src/zlib/deflate.h

   1 /* deflate.h -- internal compression state
   2  * Copyright (C) 1995-2004 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
  12 #ifndef DEFLATE_H
  13 #define DEFLATE_H
  14
  15 #include "zutil.h"
  16
  17 /* define NO_GZIP when compiling if you want to disable gzip header and
  18    trailer creation by deflate().  NO_GZIP would be used to avoid linking in
  19    the crc code when it is not needed.  For shared libraries, gzip encoding
  20    should be left enabled. */
  21 #ifndef NO_GZIP
  22 #  define GZIP
  23 #endif
  24
  25 /* ===========================================================================
  26  * Internal compression state.
  27  */
  28
  29 #define LENGTH_CODES 29
  30 /* number of length codes, not counting the special END_BLOCK code */
  31
  32 #define LITERALS  256
  33 /* number of literal bytes 0..255 */
  34
  35 #define L_CODES (LITERALS+1+LENGTH_CODES)
  36 /* number of Literal or Length codes, including the END_BLOCK code */
  37
  38 #define D_CODES   30
  39 /* number of distance codes */
  40
  41 #define BL_CODES  19
  42 /* number of codes used to transfer the bit lengths */
  43
  44 #define HEAP_SIZE (2*L_CODES+1)
  45 /* maximum heap size */
  46
  47 #define MAX_BITS 15
  48 /* All codes must not exceed MAX_BITS bits */
  49
  50 #define INIT_STATE    42
  51 #define EXTRA_STATE   69
  52 #define NAME_STATE    73
  53 #define COMMENT_STATE 91
  54 #define HCRC_STATE   103
  55 #define BUSY_STATE   113
  56 #define FINISH_STATE 666
  57 /* Stream status */
  58
  59
  60 /* Data structure describing a single value and its code string. */
  61 typedef struct ct_data_s {
  62     union {
  63         ush  freq;       /* frequency count */
  64         ush  code;       /* bit string */
  65     } fc;
  66     union {
  67         ush  dad;        /* father node in Huffman tree */
  68         ush  len;        /* length of bit string */
  69     } dl;
  70 } FAR ct_data;
  71
  72 #define Freq fc.freq
  73 #define Code fc.code
  74 #define Dad  dl.dad
  75 #define Len  dl.len
  76
  77 typedef struct static_tree_desc_s  static_tree_desc;
  78
  79 typedef struct tree_desc_s {
  80     ct_data *dyn_tree;           /* the dynamic tree */
  81     int     max_code;            /* largest code with non zero frequency */
  82     static_tree_desc *stat_desc; /* the corresponding static tree */
  83 } FAR tree_desc;
  84
  85 typedef ush Pos;
  86 typedef Pos FAR Posf;
  87 typedef unsigned IPos;
  88
  89 /* A Pos is an index in the character window. We use short instead of int to
  90  * save space in the various tables. IPos is used only for parameter passing.
  91  */
  92
  93 typedef struct internal_state {
  94     z_streamp strm;      /* pointer back to this zlib stream */
  95     int   status;        /* as the name implies */
  96     Bytef *pending_buf;  /* output still pending */
  97     ulg   pending_buf_size; /* size of pending_buf */
  98     Bytef *pending_out;  /* next pending byte to output to the stream */
  99     uInt   pending;      /* nb of bytes in the pending buffer */
 100     int   wrap;          /* bit 0 true for zlib, bit 1 true for gzip */
 101     gz_headerp  gzhead;  /* gzip header information to write */
 102     uInt   gzindex;      /* where in extra, name, or comment */
 103     Byte  method;        /* STORED (for zip only) or DEFLATED */
 104     int   last_flush;    /* value of flush param for previous deflate call */
 105
 106                 /* used by deflate.c: */
 107
 108     uInt  w_size;        /* LZ77 window size (32K by default) */
 109     uInt  w_bits;        /* log2(w_size)  (8..16) */
 110     uInt  w_mask;        /* w_size - 1 */
 111
 112     Bytef *window;
 113     /* Sliding window. Input bytes are read into the second half of the window,
 114      * and move to the first half later to keep a dictionary of at least wSize
 115      * bytes. With this organization, matches are limited to a distance of
 116      * wSize-MAX_MATCH bytes, but this ensures that IO is always
 117      * performed with a length multiple of the block size. Also, it limits
 118      * the window size to 64K, which is quite useful on MSDOS.
 119      * To do: use the user input buffer as sliding window.
 120      */
 121
 122     ulg window_size;
 123     /* Actual size of window: 2*wSize, except when the user input buffer
 124      * is directly used as sliding window.
 125      */
 126
 127     Posf *prev;
 128     /* Link to older string with same hash index. To limit the size of this
 129      * array to 64K, this link is maintained only for the last 32K strings.
 130      * An index in this array is thus a window index modulo 32K.
 131      */
 132
 133     Posf *head; /* Heads of the hash chains or NIL. */
 134
 135     uInt  ins_h;          /* hash index of string to be inserted */
 136     uInt  hash_size;      /* number of elements in hash table */
 137     uInt  hash_bits;      /* log2(hash_size) */
 138     uInt  hash_mask;      /* hash_size-1 */
 139
 140     uInt  hash_shift;
 141     /* Number of bits by which ins_h must be shifted at each input
 142      * step. It must be such that after MIN_MATCH steps, the oldest
 143      * byte no longer takes part in the hash key, that is:
 144      *   hash_shift * MIN_MATCH >= hash_bits
 145      */
 146
 147     long block_start;
 148     /* Window position at the beginning of the current output block. Gets
 149      * negative when the window is moved backwards.
 150      */
 151
 152     uInt match_length;           /* length of best match */
 153     IPos prev_match;             /* previous match */
 154     int match_available;         /* set if previous match exists */
 155     uInt strstart;               /* start of string to insert */
 156     uInt match_start;            /* start of matching string */
 157     uInt lookahead;              /* number of valid bytes ahead in window */
 158
 159     uInt prev_length;
 160     /* Length of the best match at previous step. Matches not greater than this
 161      * are discarded. This is used in the lazy match evaluation.
 162      */
 163
 164     uInt max_chain_length;
 165     /* To speed up deflation, hash chains are never searched beyond this
 166      * length.  A higher limit improves compression ratio but degrades the
 167      * speed.
 168      */
 169
 170     uInt max_lazy_match;
 171     /* Attempt to find a better match only when the current match is strictly
 172      * smaller than this value. This mechanism is used only for compression
 173      * levels >= 4.
 174      */
 175 #   define max_insert_length  max_lazy_match
 176     /* Insert new strings in the hash table only if the match length is not
 177      * greater than this length. This saves time but degrades compression.
 178      * max_insert_length is used only for compression levels <= 3.
 179      */
 180
 181     int level;    /* compression level (1..9) */
 182     int strategy; /* favor or force Huffman coding*/
 183
 184     uInt good_match;
 185     /* Use a faster search when the previous match is longer than this */
 186
 187     int nice_match; /* Stop searching when current match exceeds this */
 188
 189                 /* used by trees.c: */
 190     /* Didn't use ct_data typedef below to supress compiler warning */
 191     struct ct_data_s dyn_ltree[HEAP_SIZE];   /* literal and length tree */
 192     struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */
 193     struct ct_data_s bl_tree[2*BL_CODES+1];  /* Huffman tree for bit lengths */
 194
 195     struct tree_desc_s l_desc;               /* desc. for literal tree */
 196     struct tree_desc_s d_desc;               /* desc. for distance tree */
 197     struct tree_desc_s bl_desc;              /* desc. for bit length tree */
 198
 199     ush bl_count[MAX_BITS+1];
 200     /* number of codes at each bit length for an optimal tree */
 201
 202     int heap[2*L_CODES+1];      /* heap used to build the Huffman trees */
 203     int heap_len;               /* number of elements in the heap */
 204     int heap_max;               /* element of largest frequency */
 205     /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
 206      * The same heap array is used to build all trees.
 207      */
 208
 209     uch depth[2*L_CODES+1];
 210     /* Depth of each subtree used as tie breaker for trees of equal frequency
 211      */
 212
 213     uchf *l_buf;          /* buffer for literals or lengths */
 214
 215     uInt  lit_bufsize;
 216     /* Size of match buffer for literals/lengths.  There are 4 reasons for
 217      * limiting lit_bufsize to 64K:
 218      *   - frequencies can be kept in 16 bit counters
 219      *   - if compression is not successful for the first block, all input
 220      *     data is still in the window so we can still emit a stored block even
 221      *     when input comes from standard input.  (This can also be done for
 222      *     all blocks if lit_bufsize is not greater than 32K.)
 223      *   - if compression is not successful for a file smaller than 64K, we can
 224      *     even emit a stored file instead of a stored block (saving 5 bytes).
 225      *     This is applicable only for zip (not gzip or zlib).
 226      *   - creating new Huffman trees less frequently may not provide fast
 227      *     adaptation to changes in the input data statistics. (Take for
 228      *     example a binary file with poorly compressible code followed by
 229      *     a highly compressible string table.) Smaller buffer sizes give
 230      *     fast adaptation but have of course the overhead of transmitting
 231      *     trees more frequently.
 232      *   - I can't count above 4
 233      */
 234
 235     uInt last_lit;      /* running index in l_buf */
 236
 237     ushf *d_buf;
 238     /* Buffer for distances. To simplify the code, d_buf and l_buf have
 239      * the same number of elements. To use different lengths, an extra flag
 240      * array would be necessary.
 241      */
 242
 243     ulg opt_len;        /* bit length of current block with optimal trees */
 244     ulg static_len;     /* bit length of current block with static trees */
 245     uInt matches;       /* number of string matches in current block */
 246     int last_eob_len;   /* bit length of EOB code for last block */
 247
 248 #ifdef DEBUG
 249     ulg compressed_len; /* total bit length of compressed file mod 2^32 */
 250     ulg bits_sent;      /* bit length of compressed data sent mod 2^32 */
 251 #endif
 252
 253     ush bi_buf;
 254     /* Output buffer. bits are inserted starting at the bottom (least
 255      * significant bits).
 256      */
 257     int bi_valid;
 258     /* Number of valid bits in bi_buf.  All bits above the last valid bit
 259      * are always zero.
 260      */
 261
 262 } FAR deflate_state;
 263
 264 /* Output a byte on the stream.
 265  * IN assertion: there is enough room in pending_buf.
 266  */
 267 #define put_byte(s, c) {s->pending_buf[s->pending++] = (c);}
 268
 269
 270 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
 271 /* Minimum amount of lookahead, except at the end of the input file.
 272  * See deflate.c for comments about the MIN_MATCH+1.
 273  */
 274
 275 #define MAX_DIST(s)  ((s)->w_size-MIN_LOOKAHEAD)
 276 /* In order to simplify the code, particularly on 16 bit machines, match
 277  * distances are limited to MAX_DIST instead of WSIZE.
 278  */
 279
 280         /* in trees.c */
 281 void _tr_init         OF((deflate_state *s));
 282 int  _tr_tally        OF((deflate_state *s, unsigned dist, unsigned lc));
 283 void _tr_flush_block  OF((deflate_state *s, charf *buf, ulg stored_len,
 284                           int eof));
 285 void _tr_align        OF((deflate_state *s));
 286 void _tr_stored_block OF((deflate_state *s, charf *buf, ulg stored_len,
 287                           int eof));
 288
 289 #define d_code(dist) \
 290    ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)])
 291 /* Mapping from a distance to a distance code. dist is the distance - 1 and
 292  * must not have side effects. _dist_code[256] and _dist_code[257] are never
 293  * used.
 294  */
 295
 296 #ifndef DEBUG
 297 /* Inline versions of _tr_tally for speed: */
 298
 299 #if defined(GEN_TREES_H) || !defined(STDC)
 300   extern uch _length_code[];
 301   extern uch _dist_code[];
 302 #else
 303   extern const uch _length_code[];
 304   extern const uch _dist_code[];
 305 #endif
 306
 307 # define _tr_tally_lit(s, c, flush) \
 308   { uch cc = (c); \
 309     s->d_buf[s->last_lit] = 0; \
 310     s->l_buf[s->last_lit++] = cc; \
 311     s->dyn_ltree[cc].Freq++; \
 312     flush = (s->last_lit == s->lit_bufsize-1); \
 313    }
 314 # define _tr_tally_dist(s, distance, length, flush) \
 315   { uch len = (length); \
 316     ush dist = (distance); \
 317     s->d_buf[s->last_lit] = dist; \
 318     s->l_buf[s->last_lit++] = len; \
 319     dist--; \
 320     s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \
 321     s->dyn_dtree[d_code(dist)].Freq++; \
 322     flush = (s->last_lit == s->lit_bufsize-1); \
 323   }
 324 #else
 325 # define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c)
 326 # define _tr_tally_dist(s, distance, length, flush) \
 327               flush = _tr_tally(s, distance, length)
 328 #endif
 329
 330 #endif /* DEFLATE_H */