deflate.h revision 205471
117651Speter/* deflate.h -- internal compression state 2205471Sdelphij * Copyright (C) 1995-2009 Jean-loup Gailly 3131380Stjr * For conditions of distribution and use, see copyright notice in zlib.h 417651Speter */ 517651Speter 617651Speter/* WARNING: this file should *not* be used by applications. It is 717651Speter part of the implementation of the compression library and is 817651Speter subject to change. Applications should only use zlib.h. 917651Speter */ 1017651Speter 11146081Skientzle/* @(#) $Id$ */ 1217651Speter 13131380Stjr#ifndef DEFLATE_H 14131380Stjr#define DEFLATE_H 1517651Speter 1617651Speter#include "zutil.h" 1717651Speter 18131380Stjr/* define NO_GZIP when compiling if you want to disable gzip header and 19131380Stjr trailer creation by deflate(). NO_GZIP would be used to avoid linking in 20131380Stjr the crc code when it is not needed. For shared libraries, gzip encoding 21131380Stjr should be left enabled. */ 22131380Stjr#ifndef NO_GZIP 23131380Stjr# define GZIP 24131380Stjr#endif 25131380Stjr 2617651Speter/* =========================================================================== 2717651Speter * Internal compression state. 2817651Speter */ 2917651Speter 3017651Speter#define LENGTH_CODES 29 3117651Speter/* number of length codes, not counting the special END_BLOCK code */ 3217651Speter 3317651Speter#define LITERALS 256 3417651Speter/* number of literal bytes 0..255 */ 3517651Speter 3617651Speter#define L_CODES (LITERALS+1+LENGTH_CODES) 3717651Speter/* number of Literal or Length codes, including the END_BLOCK code */ 3817651Speter 3917651Speter#define D_CODES 30 4017651Speter/* number of distance codes */ 4117651Speter 4217651Speter#define BL_CODES 19 4317651Speter/* number of codes used to transfer the bit lengths */ 4417651Speter 4517651Speter#define HEAP_SIZE (2*L_CODES+1) 4617651Speter/* maximum heap size */ 4717651Speter 4817651Speter#define MAX_BITS 15 4917651Speter/* All codes must not exceed MAX_BITS bits */ 5017651Speter 5117651Speter#define INIT_STATE 42 52157046Sdes#define EXTRA_STATE 69 53157046Sdes#define NAME_STATE 73 54157046Sdes#define COMMENT_STATE 91 55157046Sdes#define HCRC_STATE 103 5617651Speter#define BUSY_STATE 113 5717651Speter#define FINISH_STATE 666 5817651Speter/* Stream status */ 5917651Speter 6017651Speter 6117651Speter/* Data structure describing a single value and its code string. */ 6217651Spetertypedef struct ct_data_s { 6317651Speter union { 6417651Speter ush freq; /* frequency count */ 6517651Speter ush code; /* bit string */ 6617651Speter } fc; 6717651Speter union { 6817651Speter ush dad; /* father node in Huffman tree */ 6917651Speter ush len; /* length of bit string */ 7017651Speter } dl; 7117651Speter} FAR ct_data; 7217651Speter 7317651Speter#define Freq fc.freq 7417651Speter#define Code fc.code 7517651Speter#define Dad dl.dad 7617651Speter#define Len dl.len 7717651Speter 7817651Spetertypedef struct static_tree_desc_s static_tree_desc; 7917651Speter 8017651Spetertypedef struct tree_desc_s { 8117651Speter ct_data *dyn_tree; /* the dynamic tree */ 8217651Speter int max_code; /* largest code with non zero frequency */ 8317651Speter static_tree_desc *stat_desc; /* the corresponding static tree */ 8417651Speter} FAR tree_desc; 8517651Speter 8617651Spetertypedef ush Pos; 8717651Spetertypedef Pos FAR Posf; 8817651Spetertypedef unsigned IPos; 8917651Speter 9017651Speter/* A Pos is an index in the character window. We use short instead of int to 9117651Speter * save space in the various tables. IPos is used only for parameter passing. 9217651Speter */ 9317651Speter 9417651Spetertypedef struct internal_state { 9517651Speter z_streamp strm; /* pointer back to this zlib stream */ 9617651Speter int status; /* as the name implies */ 9717651Speter Bytef *pending_buf; /* output still pending */ 9833908Ssteve ulg pending_buf_size; /* size of pending_buf */ 9917651Speter Bytef *pending_out; /* next pending byte to output to the stream */ 100157046Sdes uInt pending; /* nb of bytes in the pending buffer */ 101131380Stjr int wrap; /* bit 0 true for zlib, bit 1 true for gzip */ 102157046Sdes gz_headerp gzhead; /* gzip header information to write */ 103157046Sdes uInt gzindex; /* where in extra, name, or comment */ 10417651Speter Byte method; /* STORED (for zip only) or DEFLATED */ 10517651Speter int last_flush; /* value of flush param for previous deflate call */ 10617651Speter 10717651Speter /* used by deflate.c: */ 10817651Speter 10917651Speter uInt w_size; /* LZ77 window size (32K by default) */ 11017651Speter uInt w_bits; /* log2(w_size) (8..16) */ 11117651Speter uInt w_mask; /* w_size - 1 */ 11217651Speter 11317651Speter Bytef *window; 11417651Speter /* Sliding window. Input bytes are read into the second half of the window, 11517651Speter * and move to the first half later to keep a dictionary of at least wSize 11617651Speter * bytes. With this organization, matches are limited to a distance of 11717651Speter * wSize-MAX_MATCH bytes, but this ensures that IO is always 11817651Speter * performed with a length multiple of the block size. Also, it limits 11917651Speter * the window size to 64K, which is quite useful on MSDOS. 12017651Speter * To do: use the user input buffer as sliding window. 12117651Speter */ 12217651Speter 12317651Speter ulg window_size; 12417651Speter /* Actual size of window: 2*wSize, except when the user input buffer 12517651Speter * is directly used as sliding window. 12617651Speter */ 12717651Speter 12817651Speter Posf *prev; 12917651Speter /* Link to older string with same hash index. To limit the size of this 13017651Speter * array to 64K, this link is maintained only for the last 32K strings. 13117651Speter * An index in this array is thus a window index modulo 32K. 13217651Speter */ 13317651Speter 13417651Speter Posf *head; /* Heads of the hash chains or NIL. */ 13517651Speter 13617651Speter uInt ins_h; /* hash index of string to be inserted */ 13717651Speter uInt hash_size; /* number of elements in hash table */ 13817651Speter uInt hash_bits; /* log2(hash_size) */ 13917651Speter uInt hash_mask; /* hash_size-1 */ 14017651Speter 14117651Speter uInt hash_shift; 14217651Speter /* Number of bits by which ins_h must be shifted at each input 14317651Speter * step. It must be such that after MIN_MATCH steps, the oldest 14417651Speter * byte no longer takes part in the hash key, that is: 14517651Speter * hash_shift * MIN_MATCH >= hash_bits 14617651Speter */ 14717651Speter 14817651Speter long block_start; 14917651Speter /* Window position at the beginning of the current output block. Gets 15017651Speter * negative when the window is moved backwards. 15117651Speter */ 15217651Speter 15317651Speter uInt match_length; /* length of best match */ 15417651Speter IPos prev_match; /* previous match */ 15517651Speter int match_available; /* set if previous match exists */ 15617651Speter uInt strstart; /* start of string to insert */ 15717651Speter uInt match_start; /* start of matching string */ 15817651Speter uInt lookahead; /* number of valid bytes ahead in window */ 15917651Speter 16017651Speter uInt prev_length; 16117651Speter /* Length of the best match at previous step. Matches not greater than this 16217651Speter * are discarded. This is used in the lazy match evaluation. 16317651Speter */ 16417651Speter 16517651Speter uInt max_chain_length; 16617651Speter /* To speed up deflation, hash chains are never searched beyond this 16717651Speter * length. A higher limit improves compression ratio but degrades the 16817651Speter * speed. 16917651Speter */ 17017651Speter 17117651Speter uInt max_lazy_match; 17217651Speter /* Attempt to find a better match only when the current match is strictly 17317651Speter * smaller than this value. This mechanism is used only for compression 17417651Speter * levels >= 4. 17517651Speter */ 17617651Speter# define max_insert_length max_lazy_match 17717651Speter /* Insert new strings in the hash table only if the match length is not 17817651Speter * greater than this length. This saves time but degrades compression. 17917651Speter * max_insert_length is used only for compression levels <= 3. 18017651Speter */ 18117651Speter 18217651Speter int level; /* compression level (1..9) */ 18317651Speter int strategy; /* favor or force Huffman coding*/ 18417651Speter 18517651Speter uInt good_match; 18617651Speter /* Use a faster search when the previous match is longer than this */ 18717651Speter 18817651Speter int nice_match; /* Stop searching when current match exceeds this */ 18917651Speter 19017651Speter /* used by trees.c: */ 19117651Speter /* Didn't use ct_data typedef below to supress compiler warning */ 19217651Speter struct ct_data_s dyn_ltree[HEAP_SIZE]; /* literal and length tree */ 19317651Speter struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */ 19417651Speter struct ct_data_s bl_tree[2*BL_CODES+1]; /* Huffman tree for bit lengths */ 19517651Speter 19617651Speter struct tree_desc_s l_desc; /* desc. for literal tree */ 19717651Speter struct tree_desc_s d_desc; /* desc. for distance tree */ 19817651Speter struct tree_desc_s bl_desc; /* desc. for bit length tree */ 19917651Speter 20017651Speter ush bl_count[MAX_BITS+1]; 20117651Speter /* number of codes at each bit length for an optimal tree */ 20217651Speter 20317651Speter int heap[2*L_CODES+1]; /* heap used to build the Huffman trees */ 20417651Speter int heap_len; /* number of elements in the heap */ 20517651Speter int heap_max; /* element of largest frequency */ 20617651Speter /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used. 20717651Speter * The same heap array is used to build all trees. 20817651Speter */ 20917651Speter 21017651Speter uch depth[2*L_CODES+1]; 21117651Speter /* Depth of each subtree used as tie breaker for trees of equal frequency 21217651Speter */ 21317651Speter 21417651Speter uchf *l_buf; /* buffer for literals or lengths */ 21517651Speter 21617651Speter uInt lit_bufsize; 21717651Speter /* Size of match buffer for literals/lengths. There are 4 reasons for 21817651Speter * limiting lit_bufsize to 64K: 21917651Speter * - frequencies can be kept in 16 bit counters 22017651Speter * - if compression is not successful for the first block, all input 22117651Speter * data is still in the window so we can still emit a stored block even 22217651Speter * when input comes from standard input. (This can also be done for 22317651Speter * all blocks if lit_bufsize is not greater than 32K.) 22417651Speter * - if compression is not successful for a file smaller than 64K, we can 22517651Speter * even emit a stored file instead of a stored block (saving 5 bytes). 22617651Speter * This is applicable only for zip (not gzip or zlib). 22717651Speter * - creating new Huffman trees less frequently may not provide fast 22817651Speter * adaptation to changes in the input data statistics. (Take for 22917651Speter * example a binary file with poorly compressible code followed by 23017651Speter * a highly compressible string table.) Smaller buffer sizes give 23117651Speter * fast adaptation but have of course the overhead of transmitting 23217651Speter * trees more frequently. 23317651Speter * - I can't count above 4 23417651Speter */ 23517651Speter 23617651Speter uInt last_lit; /* running index in l_buf */ 23717651Speter 23817651Speter ushf *d_buf; 23917651Speter /* Buffer for distances. To simplify the code, d_buf and l_buf have 24017651Speter * the same number of elements. To use different lengths, an extra flag 24117651Speter * array would be necessary. 24217651Speter */ 24317651Speter 24417651Speter ulg opt_len; /* bit length of current block with optimal trees */ 24517651Speter ulg static_len; /* bit length of current block with static trees */ 24617651Speter uInt matches; /* number of string matches in current block */ 24717651Speter int last_eob_len; /* bit length of EOB code for last block */ 24817651Speter 24917651Speter#ifdef DEBUG 25042471Speter ulg compressed_len; /* total bit length of compressed file mod 2^32 */ 25142471Speter ulg bits_sent; /* bit length of compressed data sent mod 2^32 */ 25217651Speter#endif 25317651Speter 25417651Speter ush bi_buf; 25517651Speter /* Output buffer. bits are inserted starting at the bottom (least 25617651Speter * significant bits). 25717651Speter */ 25817651Speter int bi_valid; 25917651Speter /* Number of valid bits in bi_buf. All bits above the last valid bit 26017651Speter * are always zero. 26117651Speter */ 26217651Speter 263205471Sdelphij ulg high_water; 264205471Sdelphij /* High water mark offset in window for initialized bytes -- bytes above 265205471Sdelphij * this are set to zero in order to avoid memory check warnings when 266205471Sdelphij * longest match routines access bytes past the input. This is then 267205471Sdelphij * updated to the new high water mark. 268205471Sdelphij */ 269205471Sdelphij 27017651Speter} FAR deflate_state; 27117651Speter 27217651Speter/* Output a byte on the stream. 27317651Speter * IN assertion: there is enough room in pending_buf. 27417651Speter */ 27517651Speter#define put_byte(s, c) {s->pending_buf[s->pending++] = (c);} 27617651Speter 27717651Speter 27817651Speter#define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1) 27917651Speter/* Minimum amount of lookahead, except at the end of the input file. 28017651Speter * See deflate.c for comments about the MIN_MATCH+1. 28117651Speter */ 28217651Speter 28317651Speter#define MAX_DIST(s) ((s)->w_size-MIN_LOOKAHEAD) 28417651Speter/* In order to simplify the code, particularly on 16 bit machines, match 28517651Speter * distances are limited to MAX_DIST instead of WSIZE. 28617651Speter */ 28717651Speter 288205471Sdelphij#define WIN_INIT MAX_MATCH 289205471Sdelphij/* Number of bytes after end of data in window to initialize in order to avoid 290205471Sdelphij memory checker errors from longest match routines */ 291205471Sdelphij 29217651Speter /* in trees.c */ 29317651Spetervoid _tr_init OF((deflate_state *s)); 29417651Speterint _tr_tally OF((deflate_state *s, unsigned dist, unsigned lc)); 29542471Spetervoid _tr_flush_block OF((deflate_state *s, charf *buf, ulg stored_len, 296205471Sdelphij int last)); 29717651Spetervoid _tr_align OF((deflate_state *s)); 29817651Spetervoid _tr_stored_block OF((deflate_state *s, charf *buf, ulg stored_len, 299205471Sdelphij int last)); 30033908Ssteve 30133908Ssteve#define d_code(dist) \ 30233908Ssteve ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)]) 30333908Ssteve/* Mapping from a distance to a distance code. dist is the distance - 1 and 30433908Ssteve * must not have side effects. _dist_code[256] and _dist_code[257] are never 30533908Ssteve * used. 30633908Ssteve */ 30733908Ssteve 30833908Ssteve#ifndef DEBUG 30933908Ssteve/* Inline versions of _tr_tally for speed: */ 31033908Ssteve 31133908Ssteve#if defined(GEN_TREES_H) || !defined(STDC) 31233908Ssteve extern uch _length_code[]; 31333908Ssteve extern uch _dist_code[]; 31433908Ssteve#else 31533908Ssteve extern const uch _length_code[]; 31633908Ssteve extern const uch _dist_code[]; 31717651Speter#endif 31833908Ssteve 31933908Ssteve# define _tr_tally_lit(s, c, flush) \ 32033908Ssteve { uch cc = (c); \ 32133908Ssteve s->d_buf[s->last_lit] = 0; \ 32233908Ssteve s->l_buf[s->last_lit++] = cc; \ 32333908Ssteve s->dyn_ltree[cc].Freq++; \ 32433908Ssteve flush = (s->last_lit == s->lit_bufsize-1); \ 32533908Ssteve } 32633908Ssteve# define _tr_tally_dist(s, distance, length, flush) \ 32733908Ssteve { uch len = (length); \ 32833908Ssteve ush dist = (distance); \ 32933908Ssteve s->d_buf[s->last_lit] = dist; \ 33033908Ssteve s->l_buf[s->last_lit++] = len; \ 33133908Ssteve dist--; \ 33233908Ssteve s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \ 33333908Ssteve s->dyn_dtree[d_code(dist)].Freq++; \ 33433908Ssteve flush = (s->last_lit == s->lit_bufsize-1); \ 33533908Ssteve } 33633908Ssteve#else 33733908Ssteve# define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c) 33833908Ssteve# define _tr_tally_dist(s, distance, length, flush) \ 339131380Stjr flush = _tr_tally(s, distance, length) 34033908Ssteve#endif 34133908Ssteve 342131380Stjr#endif /* DEFLATE_H */ 343