zlib/deflate.h

   1 /* deflate.h -- internal compression state
   2  * Copyright (C) 1995-2002 Jean-loup Gailly
   3  * For conditions of distribution and use, see copyright notice in zlib.h
   4  */
   5
   6 /* WARNING: this file should *not* be used by applications. It is
   7    part of the implementation of the compression library and is
   8    subject to change. Applications should only use zlib.h.
   9  */
  10
  11 /* @(#) $Id$ */
  12
  13 #ifndef _DEFLATE_H
  14 #define _DEFLATE_H
  15
  16 #include "zutil.h"
  17
  18 /* ===========================================================================
  19  * Internal compression state.
  20  */
  21
  22 #define LENGTH_CODES 29
  23 /* number of length codes, not counting the special END_BLOCK code */
  24
  25 #define LITERALS  256
  26 /* number of literal bytes 0..255 */
  27
  28 #define L_CODES (LITERALS+1+LENGTH_CODES)
  29 /* number of Literal or Length codes, including the END_BLOCK code */
  30
  31 #define D_CODES   30
  32 /* number of distance codes */
  33
  34 #define BL_CODES  19
  35 /* number of codes used to transfer the bit lengths */
  36
  37 #define HEAP_SIZE (2*L_CODES+1)
  38 /* maximum heap size */
  39
  40 #define MAX_BITS 15
  41 /* All codes must not exceed MAX_BITS bits */
  42
  43 #define INIT_STATE    42
  44 #define BUSY_STATE   113
  45 #define FINISH_STATE 666
  46 /* Stream status */
  47
  48
  49 /* Data structure describing a single value and its code string. */
  50 typedef struct ct_data_s {
  51     union {
  52         ush  freq;       /* frequency count */
  53         ush  code;       /* bit string */
  54     } fc;
  55     union {
  56         ush  dad;        /* father node in Huffman tree */
  57         ush  len;        /* length of bit string */
  58     } dl;
  59 } FAR ct_data;
  60
  61 #define Freq fc.freq
  62 #define Code fc.code
  63 #define Dad  dl.dad
  64 #define Len  dl.len
  65
  66 typedef struct static_tree_desc_s  static_tree_desc;
  67
  68 typedef struct tree_desc_s {
  69     ct_data *dyn_tree;           /* the dynamic tree */
  70     int     max_code;            /* largest code with non zero frequency */
  71     static_tree_desc *stat_desc; /* the corresponding static tree */
  72 } FAR tree_desc;
  73
  74 typedef ush Pos;
  75 typedef Pos FAR Posf;
  76 typedef unsigned IPos;
  77
  78 /* A Pos is an index in the character window. We use short instead of int to
  79  * save space in the various tables. IPos is used only for parameter passing.
  80  */
  81
  82 typedef struct internal_state {
  83     z_streamp strm;      /* pointer back to this zlib stream */
  84     int   status;        /* as the name implies */
  85     Bytef *pending_buf;  /* output still pending */
  86     ulg   pending_buf_size; /* size of pending_buf */
  87     Bytef *pending_out;  /* next pending byte to output to the stream */
  88     int   pending;       /* nb of bytes in the pending buffer */
  89     int   noheader;      /* suppress zlib header and adler32 */
  90     Byte  data_type;     /* UNKNOWN, BINARY or ASCII */
  91     Byte  method;        /* STORED (for zip only) or DEFLATED */
  92     int   last_flush;    /* value of flush param for previous deflate call */
  93
  94                 /* used by deflate.c: */
  95
  96     uInt  w_size;        /* LZ77 window size (32K by default) */
  97     uInt  w_bits;        /* log2(w_size)  (8..16) */
  98     uInt  w_mask;        /* w_size - 1 */
  99
 100     Bytef *window;
 101     /* Sliding window. Input bytes are read into the second half of the window,
 102      * and move to the first half later to keep a dictionary of at least wSize
 103      * bytes. With this organization, matches are limited to a distance of
 104      * wSize-MAX_MATCH bytes, but this ensures that IO is always
 105      * performed with a length multiple of the block size. Also, it limits
 106      * the window size to 64K, which is quite useful on MSDOS.
 107      * To do: use the user input buffer as sliding window.
 108      */
 109
 110     ulg window_size;
 111     /* Actual size of window: 2*wSize, except when the user input buffer
 112      * is directly used as sliding window.
 113      */
 114
 115     Posf *prev;
 116     /* Link to older string with same hash index. To limit the size of this
 117      * array to 64K, this link is maintained only for the last 32K strings.
 118      * An index in this array is thus a window index modulo 32K.
 119      */
 120
 121     Posf *head; /* Heads of the hash chains or NIL. */
 122
 123     uInt  ins_h;          /* hash index of string to be inserted */
 124     uInt  hash_size;      /* number of elements in hash table */
 125     uInt  hash_bits;      /* log2(hash_size) */
 126     uInt  hash_mask;      /* hash_size-1 */
 127
 128     uInt  hash_shift;
 129     /* Number of bits by which ins_h must be shifted at each input
 130      * step. It must be such that after MIN_MATCH steps, the oldest
 131      * byte no longer takes part in the hash key, that is:
 132      *   hash_shift * MIN_MATCH >= hash_bits
 133      */
 134
 135     long block_start;
 136     /* Window position at the beginning of the current output block. Gets
 137      * negative when the window is moved backwards.
 138      */
 139
 140     uInt match_length;           /* length of best match */
 141     IPos prev_match;             /* previous match */
 142     int match_available;         /* set if previous match exists */
 143     uInt strstart;               /* start of string to insert */
 144     uInt match_start;            /* start of matching string */
 145     uInt lookahead;              /* number of valid bytes ahead in window */
 146
 147     uInt prev_length;
 148     /* Length of the best match at previous step. Matches not greater than this
 149      * are discarded. This is used in the lazy match evaluation.
 150      */
 151
 152     uInt max_chain_length;
 153     /* To speed up deflation, hash chains are never searched beyond this
 154      * length.  A higher limit improves compression ratio but degrades the
 155      * speed.
 156      */
 157
 158     uInt max_lazy_match;
 159     /* Attempt to find a better match only when the current match is strictly
 160      * smaller than this value. This mechanism is used only for compression
 161      * levels >= 4.
 162      */
 163 #   define max_insert_length  max_lazy_match
 164     /* Insert new strings in the hash table only if the match length is not
 165      * greater than this length. This saves time but degrades compression.
 166      * max_insert_length is used only for compression levels <= 3.
 167      */
 168
 169     int level;    /* compression level (1..9) */
 170     int strategy; /* favor or force Huffman coding*/
 171
 172     uInt good_match;
 173     /* Use a faster search when the previous match is longer than this */
 174
 175     int nice_match; /* Stop searching when current match exceeds this */
 176
 177                 /* used by trees.c: */
 178     /* Didn't use ct_data typedef below to supress compiler warning */
 179     struct ct_data_s dyn_ltree[HEAP_SIZE];   /* literal and length tree */
 180     struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */
 181     struct ct_data_s bl_tree[2*BL_CODES+1];  /* Huffman tree for bit lengths */
 182
 183     struct tree_desc_s l_desc;               /* desc. for literal tree */
 184     struct tree_desc_s d_desc;               /* desc. for distance tree */
 185     struct tree_desc_s bl_desc;              /* desc. for bit length tree */
 186
 187     ush bl_count[MAX_BITS+1];
 188     /* number of codes at each bit length for an optimal tree */
 189
 190     int heap[2*L_CODES+1];      /* heap used to build the Huffman trees */
 191     int heap_len;               /* number of elements in the heap */
 192     int heap_max;               /* element of largest frequency */
 193     /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
 194      * The same heap array is used to build all trees.
 195      */
 196
 197     uch depth[2*L_CODES+1];
 198     /* Depth of each subtree used as tie breaker for trees of equal frequency
 199      */
 200
 201     uchf *l_buf;          /* buffer for literals or lengths */
 202
 203     uInt  lit_bufsize;
 204     /* Size of match buffer for literals/lengths.  There are 4 reasons for
 205      * limiting lit_bufsize to 64K:
 206      *   - frequencies can be kept in 16 bit counters
 207      *   - if compression is not successful for the first block, all input
 208      *     data is still in the window so we can still emit a stored block even
 209      *     when input comes from standard input.  (This can also be done for
 210      *     all blocks if lit_bufsize is not greater than 32K.)
 211      *   - if compression is not successful for a file smaller than 64K, we can
 212      *     even emit a stored file instead of a stored block (saving 5 bytes).
 213      *     This is applicable only for zip (not gzip or zlib).
 214      *   - creating new Huffman trees less frequently may not provide fast
 215      *     adaptation to changes in the input data statistics. (Take for
 216      *     example a binary file with poorly compressible code followed by
 217      *     a highly compressible string table.) Smaller buffer sizes give
 218      *     fast adaptation but have of course the overhead of transmitting
 219      *     trees more frequently.
 220      *   - I can't count above 4
 221      */
 222
 223     uInt last_lit;      /* running index in l_buf */
 224
 225     ushf *d_buf;
 226     /* Buffer for distances. To simplify the code, d_buf and l_buf have
 227      * the same number of elements. To use different lengths, an extra flag
 228      * array would be necessary.
 229      */
 230
 231     ulg opt_len;        /* bit length of current block with optimal trees */
 232     ulg static_len;     /* bit length of current block with static trees */
 233     uInt matches;       /* number of string matches in current block */
 234     int last_eob_len;   /* bit length of EOB code for last block */
 235
 236 #ifdef DEBUG
 237     ulg compressed_len; /* total bit length of compressed file mod 2^32 */
 238     ulg bits_sent;      /* bit length of compressed data sent mod 2^32 */
 239 #endif
 240
 241     ush bi_buf;
 242     /* Output buffer. bits are inserted starting at the bottom (least
 243      * significant bits).
 244      */
 245     int bi_valid;
 246     /* Number of valid bits in bi_buf.  All bits above the last valid bit
 247      * are always zero.
 248      */
 249
 250 } FAR deflate_state;
 251
 252 /* Output a byte on the stream.
 253  * IN assertion: there is enough room in pending_buf.
 254  */
 255 #define put_byte(s, c) {s->pending_buf[s->pending++] = (c);}
 256
 257
 258 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
 259 /* Minimum amount of lookahead, except at the end of the input file.
 260  * See deflate.c for comments about the MIN_MATCH+1.
 261  */
 262
 263 #define MAX_DIST(s)  ((s)->w_size-MIN_LOOKAHEAD)
 264 /* In order to simplify the code, particularly on 16 bit machines, match
 265  * distances are limited to MAX_DIST instead of WSIZE.
 266  */
 267
 268         /* in trees.c */
 269 void _tr_init         OF((deflate_state *s));
 270 int  _tr_tally        OF((deflate_state *s, unsigned dist, unsigned lc));
 271 void _tr_flush_block  OF((deflate_state *s, charf *buf, ulg stored_len,
 272                           int eof));
 273 void _tr_align        OF((deflate_state *s));
 274 void _tr_stored_block OF((deflate_state *s, charf *buf, ulg stored_len,
 275                           int eof));
 276
 277 #define d_code(dist) \
 278    ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)])
 279 /* Mapping from a distance to a distance code. dist is the distance - 1 and
 280  * must not have side effects. _dist_code[256] and _dist_code[257] are never
 281  * used.
 282  */
 283
 284 #ifndef DEBUG
 285 /* Inline versions of _tr_tally for speed: */
 286
 287 #if defined(GEN_TREES_H) || !defined(STDC)
 288   extern uch _length_code[];
 289   extern uch _dist_code[];
 290 #else
 291   extern const uch _length_code[];
 292   extern const uch _dist_code[];
 293 #endif
 294
 295 # define _tr_tally_lit(s, c, flush) \
 296   { uch cc = (c); \
 297     s->d_buf[s->last_lit] = 0; \
 298     s->l_buf[s->last_lit++] = cc; \
 299     s->dyn_ltree[cc].Freq++; \
 300     flush = (s->last_lit == s->lit_bufsize-1); \
 301    }
 302 # define _tr_tally_dist(s, distance, length, flush) \
 303   { uch len = (length); \
 304     ush dist = (distance); \
 305     s->d_buf[s->last_lit] = dist; \
 306     s->l_buf[s->last_lit++] = len; \
 307     dist--; \
 308     s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \
 309     s->dyn_dtree[d_code(dist)].Freq++; \
 310     flush = (s->last_lit == s->lit_bufsize-1); \
 311   }
 312 #else
 313 # define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c)
 314 # define _tr_tally_dist(s, distance, length, flush) \
 315               flush = _tr_tally(s, distance, length)
 316 #endif
 317
 318 #endif