403
    deflate_state *s;
404
    uInt b;
405
{
406
    put_byte(s, (Byte)(b >> 8));
407
    put_byte(s, (Byte)(b & 0xff));
408
}  
409
 
410
/* =========================================================================
411
 * Flush as much pending output as possible. All deflate() output goes
412
 * through this function so some applications may wish to modify it
413
 * to avoid allocating a large strm->next_out buffer and copying into it.
414
 * (See also read_buf()).
415
 */
416
local void flush_pending(strm)
417
    z_streamp strm;
418
{
419
    unsigned len = strm->state->pending;
420
 
421
    if (len > strm->avail_out) len = strm->avail_out;
422
    if (len == 0) return;
423
 
424
    zmemcpy(strm->next_out, strm->state->pending_out, len);
425
    strm->next_out  += len;
426
    strm->state->pending_out  += len;
427
    strm->total_out += len;
428
    strm->avail_out  -= len;
429
    strm->state->pending -= len;
430
    if (strm->state->pending == 0) {
431
        strm->state->pending_out = strm->state->pending_buf;
432
    }
433
}
434
 
435
/* ========================================================================= */
436
int ZEXPORT deflate (strm, flush)
437
    z_streamp strm;
438
    int flush;
439
{
440
    int old_flush; /* value of flush param for previous deflate call */
441
    deflate_state *s;
442
 
443
    if (strm == Z_NULL || strm->state == Z_NULL ||
444
        flush > Z_FINISH || flush < 0) {
445
        return Z_STREAM_ERROR;
446
    }
447
    s = strm->state;
448
 
449
    if (strm->next_out == Z_NULL ||
450
        (strm->next_in == Z_NULL && strm->avail_in != 0) ||
451
        (s->status == FINISH_STATE && flush != Z_FINISH)) {
452
        ERR_RETURN(strm, Z_STREAM_ERROR);
453
    }
454
    if (strm->avail_out == 0) ERR_RETURN(strm, Z_BUF_ERROR);
455
 
456
    s->strm = strm; /* just in case */
457
    old_flush = s->last_flush;
458
    s->last_flush = flush;
459
 
460
    /* Write the zlib header */
461
    if (s->status == INIT_STATE) {
462
 
463
        uInt header = (Z_DEFLATED + ((s->w_bits-8)<<4)) << 8;
464
        uInt level_flags = (s->level-1) >> 1;
465
 
466
        if (level_flags > 3) level_flags = 3;
467
        header |= (level_flags << 6);
468
        if (s->strstart != 0) header |= PRESET_DICT;
469
        header += 31 - (header % 31);
470
 
471
        s->status = BUSY_STATE;
472
        putShortMSB(s, header);
473
 
474
        /* Save the adler32 of the preset dictionary: */
475
        if (s->strstart != 0) {
476
            putShortMSB(s, (uInt)(strm->adler >> 16));
477
            putShortMSB(s, (uInt)(strm->adler & 0xffff));
478
        }
479
        strm->adler = 1L;
480
    }
481
 
482
    /* Flush as much pending output as possible */
483
    if (s->pending != 0) {
484
        flush_pending(strm);
485
        if (strm->avail_out == 0) {
486
            /* Since avail_out is 0, deflate will be called again with
487
             * more output space, but possibly with both pending and
488
             * avail_in equal to zero. There won't be anything to do,
489
             * but this is not an error situation so make sure we
490
             * return OK instead of BUF_ERROR at next call of deflate:
491
             */
492
            s->last_flush = -1;
493
            return Z_OK;
494
        }
495
 
496
    /* Make sure there is something to do and avoid duplicate consecutive
497
     * flushes. For repeated and useless calls with Z_FINISH, we keep
498
     * returning Z_STREAM_END instead of Z_BUFF_ERROR.
499
     */
500
    } else if (strm->avail_in == 0 && flush <= old_flush &&
501
               flush != Z_FINISH) {
502
        ERR_RETURN(strm, Z_BUF_ERROR);
503
    }
504
 
505
    /* User must not provide more input after the first FINISH: */
506
    if (s->status == FINISH_STATE && strm->avail_in != 0) {
507
        ERR_RETURN(strm, Z_BUF_ERROR);
508
    }
509
 
510
    /* Start a new block or continue the current one.
511
     */
512
    if (strm->avail_in != 0 || s->lookahead != 0 ||
513
        (flush != Z_NO_FLUSH && s->status != FINISH_STATE)) {
514
        block_state bstate;
515
 
516
        bstate = (*(configuration_table[s->level].func))(s, flush);
517
 
518
        if (bstate == finish_started || bstate == finish_done) {
519
            s->status = FINISH_STATE;
520
        }
521
        if (bstate == need_more || bstate == finish_started) {
522
            if (strm->avail_out == 0) {
523
                s->last_flush = -1; /* avoid BUF_ERROR next call, see above */
524
            }
525
            return Z_OK;
526
            /* If flush != Z_NO_FLUSH && avail_out == 0, the next call
527
             * of deflate should use the same flush parameter to make sure
528
             * that the flush is complete. So we don't have to output an
529
             * empty block here, this will be done at next call. This also
530
             * ensures that for a very small output buffer, we emit at most
531
             * one empty block.
532
             */
533
        }
534
        if (bstate == block_done) {
535
            if (flush == Z_PARTIAL_FLUSH) {
536
                _tr_align(s);
537
            } else { /* FULL_FLUSH or SYNC_FLUSH */
538
                _tr_stored_block(s, (char*)0, 0L, 0);
539
                /* For a full flush, this empty block will be recognized
540
                 * as a special marker by inflate_sync().
541
                 */
542
                if (flush == Z_FULL_FLUSH) {
543
                    CLEAR_HASH(s);             /* forget history */
544
                }
545
            }
546
            flush_pending(strm);
547
            if (strm->avail_out == 0) {
548
              s->last_flush = -1; /* avoid BUF_ERROR at next call, see above */
549
              return Z_OK;
550
            }
551
        }
552
    }
553
    Assert(strm->avail_out > 0, "bug2");
554
 
555
    if (flush != Z_FINISH) return Z_OK;
556
    if (s->noheader) return Z_STREAM_END;
557
 
558
    /* Write the zlib trailer (adler32) */
559
    putShortMSB(s, (uInt)(strm->adler >> 16));
560
    putShortMSB(s, (uInt)(strm->adler & 0xffff));
561
    flush_pending(strm);
562
    /* If avail_out is zero, the application will call deflate again
563
     * to flush the rest.
564
     */
565
    s->noheader = -1; /* write the trailer only once! */
566
    return s->pending != 0 ? Z_OK : Z_STREAM_END;
567
}
568
 
569
/* ========================================================================= */
570
int ZEXPORT deflateEnd (strm)
571
    z_streamp strm;
572
{
573
    int status;
574
 
575
    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
576
 
577
    status = strm->state->status;
578
    if (status != INIT_STATE && status != BUSY_STATE &&
579
        status != FINISH_STATE) {
580
      return Z_STREAM_ERROR;
581
    }
582
 
583
    /* Deallocate in reverse order of allocations: */
584
    TRY_FREE(strm, strm->state->pending_buf);
585
    TRY_FREE(strm, strm->state->head);
586
    TRY_FREE(strm, strm->state->prev);
587
    TRY_FREE(strm, strm->state->window);
588
 
589
    ZFREE(strm, strm->state);
590
    strm->state = Z_NULL;
591
 
592
    return status == BUSY_STATE ? Z_DATA_ERROR : Z_OK;
593
}
594
 
595
/* =========================================================================
596
 * Copy the source state to the destination state.
597
 * To simplify the source, this is not supported for 16-bit MSDOS (which
598
 * doesn't have enough memory anyway to duplicate compression states).
599
 */
600
int ZEXPORT deflateCopy (dest, source)
601
    z_streamp dest;
602
    z_streamp source;
603
{
604
#ifdef MAXSEG_64K
605
    return Z_STREAM_ERROR;
606
#else
607
    deflate_state *ds;
608
    deflate_state *ss;
609
    ushf *overlay;
610
 
611
 
612
    if (source == Z_NULL || dest == Z_NULL || source->state == Z_NULL) {
613
        return Z_STREAM_ERROR;
614
    }
615
 
616
    ss = source->state;
617
 
618
    *dest = *source;
619
 
620
    ds = (deflate_state *) ZALLOC(dest, 1, sizeof(deflate_state));
621
    if (ds == Z_NULL) return Z_MEM_ERROR;
622
    dest->state = (struct internal_state FAR *) ds;
623
    *ds = *ss;
624
    ds->strm = dest;
625
 
626
    ds->window = (Bytef *) ZALLOC(dest, ds->w_size, 2*sizeof(Byte));
627
    ds->prev   = (Posf *)  ZALLOC(dest, ds->w_size, sizeof(Pos));
628
    ds->head   = (Posf *)  ZALLOC(dest, ds->hash_size, sizeof(Pos));
629
    overlay = (ushf *) ZALLOC(dest, ds->lit_bufsize, sizeof(ush)+2);
630
    ds->pending_buf = (uchf *) overlay;
631
 
632
    if (ds->window == Z_NULL || ds->prev == Z_NULL || ds->head == Z_NULL ||
633
        ds->pending_buf == Z_NULL) {
634
        deflateEnd (dest);
635
        return Z_MEM_ERROR;
636
    }
637
    /* following zmemcpy do not work for 16-bit MSDOS */
638
    zmemcpy(ds->window, ss->window, ds->w_size * 2 * sizeof(Byte));
639
    zmemcpy(ds->prev, ss->prev, ds->w_size * sizeof(Pos));
640
    zmemcpy(ds->head, ss->head, ds->hash_size * sizeof(Pos));
641
    zmemcpy(ds->pending_buf, ss->pending_buf, (uInt)ds->pending_buf_size);
642
 
643
    ds->pending_out = ds->pending_buf + (ss->pending_out - ss->pending_buf);
644
    ds->d_buf = overlay + ds->lit_bufsize/sizeof(ush);
645
    ds->l_buf = ds->pending_buf + (1+sizeof(ush))*ds->lit_bufsize;
646
 
647
    ds->l_desc.dyn_tree = ds->dyn_ltree;
648
    ds->d_desc.dyn_tree = ds->dyn_dtree;
649
    ds->bl_desc.dyn_tree = ds->bl_tree;
650
 
651
    return Z_OK;
652
#endif
653
}
654
 
655
/* ===========================================================================
656
 * Read a new buffer from the current input stream, update the adler32
657
 * and total number of bytes read.  All deflate() input goes through
658
 * this function so some applications may wish to modify it to avoid
659
 * allocating a large strm->next_in buffer and copying from it.
660
 * (See also flush_pending()).
661
 */
662
local int read_buf(strm, buf, size)
663
    z_streamp strm;
664
    Bytef *buf;
665
    unsigned size;
666
{
667
    unsigned len = strm->avail_in;
668
 
669
    if (len > size) len = size;
670
    if (len == 0) return 0;
671
 
672
    strm->avail_in  -= len;
673
 
674
    if (!strm->state->noheader) {
675
        strm->adler = adler32(strm->adler, strm->next_in, len);
676
    }
677
    zmemcpy(buf, strm->next_in, len);
678
    strm->next_in  += len;
679
    strm->total_in += len;
680
 
681
    return (int)len;
682
}
683
 
684
/* ===========================================================================
685
 * Initialize the "longest match" routines for a new zlib stream
686
 */
687
local void lm_init (s)
688
    deflate_state *s;
689
{
690
    s->window_size = (ulg)2L*s->w_size;
691
 
692
    CLEAR_HASH(s);
693
 
694
    /* Set the default configuration parameters:
695
     */
696
    s->max_lazy_match   = configuration_table[s->level].max_lazy;
697
    s->good_match       = configuration_table[s->level].good_length;
698
    s->nice_match       = configuration_table[s->level].nice_length;
699
    s->max_chain_length = configuration_table[s->level].max_chain;
700
 
701
    s->strstart = 0;
702
    s->block_start = 0L;
703
    s->lookahead = 0;
704
    s->match_length = s->prev_length = MIN_MATCH-1;
705
    s->match_available = 0;
706
    s->ins_h = 0;
707
#ifdef ASMV
708
    match_init(); /* initialize the asm code */
709
#endif
710
}
711
 
712
/* ===========================================================================
713
 * Set match_start to the longest match starting at the given string and
714
 * return its length. Matches shorter or equal to prev_length are discarded,
715
 * in which case the result is equal to prev_length and match_start is
716
 * garbage.
717
 * IN assertions: cur_match is the head of the hash chain for the current
718
 *   string (strstart) and its distance is <= MAX_DIST, and prev_length >= 1
719
 * OUT assertion: the match length is not greater than s->lookahead.
720
 */
721
#ifndef ASMV
722
/* For 80x86 and 680x0, an optimized version will be provided in match.asm or
723
 * match.S. The code will be functionally equivalent.
724
 */
725
#ifndef FASTEST
726
local uInt longest_match(s, cur_match)
727
    deflate_state *s;
728
    IPos cur_match;                             /* current match */
729
{
730
    unsigned chain_length = s->max_chain_length;/* max hash chain length */
731
    register Bytef *scan = s->window + s->strstart; /* current string */
732
    register Bytef *match;                       /* matched string */
733
    register int len;                           /* length of current match */
734
    int best_len = s->prev_length;              /* best match length so far */
735
    int nice_match = s->nice_match;             /* stop if match long enough */
736
    IPos limit = s->strstart > (IPos)MAX_DIST(s) ?
737
        s->strstart - (IPos)MAX_DIST(s) : NIL;
738
    /* Stop when cur_match becomes <= limit. To simplify the code,
739
     * we prevent matches with the string of window index 0.
740
     */
741
    Posf *prev = s->prev;
742
    uInt wmask = s->w_mask;
743
 
744
#ifdef UNALIGNED_OK
745
    /* Compare two bytes at a time. Note: this is not always beneficial.
746
     * Try with and without -DUNALIGNED_OK to check.
747
     */
748
    register Bytef *strend = s->window + s->strstart + MAX_MATCH - 1;
749
    register ush scan_start = *(ushf*)scan;
750
    register ush scan_end   = *(ushf*)(scan+best_len-1);
751
#else
752
    register Bytef *strend = s->window + s->strstart + MAX_MATCH;
753
    register Byte scan_end1  = scan[best_len-1];
754
    register Byte scan_end   = scan[best_len];
755
#endif
756
 
757
    /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16.
758
     * It is easy to get rid of this optimization if necessary.
759
     */
760
    Assert(s->hash_bits >= 8 && MAX_MATCH == 258, "Code too clever");
761
 
762
    /* Do not waste too much time if we already have a good match: */
763
    if (s->prev_length >= s->good_match) {
764
        chain_length >>= 2;
765
    }
766
    /* Do not look for matches beyond the end of the input. This is necessary
767
     * to make deflate deterministic.
768
     */
769
    if ((uInt)nice_match > s->lookahead) nice_match = s->lookahead;
770
 
771
    Assert((ulg)s->strstart <= s->window_size-MIN_LOOKAHEAD, "need lookahead");
772
 
773
    do {
774
        Assert(cur_match < s->strstart, "no future");
775
        match = s->window + cur_match;
776
 
777
        /* Skip to next match if the match length cannot increase
778
         * or if the match length is less than 2:
779
         */
780
#if (defined(UNALIGNED_OK) && MAX_MATCH == 258)
781
        /* This code assumes sizeof(unsigned short) == 2. Do not use
782
         * UNALIGNED_OK if your compiler uses a different size.
783
         */
784
        if (*(ushf*)(match+best_len-1) != scan_end ||
785
            *(ushf*)match != scan_start) continue;
786
 
787
        /* It is not necessary to compare scan[2] and match[2] since they are
788
         * always equal when the other bytes match, given that the hash keys
789
         * are equal and that HASH_BITS >= 8. Compare 2 bytes at a time at
790
         * strstart+3, +5, ... up to strstart+257. We check for insufficient
791
         * lookahead only every 4th comparison; the 128th check will be made
792
         * at strstart+257. If MAX_MATCH-2 is not a multiple of 8, it is
793
         * necessary to put more guard bytes at the end of the window, or
794
         * to check more often for insufficient lookahead.
795
         */
796
        Assert(scan[2] == match[2], "scan[2]?");
797
        scan++, match++;
798
        do {
799
        } while (*(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
800
                 *(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
801
                 *(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
802
                 *(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
803
                 scan < strend);
804
        /* The funny "do {}" generates better code on most compilers */
805
 
806
        /* Here, scan <= window+strstart+257 */
807
        Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan");
808
        if (*scan == *match) scan++;
809
 
810
        len = (MAX_MATCH - 1) - (int)(strend-scan);
811
        scan = strend - (MAX_MATCH-1);
812
 
813
#else /* UNALIGNED_OK */
814
 
815
        if (match[best_len]   != scan_end  ||
816
            match[best_len-1] != scan_end1 ||
817
            *match            != *scan     ||
818
            *++match          != scan[1])      continue;
819
 
820
        /* The check at best_len-1 can be removed because it will be made
821
         * again later. (This heuristic is not always a win.)
822
         * It is not necessary to compare scan[2] and match[2] since they
823
         * are always equal when the other bytes match, given that
824
         * the hash keys are equal and that HASH_BITS >= 8.
825
         */
826
        scan += 2, match++;
827
        Assert(*scan == *match, "match[2]?");
828
 
829
        /* We check for insufficient lookahead only every 8th comparison;
830
         * the 256th check will be made at strstart+258.
831
         */
832
        do {
833
        } while (*++scan == *++match && *++scan == *++match &&
834
                 *++scan == *++match && *++scan == *++match &&
835
                 *++scan == *++match && *++scan == *++match &&
836
                 *++scan == *++match && *++scan == *++match &&
837
                 scan < strend);
838
 
839
        Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan");
840
 
841
        len = MAX_MATCH - (int)(strend - scan);
842
        scan = strend - MAX_MATCH;
843
 
844
#endif /* UNALIGNED_OK */
845
 
846
        if (len > best_len) {
847
            s->match_start = cur_match;
848
            best_len = len;
849
            if (len >= nice_match) break;
850
#ifdef UNALIGNED_OK
851
            scan_end = *(ushf*)(scan+best_len-1);
852
#else
853
            scan_end1  = scan[best_len-1];
854
            scan_end   = scan[best_len];
855
#endif
856
        }
857
    } while ((cur_match = prev[cur_match & wmask]) > limit
858
             && --chain_length != 0);
859
 
860
    if ((uInt)best_len <= s->lookahead) return (uInt)best_len;
861
    return s->lookahead;
862
}
863
 
864
#else /* FASTEST */
865
/* ---------------------------------------------------------------------------
866
 * Optimized version for level == 1 only
867
 */
868
local uInt longest_match(s, cur_match)
869
    deflate_state *s;
870
    IPos cur_match;                             /* current match */
871
{
872
    register Bytef *scan = s->window + s->strstart; /* current string */
873
    register Bytef *match;                       /* matched string */
874
    register int len;                           /* length of current match */
875
    register Bytef *strend = s->window + s->strstart + MAX_MATCH;
876
 
877
    /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16.
878
     * It is easy to get rid of this optimization if necessary.
879
     */
880
    Assert(s->hash_bits >= 8 && MAX_MATCH == 258, "Code too clever");
881
 
882
    Assert((ulg)s->strstart <= s->window_size-MIN_LOOKAHEAD, "need lookahead");
883
 
884
    Assert(cur_match < s->strstart, "no future");
885
 
886
    match = s->window + cur_match;
887
 
888
    /* Return failure if the match length is less than 2:
889
     */
890
    if (match[0] != scan[0] || match[1] != scan[1]) return MIN_MATCH-1;
891
 
892
    /* The check at best_len-1 can be removed because it will be made
893
     * again later. (This heuristic is not always a win.)
894
     * It is not necessary to compare scan[2] and match[2] since they
895
     * are always equal when the other bytes match, given that
896
     * the hash keys are equal and that HASH_BITS >= 8.
897
     */
898
    scan += 2, match += 2;
899
    Assert(*scan == *match, "match[2]?");
900
 
901
    /* We check for insufficient lookahead only every 8th comparison;
902
     * the 256th check will be made at strstart+258.
903
     */
904
    do {
905
    } while (*++scan == *++match && *++scan == *++match &&
906
             *++scan == *++match && *++scan == *++match &&
907
             *++scan == *++match && *++scan == *++match &&
908
             *++scan == *++match && *++scan == *++match &&
909
             scan < strend);
910
 
911
    Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan");
912
 
913
    len = MAX_MATCH - (int)(strend - scan);
914
 
915
    if (len < MIN_MATCH) return MIN_MATCH - 1;
916
 
917
    s->match_start = cur_match;
918
    return len <= s->lookahead ? len : s->lookahead;
919
}
920
#endif /* FASTEST */
921
#endif /* ASMV */
922
 
923
#ifdef DEBUG
924
/* ===========================================================================
925
 * Check that the match at match_start is indeed a match.
926
 */
927
local void check_match(s, start, match, length)
928
    deflate_state *s;
929
    IPos start, match;
930
    int length;
931
{
932
    /* check that the match is indeed a match */
933
    if (zmemcmp(s->window + match,
934
                s->window + start, length) != EQUAL) {
935
        cprintf(" start %u, match %u, length %d\n",
936
                start, match, length);
937
        do {
938
            cprintf("%c%c", s->window[match++], s->window[start++]);
939
        } while (--length != 0);
940
        z_error("invalid match");
941
    }
942
    if (z_verbose > 1) {
943
        cprintf("\\[%d,%d]", start-match, length);
944
        do { putc(s->window[start++], stderr); } while (--length != 0);
945
    }
946
}
947
#else
948
#  define check_match(s, start, match, length)
949
#endif
950
 
951
/* ===========================================================================
952
 * Fill the window when the lookahead becomes insufficient.
953
 * Updates strstart and lookahead.
954
 *
955
 * IN assertion: lookahead < MIN_LOOKAHEAD
956
 * OUT assertions: strstart <= window_size-MIN_LOOKAHEAD
957
 *    At least one byte has been read, or avail_in == 0; reads are
958
 *    performed for at least two bytes (required for the zip translate_eol
959
 *    option -- not supported here).
960
 */
961
local void fill_window(s)
962
    deflate_state *s;
963
{
964
    register unsigned n, m;
965
    register Posf *p;
966
    unsigned more;    /* Amount of free space at the end of the window. */
967
    uInt wsize = s->w_size;
968
 
969
    do {
970
        more = (unsigned)(s->window_size -(ulg)s->lookahead -(ulg)s->strstart);
971
 
972
        /* Deal with !@#$% 64K limit: */
973
        if (more == 0 && s->strstart == 0 && s->lookahead == 0) {
974
            more = wsize;
975
 
976
        } else if (more == (unsigned)(-1)) {
977
            /* Very unlikely, but possible on 16 bit machine if strstart == 0
978
             * and lookahead == 1 (input done one byte at time)
979
             */
980
            more--;
981
 
982
        /* If the window is almost full and there is insufficient lookahead,
983
         * move the upper half to the lower one to make room in the upper half.
984
         */
985
        } else if (s->strstart >= wsize+MAX_DIST(s)) {
986
 
987
            zmemcpy(s->window, s->window+wsize, (unsigned)wsize);
988
            s->match_start -= wsize;
989
            s->strstart    -= wsize; /* we now have strstart >= MAX_DIST */
990
            s->block_start -= (long) wsize;
991
 
992
            /* Slide the hash table (could be avoided with 32 bit values
993
               at the expense of memory usage). We slide even when level == 0
994
               to keep the hash table consistent if we switch back to level > 0
995
               later. (Using level 0 permanently is not an optimal usage of
996
               zlib, so we don't care about this pathological case.)
997
             */
998
            n = s->hash_size;
999
            p = &s->head[n];
1000
            do {
1001
                m = *--p;
1002
                *p = (Pos)(m >= wsize ? m-wsize : NIL);
1003
            } while (--n);
1004
 
1005
            n = wsize;
1006
#ifndef FASTEST
1007
            p = &s->prev[n];
1008
            do {
1009
                m = *--p;
1010
                *p = (Pos)(m >= wsize ? m-wsize : NIL);
1011
                /* If n is not on any hash chain, prev[n] is garbage but
1012
                 * its value will never be used.
1013
                 */
1014
            } while (--n);
1015
#endif
1016
            more += wsize;
1017
        }
1018
        if (s->strm->avail_in == 0) return;
1019
 
1020
        /* If there was no sliding:
1021
         *    strstart <= WSIZE+MAX_DIST-1 && lookahead <= MIN_LOOKAHEAD - 1 &&
1022
         *    more == window_size - lookahead - strstart
1023
         * => more >= window_size - (MIN_LOOKAHEAD-1 + WSIZE + MAX_DIST-1)
1024
         * => more >= window_size - 2*WSIZE + 2
1025
         * In the BIG_MEM or MMAP case (not yet supported),
1026
         *   window_size == input_size + MIN_LOOKAHEAD  &&
1027
         *   strstart + s->lookahead <= input_size => more >= MIN_LOOKAHEAD.
1028
         * Otherwise, window_size == 2*WSIZE so more >= 2.
1029
         * If there was sliding, more >= WSIZE. So in all cases, more >= 2.
1030
         */
1031
        Assert(more >= 2, "more < 2");
1032
 
1033
        n = read_buf(s->strm, s->window + s->strstart + s->lookahead, more);
1034
        s->lookahead += n;
1035
 
1036
        /* Initialize the hash value now that we have some input: */
1037
        if (s->lookahead >= MIN_MATCH) {
1038
            s->ins_h = s->window[s->strstart];
1039
            UPDATE_HASH(s, s->ins_h, s->window[s->strstart+1]);
1040
#if MIN_MATCH != 3
1041
            Call UPDATE_HASH() MIN_MATCH-3 more times
1042
#endif
1043
        }
1044
        /* If the whole input has less than MIN_MATCH bytes, ins_h is garbage,
1045
         * but this is not important since only literal bytes will be emitted.
1046
         */
1047
 
1048
    } while (s->lookahead < MIN_LOOKAHEAD && s->strm->avail_in != 0);
1049
}
1050
 
1051
/* ===========================================================================
1052
 * Flush the current block, with given end-of-file flag.
1053
 * IN assertion: strstart is set to the end of the current match.
1054
 */
1055
#define FLUSH_BLOCK_ONLY(s, eof) { \
1056
   _tr_flush_block(s, (s->block_start >= 0L ? \
1057
                   (charf *)&s->window[(unsigned)s->block_start] : \
1058
                   (charf *)Z_NULL), \
1059
                (ulg)((long)s->strstart - s->block_start), \
1060
                (eof)); \
1061
   s->block_start = s->strstart; \
1062
   flush_pending(s->strm); \
1063
   Tracev((stderr,"[FLUSH]")); \
1064
}
1065
 
1066
/* Same but force premature exit if necessary. */
1067
#define FLUSH_BLOCK(s, eof) { \
1068
   FLUSH_BLOCK_ONLY(s, eof); \
1069
   if (s->strm->avail_out == 0) return (eof) ? finish_started : need_more; \
1070
}
1071
 
1072
/* ===========================================================================
1073
 * Copy without compression as much as possible from the input stream, return
1074
 * the current block state.
1075
 * This function does not insert new strings in the dictionary since
1076
 * uncompressible data is probably not useful. This function is used
1077
 * only for the level=0 compression option.
1078
 * NOTE: this function should be optimized to avoid extra copying from
1079
 * window to pending_buf.
1080
 */
1081
local block_state deflate_stored(s, flush)
1082
    deflate_state *s;
1083
    int flush;
1084
{
1085
    /* Stored blocks are limited to 0xffff bytes, pending_buf is limited
1086
     * to pending_buf_size, and each stored block has a 5 byte header:
1087
     */
1088
    ulg max_block_size = 0xffff;
1089
    ulg max_start;
1090
 
1091
    if (max_block_size > s->pending_buf_size - 5) {
1092
        max_block_size = s->pending_buf_size - 5;
1093
    }
1094
 
1095
    /* Copy as much as possible from input to output: */
1096
    for (;;) {
1097
        /* Fill the window as much as possible: */
1098
        if (s->lookahead <= 1) {
1099
 
1100
            Assert(s->strstart < s->w_size+MAX_DIST(s) ||
1101
                   s->block_start >= (long)s->w_size, "slide too late");
1102
 
1103
            fill_window(s);
1104
            if (s->lookahead == 0 && flush == Z_NO_FLUSH) return need_more;
1105
 
1106
            if (s->lookahead == 0) break; /* flush the current block */
1107
        }
1108
        Assert(s->block_start >= 0L, "block gone");
1109
 
1110
        s->strstart += s->lookahead;
1111
        s->lookahead = 0;
1112
 
1113
        /* Emit a stored block if pending_buf will be full: */
1114
        max_start = s->block_start + max_block_size;
1115
        if (s->strstart == 0 || (ulg)s->strstart >= max_start) {
1116
            /* strstart == 0 is possible when wraparound on 16-bit machine */
1117
            s->lookahead = (uInt)(s->strstart - max_start);
1118
            s->strstart = (uInt)max_start;
1119
            FLUSH_BLOCK(s, 0);
1120
        }
1121
        /* Flush if we may have to slide, otherwise block_start may become
1122
         * negative and the data will be gone:
1123
         */
1124
        if (s->strstart - (uInt)s->block_start >= MAX_DIST(s)) {
1125
            FLUSH_BLOCK(s, 0);
1126
        }
1127
    }
1128
    FLUSH_BLOCK(s, flush == Z_FINISH);
1129
    return flush == Z_FINISH ? finish_done : block_done;
1130
}
1131
 
1132
/* ===========================================================================
1133
 * Compress as much as possible from the input stream, return the current
1134
 * block state.
1135
 * This function does not perform lazy evaluation of matches and inserts
1136
 * new strings in the dictionary only for unmatched strings or for short
1137
 * matches. It is used only for the fast compression options.
1138
 */
1139
local block_state deflate_fast(s, flush)
1140
    deflate_state *s;
1141
    int flush;
1142
{
1143
    IPos hash_head = NIL; /* head of the hash chain */
1144
    int bflush;           /* set if current block must be flushed */
1145
 
1146
    for (;;) {
1147
        /* Make sure that we always have enough lookahead, except
1148
         * at the end of the input file. We need MAX_MATCH bytes
1149
         * for the next match, plus MIN_MATCH bytes to insert the
1150
         * string following the next match.
1151
         */
1152
        if (s->lookahead < MIN_LOOKAHEAD) {
1153
            fill_window(s);
1154
            if (s->lookahead < MIN_LOOKAHEAD && flush == Z_NO_FLUSH) {
1155
                return need_more;
1156
            }
1157
            if (s->lookahead == 0) break; /* flush the current block */
1158
        }
1159
 
1160
        /* Insert the string window[strstart .. strstart+2] in the
1161
         * dictionary, and set hash_head to the head of the hash chain:
1162
         */
1163
        if (s->lookahead >= MIN_MATCH) {
1164
            INSERT_STRING(s, s->strstart, hash_head);
1165
        }
1166
 
1167
        /* Find the longest match, discarding those <= prev_length.
1168
         * At this point we have always match_length < MIN_MATCH
1169
         */
1170
        if (hash_head != NIL && s->strstart - hash_head <= MAX_DIST(s)) {
1171
            /* To simplify the code, we prevent matches with the string
1172
             * of window index 0 (in particular we have to avoid a match
1173
             * of the string with itself at the start of the input file).
1174
             */
1175
            if (s->strategy != Z_HUFFMAN_ONLY) {
1176
                s->match_length = longest_match (s, hash_head);
1177
            }
1178
            /* longest_match() sets match_start */
1179
        }
1180
        if (s->match_length >= MIN_MATCH) {
1181
            check_match(s, s->strstart, s->match_start, s->match_length);
1182
 
1183
            _tr_tally_dist(s, s->strstart - s->match_start,
1184
                           s->match_length - MIN_MATCH, bflush);
1185
 
1186
            s->lookahead -= s->match_length;
1187
 
1188
            /* Insert new strings in the hash table only if the match length
1189
             * is not too large. This saves time but degrades compression.
1190
             */
1191
#ifndef FASTEST
1192
            if (s->match_length <= s->max_insert_length &&
1193
                s->lookahead >= MIN_MATCH) {
1194
                s->match_length--; /* string at strstart already in hash table */
1195
                do {
1196
                    s->strstart++;
1197
                    INSERT_STRING(s, s->strstart, hash_head);
1198
                    /* strstart never exceeds WSIZE-MAX_MATCH, so there are
1199
                     * always MIN_MATCH bytes ahead.
1200
                     */
1201
                } while (--s->match_length != 0);
1202
                s->strstart++;
1203
            } else
1204
#endif
1205
            {
1206
                s->strstart += s->match_length;
1207
                s->match_length = 0;
1208
                s->ins_h = s->window[s->strstart];
1209
                UPDATE_HASH(s, s->ins_h, s->window[s->strstart+1]);
1210
#if MIN_MATCH != 3
1211
                Call UPDATE_HASH() MIN_MATCH-3 more times
1212
#endif
1213
                /* If lookahead < MIN_MATCH, ins_h is garbage, but it does not
1214
                 * matter since it will be recomputed at next deflate call.
1215
                 */
1216
            }
1217
        } else {
1218
            /* No match, output a literal byte */
1219
            Tracevv((stderr,"%c", s->window[s->strstart]));
1220
            _tr_tally_lit (s, s->window[s->strstart], bflush);
1221
            s->lookahead--;
1222
            s->strstart++;
1223
        }
1224
        if (bflush) FLUSH_BLOCK(s, 0);
1225
    }
1226
    FLUSH_BLOCK(s, flush == Z_FINISH);
1227
    return flush == Z_FINISH ? finish_done : block_done;
1228
}
1229
 
1230
/* ===========================================================================
1231
 * Same as above, but achieves better compression. We use a lazy
1232
 * evaluation for matches: a match is finally adopted only if there is
1233
 * no better match at the next window position.
1234
 */
1235
local block_state deflate_slow(s, flush)
1236
    deflate_state *s;
1237
    int flush;
1238
{
1239
    IPos hash_head = NIL;    /* head of hash chain */
1240
    int bflush;              /* set if current block must be flushed */
1241
 
1242
    /* Process the input block. */
1243
    for (;;) {
1244
        /* Make sure that we always have enough lookahead, except
1245
         * at the end of the input file. We need MAX_MATCH bytes
1246
         * for the next match, plus MIN_MATCH bytes to insert the
1247
         * string following the next match.
1248
         */
1249
        if (s->lookahead < MIN_LOOKAHEAD) {
1250
            fill_window(s);
1251
            if (s->lookahead < MIN_LOOKAHEAD && flush == Z_NO_FLUSH) {
1252
                return need_more;
1253
            }
1254
            if (s->lookahead == 0) break; /* flush the current block */
1255
        }
1256
 
1257
        /* Insert the string window[strstart .. strstart+2] in the
1258
         * dictionary, and set hash_head to the head of the hash chain:
1259
         */
1260
        if (s->lookahead >= MIN_MATCH) {
1261
            INSERT_STRING(s, s->strstart, hash_head);
1262
        }
1263
 
1264
        /* Find the longest match, discarding those <= prev_length.
1265
         */
1266
        s->prev_length = s->match_length, s->prev_match = s->match_start;
1267
        s->match_length = MIN_MATCH-1;
1268
 
1269
        if (hash_head != NIL && s->prev_length < s->max_lazy_match &&
1270
            s->strstart - hash_head <= MAX_DIST(s)) {
1271
            /* To simplify the code, we prevent matches with the string
1272
             * of window index 0 (in particular we have to avoid a match
1273
             * of the string with itself at the start of the input file).
1274
             */
1275
            if (s->strategy != Z_HUFFMAN_ONLY) {
1276
                s->match_length = longest_match (s, hash_head);
1277
            }
1278
            /* longest_match() sets match_start */
1279
 
1280
            if (s->match_length <= 5 && (s->strategy == Z_FILTERED ||
1281
                 (s->match_length == MIN_MATCH &&
1282
                  s->strstart - s->match_start > TOO_FAR))) {
1283
 
1284
                /* If prev_match is also MIN_MATCH, match_start is garbage
1285
                 * but we will ignore the current match anyway.
1286
                 */
1287
                s->match_length = MIN_MATCH-1;
1288
            }
1289
        }
1290
        /* If there was a match at the previous step and the current
1291
         * match is not better, output the previous match:
1292
         */
1293
        if (s->prev_length >= MIN_MATCH && s->match_length <= s->prev_length) {
1294
            uInt max_insert = s->strstart + s->lookahead - MIN_MATCH;
1295
            /* Do not insert strings in hash table beyond this. */
1296
 
1297
            check_match(s, s->strstart-1, s->prev_match, s->prev_length);
1298
 
1299
            _tr_tally_dist(s, s->strstart -1 - s->prev_match,
1300
                           s->prev_length - MIN_MATCH, bflush);
1301
 
1302
            /* Insert in hash table all strings up to the end of the match.
1303
             * strstart-1 and strstart are already inserted. If there is not
1304
             * enough lookahead, the last two strings are not inserted in
1305
             * the hash table.
1306
             */
1307
            s->lookahead -= s->prev_length-1;
1308
            s->prev_length -= 2;
1309
            do {
1310
                if (++s->strstart <= max_insert) {
1311
                    INSERT_STRING(s, s->strstart, hash_head);
1312
                }
1313
            } while (--s->prev_length != 0);
1314
            s->match_available = 0;
1315
            s->match_length = MIN_MATCH-1;
1316
            s->strstart++;
1317
 
1318
            if (bflush) FLUSH_BLOCK(s, 0);
1319
 
1320
        } else if (s->match_available) {
1321
            /* If there was no match at the previous position, output a
1322
             * single literal. If there was a match but the current match
1323
             * is longer, truncate the previous match to a single literal.
1324
             */
1325
            Tracevv((stderr,"%c", s->window[s->strstart-1]));
1326
            _tr_tally_lit(s, s->window[s->strstart-1], bflush);
1327
            if (bflush) {
1328
                FLUSH_BLOCK_ONLY(s, 0);
1329
            }
1330
            s->strstart++;
1331
            s->lookahead--;
1332
            if (s->strm->avail_out == 0) return need_more;
1333
        } else {
1334
            /* There is no previous match to compare with, wait for
1335
             * the next step to decide.
1336
             */
1337
            s->match_available = 1;
1338
            s->strstart++;
1339
            s->lookahead--;
1340
        }
1341
    }
1342
    Assert (flush != Z_NO_FLUSH, "no flush?");
1343
    if (s->match_available) {
1344
        Tracevv((stderr,"%c", s->window[s->strstart-1]));
1345
        _tr_tally_lit(s, s->window[s->strstart-1], bflush);
1346
        s->match_available = 0;
1347
    }
1348
    FLUSH_BLOCK(s, flush == Z_FINISH);
1349
    return flush == Z_FINISH ? finish_done : block_done;
1350
}