src/zlib/deflate.h

   1 /* deflate.h -- internal compression state
   2  * Copyright (C) 1995-2012 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 Buf_size 16
  51 /* size of bit buffer in bi_buf */
  52
  53 #define INIT_STATE    42
  54 #define EXTRA_STATE   69
  55 #define NAME_STATE    73
  56 #define COMMENT_STATE 91
  57 #define HCRC_STATE   103
  58 #define BUSY_STATE   113
  59 #define FINISH_STATE 666
  60 /* Stream status */
  61
  62
  63 /* Data structure describing a single value and its code string. */
  64 typedef struct ct_data_s {
  65     union {
  66         ush  freq;       /* frequency count */
  67         ush  code;       /* bit string */
  68     } fc;
  69     union {
  70         ush  dad;        /* father node in Huffman tree */
  71         ush  len;        /* length of bit string */
  72     } dl;
  73 } FAR ct_data;
  74
  75 #define Freq fc.freq
  76 #define Code fc.code
  77 #define Dad  dl.dad
  78 #define Len  dl.len
  79
  80 typedef struct static_tree_desc_s  static_tree_desc;
  81
  82 typedef struct tree_desc_s {
  83     ct_data *dyn_tree;           /* the dynamic tree */
  84     int     max_code;            /* largest code with non zero frequency */
  85     static_tree_desc *stat_desc; /* the corresponding static tree */
  86 } FAR tree_desc;
  87
  88 typedef ush Pos;
  89 typedef Pos FAR Posf;
  90 typedef unsigned IPos;
  91
  92 /* A Pos is an index in the character window. We use short instead of int to
  93  * save space in the various tables. IPos is used only for parameter passing.
  94  */
  95
  96 typedef struct internal_state {
  97     z_streamp strm;      /* pointer back to this zlib stream */
  98     int   status;        /* as the name implies */
  99     Bytef *pending_buf;  /* output still pending */
 100     ulg   pending_buf_size; /* size of pending_buf */
 101     Bytef *pending_out;  /* next pending byte to output to the stream */
 102     uInt   pending;      /* nb of bytes in the pending buffer */
 103     int   wrap;          /* bit 0 true for zlib, bit 1 true for gzip */
 104     gz_headerp  gzhead;  /* gzip header information to write */
 105     uInt   gzindex;      /* where in extra, name, or comment */
 106     Byte  method;        /* can only be DEFLATED */
 107     int   last_flush;    /* value of flush param for previous deflate call */
 108
 109                 /* used by deflate.c: */
 110
 111     uInt  w_size;        /* LZ77 window size (32K by default) */
 112     uInt  w_bits;        /* log2(w_size)  (8..16) */
 113     uInt  w_mask;        /* w_size - 1 */
 114
 115     Bytef *window;
 116     /* Sliding window. Input bytes are read into the second half of the window,
 117      * and move to the first half later to keep a dictionary of at least wSize
 118      * bytes. With this organization, matches are limited to a distance of
 119      * wSize-MAX_MATCH bytes, but this ensures that IO is always
 120      * performed with a length multiple of the block size. Also, it limits
 121      * the window size to 64K, which is quite useful on MSDOS.
 122      * To do: use the user input buffer as sliding window.
 123      */
 124
 125     ulg window_size;
 126     /* Actual size of window: 2*wSize, except when the user input buffer
 127      * is directly used as sliding window.
 128      */
 129
 130     Posf *prev;
 131     /* Link to older string with same hash index. To limit the size of this
 132      * array to 64K, this link is maintained only for the last 32K strings.
 133      * An index in this array is thus a window index modulo 32K.
 134      */
 135
 136     Posf *head; /* Heads of the hash chains or NIL. */
 137
 138     uInt  ins_h;          /* hash index of string to be inserted */
 139     uInt  hash_size;      /* number of elements in hash table */
 140     uInt  hash_bits;      /* log2(hash_size) */
 141     uInt  hash_mask;      /* hash_size-1 */
 142
 143     uInt  hash_shift;
 144     /* Number of bits by which ins_h must be shifted at each input
 145      * step. It must be such that after MIN_MATCH steps, the oldest
 146      * byte no longer takes part in the hash key, that is:
 147      *   hash_shift * MIN_MATCH >= hash_bits
 148      */
 149
 150     long block_start;
 151     /* Window position at the beginning of the current output block. Gets
 152      * negative when the window is moved backwards.
 153      */
 154
 155     uInt match_length;           /* length of best match */
 156     IPos prev_match;             /* previous match */
 157     int match_available;         /* set if previous match exists */
 158     uInt strstart;               /* start of string to insert */
 159     uInt match_start;            /* start of matching string */
 160     uInt lookahead;              /* number of valid bytes ahead in window */
 161
 162     uInt prev_length;
 163     /* Length of the best match at previous step. Matches not greater than this
 164      * are discarded. This is used in the lazy match evaluation.
 165      */
 166
 167     uInt max_chain_length;
 168     /* To speed up deflation, hash chains are never searched beyond this
 169      * length.  A higher limit improves compression ratio but degrades the
 170      * speed.
 171      */
 172
 173     uInt max_lazy_match;
 174     /* Attempt to find a better match only when the current match is strictly
 175      * smaller than this value. This mechanism is used only for compression
 176      * levels >= 4.
 177      */
 178 #   define max_insert_length  max_lazy_match
 179     /* Insert new strings in the hash table only if the match length is not
 180      * greater than this length. This saves time but degrades compression.
 181      * max_insert_length is used only for compression levels <= 3.
 182      */
 183
 184     int level;    /* compression level (1..9) */
 185     int strategy; /* favor or force Huffman coding*/
 186
 187     uInt good_match;
 188     /* Use a faster search when the previous match is longer than this */
 189
 190     int nice_match; /* Stop searching when current match exceeds this */
 191
 192                 /* used by trees.c: */
 193     /* Didn't use ct_data typedef below to suppress compiler warning */
 194     struct ct_data_s dyn_ltree[HEAP_SIZE];   /* literal and length tree */
 195     struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */
 196     struct ct_data_s bl_tree[2*BL_CODES+1];  /* Huffman tree for bit lengths */
 197
 198     struct tree_desc_s l_desc;               /* desc. for literal tree */
 199     struct tree_desc_s d_desc;               /* desc. for distance tree */
 200     struct tree_desc_s bl_desc;              /* desc. for bit length tree */
 201
 202     ush bl_count[MAX_BITS+1];
 203     /* number of codes at each bit length for an optimal tree */
 204
 205     int heap[2*L_CODES+1];      /* heap used to build the Huffman trees */
 206     int heap_len;               /* number of elements in the heap */
 207     int heap_max;               /* element of largest frequency */
 208     /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
 209      * The same heap array is used to build all trees.
 210      */
 211
 212     uch depth[2*L_CODES+1];
 213     /* Depth of each subtree used as tie breaker for trees of equal frequency
 214      */
 215
 216     uchf *l_buf;          /* buffer for literals or lengths */
 217
 218     uInt  lit_bufsize;
 219     /* Size of match buffer for literals/lengths.  There are 4 reasons for
 220      * limiting lit_bufsize to 64K:
 221      *   - frequencies can be kept in 16 bit counters
 222      *   - if compression is not successful for the first block, all input
 223      *     data is still in the window so we can still emit a stored block even
 224      *     when input comes from standard input.  (This can also be done for
 225      *     all blocks if lit_bufsize is not greater than 32K.)
 226      *   - if compression is not successful for a file smaller than 64K, we can
 227      *     even emit a stored file instead of a stored block (saving 5 bytes).
 228      *     This is applicable only for zip (not gzip or zlib).
 229      *   - creating new Huffman trees less frequently may not provide fast
 230      *     adaptation to changes in the input data statistics. (Take for
 231      *     example a binary file with poorly compressible code followed by
 232      *     a highly compressible string table.) Smaller buffer sizes give
 233      *     fast adaptation but have of course the overhead of transmitting
 234      *     trees more frequently.
 235      *   - I can't count above 4
 236      */
 237
 238     uInt last_lit;      /* running index in l_buf */
 239
 240     ushf *d_buf;
 241     /* Buffer for distances. To simplify the code, d_buf and l_buf have
 242      * the same number of elements. To use different lengths, an extra flag
 243      * array would be necessary.
 244      */
 245
 246     ulg opt_len;        /* bit length of current block with optimal trees */
 247     ulg static_len;     /* bit length of current block with static trees */
 248     uInt matches;       /* number of string matches in current block */
 249     uInt insert;        /* bytes at end of window left to insert */
 250
 251 #ifdef DEBUG
 252     ulg compressed_len; /* total bit length of compressed file mod 2^32 */
 253     ulg bits_sent;      /* bit length of compressed data sent mod 2^32 */
 254 #endif
 255
 256     ush bi_buf;
 257     /* Output buffer. bits are inserted starting at the bottom (least
 258      * significant bits).
 259      */
 260     int bi_valid;
 261     /* Number of valid bits in bi_buf.  All bits above the last valid bit
 262      * are always zero.
 263      */
 264
 265     ulg high_water;
 266     /* High water mark offset in window for initialized bytes -- bytes above
 267      * this are set to zero in order to avoid memory check warnings when
 268      * longest match routines access bytes past the input.  This is then
 269      * updated to the new high water mark.
 270      */
 271
 272 } FAR deflate_state;
 273
 274 /* Output a byte on the stream.
 275  * IN assertion: there is enough room in pending_buf.
 276  */
 277 #define put_byte(s, c) {s->pending_buf[s->pending++] = (c);}
 278
 279
 280 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
 281 /* Minimum amount of lookahead, except at the end of the input file.
 282  * See deflate.c for comments about the MIN_MATCH+1.
 283  */
 284
 285 #define MAX_DIST(s)  ((s)->w_size-MIN_LOOKAHEAD)
 286 /* In order to simplify the code, particularly on 16 bit machines, match
 287  * distances are limited to MAX_DIST instead of WSIZE.
 288  */
 289
 290 #define WIN_INIT MAX_MATCH
 291 /* Number of bytes after end of data in window to initialize in order to avoid
 292    memory checker errors from longest match routines */
 293
 294         /* in trees.c */
 295 void ZLIB_INTERNAL _tr_init OF((deflate_state *s));
 296 int ZLIB_INTERNAL _tr_tally OF((deflate_state *s, unsigned dist, unsigned lc));
 297 void ZLIB_INTERNAL _tr_flush_block OF((deflate_state *s, charf *buf,
 298                         ulg stored_len, int last));
 299 void ZLIB_INTERNAL _tr_flush_bits OF((deflate_state *s));
 300 void ZLIB_INTERNAL _tr_align OF((deflate_state *s));
 301 void ZLIB_INTERNAL _tr_stored_block OF((deflate_state *s, charf *buf,
 302                         ulg stored_len, int last));
 303
 304 #define d_code(dist) \
 305    ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)])
 306 /* Mapping from a distance to a distance code. dist is the distance - 1 and
 307  * must not have side effects. _dist_code[256] and _dist_code[257] are never
 308  * used.
 309  */
 310
 311 #ifndef DEBUG
 312 /* Inline versions of _tr_tally for speed: */
 313
 314 #if defined(GEN_TREES_H) || !defined(STDC)
 315   extern uch ZLIB_INTERNAL _length_code[];
 316   extern uch ZLIB_INTERNAL _dist_code[];
 317 #else
 318   extern const uch ZLIB_INTERNAL _length_code[];
 319   extern const uch ZLIB_INTERNAL _dist_code[];
 320 #endif
 321
 322 # define _tr_tally_lit(s, c, flush) \
 323   { uch cc = (c); \
 324     s->d_buf[s->last_lit] = 0; \
 325     s->l_buf[s->last_lit++] = cc; \
 326     s->dyn_ltree[cc].Freq++; \
 327     flush = (s->last_lit == s->lit_bufsize-1); \
 328    }
 329 # define _tr_tally_dist(s, distance, length, flush) \
 330   { uch len = (length); \
 331     ush dist = (distance); \
 332     s->d_buf[s->last_lit] = dist; \
 333     s->l_buf[s->last_lit++] = len; \
 334     dist--; \
 335     s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \
 336     s->dyn_dtree[d_code(dist)].Freq++; \
 337     flush = (s->last_lit == s->lit_bufsize-1); \
 338   }
 339 #else
 340 # define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c)
 341 # define _tr_tally_dist(s, distance, length, flush) \
 342               flush = _tr_tally(s, distance, length)
 343 #endif
 344
 345 #endif /* DEFLATE_H */