1 /* deflate.c -- compress data using the deflation algorithm
2 * Copyright (C) 1995-2013 Jean-loup Gailly and Mark Adler
3 * For conditions of distribution and use, see copyright notice in zlib.h
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).
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.
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.
39 * Deutsch, L.P.,"DEFLATE Compressed Data Format Specification".
40 * Available in http://tools.ietf.org/html/rfc1951
42 * A description of the Rabin and Karp algorithm is given in the book
43 * "Algorithms" by R. Sedgewick, Addison-Wesley, p252.
45 * Fiala,E.R., and Greene,D.H.
46 * Data Compression with Finite Windows, Comm.ACM, 32,4 (1989) 490-595
54 const char deflate_copyright
[] =
55 " deflate 1.2.8.f-Proxmark3 Copyright 1995-2013 Jean-loup Gailly and Mark Adler ";
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.
63 //-----------------------------------------------------------------------------
64 // This version of zlib is modified for use within the Proxmark3 project.
65 // Files from the original distribution which are not required for this
66 // purpose are not included. All modifications can easily be found
67 // by searching for #ifdef ZLIB_PM3_TUNED and #ifndef ZLIB_PM3_TUNED.
68 //-----------------------------------------------------------------------------
72 /* ===========================================================================
73 * Function prototypes.
76 need_more
, /* block not completed, need more input or more output */
77 block_done
, /* block flush performed */
78 finish_started
, /* finish started, need only more output at next deflate */
79 finish_done
/* finish done, accept no more input or output */
82 typedef block_state (*compress_func
) OF((deflate_state
*s
, int flush
));
83 /* Compression function. Returns the block state after the call. */
85 local
void fill_window
OF((deflate_state
*s
));
86 local block_state deflate_stored
OF((deflate_state
*s
, int flush
));
87 local block_state deflate_fast
OF((deflate_state
*s
, int flush
));
89 local block_state deflate_slow
OF((deflate_state
*s
, int flush
));
91 local block_state deflate_rle
OF((deflate_state
*s
, int flush
));
92 local block_state deflate_huff
OF((deflate_state
*s
, int flush
));
93 local
void lm_init
OF((deflate_state
*s
));
94 local
void putShortMSB
OF((deflate_state
*s
, uInt b
));
95 local
void flush_pending
OF((z_streamp strm
));
96 local
int read_buf
OF((z_streamp strm
, Bytef
*buf
, unsigned size
));
98 void match_init
OF((void)); /* asm code initialization */
99 uInt longest_match
OF((deflate_state
*s
, IPos cur_match
));
101 local uInt longest_match
OF((deflate_state
*s
, IPos cur_match
));
105 local
void check_match
OF((deflate_state
*s
, IPos start
, IPos match
,
109 /* ===========================================================================
114 /* Tail of hash chains */
117 # define TOO_FAR 4096
119 /* Matches of length 3 are discarded if their distance exceeds TOO_FAR */
121 /* Values for max_lazy_match, good_match and max_chain_length, depending on
122 * the desired pack level (0..9). The values given below have been tuned to
123 * exclude worst case performance for pathological files. Better values may be
124 * found for specific files.
126 typedef struct config_s
{
127 ush good_length
; /* reduce lazy search above this match length */
128 ush max_lazy
; /* do not perform lazy search above this match length */
129 ush nice_length
; /* quit search above this match length */
135 local
const config configuration_table
[2] = {
136 /* good lazy nice chain */
137 /* 0 */ {0, 0, 0, 0, deflate_stored
}, /* store only */
138 /* 1 */ {4, 4, 8, 4, deflate_fast
}}; /* max speed, no lazy matches */
140 local
const config configuration_table
[10] = {
141 /* good lazy nice chain */
142 /* 0 */ {0, 0, 0, 0, deflate_stored
}, /* store only */
143 /* 1 */ {4, 4, 8, 4, deflate_fast
}, /* max speed, no lazy matches */
144 /* 2 */ {4, 5, 16, 8, deflate_fast
},
145 /* 3 */ {4, 6, 32, 32, deflate_fast
},
147 /* 4 */ {4, 4, 16, 16, deflate_slow
}, /* lazy matches */
148 /* 5 */ {8, 16, 32, 32, deflate_slow
},
149 /* 6 */ {8, 16, 128, 128, deflate_slow
},
150 /* 7 */ {8, 32, 128, 256, deflate_slow
},
151 /* 8 */ {32, 128, 258, 1024, deflate_slow
},
152 /* 9 */ {32, 258, 258, 4096, deflate_slow
}}; /* max compression */
155 /* Note: the deflate() code requires max_lazy >= MIN_MATCH and max_chain >= 4
156 * For deflate_fast() (levels <= 3) good is ignored and lazy has a different
161 /* result of memcmp for equal strings */
163 #ifndef NO_DUMMY_DECL
164 struct static_tree_desc_s
{int dummy
;}; /* for buggy compilers */
167 /* rank Z_BLOCK between Z_NO_FLUSH and Z_PARTIAL_FLUSH */
168 #define RANK(f) (((f) << 1) - ((f) > 4 ? 9 : 0))
170 /* ===========================================================================
171 * Update a hash value with the given input byte
172 * IN assertion: all calls to to UPDATE_HASH are made with consecutive
173 * input characters, so that a running hash key can be computed from the
174 * previous key instead of complete recalculation each time.
176 #define UPDATE_HASH(s,h,c) (h = (((h)<<s->hash_shift) ^ (c)) & s->hash_mask)
179 /* ===========================================================================
180 * Insert string str in the dictionary and set match_head to the previous head
181 * of the hash chain (the most recent string with same hash key). Return
182 * the previous length of the hash chain.
183 * If this file is compiled with -DFASTEST, the compression level is forced
184 * to 1, and no hash chains are maintained.
185 * IN assertion: all calls to to INSERT_STRING are made with consecutive
186 * input characters and the first MIN_MATCH bytes of str are valid
187 * (except for the last MIN_MATCH-1 bytes of the input file).
190 #define INSERT_STRING(s, str, match_head) \
191 (UPDATE_HASH(s, s->ins_h, s->window[(str) + (MIN_MATCH-1)]), \
192 match_head = s->head[s->ins_h], \
193 s->head[s->ins_h] = (Pos)(str))
195 #define INSERT_STRING(s, str, match_head) \
196 (UPDATE_HASH(s, s->ins_h, s->window[(str) + (MIN_MATCH-1)]), \
197 match_head = s->prev[(str) & s->w_mask] = s->head[s->ins_h], \
198 s->head[s->ins_h] = (Pos)(str))
201 /* ===========================================================================
202 * Initialize the hash table (avoiding 64K overflow for 16 bit systems).
203 * prev[] will be initialized on the fly.
205 #define CLEAR_HASH(s) \
206 s->head[s->hash_size-1] = NIL; \
207 zmemzero((Bytef *)s->head, (unsigned)(s->hash_size-1)*sizeof(*s->head));
209 /* ========================================================================= */
210 int ZEXPORT
deflateInit_(strm
, level
, version
, stream_size
)
216 return deflateInit2_(strm
, level
, Z_DEFLATED
, MAX_WBITS
, DEF_MEM_LEVEL
,
217 Z_DEFAULT_STRATEGY
, version
, stream_size
);
218 /* To do: ignore strm->next_in if we use it as window */
221 /* ========================================================================= */
222 int ZEXPORT
deflateInit2_(strm
, level
, method
, windowBits
, memLevel
, strategy
,
223 version
, stream_size
)
235 static const char my_version
[] = ZLIB_VERSION
;
238 /* We overlay pending_buf and d_buf+l_buf. This works since the average
239 * output size for (length,distance) codes is <= 24 bits.
242 if (version
== Z_NULL
|| version
[0] != my_version
[0] ||
243 stream_size
!= sizeof(z_stream
)) {
244 return Z_VERSION_ERROR
;
246 if (strm
== Z_NULL
) return Z_STREAM_ERROR
;
249 if (strm
->zalloc
== (alloc_func
)0) {
251 return Z_STREAM_ERROR
;
253 strm
->zalloc
= zcalloc
;
254 strm
->opaque
= (voidpf
)0;
257 if (strm
->zfree
== (free_func
)0)
259 return Z_STREAM_ERROR
;
261 strm
->zfree
= zcfree
;
265 if (level
!= 0) level
= 1;
267 if (level
== Z_DEFAULT_COMPRESSION
) level
= 6;
270 if (windowBits
< 0) { /* suppress zlib wrapper */
272 windowBits
= -windowBits
;
275 else if (windowBits
> 15) {
276 wrap
= 2; /* write gzip wrapper instead */
280 if (memLevel
< 1 || memLevel
> MAX_MEM_LEVEL
|| method
!= Z_DEFLATED
||
281 windowBits
< 8 || windowBits
> 15 || level
< 0 || level
> 9 ||
282 strategy
< 0 || strategy
> Z_FIXED
) {
283 return Z_STREAM_ERROR
;
285 if (windowBits
== 8) windowBits
= 9; /* until 256-byte window bug fixed */
286 s
= (deflate_state
*) ZALLOC(strm
, 1, sizeof(deflate_state
));
287 if (s
== Z_NULL
) return Z_MEM_ERROR
;
288 strm
->state
= (struct internal_state FAR
*)s
;
293 s
->w_bits
= windowBits
;
294 s
->w_size
= 1 << s
->w_bits
;
295 s
->w_mask
= s
->w_size
- 1;
297 s
->hash_bits
= memLevel
+ 7;
298 s
->hash_size
= 1 << s
->hash_bits
;
299 s
->hash_mask
= s
->hash_size
- 1;
300 s
->hash_shift
= ((s
->hash_bits
+MIN_MATCH
-1)/MIN_MATCH
);
302 s
->window
= (Bytef
*) ZALLOC(strm
, s
->w_size
, 2*sizeof(Byte
));
303 s
->prev
= (Posf
*) ZALLOC(strm
, s
->w_size
, sizeof(Pos
));
304 s
->head
= (Posf
*) ZALLOC(strm
, s
->hash_size
, sizeof(Pos
));
306 s
->high_water
= 0; /* nothing written to s->window yet */
308 s
->lit_bufsize
= 1 << (memLevel
+ 6); /* 16K elements by default */
310 overlay
= (ushf
*) ZALLOC(strm
, s
->lit_bufsize
, sizeof(ush
)+2);
311 s
->pending_buf
= (uchf
*) overlay
;
312 s
->pending_buf_size
= (ulg
)s
->lit_bufsize
* (sizeof(ush
)+2L);
314 if (s
->window
== Z_NULL
|| s
->prev
== Z_NULL
|| s
->head
== Z_NULL
||
315 s
->pending_buf
== Z_NULL
) {
316 s
->status
= FINISH_STATE
;
317 strm
->msg
= ERR_MSG(Z_MEM_ERROR
);
321 s
->d_buf
= overlay
+ s
->lit_bufsize
/sizeof(ush
);
322 s
->l_buf
= s
->pending_buf
+ (1+sizeof(ush
))*s
->lit_bufsize
;
325 s
->strategy
= strategy
;
326 s
->method
= (Byte
)method
;
328 return deflateReset(strm
);
331 /* ========================================================================= */
332 int ZEXPORT
deflateSetDictionary (strm
, dictionary
, dictLength
)
334 const Bytef
*dictionary
;
341 z_const
unsigned char *next
;
343 if (strm
== Z_NULL
|| strm
->state
== Z_NULL
|| dictionary
== Z_NULL
)
344 return Z_STREAM_ERROR
;
347 if (wrap
== 2 || (wrap
== 1 && s
->status
!= INIT_STATE
) || s
->lookahead
)
348 return Z_STREAM_ERROR
;
350 /* when using zlib wrappers, compute Adler-32 for provided dictionary */
352 strm
->adler
= adler32(strm
->adler
, dictionary
, dictLength
);
353 s
->wrap
= 0; /* avoid computing Adler-32 in read_buf */
355 /* if dictionary would fill window, just replace the history */
356 if (dictLength
>= s
->w_size
) {
357 if (wrap
== 0) { /* already empty otherwise */
363 dictionary
+= dictLength
- s
->w_size
; /* use the tail */
364 dictLength
= s
->w_size
;
367 /* insert dictionary into window and hash */
368 avail
= strm
->avail_in
;
369 next
= strm
->next_in
;
370 strm
->avail_in
= dictLength
;
371 strm
->next_in
= (z_const Bytef
*)dictionary
;
373 while (s
->lookahead
>= MIN_MATCH
) {
375 n
= s
->lookahead
- (MIN_MATCH
-1);
377 UPDATE_HASH(s
, s
->ins_h
, s
->window
[str
+ MIN_MATCH
-1]);
379 s
->prev
[str
& s
->w_mask
] = s
->head
[s
->ins_h
];
381 s
->head
[s
->ins_h
] = (Pos
)str
;
385 s
->lookahead
= MIN_MATCH
-1;
388 s
->strstart
+= s
->lookahead
;
389 s
->block_start
= (long)s
->strstart
;
390 s
->insert
= s
->lookahead
;
392 s
->match_length
= s
->prev_length
= MIN_MATCH
-1;
393 s
->match_available
= 0;
394 strm
->next_in
= next
;
395 strm
->avail_in
= avail
;
400 /* ========================================================================= */
401 int ZEXPORT
deflateResetKeep (strm
)
406 if (strm
== Z_NULL
|| strm
->state
== Z_NULL
||
407 strm
->zalloc
== (alloc_func
)0 || strm
->zfree
== (free_func
)0) {
408 return Z_STREAM_ERROR
;
411 strm
->total_in
= strm
->total_out
= 0;
412 strm
->msg
= Z_NULL
; /* use zfree if we ever allocate msg dynamically */
413 strm
->data_type
= Z_UNKNOWN
;
415 s
= (deflate_state
*)strm
->state
;
417 s
->pending_out
= s
->pending_buf
;
420 s
->wrap
= -s
->wrap
; /* was made negative by deflate(..., Z_FINISH); */
422 s
->status
= s
->wrap
? INIT_STATE
: BUSY_STATE
;
425 s
->wrap
== 2 ? crc32(0L, Z_NULL
, 0) :
427 adler32(0L, Z_NULL
, 0);
428 s
->last_flush
= Z_NO_FLUSH
;
435 /* ========================================================================= */
436 int ZEXPORT
deflateReset (strm
)
441 ret
= deflateResetKeep(strm
);
443 lm_init(strm
->state
);
447 /* ========================================================================= */
448 int ZEXPORT
deflateSetHeader (strm
, head
)
452 if (strm
== Z_NULL
|| strm
->state
== Z_NULL
) return Z_STREAM_ERROR
;
453 if (strm
->state
->wrap
!= 2) return Z_STREAM_ERROR
;
454 strm
->state
->gzhead
= head
;
458 /* ========================================================================= */
459 int ZEXPORT
deflatePending (strm
, pending
, bits
)
464 if (strm
== Z_NULL
|| strm
->state
== Z_NULL
) return Z_STREAM_ERROR
;
465 if (pending
!= Z_NULL
)
466 *pending
= strm
->state
->pending
;
468 *bits
= strm
->state
->bi_valid
;
472 /* ========================================================================= */
473 int ZEXPORT
deflatePrime (strm
, bits
, value
)
481 if (strm
== Z_NULL
|| strm
->state
== Z_NULL
) return Z_STREAM_ERROR
;
483 if ((Bytef
*)(s
->d_buf
) < s
->pending_out
+ ((Buf_size
+ 7) >> 3))
486 put
= Buf_size
- s
->bi_valid
;
489 s
->bi_buf
|= (ush
)((value
& ((1 << put
) - 1)) << s
->bi_valid
);
498 /* ========================================================================= */
499 int ZEXPORT
deflateParams(strm
, level
, strategy
)
508 if (strm
== Z_NULL
|| strm
->state
== Z_NULL
) return Z_STREAM_ERROR
;
512 if (level
!= 0) level
= 1;
514 if (level
== Z_DEFAULT_COMPRESSION
) level
= 6;
516 if (level
< 0 || level
> 9 || strategy
< 0 || strategy
> Z_FIXED
) {
517 return Z_STREAM_ERROR
;
519 func
= configuration_table
[s
->level
].func
;
521 if ((strategy
!= s
->strategy
|| func
!= configuration_table
[level
].func
) &&
522 strm
->total_in
!= 0) {
523 /* Flush the last buffer: */
524 err
= deflate(strm
, Z_BLOCK
);
525 if (err
== Z_BUF_ERROR
&& s
->pending
== 0)
528 if (s
->level
!= level
) {
530 s
->max_lazy_match
= configuration_table
[level
].max_lazy
;
531 s
->good_match
= configuration_table
[level
].good_length
;
532 s
->nice_match
= configuration_table
[level
].nice_length
;
533 s
->max_chain_length
= configuration_table
[level
].max_chain
;
535 s
->strategy
= strategy
;
539 /* ========================================================================= */
540 int ZEXPORT
deflateTune(strm
, good_length
, max_lazy
, nice_length
, max_chain
)
549 if (strm
== Z_NULL
|| strm
->state
== Z_NULL
) return Z_STREAM_ERROR
;
551 s
->good_match
= good_length
;
552 s
->max_lazy_match
= max_lazy
;
553 s
->nice_match
= nice_length
;
554 s
->max_chain_length
= max_chain
;
558 /* =========================================================================
559 * For the default windowBits of 15 and memLevel of 8, this function returns
560 * a close to exact, as well as small, upper bound on the compressed size.
561 * They are coded as constants here for a reason--if the #define's are
562 * changed, then this function needs to be changed as well. The return
563 * value for 15 and 8 only works for those exact settings.
565 * For any setting other than those defaults for windowBits and memLevel,
566 * the value returned is a conservative worst case for the maximum expansion
567 * resulting from using fixed blocks instead of stored blocks, which deflate
568 * can emit on compressed data for some combinations of the parameters.
570 * This function could be more sophisticated to provide closer upper bounds for
571 * every combination of windowBits and memLevel. But even the conservative
572 * upper bound of about 14% expansion does not seem onerous for output buffer
575 uLong ZEXPORT
deflateBound(strm
, sourceLen
)
580 uLong complen
, wraplen
;
583 /* conservative upper bound for compressed data */
584 complen
= sourceLen
+
585 ((sourceLen
+ 7) >> 3) + ((sourceLen
+ 63) >> 6) + 5;
587 /* if can't get parameters, return conservative bound plus zlib wrapper */
588 if (strm
== Z_NULL
|| strm
->state
== Z_NULL
)
591 /* compute wrapper length */
594 case 0: /* raw deflate */
597 case 1: /* zlib wrapper */
598 wraplen
= 6 + (s
->strstart
? 4 : 0);
600 case 2: /* gzip wrapper */
602 if (s
->gzhead
!= Z_NULL
) { /* user-supplied gzip header */
603 if (s
->gzhead
->extra
!= Z_NULL
)
604 wraplen
+= 2 + s
->gzhead
->extra_len
;
605 str
= s
->gzhead
->name
;
610 str
= s
->gzhead
->comment
;
619 default: /* for compiler happiness */
623 /* if not default parameters, return conservative bound */
624 if (s
->w_bits
!= 15 || s
->hash_bits
!= 8 + 7)
625 return complen
+ wraplen
;
627 /* default settings: return tight bound for that case */
628 return sourceLen
+ (sourceLen
>> 12) + (sourceLen
>> 14) +
629 (sourceLen
>> 25) + 13 - 6 + wraplen
;
632 /* =========================================================================
633 * Put a short in the pending buffer. The 16-bit value is put in MSB order.
634 * IN assertion: the stream state is correct and there is enough room in
637 local
void putShortMSB (s
, b
)
641 put_byte(s
, (Byte
)(b
>> 8));
642 put_byte(s
, (Byte
)(b
& 0xff));
645 /* =========================================================================
646 * Flush as much pending output as possible. All deflate() output goes
647 * through this function so some applications may wish to modify it
648 * to avoid allocating a large strm->next_out buffer and copying into it.
649 * (See also read_buf()).
651 local
void flush_pending(strm
)
655 deflate_state
*s
= strm
->state
;
659 if (len
> strm
->avail_out
) len
= strm
->avail_out
;
660 if (len
== 0) return;
662 zmemcpy(strm
->next_out
, s
->pending_out
, len
);
663 strm
->next_out
+= len
;
664 s
->pending_out
+= len
;
665 strm
->total_out
+= len
;
666 strm
->avail_out
-= len
;
668 if (s
->pending
== 0) {
669 s
->pending_out
= s
->pending_buf
;
673 /* ========================================================================= */
674 int ZEXPORT
deflate (strm
, flush
)
678 int old_flush
; /* value of flush param for previous deflate call */
681 if (strm
== Z_NULL
|| strm
->state
== Z_NULL
||
682 flush
> Z_BLOCK
|| flush
< 0) {
683 return Z_STREAM_ERROR
;
687 if (strm
->next_out
== Z_NULL
||
688 (strm
->next_in
== Z_NULL
&& strm
->avail_in
!= 0) ||
689 (s
->status
== FINISH_STATE
&& flush
!= Z_FINISH
)) {
690 ERR_RETURN(strm
, Z_STREAM_ERROR
);
692 if (strm
->avail_out
== 0) ERR_RETURN(strm
, Z_BUF_ERROR
);
694 s
->strm
= strm
; /* just in case */
695 old_flush
= s
->last_flush
;
696 s
->last_flush
= flush
;
698 /* Write the header */
699 if (s
->status
== INIT_STATE
) {
702 strm
->adler
= crc32(0L, Z_NULL
, 0);
706 if (s
->gzhead
== Z_NULL
) {
712 put_byte(s
, s
->level
== 9 ? 2 :
713 (s
->strategy
>= Z_HUFFMAN_ONLY
|| s
->level
< 2 ?
715 put_byte(s
, OS_CODE
);
716 s
->status
= BUSY_STATE
;
719 put_byte(s
, (s
->gzhead
->text
? 1 : 0) +
720 (s
->gzhead
->hcrc
? 2 : 0) +
721 (s
->gzhead
->extra
== Z_NULL
? 0 : 4) +
722 (s
->gzhead
->name
== Z_NULL
? 0 : 8) +
723 (s
->gzhead
->comment
== Z_NULL
? 0 : 16)
725 put_byte(s
, (Byte
)(s
->gzhead
->time
& 0xff));
726 put_byte(s
, (Byte
)((s
->gzhead
->time
>> 8) & 0xff));
727 put_byte(s
, (Byte
)((s
->gzhead
->time
>> 16) & 0xff));
728 put_byte(s
, (Byte
)((s
->gzhead
->time
>> 24) & 0xff));
729 put_byte(s
, s
->level
== 9 ? 2 :
730 (s
->strategy
>= Z_HUFFMAN_ONLY
|| s
->level
< 2 ?
732 put_byte(s
, s
->gzhead
->os
& 0xff);
733 if (s
->gzhead
->extra
!= Z_NULL
) {
734 put_byte(s
, s
->gzhead
->extra_len
& 0xff);
735 put_byte(s
, (s
->gzhead
->extra_len
>> 8) & 0xff);
738 strm
->adler
= crc32(strm
->adler
, s
->pending_buf
,
741 s
->status
= EXTRA_STATE
;
747 uInt header
= (Z_DEFLATED
+ ((s
->w_bits
-8)<<4)) << 8;
750 if (s
->strategy
>= Z_HUFFMAN_ONLY
|| s
->level
< 2)
752 else if (s
->level
< 6)
754 else if (s
->level
== 6)
758 header
|= (level_flags
<< 6);
759 if (s
->strstart
!= 0) header
|= PRESET_DICT
;
760 header
+= 31 - (header
% 31);
762 s
->status
= BUSY_STATE
;
763 putShortMSB(s
, header
);
765 /* Save the adler32 of the preset dictionary: */
766 if (s
->strstart
!= 0) {
767 putShortMSB(s
, (uInt
)(strm
->adler
>> 16));
768 putShortMSB(s
, (uInt
)(strm
->adler
& 0xffff));
770 strm
->adler
= adler32(0L, Z_NULL
, 0);
774 if (s
->status
== EXTRA_STATE
) {
775 if (s
->gzhead
->extra
!= Z_NULL
) {
776 uInt beg
= s
->pending
; /* start of bytes to update crc */
778 while (s
->gzindex
< (s
->gzhead
->extra_len
& 0xffff)) {
779 if (s
->pending
== s
->pending_buf_size
) {
780 if (s
->gzhead
->hcrc
&& s
->pending
> beg
)
781 strm
->adler
= crc32(strm
->adler
, s
->pending_buf
+ beg
,
785 if (s
->pending
== s
->pending_buf_size
)
788 put_byte(s
, s
->gzhead
->extra
[s
->gzindex
]);
791 if (s
->gzhead
->hcrc
&& s
->pending
> beg
)
792 strm
->adler
= crc32(strm
->adler
, s
->pending_buf
+ beg
,
794 if (s
->gzindex
== s
->gzhead
->extra_len
) {
796 s
->status
= NAME_STATE
;
800 s
->status
= NAME_STATE
;
802 if (s
->status
== NAME_STATE
) {
803 if (s
->gzhead
->name
!= Z_NULL
) {
804 uInt beg
= s
->pending
; /* start of bytes to update crc */
808 if (s
->pending
== s
->pending_buf_size
) {
809 if (s
->gzhead
->hcrc
&& s
->pending
> beg
)
810 strm
->adler
= crc32(strm
->adler
, s
->pending_buf
+ beg
,
814 if (s
->pending
== s
->pending_buf_size
) {
819 val
= s
->gzhead
->name
[s
->gzindex
++];
822 if (s
->gzhead
->hcrc
&& s
->pending
> beg
)
823 strm
->adler
= crc32(strm
->adler
, s
->pending_buf
+ beg
,
827 s
->status
= COMMENT_STATE
;
831 s
->status
= COMMENT_STATE
;
833 if (s
->status
== COMMENT_STATE
) {
834 if (s
->gzhead
->comment
!= Z_NULL
) {
835 uInt beg
= s
->pending
; /* start of bytes to update crc */
839 if (s
->pending
== s
->pending_buf_size
) {
840 if (s
->gzhead
->hcrc
&& s
->pending
> beg
)
841 strm
->adler
= crc32(strm
->adler
, s
->pending_buf
+ beg
,
845 if (s
->pending
== s
->pending_buf_size
) {
850 val
= s
->gzhead
->comment
[s
->gzindex
++];
853 if (s
->gzhead
->hcrc
&& s
->pending
> beg
)
854 strm
->adler
= crc32(strm
->adler
, s
->pending_buf
+ beg
,
857 s
->status
= HCRC_STATE
;
860 s
->status
= HCRC_STATE
;
862 if (s
->status
== HCRC_STATE
) {
863 if (s
->gzhead
->hcrc
) {
864 if (s
->pending
+ 2 > s
->pending_buf_size
)
866 if (s
->pending
+ 2 <= s
->pending_buf_size
) {
867 put_byte(s
, (Byte
)(strm
->adler
& 0xff));
868 put_byte(s
, (Byte
)((strm
->adler
>> 8) & 0xff));
869 strm
->adler
= crc32(0L, Z_NULL
, 0);
870 s
->status
= BUSY_STATE
;
874 s
->status
= BUSY_STATE
;
878 /* Flush as much pending output as possible */
879 if (s
->pending
!= 0) {
881 if (strm
->avail_out
== 0) {
882 /* Since avail_out is 0, deflate will be called again with
883 * more output space, but possibly with both pending and
884 * avail_in equal to zero. There won't be anything to do,
885 * but this is not an error situation so make sure we
886 * return OK instead of BUF_ERROR at next call of deflate:
892 /* Make sure there is something to do and avoid duplicate consecutive
893 * flushes. For repeated and useless calls with Z_FINISH, we keep
894 * returning Z_STREAM_END instead of Z_BUF_ERROR.
896 } else if (strm
->avail_in
== 0 && RANK(flush
) <= RANK(old_flush
) &&
898 ERR_RETURN(strm
, Z_BUF_ERROR
);
901 /* User must not provide more input after the first FINISH: */
902 if (s
->status
== FINISH_STATE
&& strm
->avail_in
!= 0) {
903 ERR_RETURN(strm
, Z_BUF_ERROR
);
906 /* Start a new block or continue the current one.
908 if (strm
->avail_in
!= 0 || s
->lookahead
!= 0 ||
909 (flush
!= Z_NO_FLUSH
&& s
->status
!= FINISH_STATE
)) {
912 bstate
= s
->strategy
== Z_HUFFMAN_ONLY
? deflate_huff(s
, flush
) :
913 (s
->strategy
== Z_RLE
? deflate_rle(s
, flush
) :
914 (*(configuration_table
[s
->level
].func
))(s
, flush
));
916 if (bstate
== finish_started
|| bstate
== finish_done
) {
917 s
->status
= FINISH_STATE
;
919 if (bstate
== need_more
|| bstate
== finish_started
) {
920 if (strm
->avail_out
== 0) {
921 s
->last_flush
= -1; /* avoid BUF_ERROR next call, see above */
924 /* If flush != Z_NO_FLUSH && avail_out == 0, the next call
925 * of deflate should use the same flush parameter to make sure
926 * that the flush is complete. So we don't have to output an
927 * empty block here, this will be done at next call. This also
928 * ensures that for a very small output buffer, we emit at most
932 if (bstate
== block_done
) {
933 if (flush
== Z_PARTIAL_FLUSH
) {
935 } else if (flush
!= Z_BLOCK
) { /* FULL_FLUSH or SYNC_FLUSH */
936 _tr_stored_block(s
, (char*)0, 0L, 0);
937 /* For a full flush, this empty block will be recognized
938 * as a special marker by inflate_sync().
940 if (flush
== Z_FULL_FLUSH
) {
941 CLEAR_HASH(s
); /* forget history */
942 if (s
->lookahead
== 0) {
950 if (strm
->avail_out
== 0) {
951 s
->last_flush
= -1; /* avoid BUF_ERROR at next call, see above */
956 Assert(strm
->avail_out
> 0, "bug2");
958 if (flush
!= Z_FINISH
) return Z_OK
;
959 if (s
->wrap
<= 0) return Z_STREAM_END
;
961 /* Write the trailer */
964 put_byte(s
, (Byte
)(strm
->adler
& 0xff));
965 put_byte(s
, (Byte
)((strm
->adler
>> 8) & 0xff));
966 put_byte(s
, (Byte
)((strm
->adler
>> 16) & 0xff));
967 put_byte(s
, (Byte
)((strm
->adler
>> 24) & 0xff));
968 put_byte(s
, (Byte
)(strm
->total_in
& 0xff));
969 put_byte(s
, (Byte
)((strm
->total_in
>> 8) & 0xff));
970 put_byte(s
, (Byte
)((strm
->total_in
>> 16) & 0xff));
971 put_byte(s
, (Byte
)((strm
->total_in
>> 24) & 0xff));
976 putShortMSB(s
, (uInt
)(strm
->adler
>> 16));
977 putShortMSB(s
, (uInt
)(strm
->adler
& 0xffff));
980 /* If avail_out is zero, the application will call deflate again
983 if (s
->wrap
> 0) s
->wrap
= -s
->wrap
; /* write the trailer only once! */
984 return s
->pending
!= 0 ? Z_OK
: Z_STREAM_END
;
987 /* ========================================================================= */
988 int ZEXPORT
deflateEnd (strm
)
993 if (strm
== Z_NULL
|| strm
->state
== Z_NULL
) return Z_STREAM_ERROR
;
995 status
= strm
->state
->status
;
996 if (status
!= INIT_STATE
&&
997 status
!= EXTRA_STATE
&&
998 status
!= NAME_STATE
&&
999 status
!= COMMENT_STATE
&&
1000 status
!= HCRC_STATE
&&
1001 status
!= BUSY_STATE
&&
1002 status
!= FINISH_STATE
) {
1003 return Z_STREAM_ERROR
;
1006 /* Deallocate in reverse order of allocations: */
1007 TRY_FREE(strm
, strm
->state
->pending_buf
);
1008 TRY_FREE(strm
, strm
->state
->head
);
1009 TRY_FREE(strm
, strm
->state
->prev
);
1010 TRY_FREE(strm
, strm
->state
->window
);
1012 ZFREE(strm
, strm
->state
);
1013 strm
->state
= Z_NULL
;
1015 return status
== BUSY_STATE
? Z_DATA_ERROR
: Z_OK
;
1018 /* =========================================================================
1019 * Copy the source state to the destination state.
1020 * To simplify the source, this is not supported for 16-bit MSDOS (which
1021 * doesn't have enough memory anyway to duplicate compression states).
1023 int ZEXPORT
deflateCopy (dest
, source
)
1028 return Z_STREAM_ERROR
;
1035 if (source
== Z_NULL
|| dest
== Z_NULL
|| source
->state
== Z_NULL
) {
1036 return Z_STREAM_ERROR
;
1041 zmemcpy((voidpf
)dest
, (voidpf
)source
, sizeof(z_stream
));
1043 ds
= (deflate_state
*) ZALLOC(dest
, 1, sizeof(deflate_state
));
1044 if (ds
== Z_NULL
) return Z_MEM_ERROR
;
1045 dest
->state
= (struct internal_state FAR
*) ds
;
1046 zmemcpy((voidpf
)ds
, (voidpf
)ss
, sizeof(deflate_state
));
1049 ds
->window
= (Bytef
*) ZALLOC(dest
, ds
->w_size
, 2*sizeof(Byte
));
1050 ds
->prev
= (Posf
*) ZALLOC(dest
, ds
->w_size
, sizeof(Pos
));
1051 ds
->head
= (Posf
*) ZALLOC(dest
, ds
->hash_size
, sizeof(Pos
));
1052 overlay
= (ushf
*) ZALLOC(dest
, ds
->lit_bufsize
, sizeof(ush
)+2);
1053 ds
->pending_buf
= (uchf
*) overlay
;
1055 if (ds
->window
== Z_NULL
|| ds
->prev
== Z_NULL
|| ds
->head
== Z_NULL
||
1056 ds
->pending_buf
== Z_NULL
) {
1060 /* following zmemcpy do not work for 16-bit MSDOS */
1061 zmemcpy(ds
->window
, ss
->window
, ds
->w_size
* 2 * sizeof(Byte
));
1062 zmemcpy((voidpf
)ds
->prev
, (voidpf
)ss
->prev
, ds
->w_size
* sizeof(Pos
));
1063 zmemcpy((voidpf
)ds
->head
, (voidpf
)ss
->head
, ds
->hash_size
* sizeof(Pos
));
1064 zmemcpy(ds
->pending_buf
, ss
->pending_buf
, (uInt
)ds
->pending_buf_size
);
1066 ds
->pending_out
= ds
->pending_buf
+ (ss
->pending_out
- ss
->pending_buf
);
1067 ds
->d_buf
= overlay
+ ds
->lit_bufsize
/sizeof(ush
);
1068 ds
->l_buf
= ds
->pending_buf
+ (1+sizeof(ush
))*ds
->lit_bufsize
;
1070 ds
->l_desc
.dyn_tree
= ds
->dyn_ltree
;
1071 ds
->d_desc
.dyn_tree
= ds
->dyn_dtree
;
1072 ds
->bl_desc
.dyn_tree
= ds
->bl_tree
;
1075 #endif /* MAXSEG_64K */
1078 /* ===========================================================================
1079 * Read a new buffer from the current input stream, update the adler32
1080 * and total number of bytes read. All deflate() input goes through
1081 * this function so some applications may wish to modify it to avoid
1082 * allocating a large strm->next_in buffer and copying from it.
1083 * (See also flush_pending()).
1085 local
int read_buf(strm
, buf
, size
)
1090 unsigned len
= strm
->avail_in
;
1092 if (len
> size
) len
= size
;
1093 if (len
== 0) return 0;
1095 strm
->avail_in
-= len
;
1097 zmemcpy(buf
, strm
->next_in
, len
);
1098 if (strm
->state
->wrap
== 1) {
1099 strm
->adler
= adler32(strm
->adler
, buf
, len
);
1102 else if (strm
->state
->wrap
== 2) {
1103 strm
->adler
= crc32(strm
->adler
, buf
, len
);
1106 strm
->next_in
+= len
;
1107 strm
->total_in
+= len
;
1112 /* ===========================================================================
1113 * Initialize the "longest match" routines for a new zlib stream
1115 local
void lm_init (s
)
1118 s
->window_size
= (ulg
)2L*s
->w_size
;
1122 /* Set the default configuration parameters:
1124 s
->max_lazy_match
= configuration_table
[s
->level
].max_lazy
;
1125 s
->good_match
= configuration_table
[s
->level
].good_length
;
1126 s
->nice_match
= configuration_table
[s
->level
].nice_length
;
1127 s
->max_chain_length
= configuration_table
[s
->level
].max_chain
;
1130 s
->block_start
= 0L;
1133 s
->match_length
= s
->prev_length
= MIN_MATCH
-1;
1134 s
->match_available
= 0;
1138 match_init(); /* initialize the asm code */
1144 /* ===========================================================================
1145 * Set match_start to the longest match starting at the given string and
1146 * return its length. Matches shorter or equal to prev_length are discarded,
1147 * in which case the result is equal to prev_length and match_start is
1149 * IN assertions: cur_match is the head of the hash chain for the current
1150 * string (strstart) and its distance is <= MAX_DIST, and prev_length >= 1
1151 * OUT assertion: the match length is not greater than s->lookahead.
1154 /* For 80x86 and 680x0, an optimized version will be provided in match.asm or
1155 * match.S. The code will be functionally equivalent.
1157 local uInt
longest_match(s
, cur_match
)
1159 IPos cur_match
; /* current match */
1161 unsigned chain_length
= s
->max_chain_length
;/* max hash chain length */
1162 register Bytef
*scan
= s
->window
+ s
->strstart
; /* current string */
1163 register Bytef
*match
; /* matched string */
1164 register int len
; /* length of current match */
1165 #ifdef ZLIB_PM3_TUNED
1166 int best_len
= MIN_MATCH
-1; // lift the restriction on prev-length
1168 int best_len
= s
->prev_length
; /* best match length so far */
1170 int nice_match
= s
->nice_match
; /* stop if match long enough */
1171 IPos limit
= s
->strstart
> (IPos
)MAX_DIST(s
) ?
1172 s
->strstart
- (IPos
)MAX_DIST(s
) : NIL
;
1173 /* Stop when cur_match becomes <= limit. To simplify the code,
1174 * we prevent matches with the string of window index 0.
1176 Posf
*prev
= s
->prev
;
1177 uInt wmask
= s
->w_mask
;
1180 /* Compare two bytes at a time. Note: this is not always beneficial.
1181 * Try with and without -DUNALIGNED_OK to check.
1183 register Bytef
*strend
= s
->window
+ s
->strstart
+ MAX_MATCH
- 1;
1184 register ush scan_start
= *(ushf
*)scan
;
1185 register ush scan_end
= *(ushf
*)(scan
+best_len
-1);
1187 register Bytef
*strend
= s
->window
+ s
->strstart
+ MAX_MATCH
;
1188 register Byte scan_end1
= scan
[best_len
-1];
1189 register Byte scan_end
= scan
[best_len
];
1192 /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16.
1193 * It is easy to get rid of this optimization if necessary.
1195 Assert(s
->hash_bits
>= 8 && MAX_MATCH
== 258, "Code too clever");
1197 /* Do not waste too much time if we already have a good match: */
1198 if (s
->prev_length
>= s
->good_match
) {
1201 /* Do not look for matches beyond the end of the input. This is necessary
1202 * to make deflate deterministic.
1204 if ((uInt
)nice_match
> s
->lookahead
) nice_match
= s
->lookahead
;
1206 Assert((ulg
)s
->strstart
<= s
->window_size
-MIN_LOOKAHEAD
, "need lookahead");
1209 Assert(cur_match
< s
->strstart
, "no future");
1210 match
= s
->window
+ cur_match
;
1212 /* Skip to next match if the match length cannot increase
1213 * or if the match length is less than 2. Note that the checks below
1214 * for insufficient lookahead only occur occasionally for performance
1215 * reasons. Therefore uninitialized memory will be accessed, and
1216 * conditional jumps will be made that depend on those values.
1217 * However the length of the match is limited to the lookahead, so
1218 * the output of deflate is not affected by the uninitialized values.
1220 #if (defined(UNALIGNED_OK) && MAX_MATCH == 258)
1221 /* This code assumes sizeof(unsigned short) == 2. Do not use
1222 * UNALIGNED_OK if your compiler uses a different size.
1224 if (*(ushf
*)(match
+best_len
-1) != scan_end
||
1225 *(ushf
*)match
!= scan_start
) continue;
1227 /* It is not necessary to compare scan[2] and match[2] since they are
1228 * always equal when the other bytes match, given that the hash keys
1229 * are equal and that HASH_BITS >= 8. Compare 2 bytes at a time at
1230 * strstart+3, +5, ... up to strstart+257. We check for insufficient
1231 * lookahead only every 4th comparison; the 128th check will be made
1232 * at strstart+257. If MAX_MATCH-2 is not a multiple of 8, it is
1233 * necessary to put more guard bytes at the end of the window, or
1234 * to check more often for insufficient lookahead.
1236 Assert(scan
[2] == match
[2], "scan[2]?");
1239 } while (*(ushf
*)(scan
+=2) == *(ushf
*)(match
+=2) &&
1240 *(ushf
*)(scan
+=2) == *(ushf
*)(match
+=2) &&
1241 *(ushf
*)(scan
+=2) == *(ushf
*)(match
+=2) &&
1242 *(ushf
*)(scan
+=2) == *(ushf
*)(match
+=2) &&
1244 /* The funny "do {}" generates better code on most compilers */
1246 /* Here, scan <= window+strstart+257 */
1247 Assert(scan
<= s
->window
+(unsigned)(s
->window_size
-1), "wild scan");
1248 if (*scan
== *match
) scan
++;
1250 len
= (MAX_MATCH
- 1) - (int)(strend
-scan
);
1251 scan
= strend
- (MAX_MATCH
-1);
1253 #else /* UNALIGNED_OK */
1255 if (match
[best_len
] != scan_end
||
1256 match
[best_len
-1] != scan_end1
||
1258 *++match
!= scan
[1]) continue;
1260 /* The check at best_len-1 can be removed because it will be made
1261 * again later. (This heuristic is not always a win.)
1262 * It is not necessary to compare scan[2] and match[2] since they
1263 * are always equal when the other bytes match, given that
1264 * the hash keys are equal and that HASH_BITS >= 8.
1267 Assert(*scan
== *match
, "match[2]?");
1269 /* We check for insufficient lookahead only every 8th comparison;
1270 * the 256th check will be made at strstart+258.
1273 } while (*++scan
== *++match
&& *++scan
== *++match
&&
1274 *++scan
== *++match
&& *++scan
== *++match
&&
1275 *++scan
== *++match
&& *++scan
== *++match
&&
1276 *++scan
== *++match
&& *++scan
== *++match
&&
1279 Assert(scan
<= s
->window
+(unsigned)(s
->window_size
-1), "wild scan");
1281 len
= MAX_MATCH
- (int)(strend
- scan
);
1282 scan
= strend
- MAX_MATCH
;
1284 #endif /* UNALIGNED_OK */
1286 if (len
> best_len
) {
1287 s
->match_start
= cur_match
;
1289 if (len
>= nice_match
) break;
1291 scan_end
= *(ushf
*)(scan
+best_len
-1);
1293 scan_end1
= scan
[best_len
-1];
1294 scan_end
= scan
[best_len
];
1297 } while ((cur_match
= prev
[cur_match
& wmask
]) > limit
1298 && --chain_length
!= 0);
1300 if ((uInt
)best_len
<= s
->lookahead
) return (uInt
)best_len
;
1301 return s
->lookahead
;
1307 /* ---------------------------------------------------------------------------
1308 * Optimized version for FASTEST only
1310 local uInt
longest_match(s
, cur_match
)
1312 IPos cur_match
; /* current match */
1314 register Bytef
*scan
= s
->window
+ s
->strstart
; /* current string */
1315 register Bytef
*match
; /* matched string */
1316 register int len
; /* length of current match */
1317 register Bytef
*strend
= s
->window
+ s
->strstart
+ MAX_MATCH
;
1319 /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16.
1320 * It is easy to get rid of this optimization if necessary.
1322 Assert(s
->hash_bits
>= 8 && MAX_MATCH
== 258, "Code too clever");
1324 Assert((ulg
)s
->strstart
<= s
->window_size
-MIN_LOOKAHEAD
, "need lookahead");
1326 Assert(cur_match
< s
->strstart
, "no future");
1328 match
= s
->window
+ cur_match
;
1330 /* Return failure if the match length is less than 2:
1332 if (match
[0] != scan
[0] || match
[1] != scan
[1]) return MIN_MATCH
-1;
1334 /* The check at best_len-1 can be removed because it will be made
1335 * again later. (This heuristic is not always a win.)
1336 * It is not necessary to compare scan[2] and match[2] since they
1337 * are always equal when the other bytes match, given that
1338 * the hash keys are equal and that HASH_BITS >= 8.
1340 scan
+= 2, match
+= 2;
1341 Assert(*scan
== *match
, "match[2]?");
1343 /* We check for insufficient lookahead only every 8th comparison;
1344 * the 256th check will be made at strstart+258.
1347 } while (*++scan
== *++match
&& *++scan
== *++match
&&
1348 *++scan
== *++match
&& *++scan
== *++match
&&
1349 *++scan
== *++match
&& *++scan
== *++match
&&
1350 *++scan
== *++match
&& *++scan
== *++match
&&
1353 Assert(scan
<= s
->window
+(unsigned)(s
->window_size
-1), "wild scan");
1355 len
= MAX_MATCH
- (int)(strend
- scan
);
1357 if (len
< MIN_MATCH
) return MIN_MATCH
- 1;
1359 s
->match_start
= cur_match
;
1360 return (uInt
)len
<= s
->lookahead
? (uInt
)len
: s
->lookahead
;
1363 #endif /* FASTEST */
1366 /* ===========================================================================
1367 * Check that the match at match_start is indeed a match.
1369 local
void check_match(s
, start
, match
, length
)
1374 /* check that the match is indeed a match */
1375 if (zmemcmp(s
->window
+ match
,
1376 s
->window
+ start
, length
) != EQUAL
) {
1377 fprintf(stderr
, " start %u, match %u, length %d\n",
1378 start
, match
, length
);
1380 fprintf(stderr
, "%c%c", s
->window
[match
++], s
->window
[start
++]);
1381 } while (--length
!= 0);
1382 z_error("invalid match");
1384 if (z_verbose
> 1) {
1385 fprintf(stderr
,"\\[%d,%d]", start
-match
, length
);
1386 do { putc(s
->window
[start
++], stderr
); } while (--length
!= 0);
1390 # define check_match(s, start, match, length)
1393 /* ===========================================================================
1394 * Fill the window when the lookahead becomes insufficient.
1395 * Updates strstart and lookahead.
1397 * IN assertion: lookahead < MIN_LOOKAHEAD
1398 * OUT assertions: strstart <= window_size-MIN_LOOKAHEAD
1399 * At least one byte has been read, or avail_in == 0; reads are
1400 * performed for at least two bytes (required for the zip translate_eol
1401 * option -- not supported here).
1403 local
void fill_window(s
)
1406 register unsigned n
, m
;
1408 unsigned more
; /* Amount of free space at the end of the window. */
1409 uInt wsize
= s
->w_size
;
1411 Assert(s
->lookahead
< MIN_LOOKAHEAD
, "already enough lookahead");
1414 more
= (unsigned)(s
->window_size
-(ulg
)s
->lookahead
-(ulg
)s
->strstart
);
1416 /* Deal with !@#$% 64K limit: */
1417 if (sizeof(int) <= 2) {
1418 if (more
== 0 && s
->strstart
== 0 && s
->lookahead
== 0) {
1421 } else if (more
== (unsigned)(-1)) {
1422 /* Very unlikely, but possible on 16 bit machine if
1423 * strstart == 0 && lookahead == 1 (input done a byte at time)
1429 /* If the window is almost full and there is insufficient lookahead,
1430 * move the upper half to the lower one to make room in the upper half.
1432 if (s
->strstart
>= wsize
+MAX_DIST(s
)) {
1434 zmemcpy(s
->window
, s
->window
+wsize
, (unsigned)wsize
);
1435 s
->match_start
-= wsize
;
1436 s
->strstart
-= wsize
; /* we now have strstart >= MAX_DIST */
1437 s
->block_start
-= (long) wsize
;
1439 /* Slide the hash table (could be avoided with 32 bit values
1440 at the expense of memory usage). We slide even when level == 0
1441 to keep the hash table consistent if we switch back to level > 0
1442 later. (Using level 0 permanently is not an optimal usage of
1443 zlib, so we don't care about this pathological case.)
1449 *p
= (Pos
)(m
>= wsize
? m
-wsize
: NIL
);
1457 *p
= (Pos
)(m
>= wsize
? m
-wsize
: NIL
);
1458 /* If n is not on any hash chain, prev[n] is garbage but
1459 * its value will never be used.
1465 if (s
->strm
->avail_in
== 0) break;
1467 /* If there was no sliding:
1468 * strstart <= WSIZE+MAX_DIST-1 && lookahead <= MIN_LOOKAHEAD - 1 &&
1469 * more == window_size - lookahead - strstart
1470 * => more >= window_size - (MIN_LOOKAHEAD-1 + WSIZE + MAX_DIST-1)
1471 * => more >= window_size - 2*WSIZE + 2
1472 * In the BIG_MEM or MMAP case (not yet supported),
1473 * window_size == input_size + MIN_LOOKAHEAD &&
1474 * strstart + s->lookahead <= input_size => more >= MIN_LOOKAHEAD.
1475 * Otherwise, window_size == 2*WSIZE so more >= 2.
1476 * If there was sliding, more >= WSIZE. So in all cases, more >= 2.
1478 Assert(more
>= 2, "more < 2");
1480 n
= read_buf(s
->strm
, s
->window
+ s
->strstart
+ s
->lookahead
, more
);
1483 /* Initialize the hash value now that we have some input: */
1484 if (s
->lookahead
+ s
->insert
>= MIN_MATCH
) {
1485 uInt str
= s
->strstart
- s
->insert
;
1486 s
->ins_h
= s
->window
[str
];
1487 UPDATE_HASH(s
, s
->ins_h
, s
->window
[str
+ 1]);
1489 Call
UPDATE_HASH() MIN_MATCH
-3 more times
1492 UPDATE_HASH(s
, s
->ins_h
, s
->window
[str
+ MIN_MATCH
-1]);
1494 s
->prev
[str
& s
->w_mask
] = s
->head
[s
->ins_h
];
1496 s
->head
[s
->ins_h
] = (Pos
)str
;
1499 if (s
->lookahead
+ s
->insert
< MIN_MATCH
)
1503 /* If the whole input has less than MIN_MATCH bytes, ins_h is garbage,
1504 * but this is not important since only literal bytes will be emitted.
1507 } while (s
->lookahead
< MIN_LOOKAHEAD
&& s
->strm
->avail_in
!= 0);
1509 /* If the WIN_INIT bytes after the end of the current data have never been
1510 * written, then zero those bytes in order to avoid memory check reports of
1511 * the use of uninitialized (or uninitialised as Julian writes) bytes by
1512 * the longest match routines. Update the high water mark for the next
1513 * time through here. WIN_INIT is set to MAX_MATCH since the longest match
1514 * routines allow scanning to strstart + MAX_MATCH, ignoring lookahead.
1516 if (s
->high_water
< s
->window_size
) {
1517 ulg curr
= s
->strstart
+ (ulg
)(s
->lookahead
);
1520 if (s
->high_water
< curr
) {
1521 /* Previous high water mark below current data -- zero WIN_INIT
1522 * bytes or up to end of window, whichever is less.
1524 init
= s
->window_size
- curr
;
1525 if (init
> WIN_INIT
)
1527 zmemzero(s
->window
+ curr
, (unsigned)init
);
1528 s
->high_water
= curr
+ init
;
1530 else if (s
->high_water
< (ulg
)curr
+ WIN_INIT
) {
1531 /* High water mark at or above current data, but below current data
1532 * plus WIN_INIT -- zero out to current data plus WIN_INIT, or up
1533 * to end of window, whichever is less.
1535 init
= (ulg
)curr
+ WIN_INIT
- s
->high_water
;
1536 if (init
> s
->window_size
- s
->high_water
)
1537 init
= s
->window_size
- s
->high_water
;
1538 zmemzero(s
->window
+ s
->high_water
, (unsigned)init
);
1539 s
->high_water
+= init
;
1543 Assert((ulg
)s
->strstart
<= s
->window_size
- MIN_LOOKAHEAD
,
1544 "not enough room for search");
1547 /* ===========================================================================
1548 * Flush the current block, with given end-of-file flag.
1549 * IN assertion: strstart is set to the end of the current match.
1551 #define FLUSH_BLOCK_ONLY(s, last) { \
1552 _tr_flush_block(s, (s->block_start >= 0L ? \
1553 (charf *)&s->window[(unsigned)s->block_start] : \
1555 (ulg)((long)s->strstart - s->block_start), \
1557 s->block_start = s->strstart; \
1558 flush_pending(s->strm); \
1559 Tracev((stderr,"[FLUSH]")); \
1562 /* Same but force premature exit if necessary. */
1563 #define FLUSH_BLOCK(s, last) { \
1564 FLUSH_BLOCK_ONLY(s, last); \
1565 if (s->strm->avail_out == 0) return (last) ? finish_started : need_more; \
1568 /* ===========================================================================
1569 * Copy without compression as much as possible from the input stream, return
1570 * the current block state.
1571 * This function does not insert new strings in the dictionary since
1572 * uncompressible data is probably not useful. This function is used
1573 * only for the level=0 compression option.
1574 * NOTE: this function should be optimized to avoid extra copying from
1575 * window to pending_buf.
1577 local block_state
deflate_stored(s
, flush
)
1581 /* Stored blocks are limited to 0xffff bytes, pending_buf is limited
1582 * to pending_buf_size, and each stored block has a 5 byte header:
1584 ulg max_block_size
= 0xffff;
1587 if (max_block_size
> s
->pending_buf_size
- 5) {
1588 max_block_size
= s
->pending_buf_size
- 5;
1591 /* Copy as much as possible from input to output: */
1593 /* Fill the window as much as possible: */
1594 if (s
->lookahead
<= 1) {
1596 Assert(s
->strstart
< s
->w_size
+MAX_DIST(s
) ||
1597 s
->block_start
>= (long)s
->w_size
, "slide too late");
1600 if (s
->lookahead
== 0 && flush
== Z_NO_FLUSH
) return need_more
;
1602 if (s
->lookahead
== 0) break; /* flush the current block */
1604 Assert(s
->block_start
>= 0L, "block gone");
1606 s
->strstart
+= s
->lookahead
;
1609 /* Emit a stored block if pending_buf will be full: */
1610 max_start
= s
->block_start
+ max_block_size
;
1611 if (s
->strstart
== 0 || (ulg
)s
->strstart
>= max_start
) {
1612 /* strstart == 0 is possible when wraparound on 16-bit machine */
1613 s
->lookahead
= (uInt
)(s
->strstart
- max_start
);
1614 s
->strstart
= (uInt
)max_start
;
1617 /* Flush if we may have to slide, otherwise block_start may become
1618 * negative and the data will be gone:
1620 if (s
->strstart
- (uInt
)s
->block_start
>= MAX_DIST(s
)) {
1625 if (flush
== Z_FINISH
) {
1629 if ((long)s
->strstart
> s
->block_start
)
1634 /* ===========================================================================
1635 * Compress as much as possible from the input stream, return the current
1637 * This function does not perform lazy evaluation of matches and inserts
1638 * new strings in the dictionary only for unmatched strings or for short
1639 * matches. It is used only for the fast compression options.
1641 local block_state
deflate_fast(s
, flush
)
1645 IPos hash_head
; /* head of the hash chain */
1646 int bflush
; /* set if current block must be flushed */
1649 /* Make sure that we always have enough lookahead, except
1650 * at the end of the input file. We need MAX_MATCH bytes
1651 * for the next match, plus MIN_MATCH bytes to insert the
1652 * string following the next match.
1654 if (s
->lookahead
< MIN_LOOKAHEAD
) {
1656 if (s
->lookahead
< MIN_LOOKAHEAD
&& flush
== Z_NO_FLUSH
) {
1659 if (s
->lookahead
== 0) break; /* flush the current block */
1662 /* Insert the string window[strstart .. strstart+2] in the
1663 * dictionary, and set hash_head to the head of the hash chain:
1666 if (s
->lookahead
>= MIN_MATCH
) {
1667 INSERT_STRING(s
, s
->strstart
, hash_head
);
1670 /* Find the longest match, discarding those <= prev_length.
1671 * At this point we have always match_length < MIN_MATCH
1673 if (hash_head
!= NIL
&& s
->strstart
- hash_head
<= MAX_DIST(s
)) {
1674 /* To simplify the code, we prevent matches with the string
1675 * of window index 0 (in particular we have to avoid a match
1676 * of the string with itself at the start of the input file).
1678 s
->match_length
= longest_match (s
, hash_head
);
1679 /* longest_match() sets match_start */
1681 if (s
->match_length
>= MIN_MATCH
) {
1682 check_match(s
, s
->strstart
, s
->match_start
, s
->match_length
);
1684 _tr_tally_dist(s
, s
->strstart
- s
->match_start
,
1685 s
->match_length
- MIN_MATCH
, bflush
);
1687 s
->lookahead
-= s
->match_length
;
1689 /* Insert new strings in the hash table only if the match length
1690 * is not too large. This saves time but degrades compression.
1693 if (s
->match_length
<= s
->max_insert_length
&&
1694 s
->lookahead
>= MIN_MATCH
) {
1695 s
->match_length
--; /* string at strstart already in table */
1698 INSERT_STRING(s
, s
->strstart
, hash_head
);
1699 /* strstart never exceeds WSIZE-MAX_MATCH, so there are
1700 * always MIN_MATCH bytes ahead.
1702 } while (--s
->match_length
!= 0);
1707 s
->strstart
+= s
->match_length
;
1708 s
->match_length
= 0;
1709 s
->ins_h
= s
->window
[s
->strstart
];
1710 UPDATE_HASH(s
, s
->ins_h
, s
->window
[s
->strstart
+1]);
1712 Call
UPDATE_HASH() MIN_MATCH
-3 more times
1714 /* If lookahead < MIN_MATCH, ins_h is garbage, but it does not
1715 * matter since it will be recomputed at next deflate call.
1719 /* No match, output a literal byte */
1720 Tracevv((stderr
,"%c", s
->window
[s
->strstart
]));
1721 _tr_tally_lit (s
, s
->window
[s
->strstart
], bflush
);
1725 if (bflush
) FLUSH_BLOCK(s
, 0);
1727 s
->insert
= s
->strstart
< MIN_MATCH
-1 ? s
->strstart
: MIN_MATCH
-1;
1728 if (flush
== Z_FINISH
) {
1738 #ifdef ZLIB_PM3_TUNED
1739 local uInt
try_harder(s
, strstart
, lookahead
, hash_head
)
1745 uInt strstart_save
= s
->strstart
;
1746 s
->strstart
= strstart
;
1747 uInt lookahead_save
= s
->lookahead
;
1748 s
->lookahead
= lookahead
;
1749 uInt ins_h_save
= s
->ins_h
;
1751 uInt best_combined_gain
= 0;
1753 uInt prev_length
= s
->prev_length
< MIN_MATCH
? 1 : s
->prev_length
;
1754 uInt best_prev_length
= prev_length
;
1755 uInt current_match_start
= s
->match_start
;
1756 uInt current_match_length
= s
->match_length
;
1759 if (hash_head
!= NIL
&& s
->strstart
- hash_head
<= MAX_DIST(s
)) {
1760 match_length
= longest_match (s
, hash_head
);
1761 /* longest_match() sets match_start */
1763 match_length
= MIN_MATCH
- 1;
1765 #if TOO_FAR <= 32767
1766 if (match_length
== MIN_MATCH
&& s
->strstart
- s
->match_start
> TOO_FAR
) {
1767 match_length
= MIN_MATCH
-1;
1770 if (s
->strstart
== strstart
) { // store match at current position
1771 current_match_length
= match_length
;
1772 current_match_start
= s
->match_start
;
1774 if (s
->strstart
- strstart
+ 1 < MIN_MATCH
) { // previous match reduced to one or two literals
1775 combined_gain
= 0; // need one literal per byte: no gain (assuming 8 bits per literal)
1777 combined_gain
= s
->strstart
- strstart
+ 1 - MIN_MATCH
; // (possibly truncated) previous_length - 3 literals
1779 if (match_length
< MIN_MATCH
) {
1780 combined_gain
+= 0; // no gain
1782 combined_gain
+= match_length
- MIN_MATCH
; // match_length bytes are coded as three literals
1784 if (combined_gain
>= best_combined_gain
) { // in case of a tie we prefer the longer prev_length
1785 best_combined_gain
= combined_gain
;
1786 best_prev_length
= s
->strstart
- strstart
+ 1;
1790 UPDATE_HASH(s
, s
->ins_h
, s
->window
[(s
->strstart
) + (MIN_MATCH
-1)]);
1791 hash_head
= s
->head
[s
->ins_h
];
1792 } while (s
->strstart
<= strstart
-1 + prev_length
// try to truncate the previous match to 1, 3, ... prev_length
1793 && s
->strstart
<= s
->window_size
- MIN_LOOKAHEAD
); // watch out for the end of the input
1795 s
->strstart
= strstart_save
;
1796 s
->lookahead
= lookahead_save
;
1797 s
->ins_h
= ins_h_save
;
1798 s
->match_length
= current_match_length
;
1799 s
->match_start
= current_match_start
;
1800 if (best_prev_length
>= MIN_MATCH
) {
1801 s
->prev_length
= best_prev_length
;
1802 s
->match_length
= MIN_MATCH
- 1;
1804 s
->prev_length
= MIN_MATCH
- 1;
1806 return best_combined_gain
;
1813 /* ===========================================================================
1814 * Same as above, but achieves better compression. We use a lazy
1815 * evaluation for matches: a match is finally adopted only if there is
1816 * no better match at the next window position.
1818 local block_state
deflate_slow(s
, flush
)
1822 IPos hash_head
; /* head of hash chain */
1823 int bflush
; /* set if current block must be flushed */
1825 /* Process the input block. */
1827 /* Make sure that we always have enough lookahead, except
1828 * at the end of the input file. We need MAX_MATCH bytes
1829 * for the next match, plus MIN_MATCH bytes to insert the
1830 * string following the next match.
1832 if (s
->lookahead
< MIN_LOOKAHEAD
) {
1834 if (s
->lookahead
< MIN_LOOKAHEAD
&& flush
== Z_NO_FLUSH
) {
1837 if (s
->lookahead
== 0) break; /* flush the current block */
1840 /* Insert the string window[strstart .. strstart+2] in the
1841 * dictionary, and set hash_head to the head of the hash chain:
1844 if (s
->lookahead
>= MIN_MATCH
) {
1845 INSERT_STRING(s
, s
->strstart
, hash_head
);
1848 /* Find the longest match, discarding those <= prev_length. */
1849 s
->prev_length
= s
->match_length
, s
->prev_match
= s
->match_start
;
1850 s
->match_length
= MIN_MATCH
-1;
1852 #ifdef ZLIB_PM3_TUNED
1853 if (s
->prev_length
< s
->max_lazy_match
) {
1854 try_harder(s
, s
->strstart
, s
->lookahead
, hash_head
);
1858 if (hash_head
!= NIL
&& s
->prev_length
< s
->max_lazy_match
&&
1859 s
->strstart
- hash_head
<= MAX_DIST(s
)) {
1860 /* To simplify the code, we prevent matches with the string
1861 * of window index 0 (in particular we have to avoid a match
1862 * of the string with itself at the start of the input file).
1864 s
->match_length
= longest_match (s
, hash_head
);
1865 /* longest_match() sets match_start */
1867 if (s
->match_length
<= 5 && (s
->strategy
== Z_FILTERED
1868 #if TOO_FAR <= 32767
1869 || (s
->match_length
== MIN_MATCH
&&
1870 s
->strstart
- s
->match_start
> TOO_FAR
)
1874 /* If prev_match is also MIN_MATCH, match_start is garbage
1875 * but we will ignore the current match anyway.
1877 s
->match_length
= MIN_MATCH
-1;
1880 #endif /* ZLIB_PM3_TUNED */
1881 /* If there was a match at the previous step and the current
1882 * match is not better, output the previous match:
1884 if (s
->prev_length
>= MIN_MATCH
&& s
->match_length
<= s
->prev_length
) {
1885 uInt max_insert
= s
->strstart
+ s
->lookahead
- MIN_MATCH
;
1886 /* Do not insert strings in hash table beyond this. */
1888 check_match(s
, s
->strstart
-1, s
->prev_match
, s
->prev_length
);
1890 _tr_tally_dist(s
, s
->strstart
-1 - s
->prev_match
,
1891 s
->prev_length
- MIN_MATCH
, bflush
);
1893 /* Insert in hash table all strings up to the end of the match.
1894 * strstart-1 and strstart are already inserted. If there is not
1895 * enough lookahead, the last two strings are not inserted in
1898 s
->lookahead
-= s
->prev_length
-1;
1899 s
->prev_length
-= 2;
1901 if (++s
->strstart
<= max_insert
) {
1902 INSERT_STRING(s
, s
->strstart
, hash_head
);
1904 } while (--s
->prev_length
!= 0);
1905 s
->match_available
= 0;
1906 s
->match_length
= MIN_MATCH
-1;
1909 if (bflush
) FLUSH_BLOCK(s
, 0);
1911 } else if (s
->match_available
) {
1912 /* If there was no match at the previous position, output a
1913 * single literal. If there was a match but the current match
1914 * is longer, truncate the previous match to a single literal.
1916 Tracevv((stderr
,"%c", s
->window
[s
->strstart
-1]));
1917 _tr_tally_lit(s
, s
->window
[s
->strstart
-1], bflush
);
1919 FLUSH_BLOCK_ONLY(s
, 0);
1923 if (s
->strm
->avail_out
== 0) return need_more
;
1925 /* There is no previous match to compare with, wait for
1926 * the next step to decide.
1928 s
->match_available
= 1;
1933 Assert (flush
!= Z_NO_FLUSH
, "no flush?");
1934 if (s
->match_available
) {
1935 Tracevv((stderr
,"%c", s
->window
[s
->strstart
-1]));
1936 _tr_tally_lit(s
, s
->window
[s
->strstart
-1], bflush
);
1937 s
->match_available
= 0;
1939 s
->insert
= s
->strstart
< MIN_MATCH
-1 ? s
->strstart
: MIN_MATCH
-1;
1940 if (flush
== Z_FINISH
) {
1948 #endif /* FASTEST */
1950 /* ===========================================================================
1951 * For Z_RLE, simply look for runs of bytes, generate matches only of distance
1952 * one. Do not maintain a hash table. (It will be regenerated if this run of
1953 * deflate switches away from Z_RLE.)
1955 local block_state
deflate_rle(s
, flush
)
1959 int bflush
; /* set if current block must be flushed */
1960 uInt prev
; /* byte at distance one to match */
1961 Bytef
*scan
, *strend
; /* scan goes up to strend for length of run */
1964 /* Make sure that we always have enough lookahead, except
1965 * at the end of the input file. We need MAX_MATCH bytes
1966 * for the longest run, plus one for the unrolled loop.
1968 if (s
->lookahead
<= MAX_MATCH
) {
1970 if (s
->lookahead
<= MAX_MATCH
&& flush
== Z_NO_FLUSH
) {
1973 if (s
->lookahead
== 0) break; /* flush the current block */
1976 /* See how many times the previous byte repeats */
1977 s
->match_length
= 0;
1978 if (s
->lookahead
>= MIN_MATCH
&& s
->strstart
> 0) {
1979 scan
= s
->window
+ s
->strstart
- 1;
1981 if (prev
== *++scan
&& prev
== *++scan
&& prev
== *++scan
) {
1982 strend
= s
->window
+ s
->strstart
+ MAX_MATCH
;
1984 } while (prev
== *++scan
&& prev
== *++scan
&&
1985 prev
== *++scan
&& prev
== *++scan
&&
1986 prev
== *++scan
&& prev
== *++scan
&&
1987 prev
== *++scan
&& prev
== *++scan
&&
1989 s
->match_length
= MAX_MATCH
- (int)(strend
- scan
);
1990 if (s
->match_length
> s
->lookahead
)
1991 s
->match_length
= s
->lookahead
;
1993 Assert(scan
<= s
->window
+(uInt
)(s
->window_size
-1), "wild scan");
1996 /* Emit match if have run of MIN_MATCH or longer, else emit literal */
1997 if (s
->match_length
>= MIN_MATCH
) {
1998 check_match(s
, s
->strstart
, s
->strstart
- 1, s
->match_length
);
2000 _tr_tally_dist(s
, 1, s
->match_length
- MIN_MATCH
, bflush
);
2002 s
->lookahead
-= s
->match_length
;
2003 s
->strstart
+= s
->match_length
;
2004 s
->match_length
= 0;
2006 /* No match, output a literal byte */
2007 Tracevv((stderr
,"%c", s
->window
[s
->strstart
]));
2008 _tr_tally_lit (s
, s
->window
[s
->strstart
], bflush
);
2012 if (bflush
) FLUSH_BLOCK(s
, 0);
2015 if (flush
== Z_FINISH
) {
2024 /* ===========================================================================
2025 * For Z_HUFFMAN_ONLY, do not look for matches. Do not maintain a hash table.
2026 * (It will be regenerated if this run of deflate switches away from Huffman.)
2028 local block_state
deflate_huff(s
, flush
)
2032 int bflush
; /* set if current block must be flushed */
2035 /* Make sure that we have a literal to write. */
2036 if (s
->lookahead
== 0) {
2038 if (s
->lookahead
== 0) {
2039 if (flush
== Z_NO_FLUSH
)
2041 break; /* flush the current block */
2045 /* Output a literal byte */
2046 s
->match_length
= 0;
2047 Tracevv((stderr
,"%c", s
->window
[s
->strstart
]));
2048 _tr_tally_lit (s
, s
->window
[s
->strstart
], bflush
);
2051 if (bflush
) FLUSH_BLOCK(s
, 0);
2054 if (flush
== Z_FINISH
) {