1 /* $OpenBSD: deflate.c,v 1.11 2009/10/27 23:59:31 deraadt Exp $ */ 2 /* deflate.c -- compress data using the deflation algorithm 3 * Copyright (C) 1995-2005 Jean-loup Gailly. 4 * For conditions of distribution and use, see copyright notice in zlib.h 5 */ 6 7 /* 8 * ALGORITHM 9 * 10 * The "deflation" process depends on being able to identify portions 11 * of the input text which are identical to earlier input (within a 12 * sliding window trailing behind the input currently being processed). 13 * 14 * The most straightforward technique turns out to be the fastest for 15 * most input files: try all possible matches and select the longest. 16 * The key feature of this algorithm is that insertions into the string 17 * dictionary are very simple and thus fast, and deletions are avoided 18 * completely. Insertions are performed at each input character, whereas 19 * string matches are performed only when the previous match ends. So it 20 * is preferable to spend more time in matches to allow very fast string 21 * insertions and avoid deletions. The matching algorithm for small 22 * strings is inspired from that of Rabin & Karp. A brute force approach 23 * is used to find longer strings when a small match has been found. 24 * A similar algorithm is used in comic (by Jan-Mark Wams) and freeze 25 * (by Leonid Broukhis). 26 * A previous version of this file used a more sophisticated algorithm 27 * (by Fiala and Greene) which is guaranteed to run in linear amortized 28 * time, but has a larger average cost, uses more memory and is patented. 29 * However the F&G algorithm may be faster for some highly redundant 30 * files if the parameter max_chain_length (described below) is too large. 31 * 32 * ACKNOWLEDGEMENTS 33 * 34 * The idea of lazy evaluation of matches is due to Jan-Mark Wams, and 35 * I found it in 'freeze' written by Leonid Broukhis. 36 * Thanks to many people for bug reports and testing. 37 * 38 * REFERENCES 39 * 40 * Deutsch, L.P.,"DEFLATE Compressed Data Format Specification". 41 * Available in http://www.ietf.org/rfc/rfc1951.txt 42 * 43 * A description of the Rabin and Karp algorithm is given in the book 44 * "Algorithms" by R. Sedgewick, Addison-Wesley, p252. 45 * 46 * Fiala,E.R., and Greene,D.H. 47 * Data Compression with Finite Windows, Comm.ACM, 32,4 (1989) 490-595 48 * 49 */ 50 51 52 #include "deflate.h" 53 54 /* 55 If you use the zlib library in a product, an acknowledgment is welcome 56 in the documentation of your product. If for some reason you cannot 57 include such an acknowledgment, I would appreciate that you keep this 58 copyright string in the executable of your product. 59 */ 60 61 /* =========================================================================== 62 * Function prototypes. 63 */ 64 typedef enum { 65 need_more, /* block not completed, need more input or more output */ 66 block_done, /* block flush performed */ 67 finish_started, /* finish started, need only more output at next deflate */ 68 finish_done /* finish done, accept no more input or output */ 69 } block_state; 70 71 typedef block_state (*compress_func) OF((deflate_state *s, int flush)); 72 /* Compression function. Returns the block state after the call. */ 73 74 local void fill_window OF((deflate_state *s)); 75 local block_state deflate_stored OF((deflate_state *s, int flush)); 76 local block_state deflate_fast OF((deflate_state *s, int flush)); 77 #ifndef FASTEST 78 local block_state deflate_slow OF((deflate_state *s, int flush)); 79 #endif 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 #ifndef FASTEST 85 #ifdef ASMV 86 void match_init OF((void)); /* asm code initialization */ 87 uInt longest_match OF((deflate_state *s, IPos cur_match)); 88 #else 89 local uInt longest_match OF((deflate_state *s, IPos cur_match)); 90 #endif 91 #endif 92 local uInt longest_match_fast OF((deflate_state *s, IPos cur_match)); 93 94 #ifdef DEBUG 95 local void check_match OF((deflate_state *s, IPos start, IPos match, 96 int length)); 97 #endif 98 99 /* =========================================================================== 100 * Local data 101 */ 102 103 #define NIL 0 104 /* Tail of hash chains */ 105 106 #ifndef TOO_FAR 107 # define TOO_FAR 4096 108 #endif 109 /* Matches of length 3 are discarded if their distance exceeds TOO_FAR */ 110 111 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1) 112 /* Minimum amount of lookahead, except at the end of the input file. 113 * See deflate.c for comments about the MIN_MATCH+1. 114 */ 115 116 /* Values for max_lazy_match, good_match and max_chain_length, depending on 117 * the desired pack level (0..9). The values given below have been tuned to 118 * exclude worst case performance for pathological files. Better values may be 119 * found for specific files. 120 */ 121 typedef struct config_s { 122 ush good_length; /* reduce lazy search above this match length */ 123 ush max_lazy; /* do not perform lazy search above this match length */ 124 ush nice_length; /* quit search above this match length */ 125 ush max_chain; 126 compress_func func; 127 } config; 128 129 #ifdef FASTEST 130 local const config configuration_table[2] = { 131 /* good lazy nice chain */ 132 /* 0 */ {0, 0, 0, 0, deflate_stored}, /* store only */ 133 /* 1 */ {4, 4, 8, 4, deflate_fast}}; /* max speed, no lazy matches */ 134 #else 135 local const config configuration_table[10] = { 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 */ 139 /* 2 */ {4, 5, 16, 8, deflate_fast}, 140 /* 3 */ {4, 6, 32, 32, deflate_fast}, 141 142 /* 4 */ {4, 4, 16, 16, deflate_slow}, /* lazy matches */ 143 /* 5 */ {8, 16, 32, 32, deflate_slow}, 144 /* 6 */ {8, 16, 128, 128, deflate_slow}, 145 /* 7 */ {8, 32, 128, 256, deflate_slow}, 146 /* 8 */ {32, 128, 258, 1024, deflate_slow}, 147 /* 9 */ {32, 258, 258, 4096, deflate_slow}}; /* max compression */ 148 #endif 149 150 /* Note: the deflate() code requires max_lazy >= MIN_MATCH and max_chain >= 4 151 * For deflate_fast() (levels <= 3) good is ignored and lazy has a different 152 * meaning. 153 */ 154 155 #define EQUAL 0 156 /* result of memcmp for equal strings */ 157 158 #ifndef NO_DUMMY_DECL 159 struct static_tree_desc_s {int dummy;}; /* for buggy compilers */ 160 #endif 161 162 /* =========================================================================== 163 * Update a hash value with the given input byte 164 * IN assertion: all calls to UPDATE_HASH are made with consecutive 165 * input characters, so that a running hash key can be computed from the 166 * previous key instead of complete recalculation each time. 167 */ 168 #define UPDATE_HASH(s,h,c) (h = (((h)<<s->hash_shift) ^ (c)) & s->hash_mask) 169 170 171 /* =========================================================================== 172 * Insert string str in the dictionary and set match_head to the previous head 173 * of the hash chain (the most recent string with same hash key). Return 174 * the previous length of the hash chain. 175 * If this file is compiled with -DFASTEST, the compression level is forced 176 * to 1, and no hash chains are maintained. 177 * IN assertion: all calls to INSERT_STRING are made with consecutive 178 * input characters and the first MIN_MATCH bytes of str are valid 179 * (except for the last MIN_MATCH-1 bytes of the input file). 180 */ 181 #ifdef FASTEST 182 #define INSERT_STRING(s, str, match_head) \ 183 (UPDATE_HASH(s, s->ins_h, s->window[(str) + (MIN_MATCH-1)]), \ 184 match_head = s->head[s->ins_h], \ 185 s->head[s->ins_h] = (Pos)(str)) 186 #else 187 #define INSERT_STRING(s, str, match_head) \ 188 (UPDATE_HASH(s, s->ins_h, s->window[(str) + (MIN_MATCH-1)]), \ 189 match_head = s->prev[(str) & s->w_mask] = s->head[s->ins_h], \ 190 s->head[s->ins_h] = (Pos)(str)) 191 #endif 192 193 /* =========================================================================== 194 * Initialize the hash table (avoiding 64K overflow for 16 bit systems). 195 * prev[] will be initialized on the fly. 196 */ 197 #define CLEAR_HASH(s) \ 198 s->head[s->hash_size-1] = NIL; \ 199 zmemzero((Bytef *)s->head, (unsigned)(s->hash_size-1)*sizeof(*s->head)); 200 201 /* ========================================================================= */ 202 int ZEXPORT deflateInit_(strm, level, version, stream_size) 203 z_streamp strm; 204 int level; 205 const char *version; 206 int stream_size; 207 { 208 return deflateInit2_(strm, level, Z_DEFLATED, MAX_WBITS, DEF_MEM_LEVEL, 209 Z_DEFAULT_STRATEGY, version, stream_size); 210 /* To do: ignore strm->next_in if we use it as window */ 211 } 212 213 /* ========================================================================= */ 214 int ZEXPORT deflateInit2_(strm, level, method, windowBits, memLevel, strategy, 215 version, stream_size) 216 z_streamp strm; 217 int level; 218 int method; 219 int windowBits; 220 int memLevel; 221 int strategy; 222 const char *version; 223 int stream_size; 224 { 225 deflate_state *s; 226 int wrap = 1; 227 static const char my_version[] = ZLIB_VERSION; 228 229 ushf *overlay; 230 /* We overlay pending_buf and d_buf+l_buf. This works since the average 231 * output size for (length,distance) codes is <= 24 bits. 232 */ 233 234 if (version == Z_NULL || version[0] != my_version[0] || 235 stream_size != sizeof(z_stream)) { 236 return Z_VERSION_ERROR; 237 } 238 if (strm == Z_NULL) return Z_STREAM_ERROR; 239 240 strm->msg = Z_NULL; 241 if (strm->zalloc == (alloc_func)0) { 242 strm->zalloc = zcalloc; 243 strm->opaque = (voidpf)0; 244 } 245 if (strm->zfree == (free_func)0) strm->zfree = zcfree; 246 247 #ifdef FASTEST 248 if (level != 0) level = 1; 249 #else 250 if (level == Z_DEFAULT_COMPRESSION) level = 6; 251 #endif 252 253 if (windowBits < 0) { /* suppress zlib wrapper */ 254 wrap = 0; 255 windowBits = -windowBits; 256 } 257 #ifdef GZIP 258 else if (windowBits > 15) { 259 wrap = 2; /* write gzip wrapper instead */ 260 windowBits -= 16; 261 } 262 #endif 263 if (memLevel < 1 || memLevel > MAX_MEM_LEVEL || method != Z_DEFLATED || 264 windowBits < 8 || windowBits > 15 || level < 0 || level > 9 || 265 strategy < 0 || strategy > Z_FIXED) { 266 return Z_STREAM_ERROR; 267 } 268 if (windowBits == 8) windowBits = 9; /* until 256-byte window bug fixed */ 269 s = (deflate_state *) ZALLOC(strm, 1, sizeof(deflate_state)); 270 if (s == Z_NULL) return Z_MEM_ERROR; 271 strm->state = (struct internal_state FAR *)s; 272 s->strm = strm; 273 274 s->wrap = wrap; 275 s->gzhead = Z_NULL; 276 s->w_bits = windowBits; 277 s->w_size = 1 << s->w_bits; 278 s->w_mask = s->w_size - 1; 279 280 s->hash_bits = memLevel + 7; 281 s->hash_size = 1 << s->hash_bits; 282 s->hash_mask = s->hash_size - 1; 283 s->hash_shift = ((s->hash_bits+MIN_MATCH-1)/MIN_MATCH); 284 285 s->window = (Bytef *) ZALLOC(strm, s->w_size, 2*sizeof(Byte)); 286 s->prev = (Posf *) ZALLOC(strm, s->w_size, sizeof(Pos)); 287 s->head = (Posf *) ZALLOC(strm, s->hash_size, sizeof(Pos)); 288 289 s->lit_bufsize = 1 << (memLevel + 6); /* 16K elements by default */ 290 291 overlay = (ushf *) ZALLOC(strm, s->lit_bufsize, sizeof(ush)+2); 292 s->pending_buf = (uchf *) overlay; 293 s->pending_buf_size = (ulg)s->lit_bufsize * (sizeof(ush)+2L); 294 295 if (s->window == Z_NULL || s->prev == Z_NULL || s->head == Z_NULL || 296 s->pending_buf == Z_NULL) { 297 s->status = FINISH_STATE; 298 strm->msg = (char*)ERR_MSG(Z_MEM_ERROR); 299 deflateEnd (strm); 300 return Z_MEM_ERROR; 301 } 302 s->d_buf = overlay + s->lit_bufsize/sizeof(ush); 303 s->l_buf = s->pending_buf + (1+sizeof(ush))*s->lit_bufsize; 304 305 s->level = level; 306 s->strategy = strategy; 307 s->method = (Byte)method; 308 309 return deflateReset(strm); 310 } 311 312 /* ========================================================================= */ 313 int ZEXPORT deflateSetDictionary (strm, dictionary, dictLength) 314 z_streamp strm; 315 const Bytef *dictionary; 316 uInt dictLength; 317 { 318 deflate_state *s; 319 uInt length = dictLength; 320 uInt n; 321 IPos hash_head = 0; 322 323 if (strm == Z_NULL || strm->state == Z_NULL || dictionary == Z_NULL || 324 strm->state->wrap == 2 || 325 (strm->state->wrap == 1 && strm->state->status != INIT_STATE)) 326 return Z_STREAM_ERROR; 327 328 s = strm->state; 329 if (s->wrap) 330 strm->adler = adler32(strm->adler, dictionary, dictLength); 331 332 if (length < MIN_MATCH) return Z_OK; 333 if (length > MAX_DIST(s)) { 334 length = MAX_DIST(s); 335 dictionary += dictLength - length; /* use the tail of the dictionary */ 336 } 337 zmemcpy(s->window, dictionary, length); 338 s->strstart = length; 339 s->block_start = (long)length; 340 341 /* Insert all strings in the hash table (except for the last two bytes). 342 * s->lookahead stays null, so s->ins_h will be recomputed at the next 343 * call of fill_window. 344 */ 345 s->ins_h = s->window[0]; 346 UPDATE_HASH(s, s->ins_h, s->window[1]); 347 for (n = 0; n <= length - MIN_MATCH; n++) { 348 INSERT_STRING(s, n, hash_head); 349 } 350 if (hash_head) hash_head = 0; /* to make compiler happy */ 351 return Z_OK; 352 } 353 354 /* ========================================================================= */ 355 int ZEXPORT deflateReset (strm) 356 z_streamp strm; 357 { 358 deflate_state *s; 359 360 if (strm == Z_NULL || strm->state == Z_NULL || 361 strm->zalloc == (alloc_func)0 || strm->zfree == (free_func)0) { 362 return Z_STREAM_ERROR; 363 } 364 365 strm->total_in = strm->total_out = 0; 366 strm->msg = Z_NULL; /* use zfree if we ever allocate msg dynamically */ 367 strm->data_type = Z_UNKNOWN; 368 369 s = (deflate_state *)strm->state; 370 s->pending = 0; 371 s->pending_out = s->pending_buf; 372 373 if (s->wrap < 0) { 374 s->wrap = -s->wrap; /* was made negative by deflate(..., Z_FINISH); */ 375 } 376 s->status = s->wrap ? INIT_STATE : BUSY_STATE; 377 strm->adler = 378 #ifdef GZIP 379 s->wrap == 2 ? crc32(0L, Z_NULL, 0) : 380 #endif 381 adler32(0L, Z_NULL, 0); 382 s->last_flush = Z_NO_FLUSH; 383 384 _tr_init(s); 385 lm_init(s); 386 387 return Z_OK; 388 } 389 390 /* ========================================================================= */ 391 int ZEXPORT deflateSetHeader (strm, head) 392 z_streamp strm; 393 gz_headerp head; 394 { 395 if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR; 396 if (strm->state->wrap != 2) return Z_STREAM_ERROR; 397 strm->state->gzhead = head; 398 return Z_OK; 399 } 400 401 /* ========================================================================= */ 402 int ZEXPORT deflatePrime (strm, bits, value) 403 z_streamp strm; 404 int bits; 405 int value; 406 { 407 if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR; 408 strm->state->bi_valid = bits; 409 strm->state->bi_buf = (ush)(value & ((1 << bits) - 1)); 410 return Z_OK; 411 } 412 413 /* ========================================================================= */ 414 int ZEXPORT deflateParams(strm, level, strategy) 415 z_streamp strm; 416 int level; 417 int strategy; 418 { 419 deflate_state *s; 420 compress_func func; 421 int err = Z_OK; 422 423 if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR; 424 s = strm->state; 425 426 #ifdef FASTEST 427 if (level != 0) level = 1; 428 #else 429 if (level == Z_DEFAULT_COMPRESSION) level = 6; 430 #endif 431 if (level < 0 || level > 9 || strategy < 0 || strategy > Z_FIXED) { 432 return Z_STREAM_ERROR; 433 } 434 func = configuration_table[s->level].func; 435 436 if (func != configuration_table[level].func && strm->total_in != 0) { 437 /* Flush the last buffer: */ 438 err = deflate(strm, Z_PARTIAL_FLUSH); 439 } 440 if (s->level != level) { 441 s->level = level; 442 s->max_lazy_match = configuration_table[level].max_lazy; 443 s->good_match = configuration_table[level].good_length; 444 s->nice_match = configuration_table[level].nice_length; 445 s->max_chain_length = configuration_table[level].max_chain; 446 } 447 s->strategy = strategy; 448 return err; 449 } 450 451 /* ========================================================================= */ 452 int ZEXPORT deflateTune(strm, good_length, max_lazy, nice_length, max_chain) 453 z_streamp strm; 454 int good_length; 455 int max_lazy; 456 int nice_length; 457 int max_chain; 458 { 459 deflate_state *s; 460 461 if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR; 462 s = strm->state; 463 s->good_match = good_length; 464 s->max_lazy_match = max_lazy; 465 s->nice_match = nice_length; 466 s->max_chain_length = max_chain; 467 return Z_OK; 468 } 469 470 /* ========================================================================= 471 * For the default windowBits of 15 and memLevel of 8, this function returns 472 * a close to exact, as well as small, upper bound on the compressed size. 473 * They are coded as constants here for a reason--if the #define's are 474 * changed, then this function needs to be changed as well. The return 475 * value for 15 and 8 only works for those exact settings. 476 * 477 * For any setting other than those defaults for windowBits and memLevel, 478 * the value returned is a conservative worst case for the maximum expansion 479 * resulting from using fixed blocks instead of stored blocks, which deflate 480 * can emit on compressed data for some combinations of the parameters. 481 * 482 * This function could be more sophisticated to provide closer upper bounds 483 * for every combination of windowBits and memLevel, as well as wrap. 484 * But even the conservative upper bound of about 14% expansion does not 485 * seem onerous for output buffer allocation. 486 */ 487 uLong ZEXPORT deflateBound(strm, sourceLen) 488 z_streamp strm; 489 uLong sourceLen; 490 { 491 deflate_state *s; 492 uLong destLen; 493 494 /* conservative upper bound */ 495 destLen = sourceLen + 496 ((sourceLen + 7) >> 3) + ((sourceLen + 63) >> 6) + 11; 497 498 /* if can't get parameters, return conservative bound */ 499 if (strm == Z_NULL || strm->state == Z_NULL) 500 return destLen; 501 502 /* if not default parameters, return conservative bound */ 503 s = strm->state; 504 if (s->w_bits != 15 || s->hash_bits != 8 + 7) 505 return destLen; 506 507 /* default settings: return tight bound for that case */ 508 return compressBound(sourceLen); 509 } 510 511 /* ========================================================================= 512 * Put a short in the pending buffer. The 16-bit value is put in MSB order. 513 * IN assertion: the stream state is correct and there is enough room in 514 * pending_buf. 515 */ 516 local void putShortMSB (s, b) 517 deflate_state *s; 518 uInt b; 519 { 520 put_byte(s, (Byte)(b >> 8)); 521 put_byte(s, (Byte)(b & 0xff)); 522 } 523 524 /* ========================================================================= 525 * Flush as much pending output as possible. All deflate() output goes 526 * through this function so some applications may wish to modify it 527 * to avoid allocating a large strm->next_out buffer and copying into it. 528 * (See also read_buf()). 529 */ 530 local void flush_pending(strm) 531 z_streamp strm; 532 { 533 unsigned len = strm->state->pending; 534 535 if (len > strm->avail_out) len = strm->avail_out; 536 if (len == 0) return; 537 538 zmemcpy(strm->next_out, strm->state->pending_out, len); 539 strm->next_out += len; 540 strm->state->pending_out += len; 541 strm->total_out += len; 542 strm->avail_out -= len; 543 strm->state->pending -= len; 544 if (strm->state->pending == 0) { 545 strm->state->pending_out = strm->state->pending_buf; 546 } 547 } 548 549 /* ========================================================================= */ 550 int ZEXPORT deflate (strm, flush) 551 z_streamp strm; 552 int flush; 553 { 554 int old_flush; /* value of flush param for previous deflate call */ 555 deflate_state *s; 556 557 if (strm == Z_NULL || strm->state == Z_NULL || 558 flush > Z_FINISH || flush < 0) { 559 return Z_STREAM_ERROR; 560 } 561 s = strm->state; 562 563 if (strm->next_out == Z_NULL || 564 (strm->next_in == Z_NULL && strm->avail_in != 0) || 565 (s->status == FINISH_STATE && flush != Z_FINISH)) { 566 ERR_RETURN(strm, Z_STREAM_ERROR); 567 } 568 if (strm->avail_out == 0) ERR_RETURN(strm, Z_BUF_ERROR); 569 570 s->strm = strm; /* just in case */ 571 old_flush = s->last_flush; 572 s->last_flush = flush; 573 574 /* Write the header */ 575 if (s->status == INIT_STATE) { 576 #ifdef GZIP 577 if (s->wrap == 2) { 578 strm->adler = crc32(0L, Z_NULL, 0); 579 put_byte(s, 31); 580 put_byte(s, 139); 581 put_byte(s, 8); 582 if (s->gzhead == NULL) { 583 put_byte(s, 0); 584 put_byte(s, 0); 585 put_byte(s, 0); 586 put_byte(s, 0); 587 put_byte(s, 0); 588 put_byte(s, s->level == 9 ? 2 : 589 (s->strategy >= Z_HUFFMAN_ONLY || s->level < 2 ? 590 4 : 0)); 591 put_byte(s, OS_CODE); 592 s->status = BUSY_STATE; 593 } 594 else { 595 put_byte(s, (s->gzhead->text ? 1 : 0) + 596 (s->gzhead->hcrc ? 2 : 0) + 597 (s->gzhead->extra == Z_NULL ? 0 : 4) + 598 (s->gzhead->name == Z_NULL ? 0 : 8) + 599 (s->gzhead->comment == Z_NULL ? 0 : 16) 600 ); 601 put_byte(s, (Byte)(s->gzhead->time & 0xff)); 602 put_byte(s, (Byte)((s->gzhead->time >> 8) & 0xff)); 603 put_byte(s, (Byte)((s->gzhead->time >> 16) & 0xff)); 604 put_byte(s, (Byte)((s->gzhead->time >> 24) & 0xff)); 605 put_byte(s, s->level == 9 ? 2 : 606 (s->strategy >= Z_HUFFMAN_ONLY || s->level < 2 ? 607 4 : 0)); 608 put_byte(s, s->gzhead->os & 0xff); 609 if (s->gzhead->extra != NULL) { 610 put_byte(s, s->gzhead->extra_len & 0xff); 611 put_byte(s, (s->gzhead->extra_len >> 8) & 0xff); 612 } 613 if (s->gzhead->hcrc) 614 strm->adler = crc32(strm->adler, s->pending_buf, 615 s->pending); 616 s->gzindex = 0; 617 s->status = EXTRA_STATE; 618 } 619 } 620 else 621 #endif 622 { 623 uInt header = (Z_DEFLATED + ((s->w_bits-8)<<4)) << 8; 624 uInt level_flags; 625 626 if (s->strategy >= Z_HUFFMAN_ONLY || s->level < 2) 627 level_flags = 0; 628 else if (s->level < 6) 629 level_flags = 1; 630 else if (s->level == 6) 631 level_flags = 2; 632 else 633 level_flags = 3; 634 header |= (level_flags << 6); 635 if (s->strstart != 0) header |= PRESET_DICT; 636 header += 31 - (header % 31); 637 638 s->status = BUSY_STATE; 639 putShortMSB(s, header); 640 641 /* Save the adler32 of the preset dictionary: */ 642 if (s->strstart != 0) { 643 putShortMSB(s, (uInt)(strm->adler >> 16)); 644 putShortMSB(s, (uInt)(strm->adler & 0xffff)); 645 } 646 strm->adler = adler32(0L, Z_NULL, 0); 647 } 648 } 649 #ifdef GZIP 650 if (s->status == EXTRA_STATE) { 651 if (s->gzhead->extra != NULL) { 652 uInt beg = s->pending; /* start of bytes to update crc */ 653 654 while (s->gzindex < (s->gzhead->extra_len & 0xffff)) { 655 if (s->pending == s->pending_buf_size) { 656 if (s->gzhead->hcrc && s->pending > beg) 657 strm->adler = crc32(strm->adler, s->pending_buf + beg, 658 s->pending - beg); 659 flush_pending(strm); 660 beg = s->pending; 661 if (s->pending == s->pending_buf_size) 662 break; 663 } 664 put_byte(s, s->gzhead->extra[s->gzindex]); 665 s->gzindex++; 666 } 667 if (s->gzhead->hcrc && s->pending > beg) 668 strm->adler = crc32(strm->adler, s->pending_buf + beg, 669 s->pending - beg); 670 if (s->gzindex == s->gzhead->extra_len) { 671 s->gzindex = 0; 672 s->status = NAME_STATE; 673 } 674 } 675 else 676 s->status = NAME_STATE; 677 } 678 if (s->status == NAME_STATE) { 679 if (s->gzhead->name != NULL) { 680 uInt beg = s->pending; /* start of bytes to update crc */ 681 int val; 682 683 do { 684 if (s->pending == s->pending_buf_size) { 685 if (s->gzhead->hcrc && s->pending > beg) 686 strm->adler = crc32(strm->adler, s->pending_buf + beg, 687 s->pending - beg); 688 flush_pending(strm); 689 beg = s->pending; 690 if (s->pending == s->pending_buf_size) { 691 val = 1; 692 break; 693 } 694 } 695 val = s->gzhead->name[s->gzindex++]; 696 put_byte(s, val); 697 } while (val != 0); 698 if (s->gzhead->hcrc && s->pending > beg) 699 strm->adler = crc32(strm->adler, s->pending_buf + beg, 700 s->pending - beg); 701 if (val == 0) { 702 s->gzindex = 0; 703 s->status = COMMENT_STATE; 704 } 705 } 706 else 707 s->status = COMMENT_STATE; 708 } 709 if (s->status == COMMENT_STATE) { 710 if (s->gzhead->comment != NULL) { 711 uInt beg = s->pending; /* start of bytes to update crc */ 712 int val; 713 714 do { 715 if (s->pending == s->pending_buf_size) { 716 if (s->gzhead->hcrc && s->pending > beg) 717 strm->adler = crc32(strm->adler, s->pending_buf + beg, 718 s->pending - beg); 719 flush_pending(strm); 720 beg = s->pending; 721 if (s->pending == s->pending_buf_size) { 722 val = 1; 723 break; 724 } 725 } 726 val = s->gzhead->comment[s->gzindex++]; 727 put_byte(s, val); 728 } while (val != 0); 729 if (s->gzhead->hcrc && s->pending > beg) 730 strm->adler = crc32(strm->adler, s->pending_buf + beg, 731 s->pending - beg); 732 if (val == 0) 733 s->status = HCRC_STATE; 734 } 735 else 736 s->status = HCRC_STATE; 737 } 738 if (s->status == HCRC_STATE) { 739 if (s->gzhead->hcrc) { 740 if (s->pending + 2 > s->pending_buf_size) 741 flush_pending(strm); 742 if (s->pending + 2 <= s->pending_buf_size) { 743 put_byte(s, (Byte)(strm->adler & 0xff)); 744 put_byte(s, (Byte)((strm->adler >> 8) & 0xff)); 745 strm->adler = crc32(0L, Z_NULL, 0); 746 s->status = BUSY_STATE; 747 } 748 } 749 else 750 s->status = BUSY_STATE; 751 } 752 #endif 753 754 /* Flush as much pending output as possible */ 755 if (s->pending != 0) { 756 flush_pending(strm); 757 if (strm->avail_out == 0) { 758 /* Since avail_out is 0, deflate will be called again with 759 * more output space, but possibly with both pending and 760 * avail_in equal to zero. There won't be anything to do, 761 * but this is not an error situation so make sure we 762 * return OK instead of BUF_ERROR at next call of deflate: 763 */ 764 s->last_flush = -1; 765 return Z_OK; 766 } 767 768 /* Make sure there is something to do and avoid duplicate consecutive 769 * flushes. For repeated and useless calls with Z_FINISH, we keep 770 * returning Z_STREAM_END instead of Z_BUF_ERROR. 771 */ 772 } else if (strm->avail_in == 0 && flush <= old_flush && 773 flush != Z_FINISH) { 774 ERR_RETURN(strm, Z_BUF_ERROR); 775 } 776 777 /* User must not provide more input after the first FINISH: */ 778 if (s->status == FINISH_STATE && strm->avail_in != 0) { 779 ERR_RETURN(strm, Z_BUF_ERROR); 780 } 781 782 /* Start a new block or continue the current one. 783 */ 784 if (strm->avail_in != 0 || s->lookahead != 0 || 785 (flush != Z_NO_FLUSH && s->status != FINISH_STATE)) { 786 block_state bstate; 787 788 bstate = (*(configuration_table[s->level].func))(s, flush); 789 790 if (bstate == finish_started || bstate == finish_done) { 791 s->status = FINISH_STATE; 792 } 793 if (bstate == need_more || bstate == finish_started) { 794 if (strm->avail_out == 0) { 795 s->last_flush = -1; /* avoid BUF_ERROR next call, see above */ 796 } 797 return Z_OK; 798 /* If flush != Z_NO_FLUSH && avail_out == 0, the next call 799 * of deflate should use the same flush parameter to make sure 800 * that the flush is complete. So we don't have to output an 801 * empty block here, this will be done at next call. This also 802 * ensures that for a very small output buffer, we emit at most 803 * one empty block. 804 */ 805 } 806 if (bstate == block_done) { 807 if (flush == Z_PARTIAL_FLUSH) { 808 _tr_align(s); 809 } else { /* FULL_FLUSH or SYNC_FLUSH */ 810 _tr_stored_block(s, (char*)0, 0L, 0); 811 /* For a full flush, this empty block will be recognized 812 * as a special marker by inflate_sync(). 813 */ 814 if (flush == Z_FULL_FLUSH) { 815 CLEAR_HASH(s); /* forget history */ 816 } 817 } 818 flush_pending(strm); 819 if (strm->avail_out == 0) { 820 s->last_flush = -1; /* avoid BUF_ERROR at next call, see above */ 821 return Z_OK; 822 } 823 } 824 } 825 Assert(strm->avail_out > 0, "bug2"); 826 827 if (flush != Z_FINISH) return Z_OK; 828 if (s->wrap <= 0) return Z_STREAM_END; 829 830 /* Write the trailer */ 831 #ifdef GZIP 832 if (s->wrap == 2) { 833 put_byte(s, (Byte)(strm->adler & 0xff)); 834 put_byte(s, (Byte)((strm->adler >> 8) & 0xff)); 835 put_byte(s, (Byte)((strm->adler >> 16) & 0xff)); 836 put_byte(s, (Byte)((strm->adler >> 24) & 0xff)); 837 put_byte(s, (Byte)(strm->total_in & 0xff)); 838 put_byte(s, (Byte)((strm->total_in >> 8) & 0xff)); 839 put_byte(s, (Byte)((strm->total_in >> 16) & 0xff)); 840 put_byte(s, (Byte)((strm->total_in >> 24) & 0xff)); 841 } 842 else 843 #endif 844 { 845 putShortMSB(s, (uInt)(strm->adler >> 16)); 846 putShortMSB(s, (uInt)(strm->adler & 0xffff)); 847 } 848 flush_pending(strm); 849 /* If avail_out is zero, the application will call deflate again 850 * to flush the rest. 851 */ 852 if (s->wrap > 0) s->wrap = -s->wrap; /* write the trailer only once! */ 853 return s->pending != 0 ? Z_OK : Z_STREAM_END; 854 } 855 856 /* ========================================================================= */ 857 int ZEXPORT deflateEnd (strm) 858 z_streamp strm; 859 { 860 int status; 861 862 if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR; 863 864 status = strm->state->status; 865 if (status != INIT_STATE && 866 status != EXTRA_STATE && 867 status != NAME_STATE && 868 status != COMMENT_STATE && 869 status != HCRC_STATE && 870 status != BUSY_STATE && 871 status != FINISH_STATE) { 872 return Z_STREAM_ERROR; 873 } 874 875 /* Deallocate in reverse order of allocations: */ 876 TRY_FREE(strm, strm->state->pending_buf); 877 TRY_FREE(strm, strm->state->head); 878 TRY_FREE(strm, strm->state->prev); 879 TRY_FREE(strm, strm->state->window); 880 881 ZFREE(strm, strm->state); 882 strm->state = Z_NULL; 883 884 return status == BUSY_STATE ? Z_DATA_ERROR : Z_OK; 885 } 886 887 /* ========================================================================= 888 * Copy the source state to the destination state. 889 * To simplify the source, this is not supported for 16-bit MSDOS (which 890 * doesn't have enough memory anyway to duplicate compression states). 891 */ 892 int ZEXPORT deflateCopy (dest, source) 893 z_streamp dest; 894 z_streamp source; 895 { 896 #ifdef MAXSEG_64K 897 return Z_STREAM_ERROR; 898 #else 899 deflate_state *ds; 900 deflate_state *ss; 901 ushf *overlay; 902 903 904 if (source == Z_NULL || dest == Z_NULL || source->state == Z_NULL) { 905 return Z_STREAM_ERROR; 906 } 907 908 ss = source->state; 909 910 zmemcpy(dest, source, sizeof(z_stream)); 911 912 ds = (deflate_state *) ZALLOC(dest, 1, sizeof(deflate_state)); 913 if (ds == Z_NULL) return Z_MEM_ERROR; 914 dest->state = (struct internal_state FAR *) ds; 915 zmemcpy(ds, ss, sizeof(deflate_state)); 916 ds->strm = dest; 917 918 ds->window = (Bytef *) ZALLOC(dest, ds->w_size, 2*sizeof(Byte)); 919 ds->prev = (Posf *) ZALLOC(dest, ds->w_size, sizeof(Pos)); 920 ds->head = (Posf *) ZALLOC(dest, ds->hash_size, sizeof(Pos)); 921 overlay = (ushf *) ZALLOC(dest, ds->lit_bufsize, sizeof(ush)+2); 922 ds->pending_buf = (uchf *) overlay; 923 924 if (ds->window == Z_NULL || ds->prev == Z_NULL || ds->head == Z_NULL || 925 ds->pending_buf == Z_NULL) { 926 deflateEnd (dest); 927 return Z_MEM_ERROR; 928 } 929 /* following zmemcpy do not work for 16-bit MSDOS */ 930 zmemcpy(ds->window, ss->window, ds->w_size * 2 * sizeof(Byte)); 931 zmemcpy(ds->prev, ss->prev, ds->w_size * sizeof(Pos)); 932 zmemcpy(ds->head, ss->head, ds->hash_size * sizeof(Pos)); 933 zmemcpy(ds->pending_buf, ss->pending_buf, (uInt)ds->pending_buf_size); 934 935 ds->pending_out = ds->pending_buf + (ss->pending_out - ss->pending_buf); 936 ds->d_buf = overlay + ds->lit_bufsize/sizeof(ush); 937 ds->l_buf = ds->pending_buf + (1+sizeof(ush))*ds->lit_bufsize; 938 939 ds->l_desc.dyn_tree = ds->dyn_ltree; 940 ds->d_desc.dyn_tree = ds->dyn_dtree; 941 ds->bl_desc.dyn_tree = ds->bl_tree; 942 943 return Z_OK; 944 #endif /* MAXSEG_64K */ 945 } 946 947 /* =========================================================================== 948 * Read a new buffer from the current input stream, update the adler32 949 * and total number of bytes read. All deflate() input goes through 950 * this function so some applications may wish to modify it to avoid 951 * allocating a large strm->next_in buffer and copying from it. 952 * (See also flush_pending()). 953 */ 954 local int read_buf(strm, buf, size) 955 z_streamp strm; 956 Bytef *buf; 957 unsigned size; 958 { 959 unsigned len = strm->avail_in; 960 961 if (len > size) len = size; 962 if (len == 0) return 0; 963 964 strm->avail_in -= len; 965 966 if (strm->state->wrap == 1) { 967 strm->adler = adler32(strm->adler, strm->next_in, len); 968 } 969 #ifdef GZIP 970 else if (strm->state->wrap == 2) { 971 strm->adler = crc32(strm->adler, strm->next_in, len); 972 } 973 #endif 974 zmemcpy(buf, strm->next_in, len); 975 strm->next_in += len; 976 strm->total_in += len; 977 978 return (int)len; 979 } 980 981 /* =========================================================================== 982 * Initialize the "longest match" routines for a new zlib stream 983 */ 984 local void lm_init (s) 985 deflate_state *s; 986 { 987 s->window_size = (ulg)2L*s->w_size; 988 989 CLEAR_HASH(s); 990 991 /* Set the default configuration parameters: 992 */ 993 s->max_lazy_match = configuration_table[s->level].max_lazy; 994 s->good_match = configuration_table[s->level].good_length; 995 s->nice_match = configuration_table[s->level].nice_length; 996 s->max_chain_length = configuration_table[s->level].max_chain; 997 998 s->strstart = 0; 999 s->block_start = 0L; 1000 s->lookahead = 0; 1001 s->match_length = s->prev_length = MIN_MATCH-1; 1002 s->match_available = 0; 1003 s->ins_h = 0; 1004 #ifndef FASTEST 1005 #ifdef ASMV 1006 match_init(); /* initialize the asm code */ 1007 #endif 1008 #endif 1009 } 1010 1011 #ifndef FASTEST 1012 /* =========================================================================== 1013 * Set match_start to the longest match starting at the given string and 1014 * return its length. Matches shorter or equal to prev_length are discarded, 1015 * in which case the result is equal to prev_length and match_start is 1016 * garbage. 1017 * IN assertions: cur_match is the head of the hash chain for the current 1018 * string (strstart) and its distance is <= MAX_DIST, and prev_length >= 1 1019 * OUT assertion: the match length is not greater than s->lookahead. 1020 */ 1021 #ifndef ASMV 1022 /* For 80x86 and 680x0, an optimized version will be provided in match.asm or 1023 * match.S. The code will be functionally equivalent. 1024 */ 1025 local uInt longest_match(s, cur_match) 1026 deflate_state *s; 1027 IPos cur_match; /* current match */ 1028 { 1029 unsigned chain_length = s->max_chain_length;/* max hash chain length */ 1030 register Bytef *scan = s->window + s->strstart; /* current string */ 1031 register Bytef *match; /* matched string */ 1032 register int len; /* length of current match */ 1033 int best_len = s->prev_length; /* best match length so far */ 1034 int nice_match = s->nice_match; /* stop if match long enough */ 1035 IPos limit = s->strstart > (IPos)MAX_DIST(s) ? 1036 s->strstart - (IPos)MAX_DIST(s) : NIL; 1037 /* Stop when cur_match becomes <= limit. To simplify the code, 1038 * we prevent matches with the string of window index 0. 1039 */ 1040 Posf *prev = s->prev; 1041 uInt wmask = s->w_mask; 1042 1043 #ifdef UNALIGNED_OK 1044 /* Compare two bytes at a time. Note: this is not always beneficial. 1045 * Try with and without -DUNALIGNED_OK to check. 1046 */ 1047 register Bytef *strend = s->window + s->strstart + MAX_MATCH - 1; 1048 register ush scan_start = *(ushf*)scan; 1049 register ush scan_end = *(ushf*)(scan+best_len-1); 1050 #else 1051 register Bytef *strend = s->window + s->strstart + MAX_MATCH; 1052 register Byte scan_end1 = scan[best_len-1]; 1053 register Byte scan_end = scan[best_len]; 1054 #endif 1055 1056 /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16. 1057 * It is easy to get rid of this optimization if necessary. 1058 */ 1059 Assert(s->hash_bits >= 8 && MAX_MATCH == 258, "Code too clever"); 1060 1061 /* Do not waste too much time if we already have a good match: */ 1062 if (s->prev_length >= s->good_match) { 1063 chain_length >>= 2; 1064 } 1065 /* Do not look for matches beyond the end of the input. This is necessary 1066 * to make deflate deterministic. 1067 */ 1068 if ((uInt)nice_match > s->lookahead) nice_match = s->lookahead; 1069 1070 Assert((ulg)s->strstart <= s->window_size-MIN_LOOKAHEAD, "need lookahead"); 1071 1072 do { 1073 Assert(cur_match < s->strstart, "no future"); 1074 match = s->window + cur_match; 1075 1076 /* Skip to next match if the match length cannot increase 1077 * or if the match length is less than 2. Note that the checks below 1078 * for insufficient lookahead only occur occasionally for performance 1079 * reasons. Therefore uninitialized memory will be accessed, and 1080 * conditional jumps will be made that depend on those values. 1081 * However the length of the match is limited to the lookahead, so 1082 * the output of deflate is not affected by the uninitialized values. 1083 */ 1084 #if (defined(UNALIGNED_OK) && MAX_MATCH == 258) 1085 /* This code assumes sizeof(unsigned short) == 2. Do not use 1086 * UNALIGNED_OK if your compiler uses a different size. 1087 */ 1088 if (*(ushf*)(match+best_len-1) != scan_end || 1089 *(ushf*)match != scan_start) continue; 1090 1091 /* It is not necessary to compare scan[2] and match[2] since they are 1092 * always equal when the other bytes match, given that the hash keys 1093 * are equal and that HASH_BITS >= 8. Compare 2 bytes at a time at 1094 * strstart+3, +5, ... up to strstart+257. We check for insufficient 1095 * lookahead only every 4th comparison; the 128th check will be made 1096 * at strstart+257. If MAX_MATCH-2 is not a multiple of 8, it is 1097 * necessary to put more guard bytes at the end of the window, or 1098 * to check more often for insufficient lookahead. 1099 */ 1100 Assert(scan[2] == match[2], "scan[2]?"); 1101 scan++, match++; 1102 do { 1103 } while (*(ushf*)(scan+=2) == *(ushf*)(match+=2) && 1104 *(ushf*)(scan+=2) == *(ushf*)(match+=2) && 1105 *(ushf*)(scan+=2) == *(ushf*)(match+=2) && 1106 *(ushf*)(scan+=2) == *(ushf*)(match+=2) && 1107 scan < strend); 1108 /* The funny "do {}" generates better code on most compilers */ 1109 1110 /* Here, scan <= window+strstart+257 */ 1111 Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan"); 1112 if (*scan == *match) scan++; 1113 1114 len = (MAX_MATCH - 1) - (int)(strend-scan); 1115 scan = strend - (MAX_MATCH-1); 1116 1117 #else /* UNALIGNED_OK */ 1118 1119 if (match[best_len] != scan_end || 1120 match[best_len-1] != scan_end1 || 1121 *match != *scan || 1122 *++match != scan[1]) continue; 1123 1124 /* The check at best_len-1 can be removed because it will be made 1125 * again later. (This heuristic is not always a win.) 1126 * It is not necessary to compare scan[2] and match[2] since they 1127 * are always equal when the other bytes match, given that 1128 * the hash keys are equal and that HASH_BITS >= 8. 1129 */ 1130 scan += 2, match++; 1131 Assert(*scan == *match, "match[2]?"); 1132 1133 /* We check for insufficient lookahead only every 8th comparison; 1134 * the 256th check will be made at strstart+258. 1135 */ 1136 do { 1137 } while (*++scan == *++match && *++scan == *++match && 1138 *++scan == *++match && *++scan == *++match && 1139 *++scan == *++match && *++scan == *++match && 1140 *++scan == *++match && *++scan == *++match && 1141 scan < strend); 1142 1143 Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan"); 1144 1145 len = MAX_MATCH - (int)(strend - scan); 1146 scan = strend - MAX_MATCH; 1147 1148 #endif /* UNALIGNED_OK */ 1149 1150 if (len > best_len) { 1151 s->match_start = cur_match; 1152 best_len = len; 1153 if (len >= nice_match) break; 1154 #ifdef UNALIGNED_OK 1155 scan_end = *(ushf*)(scan+best_len-1); 1156 #else 1157 scan_end1 = scan[best_len-1]; 1158 scan_end = scan[best_len]; 1159 #endif 1160 } 1161 } while ((cur_match = prev[cur_match & wmask]) > limit 1162 && --chain_length != 0); 1163 1164 if ((uInt)best_len <= s->lookahead) return (uInt)best_len; 1165 return s->lookahead; 1166 } 1167 #endif /* ASMV */ 1168 #endif /* FASTEST */ 1169 1170 /* --------------------------------------------------------------------------- 1171 * Optimized version for level == 1 or strategy == Z_RLE only 1172 */ 1173 local uInt longest_match_fast(s, cur_match) 1174 deflate_state *s; 1175 IPos cur_match; /* current match */ 1176 { 1177 register Bytef *scan = s->window + s->strstart; /* current string */ 1178 register Bytef *match; /* matched string */ 1179 register int len; /* length of current match */ 1180 register Bytef *strend = s->window + s->strstart + MAX_MATCH; 1181 1182 /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16. 1183 * It is easy to get rid of this optimization if necessary. 1184 */ 1185 Assert(s->hash_bits >= 8 && MAX_MATCH == 258, "Code too clever"); 1186 1187 Assert((ulg)s->strstart <= s->window_size-MIN_LOOKAHEAD, "need lookahead"); 1188 1189 Assert(cur_match < s->strstart, "no future"); 1190 1191 match = s->window + cur_match; 1192 1193 /* Return failure if the match length is less than 2: 1194 */ 1195 if (match[0] != scan[0] || match[1] != scan[1]) return MIN_MATCH-1; 1196 1197 /* The check at best_len-1 can be removed because it will be made 1198 * again later. (This heuristic is not always a win.) 1199 * It is not necessary to compare scan[2] and match[2] since they 1200 * are always equal when the other bytes match, given that 1201 * the hash keys are equal and that HASH_BITS >= 8. 1202 */ 1203 scan += 2, match += 2; 1204 Assert(*scan == *match, "match[2]?"); 1205 1206 /* We check for insufficient lookahead only every 8th comparison; 1207 * the 256th check will be made at strstart+258. 1208 */ 1209 do { 1210 } while (*++scan == *++match && *++scan == *++match && 1211 *++scan == *++match && *++scan == *++match && 1212 *++scan == *++match && *++scan == *++match && 1213 *++scan == *++match && *++scan == *++match && 1214 scan < strend); 1215 1216 Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan"); 1217 1218 len = MAX_MATCH - (int)(strend - scan); 1219 1220 if (len < MIN_MATCH) return MIN_MATCH - 1; 1221 1222 s->match_start = cur_match; 1223 return (uInt)len <= s->lookahead ? (uInt)len : s->lookahead; 1224 } 1225 1226 #ifdef DEBUG 1227 /* =========================================================================== 1228 * Check that the match at match_start is indeed a match. 1229 */ 1230 local void check_match(s, start, match, length) 1231 deflate_state *s; 1232 IPos start, match; 1233 int length; 1234 { 1235 /* check that the match is indeed a match */ 1236 if (zmemcmp(s->window + match, 1237 s->window + start, length) != EQUAL) { 1238 fprintf(stderr, " start %u, match %u, length %d\n", 1239 start, match, length); 1240 do { 1241 fprintf(stderr, "%c%c", s->window[match++], s->window[start++]); 1242 } while (--length != 0); 1243 z_error("invalid match"); 1244 } 1245 if (z_verbose > 1) { 1246 fprintf(stderr,"\\[%d,%d]", start-match, length); 1247 do { putc(s->window[start++], stderr); } while (--length != 0); 1248 } 1249 } 1250 #else 1251 # define check_match(s, start, match, length) 1252 #endif /* DEBUG */ 1253 1254 /* =========================================================================== 1255 * Fill the window when the lookahead becomes insufficient. 1256 * Updates strstart and lookahead. 1257 * 1258 * IN assertion: lookahead < MIN_LOOKAHEAD 1259 * OUT assertions: strstart <= window_size-MIN_LOOKAHEAD 1260 * At least one byte has been read, or avail_in == 0; reads are 1261 * performed for at least two bytes (required for the zip translate_eol 1262 * option -- not supported here). 1263 */ 1264 local void fill_window(s) 1265 deflate_state *s; 1266 { 1267 register unsigned n, m; 1268 register Posf *p; 1269 unsigned more; /* Amount of free space at the end of the window. */ 1270 uInt wsize = s->w_size; 1271 1272 do { 1273 more = (unsigned)(s->window_size -(ulg)s->lookahead -(ulg)s->strstart); 1274 1275 /* Deal with !@#$% 64K limit: */ 1276 if (sizeof(int) <= 2) { 1277 if (more == 0 && s->strstart == 0 && s->lookahead == 0) { 1278 more = wsize; 1279 1280 } else if (more == (unsigned)(-1)) { 1281 /* Very unlikely, but possible on 16 bit machine if 1282 * strstart == 0 && lookahead == 1 (input done a byte at time) 1283 */ 1284 more--; 1285 } 1286 } 1287 1288 /* If the window is almost full and there is insufficient lookahead, 1289 * move the upper half to the lower one to make room in the upper half. 1290 */ 1291 if (s->strstart >= wsize+MAX_DIST(s)) { 1292 1293 zmemcpy(s->window, s->window+wsize, (unsigned)wsize); 1294 s->match_start -= wsize; 1295 s->strstart -= wsize; /* we now have strstart >= MAX_DIST */ 1296 s->block_start -= (long) wsize; 1297 1298 /* Slide the hash table (could be avoided with 32 bit values 1299 at the expense of memory usage). We slide even when level == 0 1300 to keep the hash table consistent if we switch back to level > 0 1301 later. (Using level 0 permanently is not an optimal usage of 1302 zlib, so we don't care about this pathological case.) 1303 */ 1304 /* %%% avoid this when Z_RLE */ 1305 n = s->hash_size; 1306 p = &s->head[n]; 1307 do { 1308 m = *--p; 1309 *p = (Pos)(m >= wsize ? m-wsize : NIL); 1310 } while (--n); 1311 1312 n = wsize; 1313 #ifndef FASTEST 1314 p = &s->prev[n]; 1315 do { 1316 m = *--p; 1317 *p = (Pos)(m >= wsize ? m-wsize : NIL); 1318 /* If n is not on any hash chain, prev[n] is garbage but 1319 * its value will never be used. 1320 */ 1321 } while (--n); 1322 #endif 1323 more += wsize; 1324 } 1325 if (s->strm->avail_in == 0) return; 1326 1327 /* If there was no sliding: 1328 * strstart <= WSIZE+MAX_DIST-1 && lookahead <= MIN_LOOKAHEAD - 1 && 1329 * more == window_size - lookahead - strstart 1330 * => more >= window_size - (MIN_LOOKAHEAD-1 + WSIZE + MAX_DIST-1) 1331 * => more >= window_size - 2*WSIZE + 2 1332 * In the BIG_MEM or MMAP case (not yet supported), 1333 * window_size == input_size + MIN_LOOKAHEAD && 1334 * strstart + s->lookahead <= input_size => more >= MIN_LOOKAHEAD. 1335 * Otherwise, window_size == 2*WSIZE so more >= 2. 1336 * If there was sliding, more >= WSIZE. So in all cases, more >= 2. 1337 */ 1338 Assert(more >= 2, "more < 2"); 1339 1340 n = read_buf(s->strm, s->window + s->strstart + s->lookahead, more); 1341 s->lookahead += n; 1342 1343 /* Initialize the hash value now that we have some input: */ 1344 if (s->lookahead >= MIN_MATCH) { 1345 s->ins_h = s->window[s->strstart]; 1346 UPDATE_HASH(s, s->ins_h, s->window[s->strstart+1]); 1347 #if MIN_MATCH != 3 1348 Call UPDATE_HASH() MIN_MATCH-3 more times 1349 #endif 1350 } 1351 /* If the whole input has less than MIN_MATCH bytes, ins_h is garbage, 1352 * but this is not important since only literal bytes will be emitted. 1353 */ 1354 1355 } while (s->lookahead < MIN_LOOKAHEAD && s->strm->avail_in != 0); 1356 } 1357 1358 /* =========================================================================== 1359 * Flush the current block, with given end-of-file flag. 1360 * IN assertion: strstart is set to the end of the current match. 1361 */ 1362 #define FLUSH_BLOCK_ONLY(s, eof) { \ 1363 _tr_flush_block(s, (s->block_start >= 0L ? \ 1364 (charf *)&s->window[(unsigned)s->block_start] : \ 1365 (charf *)Z_NULL), \ 1366 (ulg)((long)s->strstart - s->block_start), \ 1367 (eof)); \ 1368 s->block_start = s->strstart; \ 1369 flush_pending(s->strm); \ 1370 Tracev((stderr,"[FLUSH]")); \ 1371 } 1372 1373 /* Same but force premature exit if necessary. */ 1374 #define FLUSH_BLOCK(s, eof) { \ 1375 FLUSH_BLOCK_ONLY(s, eof); \ 1376 if (s->strm->avail_out == 0) return (eof) ? finish_started : need_more; \ 1377 } 1378 1379 /* =========================================================================== 1380 * Copy without compression as much as possible from the input stream, return 1381 * the current block state. 1382 * This function does not insert new strings in the dictionary since 1383 * uncompressible data is probably not useful. This function is used 1384 * only for the level=0 compression option. 1385 * NOTE: this function should be optimized to avoid extra copying from 1386 * window to pending_buf. 1387 */ 1388 local block_state deflate_stored(s, flush) 1389 deflate_state *s; 1390 int flush; 1391 { 1392 /* Stored blocks are limited to 0xffff bytes, pending_buf is limited 1393 * to pending_buf_size, and each stored block has a 5 byte header: 1394 */ 1395 ulg max_block_size = 0xffff; 1396 ulg max_start; 1397 1398 if (max_block_size > s->pending_buf_size - 5) { 1399 max_block_size = s->pending_buf_size - 5; 1400 } 1401 1402 /* Copy as much as possible from input to output: */ 1403 for (;;) { 1404 /* Fill the window as much as possible: */ 1405 if (s->lookahead <= 1) { 1406 1407 Assert(s->strstart < s->w_size+MAX_DIST(s) || 1408 s->block_start >= (long)s->w_size, "slide too late"); 1409 1410 fill_window(s); 1411 if (s->lookahead == 0 && flush == Z_NO_FLUSH) return need_more; 1412 1413 if (s->lookahead == 0) break; /* flush the current block */ 1414 } 1415 Assert(s->block_start >= 0L, "block gone"); 1416 1417 s->strstart += s->lookahead; 1418 s->lookahead = 0; 1419 1420 /* Emit a stored block if pending_buf will be full: */ 1421 max_start = s->block_start + max_block_size; 1422 if (s->strstart == 0 || (ulg)s->strstart >= max_start) { 1423 /* strstart == 0 is possible when wraparound on 16-bit machine */ 1424 s->lookahead = (uInt)(s->strstart - max_start); 1425 s->strstart = (uInt)max_start; 1426 FLUSH_BLOCK(s, 0); 1427 } 1428 /* Flush if we may have to slide, otherwise block_start may become 1429 * negative and the data will be gone: 1430 */ 1431 if (s->strstart - (uInt)s->block_start >= MAX_DIST(s)) { 1432 FLUSH_BLOCK(s, 0); 1433 } 1434 } 1435 FLUSH_BLOCK(s, flush == Z_FINISH); 1436 return flush == Z_FINISH ? finish_done : block_done; 1437 } 1438 1439 /* =========================================================================== 1440 * Compress as much as possible from the input stream, return the current 1441 * block state. 1442 * This function does not perform lazy evaluation of matches and inserts 1443 * new strings in the dictionary only for unmatched strings or for short 1444 * matches. It is used only for the fast compression options. 1445 */ 1446 local block_state deflate_fast(s, flush) 1447 deflate_state *s; 1448 int flush; 1449 { 1450 IPos hash_head = NIL; /* head of the hash chain */ 1451 int bflush; /* set if current block must be flushed */ 1452 1453 for (;;) { 1454 /* Make sure that we always have enough lookahead, except 1455 * at the end of the input file. We need MAX_MATCH bytes 1456 * for the next match, plus MIN_MATCH bytes to insert the 1457 * string following the next match. 1458 */ 1459 if (s->lookahead < MIN_LOOKAHEAD) { 1460 fill_window(s); 1461 if (s->lookahead < MIN_LOOKAHEAD && flush == Z_NO_FLUSH) { 1462 return need_more; 1463 } 1464 if (s->lookahead == 0) break; /* flush the current block */ 1465 } 1466 1467 /* Insert the string window[strstart .. strstart+2] in the 1468 * dictionary, and set hash_head to the head of the hash chain: 1469 */ 1470 if (s->lookahead >= MIN_MATCH) { 1471 INSERT_STRING(s, s->strstart, hash_head); 1472 } 1473 1474 /* Find the longest match, discarding those <= prev_length. 1475 * At this point we have always match_length < MIN_MATCH 1476 */ 1477 if (hash_head != NIL && s->strstart - hash_head <= MAX_DIST(s)) { 1478 /* To simplify the code, we prevent matches with the string 1479 * of window index 0 (in particular we have to avoid a match 1480 * of the string with itself at the start of the input file). 1481 */ 1482 #ifdef FASTEST 1483 if ((s->strategy != Z_HUFFMAN_ONLY && s->strategy != Z_RLE) || 1484 (s->strategy == Z_RLE && s->strstart - hash_head == 1)) { 1485 s->match_length = longest_match_fast (s, hash_head); 1486 } 1487 #else 1488 if (s->strategy != Z_HUFFMAN_ONLY && s->strategy != Z_RLE) { 1489 s->match_length = longest_match (s, hash_head); 1490 } else if (s->strategy == Z_RLE && s->strstart - hash_head == 1) { 1491 s->match_length = longest_match_fast (s, hash_head); 1492 } 1493 #endif 1494 /* longest_match() or longest_match_fast() sets match_start */ 1495 } 1496 if (s->match_length >= MIN_MATCH) { 1497 check_match(s, s->strstart, s->match_start, s->match_length); 1498 1499 _tr_tally_dist(s, s->strstart - s->match_start, 1500 s->match_length - MIN_MATCH, bflush); 1501 1502 s->lookahead -= s->match_length; 1503 1504 /* Insert new strings in the hash table only if the match length 1505 * is not too large. This saves time but degrades compression. 1506 */ 1507 #ifndef FASTEST 1508 if (s->match_length <= s->max_insert_length && 1509 s->lookahead >= MIN_MATCH) { 1510 s->match_length--; /* string at strstart already in table */ 1511 do { 1512 s->strstart++; 1513 INSERT_STRING(s, s->strstart, hash_head); 1514 /* strstart never exceeds WSIZE-MAX_MATCH, so there are 1515 * always MIN_MATCH bytes ahead. 1516 */ 1517 } while (--s->match_length != 0); 1518 s->strstart++; 1519 } else 1520 #endif 1521 { 1522 s->strstart += s->match_length; 1523 s->match_length = 0; 1524 s->ins_h = s->window[s->strstart]; 1525 UPDATE_HASH(s, s->ins_h, s->window[s->strstart+1]); 1526 #if MIN_MATCH != 3 1527 Call UPDATE_HASH() MIN_MATCH-3 more times 1528 #endif 1529 /* If lookahead < MIN_MATCH, ins_h is garbage, but it does not 1530 * matter since it will be recomputed at next deflate call. 1531 */ 1532 } 1533 } else { 1534 /* No match, output a literal byte */ 1535 Tracevv((stderr,"%c", s->window[s->strstart])); 1536 _tr_tally_lit (s, s->window[s->strstart], bflush); 1537 s->lookahead--; 1538 s->strstart++; 1539 } 1540 if (bflush) FLUSH_BLOCK(s, 0); 1541 } 1542 FLUSH_BLOCK(s, flush == Z_FINISH); 1543 return flush == Z_FINISH ? finish_done : block_done; 1544 } 1545 1546 #ifndef FASTEST 1547 /* =========================================================================== 1548 * Same as above, but achieves better compression. We use a lazy 1549 * evaluation for matches: a match is finally adopted only if there is 1550 * no better match at the next window position. 1551 */ 1552 local block_state deflate_slow(s, flush) 1553 deflate_state *s; 1554 int flush; 1555 { 1556 IPos hash_head = NIL; /* head of hash chain */ 1557 int bflush; /* set if current block must be flushed */ 1558 1559 /* Process the input block. */ 1560 for (;;) { 1561 /* Make sure that we always have enough lookahead, except 1562 * at the end of the input file. We need MAX_MATCH bytes 1563 * for the next match, plus MIN_MATCH bytes to insert the 1564 * string following the next match. 1565 */ 1566 if (s->lookahead < MIN_LOOKAHEAD) { 1567 fill_window(s); 1568 if (s->lookahead < MIN_LOOKAHEAD && flush == Z_NO_FLUSH) { 1569 return need_more; 1570 } 1571 if (s->lookahead == 0) break; /* flush the current block */ 1572 } 1573 1574 /* Insert the string window[strstart .. strstart+2] in the 1575 * dictionary, and set hash_head to the head of the hash chain: 1576 */ 1577 if (s->lookahead >= MIN_MATCH) { 1578 INSERT_STRING(s, s->strstart, hash_head); 1579 } 1580 1581 /* Find the longest match, discarding those <= prev_length. 1582 */ 1583 s->prev_length = s->match_length, s->prev_match = s->match_start; 1584 s->match_length = MIN_MATCH-1; 1585 1586 if (hash_head != NIL && s->prev_length < s->max_lazy_match && 1587 s->strstart - hash_head <= MAX_DIST(s)) { 1588 /* To simplify the code, we prevent matches with the string 1589 * of window index 0 (in particular we have to avoid a match 1590 * of the string with itself at the start of the input file). 1591 */ 1592 if (s->strategy != Z_HUFFMAN_ONLY && s->strategy != Z_RLE) { 1593 s->match_length = longest_match (s, hash_head); 1594 } else if (s->strategy == Z_RLE && s->strstart - hash_head == 1) { 1595 s->match_length = longest_match_fast (s, hash_head); 1596 } 1597 /* longest_match() or longest_match_fast() sets match_start */ 1598 1599 if (s->match_length <= 5 && (s->strategy == Z_FILTERED 1600 #if TOO_FAR <= 32767 1601 || (s->match_length == MIN_MATCH && 1602 s->strstart - s->match_start > TOO_FAR) 1603 #endif 1604 )) { 1605 1606 /* If prev_match is also MIN_MATCH, match_start is garbage 1607 * but we will ignore the current match anyway. 1608 */ 1609 s->match_length = MIN_MATCH-1; 1610 } 1611 } 1612 /* If there was a match at the previous step and the current 1613 * match is not better, output the previous match: 1614 */ 1615 if (s->prev_length >= MIN_MATCH && s->match_length <= s->prev_length) { 1616 uInt max_insert = s->strstart + s->lookahead - MIN_MATCH; 1617 /* Do not insert strings in hash table beyond this. */ 1618 1619 check_match(s, s->strstart-1, s->prev_match, s->prev_length); 1620 1621 _tr_tally_dist(s, s->strstart -1 - s->prev_match, 1622 s->prev_length - MIN_MATCH, bflush); 1623 1624 /* Insert in hash table all strings up to the end of the match. 1625 * strstart-1 and strstart are already inserted. If there is not 1626 * enough lookahead, the last two strings are not inserted in 1627 * the hash table. 1628 */ 1629 s->lookahead -= s->prev_length-1; 1630 s->prev_length -= 2; 1631 do { 1632 if (++s->strstart <= max_insert) { 1633 INSERT_STRING(s, s->strstart, hash_head); 1634 } 1635 } while (--s->prev_length != 0); 1636 s->match_available = 0; 1637 s->match_length = MIN_MATCH-1; 1638 s->strstart++; 1639 1640 if (bflush) FLUSH_BLOCK(s, 0); 1641 1642 } else if (s->match_available) { 1643 /* If there was no match at the previous position, output a 1644 * single literal. If there was a match but the current match 1645 * is longer, truncate the previous match to a single literal. 1646 */ 1647 Tracevv((stderr,"%c", s->window[s->strstart-1])); 1648 _tr_tally_lit(s, s->window[s->strstart-1], bflush); 1649 if (bflush) { 1650 FLUSH_BLOCK_ONLY(s, 0); 1651 } 1652 s->strstart++; 1653 s->lookahead--; 1654 if (s->strm->avail_out == 0) return need_more; 1655 } else { 1656 /* There is no previous match to compare with, wait for 1657 * the next step to decide. 1658 */ 1659 s->match_available = 1; 1660 s->strstart++; 1661 s->lookahead--; 1662 } 1663 } 1664 Assert (flush != Z_NO_FLUSH, "no flush?"); 1665 if (s->match_available) { 1666 Tracevv((stderr,"%c", s->window[s->strstart-1])); 1667 _tr_tally_lit(s, s->window[s->strstart-1], bflush); 1668 s->match_available = 0; 1669 } 1670 FLUSH_BLOCK(s, flush == Z_FINISH); 1671 return flush == Z_FINISH ? finish_done : block_done; 1672 } 1673 #endif /* FASTEST */ 1674 1675 #if 0 1676 /* =========================================================================== 1677 * For Z_RLE, simply look for runs of bytes, generate matches only of distance 1678 * one. Do not maintain a hash table. (It will be regenerated if this run of 1679 * deflate switches away from Z_RLE.) 1680 */ 1681 local block_state deflate_rle(s, flush) 1682 deflate_state *s; 1683 int flush; 1684 { 1685 int bflush; /* set if current block must be flushed */ 1686 uInt run; /* length of run */ 1687 uInt max; /* maximum length of run */ 1688 uInt prev; /* byte at distance one to match */ 1689 Bytef *scan; /* scan for end of run */ 1690 1691 for (;;) { 1692 /* Make sure that we always have enough lookahead, except 1693 * at the end of the input file. We need MAX_MATCH bytes 1694 * for the longest encodable run. 1695 */ 1696 if (s->lookahead < MAX_MATCH) { 1697 fill_window(s); 1698 if (s->lookahead < MAX_MATCH && flush == Z_NO_FLUSH) { 1699 return need_more; 1700 } 1701 if (s->lookahead == 0) break; /* flush the current block */ 1702 } 1703 1704 /* See how many times the previous byte repeats */ 1705 run = 0; 1706 if (s->strstart > 0) { /* if there is a previous byte, that is */ 1707 max = s->lookahead < MAX_MATCH ? s->lookahead : MAX_MATCH; 1708 scan = s->window + s->strstart - 1; 1709 prev = *scan++; 1710 do { 1711 if (*scan++ != prev) 1712 break; 1713 } while (++run < max); 1714 } 1715 1716 /* Emit match if have run of MIN_MATCH or longer, else emit literal */ 1717 if (run >= MIN_MATCH) { 1718 check_match(s, s->strstart, s->strstart - 1, run); 1719 _tr_tally_dist(s, 1, run - MIN_MATCH, bflush); 1720 s->lookahead -= run; 1721 s->strstart += run; 1722 } else { 1723 /* No match, output a literal byte */ 1724 Tracevv((stderr,"%c", s->window[s->strstart])); 1725 _tr_tally_lit (s, s->window[s->strstart], bflush); 1726 s->lookahead--; 1727 s->strstart++; 1728 } 1729 if (bflush) FLUSH_BLOCK(s, 0); 1730 } 1731 FLUSH_BLOCK(s, flush == Z_FINISH); 1732 return flush == Z_FINISH ? finish_done : block_done; 1733 } 1734 #endif 1735