117651Speter/* deflate.h -- internal compression state 2237410Sdelphij * Copyright (C) 1995-2012 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 51237410Sdelphij#define Buf_size 16 52237410Sdelphij/* size of bit buffer in bi_buf */ 53237410Sdelphij 5417651Speter#define INIT_STATE 42 55157046Sdes#define EXTRA_STATE 69 56157046Sdes#define NAME_STATE 73 57157046Sdes#define COMMENT_STATE 91 58157046Sdes#define HCRC_STATE 103 5917651Speter#define BUSY_STATE 113 6017651Speter#define FINISH_STATE 666 6117651Speter/* Stream status */ 6217651Speter 6317651Speter 6417651Speter/* Data structure describing a single value and its code string. */ 6517651Spetertypedef struct ct_data_s { 6617651Speter union { 6717651Speter ush freq; /* frequency count */ 6817651Speter ush code; /* bit string */ 6917651Speter } fc; 7017651Speter union { 7117651Speter ush dad; /* father node in Huffman tree */ 7217651Speter ush len; /* length of bit string */ 7317651Speter } dl; 7417651Speter} FAR ct_data; 7517651Speter 7617651Speter#define Freq fc.freq 7717651Speter#define Code fc.code 7817651Speter#define Dad dl.dad 7917651Speter#define Len dl.len 8017651Speter 8117651Spetertypedef struct static_tree_desc_s static_tree_desc; 8217651Speter 8317651Spetertypedef struct tree_desc_s { 8417651Speter ct_data *dyn_tree; /* the dynamic tree */ 8517651Speter int max_code; /* largest code with non zero frequency */ 8617651Speter static_tree_desc *stat_desc; /* the corresponding static tree */ 8717651Speter} FAR tree_desc; 8817651Speter 8917651Spetertypedef ush Pos; 9017651Spetertypedef Pos FAR Posf; 9117651Spetertypedef unsigned IPos; 9217651Speter 9317651Speter/* A Pos is an index in the character window. We use short instead of int to 9417651Speter * save space in the various tables. IPos is used only for parameter passing. 9517651Speter */ 9617651Speter 9717651Spetertypedef struct internal_state { 9817651Speter z_streamp strm; /* pointer back to this zlib stream */ 9917651Speter int status; /* as the name implies */ 10017651Speter Bytef *pending_buf; /* output still pending */ 10133908Ssteve ulg pending_buf_size; /* size of pending_buf */ 10217651Speter Bytef *pending_out; /* next pending byte to output to the stream */ 103157046Sdes uInt pending; /* nb of bytes in the pending buffer */ 104131380Stjr int wrap; /* bit 0 true for zlib, bit 1 true for gzip */ 105157046Sdes gz_headerp gzhead; /* gzip header information to write */ 106157046Sdes uInt gzindex; /* where in extra, name, or comment */ 107250261Sdelphij Byte method; /* can only be DEFLATED */ 10817651Speter int last_flush; /* value of flush param for previous deflate call */ 10917651Speter 11017651Speter /* used by deflate.c: */ 11117651Speter 11217651Speter uInt w_size; /* LZ77 window size (32K by default) */ 11317651Speter uInt w_bits; /* log2(w_size) (8..16) */ 11417651Speter uInt w_mask; /* w_size - 1 */ 11517651Speter 11617651Speter Bytef *window; 11717651Speter /* Sliding window. Input bytes are read into the second half of the window, 11817651Speter * and move to the first half later to keep a dictionary of at least wSize 11917651Speter * bytes. With this organization, matches are limited to a distance of 12017651Speter * wSize-MAX_MATCH bytes, but this ensures that IO is always 12117651Speter * performed with a length multiple of the block size. Also, it limits 12217651Speter * the window size to 64K, which is quite useful on MSDOS. 12317651Speter * To do: use the user input buffer as sliding window. 12417651Speter */ 12517651Speter 12617651Speter ulg window_size; 12717651Speter /* Actual size of window: 2*wSize, except when the user input buffer 12817651Speter * is directly used as sliding window. 12917651Speter */ 13017651Speter 13117651Speter Posf *prev; 13217651Speter /* Link to older string with same hash index. To limit the size of this 13317651Speter * array to 64K, this link is maintained only for the last 32K strings. 13417651Speter * An index in this array is thus a window index modulo 32K. 13517651Speter */ 13617651Speter 13717651Speter Posf *head; /* Heads of the hash chains or NIL. */ 13817651Speter 13917651Speter uInt ins_h; /* hash index of string to be inserted */ 14017651Speter uInt hash_size; /* number of elements in hash table */ 14117651Speter uInt hash_bits; /* log2(hash_size) */ 14217651Speter uInt hash_mask; /* hash_size-1 */ 14317651Speter 14417651Speter uInt hash_shift; 14517651Speter /* Number of bits by which ins_h must be shifted at each input 14617651Speter * step. It must be such that after MIN_MATCH steps, the oldest 14717651Speter * byte no longer takes part in the hash key, that is: 14817651Speter * hash_shift * MIN_MATCH >= hash_bits 14917651Speter */ 15017651Speter 15117651Speter long block_start; 15217651Speter /* Window position at the beginning of the current output block. Gets 15317651Speter * negative when the window is moved backwards. 15417651Speter */ 15517651Speter 15617651Speter uInt match_length; /* length of best match */ 15717651Speter IPos prev_match; /* previous match */ 15817651Speter int match_available; /* set if previous match exists */ 15917651Speter uInt strstart; /* start of string to insert */ 16017651Speter uInt match_start; /* start of matching string */ 16117651Speter uInt lookahead; /* number of valid bytes ahead in window */ 16217651Speter 16317651Speter uInt prev_length; 16417651Speter /* Length of the best match at previous step. Matches not greater than this 16517651Speter * are discarded. This is used in the lazy match evaluation. 16617651Speter */ 16717651Speter 16817651Speter uInt max_chain_length; 16917651Speter /* To speed up deflation, hash chains are never searched beyond this 17017651Speter * length. A higher limit improves compression ratio but degrades the 17117651Speter * speed. 17217651Speter */ 17317651Speter 17417651Speter uInt max_lazy_match; 17517651Speter /* Attempt to find a better match only when the current match is strictly 17617651Speter * smaller than this value. This mechanism is used only for compression 17717651Speter * levels >= 4. 17817651Speter */ 17917651Speter# define max_insert_length max_lazy_match 18017651Speter /* Insert new strings in the hash table only if the match length is not 18117651Speter * greater than this length. This saves time but degrades compression. 18217651Speter * max_insert_length is used only for compression levels <= 3. 18317651Speter */ 18417651Speter 18517651Speter int level; /* compression level (1..9) */ 18617651Speter int strategy; /* favor or force Huffman coding*/ 18717651Speter 18817651Speter uInt good_match; 18917651Speter /* Use a faster search when the previous match is longer than this */ 19017651Speter 19117651Speter int nice_match; /* Stop searching when current match exceeds this */ 19217651Speter 19317651Speter /* used by trees.c: */ 194237410Sdelphij /* Didn't use ct_data typedef below to suppress compiler warning */ 19517651Speter struct ct_data_s dyn_ltree[HEAP_SIZE]; /* literal and length tree */ 19617651Speter struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */ 19717651Speter struct ct_data_s bl_tree[2*BL_CODES+1]; /* Huffman tree for bit lengths */ 19817651Speter 19917651Speter struct tree_desc_s l_desc; /* desc. for literal tree */ 20017651Speter struct tree_desc_s d_desc; /* desc. for distance tree */ 20117651Speter struct tree_desc_s bl_desc; /* desc. for bit length tree */ 20217651Speter 20317651Speter ush bl_count[MAX_BITS+1]; 20417651Speter /* number of codes at each bit length for an optimal tree */ 20517651Speter 20617651Speter int heap[2*L_CODES+1]; /* heap used to build the Huffman trees */ 20717651Speter int heap_len; /* number of elements in the heap */ 20817651Speter int heap_max; /* element of largest frequency */ 20917651Speter /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used. 21017651Speter * The same heap array is used to build all trees. 21117651Speter */ 21217651Speter 21317651Speter uch depth[2*L_CODES+1]; 21417651Speter /* Depth of each subtree used as tie breaker for trees of equal frequency 21517651Speter */ 21617651Speter 21717651Speter uchf *l_buf; /* buffer for literals or lengths */ 21817651Speter 21917651Speter uInt lit_bufsize; 22017651Speter /* Size of match buffer for literals/lengths. There are 4 reasons for 22117651Speter * limiting lit_bufsize to 64K: 22217651Speter * - frequencies can be kept in 16 bit counters 22317651Speter * - if compression is not successful for the first block, all input 22417651Speter * data is still in the window so we can still emit a stored block even 22517651Speter * when input comes from standard input. (This can also be done for 22617651Speter * all blocks if lit_bufsize is not greater than 32K.) 22717651Speter * - if compression is not successful for a file smaller than 64K, we can 22817651Speter * even emit a stored file instead of a stored block (saving 5 bytes). 22917651Speter * This is applicable only for zip (not gzip or zlib). 23017651Speter * - creating new Huffman trees less frequently may not provide fast 23117651Speter * adaptation to changes in the input data statistics. (Take for 23217651Speter * example a binary file with poorly compressible code followed by 23317651Speter * a highly compressible string table.) Smaller buffer sizes give 23417651Speter * fast adaptation but have of course the overhead of transmitting 23517651Speter * trees more frequently. 23617651Speter * - I can't count above 4 23717651Speter */ 23817651Speter 23917651Speter uInt last_lit; /* running index in l_buf */ 24017651Speter 24117651Speter ushf *d_buf; 24217651Speter /* Buffer for distances. To simplify the code, d_buf and l_buf have 24317651Speter * the same number of elements. To use different lengths, an extra flag 24417651Speter * array would be necessary. 24517651Speter */ 24617651Speter 24717651Speter ulg opt_len; /* bit length of current block with optimal trees */ 24817651Speter ulg static_len; /* bit length of current block with static trees */ 24917651Speter uInt matches; /* number of string matches in current block */ 250237410Sdelphij uInt insert; /* bytes at end of window left to insert */ 25117651Speter 25217651Speter#ifdef DEBUG 25342471Speter ulg compressed_len; /* total bit length of compressed file mod 2^32 */ 25442471Speter ulg bits_sent; /* bit length of compressed data sent mod 2^32 */ 25517651Speter#endif 25617651Speter 25717651Speter ush bi_buf; 25817651Speter /* Output buffer. bits are inserted starting at the bottom (least 25917651Speter * significant bits). 26017651Speter */ 26117651Speter int bi_valid; 26217651Speter /* Number of valid bits in bi_buf. All bits above the last valid bit 26317651Speter * are always zero. 26417651Speter */ 26517651Speter 266205471Sdelphij ulg high_water; 267205471Sdelphij /* High water mark offset in window for initialized bytes -- bytes above 268205471Sdelphij * this are set to zero in order to avoid memory check warnings when 269205471Sdelphij * longest match routines access bytes past the input. This is then 270205471Sdelphij * updated to the new high water mark. 271205471Sdelphij */ 272205471Sdelphij 27317651Speter} FAR deflate_state; 27417651Speter 27517651Speter/* Output a byte on the stream. 27617651Speter * IN assertion: there is enough room in pending_buf. 27717651Speter */ 27817651Speter#define put_byte(s, c) {s->pending_buf[s->pending++] = (c);} 27917651Speter 28017651Speter 28117651Speter#define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1) 28217651Speter/* Minimum amount of lookahead, except at the end of the input file. 28317651Speter * See deflate.c for comments about the MIN_MATCH+1. 28417651Speter */ 28517651Speter 28617651Speter#define MAX_DIST(s) ((s)->w_size-MIN_LOOKAHEAD) 28717651Speter/* In order to simplify the code, particularly on 16 bit machines, match 28817651Speter * distances are limited to MAX_DIST instead of WSIZE. 28917651Speter */ 29017651Speter 291205471Sdelphij#define WIN_INIT MAX_MATCH 292205471Sdelphij/* Number of bytes after end of data in window to initialize in order to avoid 293205471Sdelphij memory checker errors from longest match routines */ 294205471Sdelphij 29517651Speter /* in trees.c */ 296206924Sdelphijvoid ZLIB_INTERNAL _tr_init OF((deflate_state *s)); 297206924Sdelphijint ZLIB_INTERNAL _tr_tally OF((deflate_state *s, unsigned dist, unsigned lc)); 298206924Sdelphijvoid ZLIB_INTERNAL _tr_flush_block OF((deflate_state *s, charf *buf, 299206924Sdelphij ulg stored_len, int last)); 300237410Sdelphijvoid ZLIB_INTERNAL _tr_flush_bits OF((deflate_state *s)); 301206924Sdelphijvoid ZLIB_INTERNAL _tr_align OF((deflate_state *s)); 302206924Sdelphijvoid ZLIB_INTERNAL _tr_stored_block OF((deflate_state *s, charf *buf, 303206924Sdelphij ulg stored_len, int last)); 30433908Ssteve 30533908Ssteve#define d_code(dist) \ 30633908Ssteve ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)]) 30733908Ssteve/* Mapping from a distance to a distance code. dist is the distance - 1 and 30833908Ssteve * must not have side effects. _dist_code[256] and _dist_code[257] are never 30933908Ssteve * used. 31033908Ssteve */ 31133908Ssteve 31233908Ssteve#ifndef DEBUG 31333908Ssteve/* Inline versions of _tr_tally for speed: */ 31433908Ssteve 31533908Ssteve#if defined(GEN_TREES_H) || !defined(STDC) 316206924Sdelphij extern uch ZLIB_INTERNAL _length_code[]; 317206924Sdelphij extern uch ZLIB_INTERNAL _dist_code[]; 31833908Ssteve#else 319206924Sdelphij extern const uch ZLIB_INTERNAL _length_code[]; 320206924Sdelphij extern const uch ZLIB_INTERNAL _dist_code[]; 32117651Speter#endif 32233908Ssteve 32333908Ssteve# define _tr_tally_lit(s, c, flush) \ 32433908Ssteve { uch cc = (c); \ 32533908Ssteve s->d_buf[s->last_lit] = 0; \ 32633908Ssteve s->l_buf[s->last_lit++] = cc; \ 32733908Ssteve s->dyn_ltree[cc].Freq++; \ 32833908Ssteve flush = (s->last_lit == s->lit_bufsize-1); \ 32933908Ssteve } 33033908Ssteve# define _tr_tally_dist(s, distance, length, flush) \ 33133908Ssteve { uch len = (length); \ 33233908Ssteve ush dist = (distance); \ 33333908Ssteve s->d_buf[s->last_lit] = dist; \ 33433908Ssteve s->l_buf[s->last_lit++] = len; \ 33533908Ssteve dist--; \ 33633908Ssteve s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \ 33733908Ssteve s->dyn_dtree[d_code(dist)].Freq++; \ 33833908Ssteve flush = (s->last_lit == s->lit_bufsize-1); \ 33933908Ssteve } 34033908Ssteve#else 34133908Ssteve# define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c) 34233908Ssteve# define _tr_tally_dist(s, distance, length, flush) \ 343131380Stjr flush = _tr_tally(s, distance, length) 34433908Ssteve#endif 34533908Ssteve 346131380Stjr#endif /* DEFLATE_H */ 347