117651Speter/* deflate.h -- internal compression state 2313796Sdelphij * Copyright (C) 1995-2016 Jean-loup Gailly 3131377Stjr * 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 1133904Ssteve/* @(#) $Id$ */ 1217651Speter 13131377Stjr#ifndef DEFLATE_H 14131377Stjr#define DEFLATE_H 1517651Speter 1617651Speter#include "zutil.h" 1717651Speter 18131377Stjr/* define NO_GZIP when compiling if you want to disable gzip header and 19131377Stjr trailer creation by deflate(). NO_GZIP would be used to avoid linking in 20131377Stjr the crc code when it is not needed. For shared libraries, gzip encoding 21131377Stjr should be left enabled. */ 22131377Stjr#ifndef NO_GZIP 23131377Stjr# define GZIP 24131377Stjr#endif 25131377Stjr 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 51230837Sdelphij#define Buf_size 16 52230837Sdelphij/* size of bit buffer in bi_buf */ 53230837Sdelphij 54313796Sdelphij#define INIT_STATE 42 /* zlib header -> BUSY_STATE */ 55313796Sdelphij#ifdef GZIP 56313796Sdelphij# define GZIP_STATE 57 /* gzip header -> BUSY_STATE | EXTRA_STATE */ 57313796Sdelphij#endif 58313796Sdelphij#define EXTRA_STATE 69 /* gzip extra block -> NAME_STATE */ 59313796Sdelphij#define NAME_STATE 73 /* gzip file name -> COMMENT_STATE */ 60313796Sdelphij#define COMMENT_STATE 91 /* gzip comment -> HCRC_STATE */ 61313796Sdelphij#define HCRC_STATE 103 /* gzip header CRC -> BUSY_STATE */ 62313796Sdelphij#define BUSY_STATE 113 /* deflate -> FINISH_STATE */ 63313796Sdelphij#define FINISH_STATE 666 /* stream complete */ 6417651Speter/* Stream status */ 6517651Speter 6617651Speter 6717651Speter/* Data structure describing a single value and its code string. */ 6817651Spetertypedef struct ct_data_s { 6917651Speter union { 7017651Speter ush freq; /* frequency count */ 7117651Speter ush code; /* bit string */ 7217651Speter } fc; 7317651Speter union { 7417651Speter ush dad; /* father node in Huffman tree */ 7517651Speter ush len; /* length of bit string */ 7617651Speter } dl; 7717651Speter} FAR ct_data; 7817651Speter 7917651Speter#define Freq fc.freq 8017651Speter#define Code fc.code 8117651Speter#define Dad dl.dad 8217651Speter#define Len dl.len 8317651Speter 8417651Spetertypedef struct static_tree_desc_s static_tree_desc; 8517651Speter 8617651Spetertypedef struct tree_desc_s { 8717651Speter ct_data *dyn_tree; /* the dynamic tree */ 8817651Speter int max_code; /* largest code with non zero frequency */ 89313796Sdelphij const static_tree_desc *stat_desc; /* the corresponding static tree */ 9017651Speter} FAR tree_desc; 9117651Speter 9217651Spetertypedef ush Pos; 9317651Spetertypedef Pos FAR Posf; 9417651Spetertypedef unsigned IPos; 9517651Speter 9617651Speter/* A Pos is an index in the character window. We use short instead of int to 9717651Speter * save space in the various tables. IPos is used only for parameter passing. 9817651Speter */ 9917651Speter 10017651Spetertypedef struct internal_state { 10117651Speter z_streamp strm; /* pointer back to this zlib stream */ 10217651Speter int status; /* as the name implies */ 10317651Speter Bytef *pending_buf; /* output still pending */ 10433904Ssteve ulg pending_buf_size; /* size of pending_buf */ 10517651Speter Bytef *pending_out; /* next pending byte to output to the stream */ 106313796Sdelphij ulg pending; /* nb of bytes in the pending buffer */ 107131377Stjr int wrap; /* bit 0 true for zlib, bit 1 true for gzip */ 108157043Sdes gz_headerp gzhead; /* gzip header information to write */ 109313796Sdelphij ulg gzindex; /* where in extra, name, or comment */ 110250224Sdelphij Byte method; /* can only be DEFLATED */ 11117651Speter int last_flush; /* value of flush param for previous deflate call */ 11217651Speter 11317651Speter /* used by deflate.c: */ 11417651Speter 11517651Speter uInt w_size; /* LZ77 window size (32K by default) */ 11617651Speter uInt w_bits; /* log2(w_size) (8..16) */ 11717651Speter uInt w_mask; /* w_size - 1 */ 11817651Speter 11917651Speter Bytef *window; 12017651Speter /* Sliding window. Input bytes are read into the second half of the window, 12117651Speter * and move to the first half later to keep a dictionary of at least wSize 12217651Speter * bytes. With this organization, matches are limited to a distance of 12317651Speter * wSize-MAX_MATCH bytes, but this ensures that IO is always 12417651Speter * performed with a length multiple of the block size. Also, it limits 12517651Speter * the window size to 64K, which is quite useful on MSDOS. 12617651Speter * To do: use the user input buffer as sliding window. 12717651Speter */ 12817651Speter 12917651Speter ulg window_size; 13017651Speter /* Actual size of window: 2*wSize, except when the user input buffer 13117651Speter * is directly used as sliding window. 13217651Speter */ 13317651Speter 13417651Speter Posf *prev; 13517651Speter /* Link to older string with same hash index. To limit the size of this 13617651Speter * array to 64K, this link is maintained only for the last 32K strings. 13717651Speter * An index in this array is thus a window index modulo 32K. 13817651Speter */ 13917651Speter 14017651Speter Posf *head; /* Heads of the hash chains or NIL. */ 14117651Speter 14217651Speter uInt ins_h; /* hash index of string to be inserted */ 14317651Speter uInt hash_size; /* number of elements in hash table */ 14417651Speter uInt hash_bits; /* log2(hash_size) */ 14517651Speter uInt hash_mask; /* hash_size-1 */ 14617651Speter 14717651Speter uInt hash_shift; 14817651Speter /* Number of bits by which ins_h must be shifted at each input 14917651Speter * step. It must be such that after MIN_MATCH steps, the oldest 15017651Speter * byte no longer takes part in the hash key, that is: 15117651Speter * hash_shift * MIN_MATCH >= hash_bits 15217651Speter */ 15317651Speter 15417651Speter long block_start; 15517651Speter /* Window position at the beginning of the current output block. Gets 15617651Speter * negative when the window is moved backwards. 15717651Speter */ 15817651Speter 15917651Speter uInt match_length; /* length of best match */ 16017651Speter IPos prev_match; /* previous match */ 16117651Speter int match_available; /* set if previous match exists */ 16217651Speter uInt strstart; /* start of string to insert */ 16317651Speter uInt match_start; /* start of matching string */ 16417651Speter uInt lookahead; /* number of valid bytes ahead in window */ 16517651Speter 16617651Speter uInt prev_length; 16717651Speter /* Length of the best match at previous step. Matches not greater than this 16817651Speter * are discarded. This is used in the lazy match evaluation. 16917651Speter */ 17017651Speter 17117651Speter uInt max_chain_length; 17217651Speter /* To speed up deflation, hash chains are never searched beyond this 17317651Speter * length. A higher limit improves compression ratio but degrades the 17417651Speter * speed. 17517651Speter */ 17617651Speter 17717651Speter uInt max_lazy_match; 17817651Speter /* Attempt to find a better match only when the current match is strictly 17917651Speter * smaller than this value. This mechanism is used only for compression 18017651Speter * levels >= 4. 18117651Speter */ 18217651Speter# define max_insert_length max_lazy_match 18317651Speter /* Insert new strings in the hash table only if the match length is not 18417651Speter * greater than this length. This saves time but degrades compression. 18517651Speter * max_insert_length is used only for compression levels <= 3. 18617651Speter */ 18717651Speter 18817651Speter int level; /* compression level (1..9) */ 18917651Speter int strategy; /* favor or force Huffman coding*/ 19017651Speter 19117651Speter uInt good_match; 19217651Speter /* Use a faster search when the previous match is longer than this */ 19317651Speter 19417651Speter int nice_match; /* Stop searching when current match exceeds this */ 19517651Speter 19617651Speter /* used by trees.c: */ 197230837Sdelphij /* Didn't use ct_data typedef below to suppress compiler warning */ 19817651Speter struct ct_data_s dyn_ltree[HEAP_SIZE]; /* literal and length tree */ 19917651Speter struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */ 20017651Speter struct ct_data_s bl_tree[2*BL_CODES+1]; /* Huffman tree for bit lengths */ 20117651Speter 20217651Speter struct tree_desc_s l_desc; /* desc. for literal tree */ 20317651Speter struct tree_desc_s d_desc; /* desc. for distance tree */ 20417651Speter struct tree_desc_s bl_desc; /* desc. for bit length tree */ 20517651Speter 20617651Speter ush bl_count[MAX_BITS+1]; 20717651Speter /* number of codes at each bit length for an optimal tree */ 20817651Speter 20917651Speter int heap[2*L_CODES+1]; /* heap used to build the Huffman trees */ 21017651Speter int heap_len; /* number of elements in the heap */ 21117651Speter int heap_max; /* element of largest frequency */ 21217651Speter /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used. 21317651Speter * The same heap array is used to build all trees. 21417651Speter */ 21517651Speter 21617651Speter uch depth[2*L_CODES+1]; 21717651Speter /* Depth of each subtree used as tie breaker for trees of equal frequency 21817651Speter */ 21917651Speter 22017651Speter uchf *l_buf; /* buffer for literals or lengths */ 22117651Speter 22217651Speter uInt lit_bufsize; 22317651Speter /* Size of match buffer for literals/lengths. There are 4 reasons for 22417651Speter * limiting lit_bufsize to 64K: 22517651Speter * - frequencies can be kept in 16 bit counters 22617651Speter * - if compression is not successful for the first block, all input 22717651Speter * data is still in the window so we can still emit a stored block even 22817651Speter * when input comes from standard input. (This can also be done for 22917651Speter * all blocks if lit_bufsize is not greater than 32K.) 23017651Speter * - if compression is not successful for a file smaller than 64K, we can 23117651Speter * even emit a stored file instead of a stored block (saving 5 bytes). 23217651Speter * This is applicable only for zip (not gzip or zlib). 23317651Speter * - creating new Huffman trees less frequently may not provide fast 23417651Speter * adaptation to changes in the input data statistics. (Take for 23517651Speter * example a binary file with poorly compressible code followed by 23617651Speter * a highly compressible string table.) Smaller buffer sizes give 23717651Speter * fast adaptation but have of course the overhead of transmitting 23817651Speter * trees more frequently. 23917651Speter * - I can't count above 4 24017651Speter */ 24117651Speter 24217651Speter uInt last_lit; /* running index in l_buf */ 24317651Speter 24417651Speter ushf *d_buf; 24517651Speter /* Buffer for distances. To simplify the code, d_buf and l_buf have 24617651Speter * the same number of elements. To use different lengths, an extra flag 24717651Speter * array would be necessary. 24817651Speter */ 24917651Speter 25017651Speter ulg opt_len; /* bit length of current block with optimal trees */ 25117651Speter ulg static_len; /* bit length of current block with static trees */ 25217651Speter uInt matches; /* number of string matches in current block */ 253230837Sdelphij uInt insert; /* bytes at end of window left to insert */ 25417651Speter 255313796Sdelphij#ifdef ZLIB_DEBUG 25642468Speter ulg compressed_len; /* total bit length of compressed file mod 2^32 */ 25742468Speter ulg bits_sent; /* bit length of compressed data sent mod 2^32 */ 25817651Speter#endif 25917651Speter 26017651Speter ush bi_buf; 26117651Speter /* Output buffer. bits are inserted starting at the bottom (least 26217651Speter * significant bits). 26317651Speter */ 26417651Speter int bi_valid; 26517651Speter /* Number of valid bits in bi_buf. All bits above the last valid bit 26617651Speter * are always zero. 26717651Speter */ 26817651Speter 269205194Sdelphij ulg high_water; 270205194Sdelphij /* High water mark offset in window for initialized bytes -- bytes above 271205194Sdelphij * this are set to zero in order to avoid memory check warnings when 272205194Sdelphij * longest match routines access bytes past the input. This is then 273205194Sdelphij * updated to the new high water mark. 274205194Sdelphij */ 275205194Sdelphij 27617651Speter} FAR deflate_state; 27717651Speter 27817651Speter/* Output a byte on the stream. 27917651Speter * IN assertion: there is enough room in pending_buf. 28017651Speter */ 281313796Sdelphij#define put_byte(s, c) {s->pending_buf[s->pending++] = (Bytef)(c);} 28217651Speter 28317651Speter 28417651Speter#define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1) 28517651Speter/* Minimum amount of lookahead, except at the end of the input file. 28617651Speter * See deflate.c for comments about the MIN_MATCH+1. 28717651Speter */ 28817651Speter 28917651Speter#define MAX_DIST(s) ((s)->w_size-MIN_LOOKAHEAD) 29017651Speter/* In order to simplify the code, particularly on 16 bit machines, match 29117651Speter * distances are limited to MAX_DIST instead of WSIZE. 29217651Speter */ 29317651Speter 294205194Sdelphij#define WIN_INIT MAX_MATCH 295205194Sdelphij/* Number of bytes after end of data in window to initialize in order to avoid 296205194Sdelphij memory checker errors from longest match routines */ 297205194Sdelphij 29817651Speter /* in trees.c */ 299206905Sdelphijvoid ZLIB_INTERNAL _tr_init OF((deflate_state *s)); 300206905Sdelphijint ZLIB_INTERNAL _tr_tally OF((deflate_state *s, unsigned dist, unsigned lc)); 301206905Sdelphijvoid ZLIB_INTERNAL _tr_flush_block OF((deflate_state *s, charf *buf, 302206905Sdelphij ulg stored_len, int last)); 303230837Sdelphijvoid ZLIB_INTERNAL _tr_flush_bits OF((deflate_state *s)); 304206905Sdelphijvoid ZLIB_INTERNAL _tr_align OF((deflate_state *s)); 305206905Sdelphijvoid ZLIB_INTERNAL _tr_stored_block OF((deflate_state *s, charf *buf, 306206905Sdelphij ulg stored_len, int last)); 30733904Ssteve 30833904Ssteve#define d_code(dist) \ 30933904Ssteve ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)]) 31033904Ssteve/* Mapping from a distance to a distance code. dist is the distance - 1 and 31133904Ssteve * must not have side effects. _dist_code[256] and _dist_code[257] are never 31233904Ssteve * used. 31333904Ssteve */ 31433904Ssteve 315313796Sdelphij#ifndef ZLIB_DEBUG 31633904Ssteve/* Inline versions of _tr_tally for speed: */ 31733904Ssteve 31833904Ssteve#if defined(GEN_TREES_H) || !defined(STDC) 319206905Sdelphij extern uch ZLIB_INTERNAL _length_code[]; 320206905Sdelphij extern uch ZLIB_INTERNAL _dist_code[]; 32133904Ssteve#else 322206905Sdelphij extern const uch ZLIB_INTERNAL _length_code[]; 323206905Sdelphij extern const uch ZLIB_INTERNAL _dist_code[]; 32417651Speter#endif 32533904Ssteve 32633904Ssteve# define _tr_tally_lit(s, c, flush) \ 32733904Ssteve { uch cc = (c); \ 32833904Ssteve s->d_buf[s->last_lit] = 0; \ 32933904Ssteve s->l_buf[s->last_lit++] = cc; \ 33033904Ssteve s->dyn_ltree[cc].Freq++; \ 33133904Ssteve flush = (s->last_lit == s->lit_bufsize-1); \ 33233904Ssteve } 33333904Ssteve# define _tr_tally_dist(s, distance, length, flush) \ 334313796Sdelphij { uch len = (uch)(length); \ 335313796Sdelphij ush dist = (ush)(distance); \ 33633904Ssteve s->d_buf[s->last_lit] = dist; \ 33733904Ssteve s->l_buf[s->last_lit++] = len; \ 33833904Ssteve dist--; \ 33933904Ssteve s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \ 34033904Ssteve s->dyn_dtree[d_code(dist)].Freq++; \ 34133904Ssteve flush = (s->last_lit == s->lit_bufsize-1); \ 34233904Ssteve } 34333904Ssteve#else 34433904Ssteve# define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c) 34533904Ssteve# define _tr_tally_dist(s, distance, length, flush) \ 346131377Stjr flush = _tr_tally(s, distance, length) 34733904Ssteve#endif 34833904Ssteve 349131377Stjr#endif /* DEFLATE_H */ 350