diff options
Diffstat (limited to 'dep/src/zlib/deflate.c')
| -rw-r--r-- | dep/src/zlib/deflate.c | 186 |
1 files changed, 186 insertions, 0 deletions
diff --git a/dep/src/zlib/deflate.c b/dep/src/zlib/deflate.c index 0651ae4cd8a..29ce1f64a57 100644 --- a/dep/src/zlib/deflate.c +++ b/dep/src/zlib/deflate.c @@ -2,6 +2,7 @@ * Copyright (C) 1995-2005 Jean-loup Gailly. * For conditions of distribution and use, see copyright notice in zlib.h */ + /* * ALGORITHM * @@ -45,8 +46,11 @@ * Data Compression with Finite Windows, Comm.ACM, 32,4 (1989) 490-595 * */ + /* @(#) $Id$ */ + #include "deflate.h" + const char deflate_copyright[] = " deflate 1.2.3 Copyright 1995-2005 Jean-loup Gailly "; /* @@ -55,6 +59,7 @@ const char deflate_copyright[] = include such an acknowledgment, I would appreciate that you keep this copyright string in the executable of your product. */ + /* =========================================================================== * Function prototypes. */ @@ -64,8 +69,10 @@ typedef enum { finish_started, /* finish started, need only more output at next deflate */ finish_done /* finish done, accept no more input or output */ } block_state; + typedef block_state (*compress_func) OF((deflate_state *s, int flush)); /* Compression function. Returns the block state after the call. */ + local void fill_window OF((deflate_state *s)); local block_state deflate_stored OF((deflate_state *s, int flush)); local block_state deflate_fast OF((deflate_state *s, int flush)); @@ -85,23 +92,29 @@ local uInt longest_match OF((deflate_state *s, IPos cur_match)); #endif #endif local uInt longest_match_fast OF((deflate_state *s, IPos cur_match)); + #ifdef DEBUG local void check_match OF((deflate_state *s, IPos start, IPos match, int length)); #endif + /* =========================================================================== * Local data */ + #define NIL 0 /* Tail of hash chains */ + #ifndef TOO_FAR # define TOO_FAR 4096 #endif /* Matches of length 3 are discarded if their distance exceeds TOO_FAR */ + #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1) /* Minimum amount of lookahead, except at the end of the input file. * See deflate.c for comments about the MIN_MATCH+1. */ + /* Values for max_lazy_match, good_match and max_chain_length, depending on * the desired pack level (0..9). The values given below have been tuned to * exclude worst case performance for pathological files. Better values may be @@ -114,6 +127,7 @@ typedef struct config_s { ush max_chain; compress_func func; } config; + #ifdef FASTEST local const config configuration_table[2] = { /* good lazy nice chain */ @@ -126,6 +140,7 @@ local const config configuration_table[10] = { /* 1 */ {4, 4, 8, 4, deflate_fast}, /* max speed, no lazy matches */ /* 2 */ {4, 5, 16, 8, deflate_fast}, /* 3 */ {4, 6, 32, 32, deflate_fast}, + /* 4 */ {4, 4, 16, 16, deflate_slow}, /* lazy matches */ /* 5 */ {8, 16, 32, 32, deflate_slow}, /* 6 */ {8, 16, 128, 128, deflate_slow}, @@ -133,15 +148,19 @@ local const config configuration_table[10] = { /* 8 */ {32, 128, 258, 1024, deflate_slow}, /* 9 */ {32, 258, 258, 4096, deflate_slow}}; /* max compression */ #endif + /* Note: the deflate() code requires max_lazy >= MIN_MATCH and max_chain >= 4 * For deflate_fast() (levels <= 3) good is ignored and lazy has a different * meaning. */ + #define EQUAL 0 /* result of memcmp for equal strings */ + #ifndef NO_DUMMY_DECL struct static_tree_desc_s {int dummy;}; /* for buggy compilers */ #endif + /* =========================================================================== * Update a hash value with the given input byte * IN assertion: all calls to to UPDATE_HASH are made with consecutive @@ -150,6 +169,7 @@ struct static_tree_desc_s {int dummy;}; /* for buggy compilers */ */ #define UPDATE_HASH(s,h,c) (h = (((h)<<s->hash_shift) ^ (c)) & s->hash_mask) + /* =========================================================================== * Insert string str in the dictionary and set match_head to the previous head * of the hash chain (the most recent string with same hash key). Return @@ -171,6 +191,7 @@ struct static_tree_desc_s {int dummy;}; /* for buggy compilers */ match_head = s->prev[(str) & s->w_mask] = s->head[s->ins_h], \ s->head[s->ins_h] = (Pos)(str)) #endif + /* =========================================================================== * Initialize the hash table (avoiding 64K overflow for 16 bit systems). * prev[] will be initialized on the fly. @@ -178,6 +199,7 @@ struct static_tree_desc_s {int dummy;}; /* for buggy compilers */ #define CLEAR_HASH(s) \ s->head[s->hash_size-1] = NIL; \ zmemzero((Bytef *)s->head, (unsigned)(s->hash_size-1)*sizeof(*s->head)); + /* ========================================================================= */ int ZEXPORT deflateInit_(strm, level, version, stream_size) z_streamp strm; @@ -189,6 +211,7 @@ int ZEXPORT deflateInit_(strm, level, version, stream_size) Z_DEFAULT_STRATEGY, version, stream_size); /* To do: ignore strm->next_in if we use it as window */ } + /* ========================================================================= */ int ZEXPORT deflateInit2_(strm, level, method, windowBits, memLevel, strategy, version, stream_size) @@ -204,26 +227,31 @@ int ZEXPORT deflateInit2_(strm, level, method, windowBits, memLevel, strategy, deflate_state *s; int wrap = 1; static const char my_version[] = ZLIB_VERSION; + ushf *overlay; /* We overlay pending_buf and d_buf+l_buf. This works since the average * output size for (length,distance) codes is <= 24 bits. */ + if (version == Z_NULL || version[0] != my_version[0] || stream_size != sizeof(z_stream)) { return Z_VERSION_ERROR; } if (strm == Z_NULL) return Z_STREAM_ERROR; + strm->msg = Z_NULL; if (strm->zalloc == (alloc_func)0) { strm->zalloc = zcalloc; strm->opaque = (voidpf)0; } if (strm->zfree == (free_func)0) strm->zfree = zcfree; + #ifdef FASTEST if (level != 0) level = 1; #else if (level == Z_DEFAULT_COMPRESSION) level = 6; #endif + if (windowBits < 0) { /* suppress zlib wrapper */ wrap = 0; windowBits = -windowBits; @@ -244,22 +272,28 @@ int ZEXPORT deflateInit2_(strm, level, method, windowBits, memLevel, strategy, if (s == Z_NULL) return Z_MEM_ERROR; strm->state = (struct internal_state FAR *)s; s->strm = strm; + s->wrap = wrap; s->gzhead = Z_NULL; s->w_bits = windowBits; s->w_size = 1 << s->w_bits; s->w_mask = s->w_size - 1; + s->hash_bits = memLevel + 7; s->hash_size = 1 << s->hash_bits; s->hash_mask = s->hash_size - 1; s->hash_shift = ((s->hash_bits+MIN_MATCH-1)/MIN_MATCH); + s->window = (Bytef *) ZALLOC(strm, s->w_size, 2*sizeof(Byte)); s->prev = (Posf *) ZALLOC(strm, s->w_size, sizeof(Pos)); s->head = (Posf *) ZALLOC(strm, s->hash_size, sizeof(Pos)); + s->lit_bufsize = 1 << (memLevel + 6); /* 16K elements by default */ + overlay = (ushf *) ZALLOC(strm, s->lit_bufsize, sizeof(ush)+2); s->pending_buf = (uchf *) overlay; s->pending_buf_size = (ulg)s->lit_bufsize * (sizeof(ush)+2L); + if (s->window == Z_NULL || s->prev == Z_NULL || s->head == Z_NULL || s->pending_buf == Z_NULL) { s->status = FINISH_STATE; @@ -269,11 +303,14 @@ int ZEXPORT deflateInit2_(strm, level, method, windowBits, memLevel, strategy, } s->d_buf = overlay + s->lit_bufsize/sizeof(ush); s->l_buf = s->pending_buf + (1+sizeof(ush))*s->lit_bufsize; + s->level = level; s->strategy = strategy; s->method = (Byte)method; + return deflateReset(strm); } + /* ========================================================================= */ int ZEXPORT deflateSetDictionary (strm, dictionary, dictLength) z_streamp strm; @@ -284,13 +321,16 @@ int ZEXPORT deflateSetDictionary (strm, dictionary, dictLength) uInt length = dictLength; uInt n; IPos hash_head = 0; + if (strm == Z_NULL || strm->state == Z_NULL || dictionary == Z_NULL || strm->state->wrap == 2 || (strm->state->wrap == 1 && strm->state->status != INIT_STATE)) return Z_STREAM_ERROR; + s = strm->state; if (s->wrap) strm->adler = adler32(strm->adler, dictionary, dictLength); + if (length < MIN_MATCH) return Z_OK; if (length > MAX_DIST(s)) { length = MAX_DIST(s); @@ -299,6 +339,7 @@ int ZEXPORT deflateSetDictionary (strm, dictionary, dictLength) zmemcpy(s->window, dictionary, length); s->strstart = length; s->block_start = (long)length; + /* Insert all strings in the hash table (except for the last two bytes). * s->lookahead stays null, so s->ins_h will be recomputed at the next * call of fill_window. @@ -311,21 +352,26 @@ int ZEXPORT deflateSetDictionary (strm, dictionary, dictLength) if (hash_head) hash_head = 0; /* to make compiler happy */ return Z_OK; } + /* ========================================================================= */ int ZEXPORT deflateReset (strm) z_streamp strm; { deflate_state *s; + if (strm == Z_NULL || strm->state == Z_NULL || strm->zalloc == (alloc_func)0 || strm->zfree == (free_func)0) { return Z_STREAM_ERROR; } + strm->total_in = strm->total_out = 0; strm->msg = Z_NULL; /* use zfree if we ever allocate msg dynamically */ strm->data_type = Z_UNKNOWN; + s = (deflate_state *)strm->state; s->pending = 0; s->pending_out = s->pending_buf; + if (s->wrap < 0) { s->wrap = -s->wrap; /* was made negative by deflate(..., Z_FINISH); */ } @@ -336,10 +382,13 @@ int ZEXPORT deflateReset (strm) #endif adler32(0L, Z_NULL, 0); s->last_flush = Z_NO_FLUSH; + _tr_init(s); lm_init(s); + return Z_OK; } + /* ========================================================================= */ int ZEXPORT deflateSetHeader (strm, head) z_streamp strm; @@ -350,6 +399,7 @@ int ZEXPORT deflateSetHeader (strm, head) strm->state->gzhead = head; return Z_OK; } + /* ========================================================================= */ int ZEXPORT deflatePrime (strm, bits, value) z_streamp strm; @@ -361,6 +411,7 @@ int ZEXPORT deflatePrime (strm, bits, value) strm->state->bi_buf = (ush)(value & ((1 << bits) - 1)); return Z_OK; } + /* ========================================================================= */ int ZEXPORT deflateParams(strm, level, strategy) z_streamp strm; @@ -370,8 +421,10 @@ int ZEXPORT deflateParams(strm, level, strategy) deflate_state *s; compress_func func; int err = Z_OK; + if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR; s = strm->state; + #ifdef FASTEST if (level != 0) level = 1; #else @@ -381,6 +434,7 @@ int ZEXPORT deflateParams(strm, level, strategy) return Z_STREAM_ERROR; } func = configuration_table[s->level].func; + if (func != configuration_table[level].func && strm->total_in != 0) { /* Flush the last buffer: */ err = deflate(strm, Z_PARTIAL_FLUSH); @@ -395,6 +449,7 @@ int ZEXPORT deflateParams(strm, level, strategy) s->strategy = strategy; return err; } + /* ========================================================================= */ int ZEXPORT deflateTune(strm, good_length, max_lazy, nice_length, max_chain) z_streamp strm; @@ -404,6 +459,7 @@ int ZEXPORT deflateTune(strm, good_length, max_lazy, nice_length, max_chain) int max_chain; { deflate_state *s; + if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR; s = strm->state; s->good_match = good_length; @@ -412,6 +468,7 @@ int ZEXPORT deflateTune(strm, good_length, max_lazy, nice_length, max_chain) s->max_chain_length = max_chain; return Z_OK; } + /* ========================================================================= * For the default windowBits of 15 and memLevel of 8, this function returns * a close to exact, as well as small, upper bound on the compressed size. @@ -435,19 +492,24 @@ uLong ZEXPORT deflateBound(strm, sourceLen) { deflate_state *s; uLong destLen; + /* conservative upper bound */ destLen = sourceLen + ((sourceLen + 7) >> 3) + ((sourceLen + 63) >> 6) + 11; + /* if can't get parameters, return conservative bound */ if (strm == Z_NULL || strm->state == Z_NULL) return destLen; + /* if not default parameters, return conservative bound */ s = strm->state; if (s->w_bits != 15 || s->hash_bits != 8 + 7) return destLen; + /* default settings: return tight bound for that case */ return compressBound(sourceLen); } + /* ========================================================================= * Put a short in the pending buffer. The 16-bit value is put in MSB order. * IN assertion: the stream state is correct and there is enough room in @@ -460,6 +522,7 @@ local void putShortMSB (s, b) put_byte(s, (Byte)(b >> 8)); put_byte(s, (Byte)(b & 0xff)); } + /* ========================================================================= * Flush as much pending output as possible. All deflate() output goes * through this function so some applications may wish to modify it @@ -470,8 +533,10 @@ local void flush_pending(strm) z_streamp strm; { unsigned len = strm->state->pending; + if (len > strm->avail_out) len = strm->avail_out; if (len == 0) return; + zmemcpy(strm->next_out, strm->state->pending_out, len); strm->next_out += len; strm->state->pending_out += len; @@ -482,6 +547,7 @@ local void flush_pending(strm) strm->state->pending_out = strm->state->pending_buf; } } + /* ========================================================================= */ int ZEXPORT deflate (strm, flush) z_streamp strm; @@ -489,20 +555,24 @@ int ZEXPORT deflate (strm, flush) { int old_flush; /* value of flush param for previous deflate call */ deflate_state *s; + if (strm == Z_NULL || strm->state == Z_NULL || flush > Z_FINISH || flush < 0) { return Z_STREAM_ERROR; } s = strm->state; + if (strm->next_out == Z_NULL || (strm->next_in == Z_NULL && strm->avail_in != 0) || (s->status == FINISH_STATE && flush != Z_FINISH)) { ERR_RETURN(strm, Z_STREAM_ERROR); } if (strm->avail_out == 0) ERR_RETURN(strm, Z_BUF_ERROR); + s->strm = strm; /* just in case */ old_flush = s->last_flush; s->last_flush = flush; + /* Write the header */ if (s->status == INIT_STATE) { #ifdef GZIP @@ -554,6 +624,7 @@ int ZEXPORT deflate (strm, flush) { uInt header = (Z_DEFLATED + ((s->w_bits-8)<<4)) << 8; uInt level_flags; + if (s->strategy >= Z_HUFFMAN_ONLY || s->level < 2) level_flags = 0; else if (s->level < 6) @@ -565,8 +636,10 @@ int ZEXPORT deflate (strm, flush) header |= (level_flags << 6); if (s->strstart != 0) header |= PRESET_DICT; header += 31 - (header % 31); + s->status = BUSY_STATE; putShortMSB(s, header); + /* Save the adler32 of the preset dictionary: */ if (s->strstart != 0) { putShortMSB(s, (uInt)(strm->adler >> 16)); @@ -579,6 +652,7 @@ int ZEXPORT deflate (strm, flush) if (s->status == EXTRA_STATE) { if (s->gzhead->extra != NULL) { uInt beg = s->pending; /* start of bytes to update crc */ + while (s->gzindex < (s->gzhead->extra_len & 0xffff)) { if (s->pending == s->pending_buf_size) { if (s->gzhead->hcrc && s->pending > beg) @@ -607,6 +681,7 @@ int ZEXPORT deflate (strm, flush) if (s->gzhead->name != NULL) { uInt beg = s->pending; /* start of bytes to update crc */ int val; + do { if (s->pending == s->pending_buf_size) { if (s->gzhead->hcrc && s->pending > beg) @@ -637,6 +712,7 @@ int ZEXPORT deflate (strm, flush) if (s->gzhead->comment != NULL) { uInt beg = s->pending; /* start of bytes to update crc */ int val; + do { if (s->pending == s->pending_buf_size) { if (s->gzhead->hcrc && s->pending > beg) @@ -676,6 +752,7 @@ int ZEXPORT deflate (strm, flush) s->status = BUSY_STATE; } #endif + /* Flush as much pending output as possible */ if (s->pending != 0) { flush_pending(strm); @@ -689,6 +766,7 @@ int ZEXPORT deflate (strm, flush) s->last_flush = -1; return Z_OK; } + /* Make sure there is something to do and avoid duplicate consecutive * flushes. For repeated and useless calls with Z_FINISH, we keep * returning Z_STREAM_END instead of Z_BUF_ERROR. @@ -697,16 +775,20 @@ int ZEXPORT deflate (strm, flush) flush != Z_FINISH) { ERR_RETURN(strm, Z_BUF_ERROR); } + /* User must not provide more input after the first FINISH: */ if (s->status == FINISH_STATE && strm->avail_in != 0) { ERR_RETURN(strm, Z_BUF_ERROR); } + /* Start a new block or continue the current one. */ if (strm->avail_in != 0 || s->lookahead != 0 || (flush != Z_NO_FLUSH && s->status != FINISH_STATE)) { block_state bstate; + bstate = (*(configuration_table[s->level].func))(s, flush); + if (bstate == finish_started || bstate == finish_done) { s->status = FINISH_STATE; } @@ -743,8 +825,10 @@ int ZEXPORT deflate (strm, flush) } } Assert(strm->avail_out > 0, "bug2"); + if (flush != Z_FINISH) return Z_OK; if (s->wrap <= 0) return Z_STREAM_END; + /* Write the trailer */ #ifdef GZIP if (s->wrap == 2) { @@ -770,12 +854,15 @@ int ZEXPORT deflate (strm, flush) if (s->wrap > 0) s->wrap = -s->wrap; /* write the trailer only once! */ return s->pending != 0 ? Z_OK : Z_STREAM_END; } + /* ========================================================================= */ int ZEXPORT deflateEnd (strm) z_streamp strm; { int status; + if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR; + status = strm->state->status; if (status != INIT_STATE && status != EXTRA_STATE && @@ -786,15 +873,19 @@ int ZEXPORT deflateEnd (strm) status != FINISH_STATE) { return Z_STREAM_ERROR; } + /* Deallocate in reverse order of allocations: */ TRY_FREE(strm, strm->state->pending_buf); TRY_FREE(strm, strm->state->head); TRY_FREE(strm, strm->state->prev); TRY_FREE(strm, strm->state->window); + ZFREE(strm, strm->state); strm->state = Z_NULL; + return status == BUSY_STATE ? Z_DATA_ERROR : Z_OK; } + /* ========================================================================= * Copy the source state to the destination state. * To simplify the source, this is not supported for 16-bit MSDOS (which @@ -811,21 +902,27 @@ int ZEXPORT deflateCopy (dest, source) deflate_state *ss; ushf *overlay; + if (source == Z_NULL || dest == Z_NULL || source->state == Z_NULL) { return Z_STREAM_ERROR; } + ss = source->state; + zmemcpy(dest, source, sizeof(z_stream)); + ds = (deflate_state *) ZALLOC(dest, 1, sizeof(deflate_state)); if (ds == Z_NULL) return Z_MEM_ERROR; dest->state = (struct internal_state FAR *) ds; zmemcpy(ds, ss, sizeof(deflate_state)); ds->strm = dest; + ds->window = (Bytef *) ZALLOC(dest, ds->w_size, 2*sizeof(Byte)); ds->prev = (Posf *) ZALLOC(dest, ds->w_size, sizeof(Pos)); ds->head = (Posf *) ZALLOC(dest, ds->hash_size, sizeof(Pos)); overlay = (ushf *) ZALLOC(dest, ds->lit_bufsize, sizeof(ush)+2); ds->pending_buf = (uchf *) overlay; + if (ds->window == Z_NULL || ds->prev == Z_NULL || ds->head == Z_NULL || ds->pending_buf == Z_NULL) { deflateEnd (dest); @@ -836,15 +933,19 @@ int ZEXPORT deflateCopy (dest, source) zmemcpy(ds->prev, ss->prev, ds->w_size * sizeof(Pos)); zmemcpy(ds->head, ss->head, ds->hash_size * sizeof(Pos)); zmemcpy(ds->pending_buf, ss->pending_buf, (uInt)ds->pending_buf_size); + ds->pending_out = ds->pending_buf + (ss->pending_out - ss->pending_buf); ds->d_buf = overlay + ds->lit_bufsize/sizeof(ush); ds->l_buf = ds->pending_buf + (1+sizeof(ush))*ds->lit_bufsize; + ds->l_desc.dyn_tree = ds->dyn_ltree; ds->d_desc.dyn_tree = ds->dyn_dtree; ds->bl_desc.dyn_tree = ds->bl_tree; + return Z_OK; #endif /* MAXSEG_64K */ } + /* =========================================================================== * Read a new buffer from the current input stream, update the adler32 * and total number of bytes read. All deflate() input goes through @@ -858,9 +959,12 @@ local int read_buf(strm, buf, size) unsigned size; { unsigned len = strm->avail_in; + if (len > size) len = size; if (len == 0) return 0; + strm->avail_in -= len; + if (strm->state->wrap == 1) { strm->adler = adler32(strm->adler, strm->next_in, len); } @@ -872,8 +976,10 @@ local int read_buf(strm, buf, size) zmemcpy(buf, strm->next_in, len); strm->next_in += len; strm->total_in += len; + return (int)len; } + /* =========================================================================== * Initialize the "longest match" routines for a new zlib stream */ @@ -881,13 +987,16 @@ local void lm_init (s) deflate_state *s; { s->window_size = (ulg)2L*s->w_size; + CLEAR_HASH(s); + /* Set the default configuration parameters: */ s->max_lazy_match = configuration_table[s->level].max_lazy; s->good_match = configuration_table[s->level].good_length; s->nice_match = configuration_table[s->level].nice_length; s->max_chain_length = configuration_table[s->level].max_chain; + s->strstart = 0; s->block_start = 0L; s->lookahead = 0; @@ -900,6 +1009,7 @@ local void lm_init (s) #endif #endif } + #ifndef FASTEST /* =========================================================================== * Set match_start to the longest match starting at the given string and @@ -931,6 +1041,7 @@ local uInt longest_match(s, cur_match) */ Posf *prev = s->prev; uInt wmask = s->w_mask; + #ifdef UNALIGNED_OK /* Compare two bytes at a time. Note: this is not always beneficial. * Try with and without -DUNALIGNED_OK to check. @@ -943,10 +1054,12 @@ local uInt longest_match(s, cur_match) register Byte scan_end1 = scan[best_len-1]; register Byte scan_end = scan[best_len]; #endif + /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16. * It is easy to get rid of this optimization if necessary. */ Assert(s->hash_bits >= 8 && MAX_MATCH == 258, "Code too clever"); + /* Do not waste too much time if we already have a good match: */ if (s->prev_length >= s->good_match) { chain_length >>= 2; @@ -955,10 +1068,13 @@ local uInt longest_match(s, cur_match) * to make deflate deterministic. */ if ((uInt)nice_match > s->lookahead) nice_match = s->lookahead; + Assert((ulg)s->strstart <= s->window_size-MIN_LOOKAHEAD, "need lookahead"); + do { Assert(cur_match < s->strstart, "no future"); match = s->window + cur_match; + /* Skip to next match if the match length cannot increase * or if the match length is less than 2. Note that the checks below * for insufficient lookahead only occur occasionally for performance @@ -973,6 +1089,7 @@ local uInt longest_match(s, cur_match) */ if (*(ushf*)(match+best_len-1) != scan_end || *(ushf*)match != scan_start) continue; + /* It is not necessary to compare scan[2] and match[2] since they are * always equal when the other bytes match, given that the hash keys * are equal and that HASH_BITS >= 8. Compare 2 bytes at a time at @@ -991,16 +1108,21 @@ local uInt longest_match(s, cur_match) *(ushf*)(scan+=2) == *(ushf*)(match+=2) && scan < strend); /* The funny "do {}" generates better code on most compilers */ + /* Here, scan <= window+strstart+257 */ Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan"); if (*scan == *match) scan++; + len = (MAX_MATCH - 1) - (int)(strend-scan); scan = strend - (MAX_MATCH-1); + #else /* UNALIGNED_OK */ + if (match[best_len] != scan_end || match[best_len-1] != scan_end1 || *match != *scan || *++match != scan[1]) continue; + /* The check at best_len-1 can be removed because it will be made * again later. (This heuristic is not always a win.) * It is not necessary to compare scan[2] and match[2] since they @@ -1009,6 +1131,7 @@ local uInt longest_match(s, cur_match) */ scan += 2, match++; Assert(*scan == *match, "match[2]?"); + /* We check for insufficient lookahead only every 8th comparison; * the 256th check will be made at strstart+258. */ @@ -1018,10 +1141,14 @@ local uInt longest_match(s, cur_match) *++scan == *++match && *++scan == *++match && *++scan == *++match && *++scan == *++match && scan < strend); + Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan"); + len = MAX_MATCH - (int)(strend - scan); scan = strend - MAX_MATCH; + #endif /* UNALIGNED_OK */ + if (len > best_len) { s->match_start = cur_match; best_len = len; @@ -1035,11 +1162,13 @@ local uInt longest_match(s, cur_match) } } while ((cur_match = prev[cur_match & wmask]) > limit && --chain_length != 0); + if ((uInt)best_len <= s->lookahead) return (uInt)best_len; return s->lookahead; } #endif /* ASMV */ #endif /* FASTEST */ + /* --------------------------------------------------------------------------- * Optimized version for level == 1 or strategy == Z_RLE only */ @@ -1051,16 +1180,22 @@ local uInt longest_match_fast(s, cur_match) register Bytef *match; /* matched string */ register int len; /* length of current match */ register Bytef *strend = s->window + s->strstart + MAX_MATCH; + /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16. * It is easy to get rid of this optimization if necessary. */ Assert(s->hash_bits >= 8 && MAX_MATCH == 258, "Code too clever"); + Assert((ulg)s->strstart <= s->window_size-MIN_LOOKAHEAD, "need lookahead"); + Assert(cur_match < s->strstart, "no future"); + match = s->window + cur_match; + /* Return failure if the match length is less than 2: */ if (match[0] != scan[0] || match[1] != scan[1]) return MIN_MATCH-1; + /* The check at best_len-1 can be removed because it will be made * again later. (This heuristic is not always a win.) * It is not necessary to compare scan[2] and match[2] since they @@ -1069,6 +1204,7 @@ local uInt longest_match_fast(s, cur_match) */ scan += 2, match += 2; Assert(*scan == *match, "match[2]?"); + /* We check for insufficient lookahead only every 8th comparison; * the 256th check will be made at strstart+258. */ @@ -1078,12 +1214,17 @@ local uInt longest_match_fast(s, cur_match) *++scan == *++match && *++scan == *++match && *++scan == *++match && *++scan == *++match && scan < strend); + Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan"); + len = MAX_MATCH - (int)(strend - scan); + if (len < MIN_MATCH) return MIN_MATCH - 1; + s->match_start = cur_match; return (uInt)len <= s->lookahead ? (uInt)len : s->lookahead; } + #ifdef DEBUG /* =========================================================================== * Check that the match at match_start is indeed a match. @@ -1111,6 +1252,7 @@ local void check_match(s, start, match, length) #else # define check_match(s, start, match, length) #endif /* DEBUG */ + /* =========================================================================== * Fill the window when the lookahead becomes insufficient. * Updates strstart and lookahead. @@ -1128,12 +1270,15 @@ local void fill_window(s) register Posf *p; unsigned more; /* Amount of free space at the end of the window. */ uInt wsize = s->w_size; + do { more = (unsigned)(s->window_size -(ulg)s->lookahead -(ulg)s->strstart); + /* Deal with !@#$% 64K limit: */ if (sizeof(int) <= 2) { if (more == 0 && s->strstart == 0 && s->lookahead == 0) { more = wsize; + } else if (more == (unsigned)(-1)) { /* Very unlikely, but possible on 16 bit machine if * strstart == 0 && lookahead == 1 (input done a byte at time) @@ -1141,14 +1286,17 @@ local void fill_window(s) more--; } } + /* If the window is almost full and there is insufficient lookahead, * move the upper half to the lower one to make room in the upper half. */ if (s->strstart >= wsize+MAX_DIST(s)) { + zmemcpy(s->window, s->window+wsize, (unsigned)wsize); s->match_start -= wsize; s->strstart -= wsize; /* we now have strstart >= MAX_DIST */ s->block_start -= (long) wsize; + /* Slide the hash table (could be avoided with 32 bit values at the expense of memory usage). We slide even when level == 0 to keep the hash table consistent if we switch back to level > 0 @@ -1162,6 +1310,7 @@ local void fill_window(s) m = *--p; *p = (Pos)(m >= wsize ? m-wsize : NIL); } while (--n); + n = wsize; #ifndef FASTEST p = &s->prev[n]; @@ -1176,6 +1325,7 @@ local void fill_window(s) more += wsize; } if (s->strm->avail_in == 0) return; + /* If there was no sliding: * strstart <= WSIZE+MAX_DIST-1 && lookahead <= MIN_LOOKAHEAD - 1 && * more == window_size - lookahead - strstart @@ -1188,8 +1338,10 @@ local void fill_window(s) * If there was sliding, more >= WSIZE. So in all cases, more >= 2. */ Assert(more >= 2, "more < 2"); + n = read_buf(s->strm, s->window + s->strstart + s->lookahead, more); s->lookahead += n; + /* Initialize the hash value now that we have some input: */ if (s->lookahead >= MIN_MATCH) { s->ins_h = s->window[s->strstart]; @@ -1201,8 +1353,10 @@ local void fill_window(s) /* If the whole input has less than MIN_MATCH bytes, ins_h is garbage, * but this is not important since only literal bytes will be emitted. */ + } while (s->lookahead < MIN_LOOKAHEAD && s->strm->avail_in != 0); } + /* =========================================================================== * Flush the current block, with given end-of-file flag. * IN assertion: strstart is set to the end of the current match. @@ -1217,11 +1371,13 @@ local void fill_window(s) flush_pending(s->strm); \ Tracev((stderr,"[FLUSH]")); \ } + /* Same but force premature exit if necessary. */ #define FLUSH_BLOCK(s, eof) { \ FLUSH_BLOCK_ONLY(s, eof); \ if (s->strm->avail_out == 0) return (eof) ? finish_started : need_more; \ } + /* =========================================================================== * Copy without compression as much as possible from the input stream, return * the current block state. @@ -1240,22 +1396,29 @@ local block_state deflate_stored(s, flush) */ ulg max_block_size = 0xffff; ulg max_start; + if (max_block_size > s->pending_buf_size - 5) { max_block_size = s->pending_buf_size - 5; } + /* Copy as much as possible from input to output: */ for (;;) { /* Fill the window as much as possible: */ if (s->lookahead <= 1) { + Assert(s->strstart < s->w_size+MAX_DIST(s) || s->block_start >= (long)s->w_size, "slide too late"); + fill_window(s); if (s->lookahead == 0 && flush == Z_NO_FLUSH) return need_more; + if (s->lookahead == 0) break; /* flush the current block */ } Assert(s->block_start >= 0L, "block gone"); + s->strstart += s->lookahead; s->lookahead = 0; + /* Emit a stored block if pending_buf will be full: */ max_start = s->block_start + max_block_size; if (s->strstart == 0 || (ulg)s->strstart >= max_start) { @@ -1274,6 +1437,7 @@ local block_state deflate_stored(s, flush) FLUSH_BLOCK(s, flush == Z_FINISH); return flush == Z_FINISH ? finish_done : block_done; } + /* =========================================================================== * Compress as much as possible from the input stream, return the current * block state. @@ -1287,6 +1451,7 @@ local block_state deflate_fast(s, flush) { IPos hash_head = NIL; /* head of the hash chain */ int bflush; /* set if current block must be flushed */ + for (;;) { /* Make sure that we always have enough lookahead, except * at the end of the input file. We need MAX_MATCH bytes @@ -1300,12 +1465,14 @@ local block_state deflate_fast(s, flush) } if (s->lookahead == 0) break; /* flush the current block */ } + /* Insert the string window[strstart .. strstart+2] in the * dictionary, and set hash_head to the head of the hash chain: */ if (s->lookahead >= MIN_MATCH) { INSERT_STRING(s, s->strstart, hash_head); } + /* Find the longest match, discarding those <= prev_length. * At this point we have always match_length < MIN_MATCH */ @@ -1330,9 +1497,12 @@ local block_state deflate_fast(s, flush) } if (s->match_length >= MIN_MATCH) { check_match(s, s->strstart, s->match_start, s->match_length); + _tr_tally_dist(s, s->strstart - s->match_start, s->match_length - MIN_MATCH, bflush); + s->lookahead -= s->match_length; + /* Insert new strings in the hash table only if the match length * is not too large. This saves time but degrades compression. */ @@ -1374,6 +1544,7 @@ local block_state deflate_fast(s, flush) FLUSH_BLOCK(s, flush == Z_FINISH); return flush == Z_FINISH ? finish_done : block_done; } + #ifndef FASTEST /* =========================================================================== * Same as above, but achieves better compression. We use a lazy @@ -1386,6 +1557,7 @@ local block_state deflate_slow(s, flush) { IPos hash_head = NIL; /* head of hash chain */ int bflush; /* set if current block must be flushed */ + /* Process the input block. */ for (;;) { /* Make sure that we always have enough lookahead, except @@ -1400,16 +1572,19 @@ local block_state deflate_slow(s, flush) } if (s->lookahead == 0) break; /* flush the current block */ } + /* Insert the string window[strstart .. strstart+2] in the * dictionary, and set hash_head to the head of the hash chain: */ if (s->lookahead >= MIN_MATCH) { INSERT_STRING(s, s->strstart, hash_head); } + /* Find the longest match, discarding those <= prev_length. */ s->prev_length = s->match_length, s->prev_match = s->match_start; s->match_length = MIN_MATCH-1; + if (hash_head != NIL && s->prev_length < s->max_lazy_match && s->strstart - hash_head <= MAX_DIST(s)) { /* To simplify the code, we prevent matches with the string @@ -1422,12 +1597,14 @@ local block_state deflate_slow(s, flush) s->match_length = longest_match_fast (s, hash_head); } /* longest_match() or longest_match_fast() sets match_start */ + if (s->match_length <= 5 && (s->strategy == Z_FILTERED #if TOO_FAR <= 32767 || (s->match_length == MIN_MATCH && s->strstart - s->match_start > TOO_FAR) #endif )) { + /* If prev_match is also MIN_MATCH, match_start is garbage * but we will ignore the current match anyway. */ @@ -1440,9 +1617,12 @@ local block_state deflate_slow(s, flush) if (s->prev_length >= MIN_MATCH && s->match_length <= s->prev_length) { uInt max_insert = s->strstart + s->lookahead - MIN_MATCH; /* Do not insert strings in hash table beyond this. */ + check_match(s, s->strstart-1, s->prev_match, s->prev_length); + _tr_tally_dist(s, s->strstart -1 - s->prev_match, s->prev_length - MIN_MATCH, bflush); + /* Insert in hash table all strings up to the end of the match. * strstart-1 and strstart are already inserted. If there is not * enough lookahead, the last two strings are not inserted in @@ -1458,7 +1638,9 @@ local block_state deflate_slow(s, flush) s->match_available = 0; s->match_length = MIN_MATCH-1; s->strstart++; + if (bflush) FLUSH_BLOCK(s, 0); + } else if (s->match_available) { /* If there was no match at the previous position, output a * single literal. If there was a match but the current match @@ -1491,6 +1673,7 @@ local block_state deflate_slow(s, flush) return flush == Z_FINISH ? finish_done : block_done; } #endif /* FASTEST */ + #if 0 /* =========================================================================== * For Z_RLE, simply look for runs of bytes, generate matches only of distance @@ -1506,6 +1689,7 @@ local block_state deflate_rle(s, flush) uInt max; /* maximum length of run */ uInt prev; /* byte at distance one to match */ Bytef *scan; /* scan for end of run */ + for (;;) { /* Make sure that we always have enough lookahead, except * at the end of the input file. We need MAX_MATCH bytes @@ -1518,6 +1702,7 @@ local block_state deflate_rle(s, flush) } if (s->lookahead == 0) break; /* flush the current block */ } + /* See how many times the previous byte repeats */ run = 0; if (s->strstart > 0) { /* if there is a previous byte, that is */ @@ -1529,6 +1714,7 @@ local block_state deflate_rle(s, flush) break; } while (++run < max); } + /* Emit match if have run of MIN_MATCH or longer, else emit literal */ if (run >= MIN_MATCH) { check_match(s, s->strstart, s->strstart - 1, run); |