zlib/deflate.h

   1 /* deflate.h -- internal compression state
   2  * Copyright (C) 1995-2018 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 Buf_size 16
  52 /* size of bit buffer in bi_buf */
  53
  54 #define INIT_STATE    42    /* zlib header -> BUSY_STATE */
  55 #ifdef GZIP
  56 #  define GZIP_STATE  57    /* gzip header -> BUSY_STATE | EXTRA_STATE */
  57 #endif
  58 #define EXTRA_STATE   69    /* gzip extra block -> NAME_STATE */
  59 #define NAME_STATE    73    /* gzip file name -> COMMENT_STATE */
  60 #define COMMENT_STATE 91    /* gzip comment -> HCRC_STATE */
  61 #define HCRC_STATE   103    /* gzip header CRC -> BUSY_STATE */
  62 #define BUSY_STATE   113    /* deflate -> FINISH_STATE */
  63 #define FINISH_STATE 666    /* stream complete */
  64 /* Stream status */
  65
  66
  67 /* Data structure describing a single value and its code string. */
  68 typedef struct ct_data_s {
  69     union {
  70         ush  freq;       /* frequency count */
  71         ush  code;       /* bit string */
  72     } fc;
  73     union {
  74         ush  dad;        /* father node in Huffman tree */
  75         ush  len;        /* length of bit string */
  76     } dl;
  77 } FAR ct_data;
  78
  79 #define Freq fc.freq
  80 #define Code fc.code
  81 #define Dad  dl.dad
  82 #define Len  dl.len
  83
  84 typedef struct static_tree_desc_s  static_tree_desc;
  85
  86 typedef struct tree_desc_s {
  87     ct_data *dyn_tree;           /* the dynamic tree */
  88     int     max_code;            /* largest code with non zero frequency */
  89     const static_tree_desc *stat_desc;  /* the corresponding static tree */
  90 } FAR tree_desc;
  91
  92 typedef ush Pos;
  93 typedef Pos FAR Posf;
  94 typedef unsigned IPos;
  95
  96 /* A Pos is an index in the character window. We use short instead of int to
  97  * save space in the various tables. IPos is used only for parameter passing.
  98  */
  99
 100 typedef struct internal_state {
 101     z_streamp strm;      /* pointer back to this zlib stream */
 102     int   status;        /* as the name implies */
 103     Bytef *pending_buf;  /* output still pending */
 104     ulg   pending_buf_size; /* size of pending_buf */
 105     Bytef *pending_out;  /* next pending byte to output to the stream */
 106     ulg   pending;       /* nb of bytes in the pending buffer */
 107     int   wrap;          /* bit 0 true for zlib, bit 1 true for gzip */
 108     gz_headerp  gzhead;  /* gzip header information to write */
 109     ulg   gzindex;       /* where in extra, name, or comment */
 110     Byte  method;        /* can only be DEFLATED */
 111     int   last_flush;    /* value of flush param for previous deflate call */
 112
 113                 /* used by deflate.c: */
 114
 115     uInt  w_size;        /* LZ77 window size (32K by default) */
 116     uInt  w_bits;        /* log2(w_size)  (8..16) */
 117     uInt  w_mask;        /* w_size - 1 */
 118
 119     Bytef *window;
 120     /* Sliding window. Input bytes are read into the second half of the window,
 121      * and move to the first half later to keep a dictionary of at least wSize
 122      * bytes. With this organization, matches are limited to a distance of
 123      * wSize-MAX_MATCH bytes, but this ensures that IO is always
 124      * performed with a length multiple of the block size. Also, it limits
 125      * the window size to 64K, which is quite useful on MSDOS.
 126      * To do: use the user input buffer as sliding window.
 127      */
 128
 129     ulg window_size;
 130     /* Actual size of window: 2*wSize, except when the user input buffer
 131      * is directly used as sliding window.
 132      */
 133
 134     Posf *prev;
 135     /* Link to older string with same hash index. To limit the size of this
 136      * array to 64K, this link is maintained only for the last 32K strings.
 137      * An index in this array is thus a window index modulo 32K.
 138      */
 139
 140     Posf *head; /* Heads of the hash chains or NIL. */
 141
 142     uInt  ins_h;          /* hash index of string to be inserted */
 143     uInt  hash_size;      /* number of elements in hash table */
 144     uInt  hash_bits;      /* log2(hash_size) */
 145     uInt  hash_mask;      /* hash_size-1 */
 146
 147     uInt  hash_shift;
 148     /* Number of bits by which ins_h must be shifted at each input
 149      * step. It must be such that after MIN_MATCH steps, the oldest
 150      * byte no longer takes part in the hash key, that is:
 151      *   hash_shift * MIN_MATCH >= hash_bits
 152      */
 153
 154     long block_start;
 155     /* Window position at the beginning of the current output block. Gets
 156      * negative when the window is moved backwards.
 157      */
 158
 159     uInt match_length;           /* length of best match */
 160     IPos prev_match;             /* previous match */
 161     int match_available;         /* set if previous match exists */
 162     uInt strstart;               /* start of string to insert */
 163     uInt match_start;            /* start of matching string */
 164     uInt lookahead;              /* number of valid bytes ahead in window */
 165
 166     uInt prev_length;
 167     /* Length of the best match at previous step. Matches not greater than this
 168      * are discarded. This is used in the lazy match evaluation.
 169      */
 170
 171     uInt max_chain_length;
 172     /* To speed up deflation, hash chains are never searched beyond this
 173      * length.  A higher limit improves compression ratio but degrades the
 174      * speed.
 175      */
 176
 177     uInt max_lazy_match;
 178     /* Attempt to find a better match only when the current match is strictly
 179      * smaller than this value. This mechanism is used only for compression
 180      * levels >= 4.
 181      */
 182 #   define max_insert_length  max_lazy_match
 183     /* Insert new strings in the hash table only if the match length is not
 184      * greater than this length. This saves time but degrades compression.
 185      * max_insert_length is used only for compression levels <= 3.
 186      */
 187
 188     int level;    /* compression level (1..9) */
 189     int strategy; /* favor or force Huffman coding*/
 190
 191     uInt good_match;
 192     /* Use a faster search when the previous match is longer than this */
 193
 194     int nice_match; /* Stop searching when current match exceeds this */
 195
 196                 /* used by trees.c: */
 197     /* Didn't use ct_data typedef below to suppress compiler warning */
 198     struct ct_data_s dyn_ltree[HEAP_SIZE];   /* literal and length tree */
 199     struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */
 200     struct ct_data_s bl_tree[2*BL_CODES+1];  /* Huffman tree for bit lengths */
 201
 202     struct tree_desc_s l_desc;               /* desc. for literal tree */
 203     struct tree_desc_s d_desc;               /* desc. for distance tree */
 204     struct tree_desc_s bl_desc;              /* desc. for bit length tree */
 205
 206     ush bl_count[MAX_BITS+1];
 207     /* number of codes at each bit length for an optimal tree */
 208
 209     int heap[2*L_CODES+1];      /* heap used to build the Huffman trees */
 210     int heap_len;               /* number of elements in the heap */
 211     int heap_max;               /* element of largest frequency */
 212     /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
 213      * The same heap array is used to build all trees.
 214      */
 215
 216     uch depth[2*L_CODES+1];
 217     /* Depth of each subtree used as tie breaker for trees of equal frequency
 218      */
 219
 220     uchf *sym_buf;        /* buffer for distances and literals/lengths */
 221
 222     uInt  lit_bufsize;
 223     /* Size of match buffer for literals/lengths.  There are 4 reasons for
 224      * limiting lit_bufsize to 64K:
 225      *   - frequencies can be kept in 16 bit counters
 226      *   - if compression is not successful for the first block, all input
 227      *     data is still in the window so we can still emit a stored block even
 228      *     when input comes from standard input.  (This can also be done for
 229      *     all blocks if lit_bufsize is not greater than 32K.)
 230      *   - if compression is not successful for a file smaller than 64K, we can
 231      *     even emit a stored file instead of a stored block (saving 5 bytes).
 232      *     This is applicable only for zip (not gzip or zlib).
 233      *   - creating new Huffman trees less frequently may not provide fast
 234      *     adaptation to changes in the input data statistics. (Take for
 235      *     example a binary file with poorly compressible code followed by
 236      *     a highly compressible string table.) Smaller buffer sizes give
 237      *     fast adaptation but have of course the overhead of transmitting
 238      *     trees more frequently.
 239      *   - I can't count above 4
 240      */
 241
 242     uInt sym_next;      /* running index in sym_buf */
 243     uInt sym_end;       /* symbol table full when sym_next reaches this */
 244
 245     ulg opt_len;        /* bit length of current block with optimal trees */
 246     ulg static_len;     /* bit length of current block with static trees */
 247     uInt matches;       /* number of string matches in current block */
 248     uInt insert;        /* bytes at end of window left to insert */
 249
 250 #ifdef ZLIB_DEBUG
 251     ulg compressed_len; /* total bit length of compressed file mod 2^32 */
 252     ulg bits_sent;      /* bit length of compressed data sent mod 2^32 */
 253 #endif
 254
 255     ush bi_buf;
 256     /* Output buffer. bits are inserted starting at the bottom (least
 257      * significant bits).
 258      */
 259     int bi_valid;
 260     /* Number of valid bits in bi_buf.  All bits above the last valid bit
 261      * are always zero.
 262      */
 263
 264     ulg high_water;
 265     /* High water mark offset in window for initialized bytes -- bytes above
 266      * this are set to zero in order to avoid memory check warnings when
 267      * longest match routines access bytes past the input.  This is then
 268      * updated to the new high water mark.
 269      */
 270
 271 } FAR deflate_state;
 272
 273 /* Output a byte on the stream.
 274  * IN assertion: there is enough room in pending_buf.
 275  */
 276 #define put_byte(s, c) {s->pending_buf[s->pending++] = (Bytef)(c);}
 277
 278
 279 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
 280 /* Minimum amount of lookahead, except at the end of the input file.
 281  * See deflate.c for comments about the MIN_MATCH+1.
 282  */
 283
 284 #define MAX_DIST(s)  ((s)->w_size-MIN_LOOKAHEAD)
 285 /* In order to simplify the code, particularly on 16 bit machines, match
 286  * distances are limited to MAX_DIST instead of WSIZE.
 287  */
 288
 289 #define WIN_INIT MAX_MATCH
 290 /* Number of bytes after end of data in window to initialize in order to avoid
 291    memory checker errors from longest match routines */
 292
 293         /* in trees.c */
 294 void ZLIB_INTERNAL _tr_init OF((deflate_state *s));
 295 int ZLIB_INTERNAL _tr_tally OF((deflate_state *s, unsigned dist, unsigned lc));
 296 void ZLIB_INTERNAL _tr_flush_block OF((deflate_state *s, charf *buf,
 297                         ulg stored_len, int last));
 298 void ZLIB_INTERNAL _tr_flush_bits OF((deflate_state *s));
 299 void ZLIB_INTERNAL _tr_align OF((deflate_state *s));
 300 void ZLIB_INTERNAL _tr_stored_block OF((deflate_state *s, charf *buf,
 301                         ulg stored_len, int last));
 302
 303 #define d_code(dist) \
 304    ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)])
 305 /* Mapping from a distance to a distance code. dist is the distance - 1 and
 306  * must not have side effects. _dist_code[256] and _dist_code[257] are never
 307  * used.
 308  */
 309
 310 #ifndef ZLIB_DEBUG
 311 /* Inline versions of _tr_tally for speed: */
 312
 313 #if defined(GEN_TREES_H) || !defined(STDC)
 314   extern uch ZLIB_INTERNAL _length_code[];
 315   extern uch ZLIB_INTERNAL _dist_code[];
 316 #else
 317   extern const uch ZLIB_INTERNAL _length_code[];
 318   extern const uch ZLIB_INTERNAL _dist_code[];
 319 #endif
 320
 321 # define _tr_tally_lit(s, c, flush) \
 322   { uch cc = (c); \
 323     s->sym_buf[s->sym_next++] = 0; \
 324     s->sym_buf[s->sym_next++] = 0; \
 325     s->sym_buf[s->sym_next++] = cc; \
 326     s->dyn_ltree[cc].Freq++; \
 327     flush = (s->sym_next == s->sym_end); \
 328    }
 329 # define _tr_tally_dist(s, distance, length, flush) \
 330   { uch len = (uch)(length); \
 331     ush dist = (ush)(distance); \
 332     s->sym_buf[s->sym_next++] = dist; \
 333     s->sym_buf[s->sym_next++] = dist >> 8; \
 334     s->sym_buf[s->sym_next++] = len; \
 335     dist--; \
 336     s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \
 337     s->dyn_dtree[d_code(dist)].Freq++; \
 338     flush = (s->sym_next == s->sym_end); \
 339   }
 340 #else
 341 # define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c)
 342 # define _tr_tally_dist(s, distance, length, flush) \
 343               flush = _tr_tally(s, distance, length)
 344 #endif
 345
 346 #endif /* DEFLATE_H */