Chris@43: /* deflate.h -- internal compression state Chris@43: * Copyright (C) 1995-2012 Jean-loup Gailly Chris@43: * For conditions of distribution and use, see copyright notice in zlib.h Chris@43: */ Chris@43: Chris@43: /* WARNING: this file should *not* be used by applications. It is Chris@43: part of the implementation of the compression library and is Chris@43: subject to change. Applications should only use zlib.h. Chris@43: */ Chris@43: Chris@43: /* @(#) $Id$ */ Chris@43: Chris@43: #ifndef DEFLATE_H Chris@43: #define DEFLATE_H Chris@43: Chris@43: #include "zutil.h" Chris@43: Chris@43: /* define NO_GZIP when compiling if you want to disable gzip header and Chris@43: trailer creation by deflate(). NO_GZIP would be used to avoid linking in Chris@43: the crc code when it is not needed. For shared libraries, gzip encoding Chris@43: should be left enabled. */ Chris@43: #ifndef NO_GZIP Chris@43: # define GZIP Chris@43: #endif Chris@43: Chris@43: /* =========================================================================== Chris@43: * Internal compression state. Chris@43: */ Chris@43: Chris@43: #define LENGTH_CODES 29 Chris@43: /* number of length codes, not counting the special END_BLOCK code */ Chris@43: Chris@43: #define LITERALS 256 Chris@43: /* number of literal bytes 0..255 */ Chris@43: Chris@43: #define L_CODES (LITERALS+1+LENGTH_CODES) Chris@43: /* number of Literal or Length codes, including the END_BLOCK code */ Chris@43: Chris@43: #define D_CODES 30 Chris@43: /* number of distance codes */ Chris@43: Chris@43: #define BL_CODES 19 Chris@43: /* number of codes used to transfer the bit lengths */ Chris@43: Chris@43: #define HEAP_SIZE (2*L_CODES+1) Chris@43: /* maximum heap size */ Chris@43: Chris@43: #define MAX_BITS 15 Chris@43: /* All codes must not exceed MAX_BITS bits */ Chris@43: Chris@43: #define Buf_size 16 Chris@43: /* size of bit buffer in bi_buf */ Chris@43: Chris@43: #define INIT_STATE 42 Chris@43: #define EXTRA_STATE 69 Chris@43: #define NAME_STATE 73 Chris@43: #define COMMENT_STATE 91 Chris@43: #define HCRC_STATE 103 Chris@43: #define BUSY_STATE 113 Chris@43: #define FINISH_STATE 666 Chris@43: /* Stream status */ Chris@43: Chris@43: Chris@43: /* Data structure describing a single value and its code string. */ Chris@43: typedef struct ct_data_s { Chris@43: union { Chris@43: ush freq; /* frequency count */ Chris@43: ush code; /* bit string */ Chris@43: } fc; Chris@43: union { Chris@43: ush dad; /* father node in Huffman tree */ Chris@43: ush len; /* length of bit string */ Chris@43: } dl; Chris@43: } FAR ct_data; Chris@43: Chris@43: #define Freq fc.freq Chris@43: #define Code fc.code Chris@43: #define Dad dl.dad Chris@43: #define Len dl.len Chris@43: Chris@43: typedef struct static_tree_desc_s static_tree_desc; Chris@43: Chris@43: typedef struct tree_desc_s { Chris@43: ct_data *dyn_tree; /* the dynamic tree */ Chris@43: int max_code; /* largest code with non zero frequency */ Chris@43: static_tree_desc *stat_desc; /* the corresponding static tree */ Chris@43: } FAR tree_desc; Chris@43: Chris@43: typedef ush Pos; Chris@43: typedef Pos FAR Posf; Chris@43: typedef unsigned IPos; Chris@43: Chris@43: /* A Pos is an index in the character window. We use short instead of int to Chris@43: * save space in the various tables. IPos is used only for parameter passing. Chris@43: */ Chris@43: Chris@43: typedef struct internal_state { Chris@43: z_streamp strm; /* pointer back to this zlib stream */ Chris@43: int status; /* as the name implies */ Chris@43: Bytef *pending_buf; /* output still pending */ Chris@43: ulg pending_buf_size; /* size of pending_buf */ Chris@43: Bytef *pending_out; /* next pending byte to output to the stream */ Chris@43: uInt pending; /* nb of bytes in the pending buffer */ Chris@43: int wrap; /* bit 0 true for zlib, bit 1 true for gzip */ Chris@43: gz_headerp gzhead; /* gzip header information to write */ Chris@43: uInt gzindex; /* where in extra, name, or comment */ Chris@43: Byte method; /* can only be DEFLATED */ Chris@43: int last_flush; /* value of flush param for previous deflate call */ Chris@43: Chris@43: /* used by deflate.c: */ Chris@43: Chris@43: uInt w_size; /* LZ77 window size (32K by default) */ Chris@43: uInt w_bits; /* log2(w_size) (8..16) */ Chris@43: uInt w_mask; /* w_size - 1 */ Chris@43: Chris@43: Bytef *window; Chris@43: /* Sliding window. Input bytes are read into the second half of the window, Chris@43: * and move to the first half later to keep a dictionary of at least wSize Chris@43: * bytes. With this organization, matches are limited to a distance of Chris@43: * wSize-MAX_MATCH bytes, but this ensures that IO is always Chris@43: * performed with a length multiple of the block size. Also, it limits Chris@43: * the window size to 64K, which is quite useful on MSDOS. Chris@43: * To do: use the user input buffer as sliding window. Chris@43: */ Chris@43: Chris@43: ulg window_size; Chris@43: /* Actual size of window: 2*wSize, except when the user input buffer Chris@43: * is directly used as sliding window. Chris@43: */ Chris@43: Chris@43: Posf *prev; Chris@43: /* Link to older string with same hash index. To limit the size of this Chris@43: * array to 64K, this link is maintained only for the last 32K strings. Chris@43: * An index in this array is thus a window index modulo 32K. Chris@43: */ Chris@43: Chris@43: Posf *head; /* Heads of the hash chains or NIL. */ Chris@43: Chris@43: uInt ins_h; /* hash index of string to be inserted */ Chris@43: uInt hash_size; /* number of elements in hash table */ Chris@43: uInt hash_bits; /* log2(hash_size) */ Chris@43: uInt hash_mask; /* hash_size-1 */ Chris@43: Chris@43: uInt hash_shift; Chris@43: /* Number of bits by which ins_h must be shifted at each input Chris@43: * step. It must be such that after MIN_MATCH steps, the oldest Chris@43: * byte no longer takes part in the hash key, that is: Chris@43: * hash_shift * MIN_MATCH >= hash_bits Chris@43: */ Chris@43: Chris@43: long block_start; Chris@43: /* Window position at the beginning of the current output block. Gets Chris@43: * negative when the window is moved backwards. Chris@43: */ Chris@43: Chris@43: uInt match_length; /* length of best match */ Chris@43: IPos prev_match; /* previous match */ Chris@43: int match_available; /* set if previous match exists */ Chris@43: uInt strstart; /* start of string to insert */ Chris@43: uInt match_start; /* start of matching string */ Chris@43: uInt lookahead; /* number of valid bytes ahead in window */ Chris@43: Chris@43: uInt prev_length; Chris@43: /* Length of the best match at previous step. Matches not greater than this Chris@43: * are discarded. This is used in the lazy match evaluation. Chris@43: */ Chris@43: Chris@43: uInt max_chain_length; Chris@43: /* To speed up deflation, hash chains are never searched beyond this Chris@43: * length. A higher limit improves compression ratio but degrades the Chris@43: * speed. Chris@43: */ Chris@43: Chris@43: uInt max_lazy_match; Chris@43: /* Attempt to find a better match only when the current match is strictly Chris@43: * smaller than this value. This mechanism is used only for compression Chris@43: * levels >= 4. Chris@43: */ Chris@43: # define max_insert_length max_lazy_match Chris@43: /* Insert new strings in the hash table only if the match length is not Chris@43: * greater than this length. This saves time but degrades compression. Chris@43: * max_insert_length is used only for compression levels <= 3. Chris@43: */ Chris@43: Chris@43: int level; /* compression level (1..9) */ Chris@43: int strategy; /* favor or force Huffman coding*/ Chris@43: Chris@43: uInt good_match; Chris@43: /* Use a faster search when the previous match is longer than this */ Chris@43: Chris@43: int nice_match; /* Stop searching when current match exceeds this */ Chris@43: Chris@43: /* used by trees.c: */ Chris@43: /* Didn't use ct_data typedef below to suppress compiler warning */ Chris@43: struct ct_data_s dyn_ltree[HEAP_SIZE]; /* literal and length tree */ Chris@43: struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */ Chris@43: struct ct_data_s bl_tree[2*BL_CODES+1]; /* Huffman tree for bit lengths */ Chris@43: Chris@43: struct tree_desc_s l_desc; /* desc. for literal tree */ Chris@43: struct tree_desc_s d_desc; /* desc. for distance tree */ Chris@43: struct tree_desc_s bl_desc; /* desc. for bit length tree */ Chris@43: Chris@43: ush bl_count[MAX_BITS+1]; Chris@43: /* number of codes at each bit length for an optimal tree */ Chris@43: Chris@43: int heap[2*L_CODES+1]; /* heap used to build the Huffman trees */ Chris@43: int heap_len; /* number of elements in the heap */ Chris@43: int heap_max; /* element of largest frequency */ Chris@43: /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used. Chris@43: * The same heap array is used to build all trees. Chris@43: */ Chris@43: Chris@43: uch depth[2*L_CODES+1]; Chris@43: /* Depth of each subtree used as tie breaker for trees of equal frequency Chris@43: */ Chris@43: Chris@43: uchf *l_buf; /* buffer for literals or lengths */ Chris@43: Chris@43: uInt lit_bufsize; Chris@43: /* Size of match buffer for literals/lengths. There are 4 reasons for Chris@43: * limiting lit_bufsize to 64K: Chris@43: * - frequencies can be kept in 16 bit counters Chris@43: * - if compression is not successful for the first block, all input Chris@43: * data is still in the window so we can still emit a stored block even Chris@43: * when input comes from standard input. (This can also be done for Chris@43: * all blocks if lit_bufsize is not greater than 32K.) Chris@43: * - if compression is not successful for a file smaller than 64K, we can Chris@43: * even emit a stored file instead of a stored block (saving 5 bytes). Chris@43: * This is applicable only for zip (not gzip or zlib). Chris@43: * - creating new Huffman trees less frequently may not provide fast Chris@43: * adaptation to changes in the input data statistics. (Take for Chris@43: * example a binary file with poorly compressible code followed by Chris@43: * a highly compressible string table.) Smaller buffer sizes give Chris@43: * fast adaptation but have of course the overhead of transmitting Chris@43: * trees more frequently. Chris@43: * - I can't count above 4 Chris@43: */ Chris@43: Chris@43: uInt last_lit; /* running index in l_buf */ Chris@43: Chris@43: ushf *d_buf; Chris@43: /* Buffer for distances. To simplify the code, d_buf and l_buf have Chris@43: * the same number of elements. To use different lengths, an extra flag Chris@43: * array would be necessary. Chris@43: */ Chris@43: Chris@43: ulg opt_len; /* bit length of current block with optimal trees */ Chris@43: ulg static_len; /* bit length of current block with static trees */ Chris@43: uInt matches; /* number of string matches in current block */ Chris@43: uInt insert; /* bytes at end of window left to insert */ Chris@43: Chris@43: #ifdef DEBUG Chris@43: ulg compressed_len; /* total bit length of compressed file mod 2^32 */ Chris@43: ulg bits_sent; /* bit length of compressed data sent mod 2^32 */ Chris@43: #endif Chris@43: Chris@43: ush bi_buf; Chris@43: /* Output buffer. bits are inserted starting at the bottom (least Chris@43: * significant bits). Chris@43: */ Chris@43: int bi_valid; Chris@43: /* Number of valid bits in bi_buf. All bits above the last valid bit Chris@43: * are always zero. Chris@43: */ Chris@43: Chris@43: ulg high_water; Chris@43: /* High water mark offset in window for initialized bytes -- bytes above Chris@43: * this are set to zero in order to avoid memory check warnings when Chris@43: * longest match routines access bytes past the input. This is then Chris@43: * updated to the new high water mark. Chris@43: */ Chris@43: Chris@43: } FAR deflate_state; Chris@43: Chris@43: /* Output a byte on the stream. Chris@43: * IN assertion: there is enough room in pending_buf. Chris@43: */ Chris@43: #define put_byte(s, c) {s->pending_buf[s->pending++] = (c);} Chris@43: Chris@43: Chris@43: #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1) Chris@43: /* Minimum amount of lookahead, except at the end of the input file. Chris@43: * See deflate.c for comments about the MIN_MATCH+1. Chris@43: */ Chris@43: Chris@43: #define MAX_DIST(s) ((s)->w_size-MIN_LOOKAHEAD) Chris@43: /* In order to simplify the code, particularly on 16 bit machines, match Chris@43: * distances are limited to MAX_DIST instead of WSIZE. Chris@43: */ Chris@43: Chris@43: #define WIN_INIT MAX_MATCH Chris@43: /* Number of bytes after end of data in window to initialize in order to avoid Chris@43: memory checker errors from longest match routines */ Chris@43: Chris@43: /* in trees.c */ Chris@43: void ZLIB_INTERNAL _tr_init OF((deflate_state *s)); Chris@43: int ZLIB_INTERNAL _tr_tally OF((deflate_state *s, unsigned dist, unsigned lc)); Chris@43: void ZLIB_INTERNAL _tr_flush_block OF((deflate_state *s, charf *buf, Chris@43: ulg stored_len, int last)); Chris@43: void ZLIB_INTERNAL _tr_flush_bits OF((deflate_state *s)); Chris@43: void ZLIB_INTERNAL _tr_align OF((deflate_state *s)); Chris@43: void ZLIB_INTERNAL _tr_stored_block OF((deflate_state *s, charf *buf, Chris@43: ulg stored_len, int last)); Chris@43: Chris@43: #define d_code(dist) \ Chris@43: ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)]) Chris@43: /* Mapping from a distance to a distance code. dist is the distance - 1 and Chris@43: * must not have side effects. _dist_code[256] and _dist_code[257] are never Chris@43: * used. Chris@43: */ Chris@43: Chris@43: #ifndef DEBUG Chris@43: /* Inline versions of _tr_tally for speed: */ Chris@43: Chris@43: #if defined(GEN_TREES_H) || !defined(STDC) Chris@43: extern uch ZLIB_INTERNAL _length_code[]; Chris@43: extern uch ZLIB_INTERNAL _dist_code[]; Chris@43: #else Chris@43: extern const uch ZLIB_INTERNAL _length_code[]; Chris@43: extern const uch ZLIB_INTERNAL _dist_code[]; Chris@43: #endif Chris@43: Chris@43: # define _tr_tally_lit(s, c, flush) \ Chris@43: { uch cc = (c); \ Chris@43: s->d_buf[s->last_lit] = 0; \ Chris@43: s->l_buf[s->last_lit++] = cc; \ Chris@43: s->dyn_ltree[cc].Freq++; \ Chris@43: flush = (s->last_lit == s->lit_bufsize-1); \ Chris@43: } Chris@43: # define _tr_tally_dist(s, distance, length, flush) \ Chris@43: { uch len = (length); \ Chris@43: ush dist = (distance); \ Chris@43: s->d_buf[s->last_lit] = dist; \ Chris@43: s->l_buf[s->last_lit++] = len; \ Chris@43: dist--; \ Chris@43: s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \ Chris@43: s->dyn_dtree[d_code(dist)].Freq++; \ Chris@43: flush = (s->last_lit == s->lit_bufsize-1); \ Chris@43: } Chris@43: #else Chris@43: # define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c) Chris@43: # define _tr_tally_dist(s, distance, length, flush) \ Chris@43: flush = _tr_tally(s, distance, length) Chris@43: #endif Chris@43: Chris@43: #endif /* DEFLATE_H */