src/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 /* define NO_GZIP when compiling if you want to disable gzip header and
  19    trailer creation by deflate().  NO_GZIP would be used to avoid linking in
  20    the crc code when it is not needed.  For shared libraries, gzip encoding
  21    should be left enabled. */
  22 #ifndef NO_GZIP
  23 #  define GZIP
  24 #endif
  25
  26 /* ===========================================================================
  27  * Internal compression state.
  28  */
  29
  30 #define LENGTH_CODES 29
  31 /* number of length codes, not counting the special END_BLOCK code */
  32
  33 #define LITERALS  256
  34 /* number of literal bytes 0..255 */
  35
  36 #define L_CODES (LITERALS+1+LENGTH_CODES)
  37 /* number of Literal or Length codes, including the END_BLOCK code */
  38
  39 #define D_CODES   30
  40 /* number of distance codes */
  41
  42 #define BL_CODES  19
  43 /* number of codes used to transfer the bit lengths */
  44
  45 #define HEAP_SIZE (2*L_CODES+1)
  46 /* maximum heap size */
  47
  48 #define MAX_BITS 15
  49 /* All codes must not exceed MAX_BITS bits */
  50
  51 #define INIT_STATE    42
  52 #define BUSY_STATE   113
  53 #define FINISH_STATE 666
  54 /* Stream status */
  55
  56
  57 /* Data structure describing a single value and its code string. */
  58 typedef struct ct_data_s {
  59     union {
  60         ush  freq;       /* frequency count */
  61         ush  code;       /* bit string */
  62     } fc;
  63     union {
  64         ush  dad;        /* father node in Huffman tree */
  65         ush  len;        /* length of bit string */
  66     } dl;
  67 } FAR ct_data;
  68
  69 #define Freq fc.freq
  70 #define Code fc.code
  71 #define Dad  dl.dad
  72 #define Len  dl.len
  73
  74 typedef struct static_tree_desc_s  static_tree_desc;
  75
  76 typedef struct tree_desc_s {
  77     ct_data *dyn_tree;           /* the dynamic tree */
  78     int     max_code;            /* largest code with non zero frequency */
  79     static_tree_desc *stat_desc; /* the corresponding static tree */
  80 } FAR tree_desc;
  81
  82 typedef ush Pos;
  83 typedef Pos FAR Posf;
  84 typedef unsigned IPos;
  85
  86 /* A Pos is an index in the character window. We use short instead of int to
  87  * save space in the various tables. IPos is used only for parameter passing.
  88  */
  89
  90 typedef struct internal_state {
  91     z_streamp strm;      /* pointer back to this zlib stream */
  92     int   status;        /* as the name implies */
  93     Bytef *pending_buf;  /* output still pending */
  94     ulg   pending_buf_size; /* size of pending_buf */
  95     Bytef *pending_out;  /* next pending byte to output to the stream */
  96     int   pending;       /* nb of bytes in the pending buffer */
  97     int   wrap;          /* bit 0 true for zlib, bit 1 true for gzip */
  98     Byte  data_type;     /* UNKNOWN, BINARY or ASCII */
  99     Byte  method;        /* STORED (for zip only) or DEFLATED */
 100     int   last_flush;    /* value of flush param for previous deflate call */
 101
 102                 /* used by deflate.c: */
 103
 104     uInt  w_size;        /* LZ77 window size (32K by default) */
 105     uInt  w_bits;        /* log2(w_size)  (8..16) */
 106     uInt  w_mask;        /* w_size - 1 */
 107
 108     Bytef *window;
 109     /* Sliding window. Input bytes are read into the second half of the window,
 110      * and move to the first half later to keep a dictionary of at least wSize
 111      * bytes. With this organization, matches are limited to a distance of
 112      * wSize-MAX_MATCH bytes, but this ensures that IO is always
 113      * performed with a length multiple of the block size. Also, it limits
 114      * the window size to 64K, which is quite useful on MSDOS.
 115      * To do: use the user input buffer as sliding window.
 116      */
 117
 118     ulg window_size;
 119     /* Actual size of window: 2*wSize, except when the user input buffer
 120      * is directly used as sliding window.
 121      */
 122
 123     Posf *prev;
 124     /* Link to older string with same hash index. To limit the size of this
 125      * array to 64K, this link is maintained only for the last 32K strings.
 126      * An index in this array is thus a window index modulo 32K.
 127      */
 128
 129     Posf *head; /* Heads of the hash chains or NIL. */
 130
 131     uInt  ins_h;          /* hash index of string to be inserted */
 132     uInt  hash_size;      /* number of elements in hash table */
 133     uInt  hash_bits;      /* log2(hash_size) */
 134     uInt  hash_mask;      /* hash_size-1 */
 135
 136     uInt  hash_shift;
 137     /* Number of bits by which ins_h must be shifted at each input
 138      * step. It must be such that after MIN_MATCH steps, the oldest
 139      * byte no longer takes part in the hash key, that is:
 140      *   hash_shift * MIN_MATCH >= hash_bits
 141      */
 142
 143     long block_start;
 144     /* Window position at the beginning of the current output block. Gets
 145      * negative when the window is moved backwards.
 146      */
 147
 148     uInt match_length;           /* length of best match */
 149     IPos prev_match;             /* previous match */
 150     int match_available;         /* set if previous match exists */
 151     uInt strstart;               /* start of string to insert */
 152     uInt match_start;            /* start of matching string */
 153     uInt lookahead;              /* number of valid bytes ahead in window */
 154
 155     uInt prev_length;
 156     /* Length of the best match at previous step. Matches not greater than this
 157      * are discarded. This is used in the lazy match evaluation.
 158      */
 159
 160     uInt max_chain_length;
 161     /* To speed up deflation, hash chains are never searched beyond this
 162      * length.  A higher limit improves compression ratio but degrades the
 163      * speed.
 164      */
 165
 166     uInt max_lazy_match;
 167     /* Attempt to find a better match only when the current match is strictly
 168      * smaller than this value. This mechanism is used only for compression
 169      * levels >= 4.
 170      */
 171 #   define max_insert_length  max_lazy_match
 172     /* Insert new strings in the hash table only if the match length is not
 173      * greater than this length. This saves time but degrades compression.
 174      * max_insert_length is used only for compression levels <= 3.
 175      */
 176
 177     int level;    /* compression level (1..9) */
 178     int strategy; /* favor or force Huffman coding*/
 179
 180     uInt good_match;
 181     /* Use a faster search when the previous match is longer than this */
 182
 183     int nice_match; /* Stop searching when current match exceeds this */
 184
 185                 /* used by trees.c: */
 186     /* Didn't use ct_data typedef below to supress compiler warning */
 187     struct ct_data_s dyn_ltree[HEAP_SIZE];   /* literal and length tree */
 188     struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */
 189     struct ct_data_s bl_tree[2*BL_CODES+1];  /* Huffman tree for bit lengths */
 190
 191     struct tree_desc_s l_desc;               /* desc. for literal tree */
 192     struct tree_desc_s d_desc;               /* desc. for distance tree */
 193     struct tree_desc_s bl_desc;              /* desc. for bit length tree */
 194
 195     ush bl_count[MAX_BITS+1];
 196     /* number of codes at each bit length for an optimal tree */
 197
 198     int heap[2*L_CODES+1];      /* heap used to build the Huffman trees */
 199     int heap_len;               /* number of elements in the heap */
 200     int heap_max;               /* element of largest frequency */
 201     /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
 202      * The same heap array is used to build all trees.
 203      */
 204
 205     uch depth[2*L_CODES+1];
 206     /* Depth of each subtree used as tie breaker for trees of equal frequency
 207      */
 208
 209     uchf *l_buf;          /* buffer for literals or lengths */
 210
 211     uInt  lit_bufsize;
 212     /* Size of match buffer for literals/lengths.  There are 4 reasons for
 213      * limiting lit_bufsize to 64K:
 214      *   - frequencies can be kept in 16 bit counters
 215      *   - if compression is not successful for the first block, all input
 216      *     data is still in the window so we can still emit a stored block even
 217      *     when input comes from standard input.  (This can also be done for
 218      *     all blocks if lit_bufsize is not greater than 32K.)
 219      *   - if compression is not successful for a file smaller than 64K, we can
 220      *     even emit a stored file instead of a stored block (saving 5 bytes).
 221      *     This is applicable only for zip (not gzip or zlib).
 222      *   - creating new Huffman trees less frequently may not provide fast
 223      *     adaptation to changes in the input data statistics. (Take for
 224      *     example a binary file with poorly compressible code followed by
 225      *     a highly compressible string table.) Smaller buffer sizes give
 226      *     fast adaptation but have of course the overhead of transmitting
 227      *     trees more frequently.
 228      *   - I can't count above 4
 229      */
 230
 231     uInt last_lit;      /* running index in l_buf */
 232
 233     ushf *d_buf;
 234     /* Buffer for distances. To simplify the code, d_buf and l_buf have
 235      * the same number of elements. To use different lengths, an extra flag
 236      * array would be necessary.
 237      */
 238
 239     ulg opt_len;        /* bit length of current block with optimal trees */
 240     ulg static_len;     /* bit length of current block with static trees */
 241     uInt matches;       /* number of string matches in current block */
 242     int last_eob_len;   /* bit length of EOB code for last block */
 243
 244 #ifdef DEBUG
 245     ulg compressed_len; /* total bit length of compressed file mod 2^32 */
 246     ulg bits_sent;      /* bit length of compressed data sent mod 2^32 */
 247 #endif
 248
 249     ush bi_buf;
 250     /* Output buffer. bits are inserted starting at the bottom (least
 251      * significant bits).
 252      */
 253     int bi_valid;
 254     /* Number of valid bits in bi_buf.  All bits above the last valid bit
 255      * are always zero.
 256      */
 257
 258 } FAR deflate_state;
 259
 260 /* Output a byte on the stream.
 261  * IN assertion: there is enough room in pending_buf.
 262  */
 263 #define put_byte(s, c) {s->pending_buf[s->pending++] = (c);}
 264
 265
 266 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
 267 /* Minimum amount of lookahead, except at the end of the input file.
 268  * See deflate.c for comments about the MIN_MATCH+1.
 269  */
 270
 271 #define MAX_DIST(s)  ((s)->w_size-MIN_LOOKAHEAD)
 272 /* In order to simplify the code, particularly on 16 bit machines, match
 273  * distances are limited to MAX_DIST instead of WSIZE.
 274  */
 275
 276         /* in trees.c */
 277 void _tr_init         OF((deflate_state *s));
 278 int  _tr_tally        OF((deflate_state *s, unsigned dist, unsigned lc));
 279 void _tr_flush_block  OF((deflate_state *s, charf *buf, ulg stored_len,
 280                           int eof));
 281 void _tr_align        OF((deflate_state *s));
 282 void _tr_stored_block OF((deflate_state *s, charf *buf, ulg stored_len,
 283                           int eof));
 284
 285 #define d_code(dist) \
 286    ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)])
 287 /* Mapping from a distance to a distance code. dist is the distance - 1 and
 288  * must not have side effects. _dist_code[256] and _dist_code[257] are never
 289  * used.
 290  */
 291
 292 #ifndef DEBUG
 293 /* Inline versions of _tr_tally for speed: */
 294
 295 #if defined(GEN_TREES_H) || !defined(STDC)
 296   extern uch _length_code[];
 297   extern uch _dist_code[];
 298 #else
 299   extern const uch _length_code[];
 300   extern const uch _dist_code[];
 301 #endif
 302
 303 # define _tr_tally_lit(s, c, flush) \
 304   { uch cc = (c); \
 305     s->d_buf[s->last_lit] = 0; \
 306     s->l_buf[s->last_lit++] = cc; \
 307     s->dyn_ltree[cc].Freq++; \
 308     flush = (s->last_lit == s->lit_bufsize-1); \
 309    }
 310 # define _tr_tally_dist(s, distance, length, flush) \
 311   { uch len = (length); \
 312     ush dist = (distance); \
 313     s->d_buf[s->last_lit] = dist; \
 314     s->l_buf[s->last_lit++] = len; \
 315     dist--; \
 316     s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \
 317     s->dyn_dtree[d_code(dist)].Freq++; \
 318     flush = (s->last_lit == s->lit_bufsize-1); \
 319   }
 320 #else
 321 # define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c)
 322 # define _tr_tally_dist(s, distance, length, flush) \
 323               flush = _tr_tally(s, distance, length)
 324 #endif
 325
 326 #endif /* DEFLATE_H */