Back to home page

LXR

 
 

    


File indexing completed on 2025-05-11 08:24:10

0001 /* deflate.h -- internal compression state
0002  * Copyright (C) 1995-2018 Jean-loup Gailly
0003  * For conditions of distribution and use, see copyright notice in zlib.h
0004  */
0005 
0006 /* WARNING: this file should *not* be used by applications. It is
0007    part of the implementation of the compression library and is
0008    subject to change. Applications should only use zlib.h.
0009  */
0010 
0011 /* @(#) $Id$ */
0012 
0013 #ifndef DEFLATE_H
0014 #define DEFLATE_H
0015 
0016 #include "zutil.h"
0017 
0018 /* define NO_GZIP when compiling if you want to disable gzip header and
0019    trailer creation by deflate().  NO_GZIP would be used to avoid linking in
0020    the crc code when it is not needed.  For shared libraries, gzip encoding
0021    should be left enabled. */
0022 #ifndef NO_GZIP
0023 #  define GZIP
0024 #endif
0025 
0026 /* ===========================================================================
0027  * Internal compression state.
0028  */
0029 
0030 #define LENGTH_CODES 29
0031 /* number of length codes, not counting the special END_BLOCK code */
0032 
0033 #define LITERALS  256
0034 /* number of literal bytes 0..255 */
0035 
0036 #define L_CODES (LITERALS+1+LENGTH_CODES)
0037 /* number of Literal or Length codes, including the END_BLOCK code */
0038 
0039 #define D_CODES   30
0040 /* number of distance codes */
0041 
0042 #define BL_CODES  19
0043 /* number of codes used to transfer the bit lengths */
0044 
0045 #define HEAP_SIZE (2*L_CODES+1)
0046 /* maximum heap size */
0047 
0048 #define MAX_BITS 15
0049 /* All codes must not exceed MAX_BITS bits */
0050 
0051 #define Buf_size 16
0052 /* size of bit buffer in bi_buf */
0053 
0054 #define INIT_STATE    42    /* zlib header -> BUSY_STATE */
0055 #ifdef GZIP
0056 #  define GZIP_STATE  57    /* gzip header -> BUSY_STATE | EXTRA_STATE */
0057 #endif
0058 #define EXTRA_STATE   69    /* gzip extra block -> NAME_STATE */
0059 #define NAME_STATE    73    /* gzip file name -> COMMENT_STATE */
0060 #define COMMENT_STATE 91    /* gzip comment -> HCRC_STATE */
0061 #define HCRC_STATE   103    /* gzip header CRC -> BUSY_STATE */
0062 #define BUSY_STATE   113    /* deflate -> FINISH_STATE */
0063 #define FINISH_STATE 666    /* stream complete */
0064 /* Stream status */
0065 
0066 
0067 /* Data structure describing a single value and its code string. */
0068 typedef struct ct_data_s {
0069     union {
0070         ush  freq;       /* frequency count */
0071         ush  code;       /* bit string */
0072     } fc;
0073     union {
0074         ush  dad;        /* father node in Huffman tree */
0075         ush  len;        /* length of bit string */
0076     } dl;
0077 } FAR ct_data;
0078 
0079 #define Freq fc.freq
0080 #define Code fc.code
0081 #define Dad  dl.dad
0082 #define Len  dl.len
0083 
0084 typedef struct static_tree_desc_s  static_tree_desc;
0085 
0086 typedef struct tree_desc_s {
0087     ct_data *dyn_tree;           /* the dynamic tree */
0088     int     max_code;            /* largest code with non zero frequency */
0089     const static_tree_desc *stat_desc;  /* the corresponding static tree */
0090 } FAR tree_desc;
0091 
0092 typedef ush Pos;
0093 typedef Pos FAR Posf;
0094 typedef unsigned IPos;
0095 
0096 /* A Pos is an index in the character window. We use short instead of int to
0097  * save space in the various tables. IPos is used only for parameter passing.
0098  */
0099 
0100 typedef struct internal_state {
0101     z_streamp strm;      /* pointer back to this zlib stream */
0102     int   status;        /* as the name implies */
0103     Bytef *pending_buf;  /* output still pending */
0104     ulg   pending_buf_size; /* size of pending_buf */
0105     Bytef *pending_out;  /* next pending byte to output to the stream */
0106     ulg   pending;       /* nb of bytes in the pending buffer */
0107     int   wrap;          /* bit 0 true for zlib, bit 1 true for gzip */
0108     gz_headerp  gzhead;  /* gzip header information to write */
0109     ulg   gzindex;       /* where in extra, name, or comment */
0110     Byte  method;        /* can only be DEFLATED */
0111     int   last_flush;    /* value of flush param for previous deflate call */
0112 
0113                 /* used by deflate.c: */
0114 
0115     uInt  w_size;        /* LZ77 window size (32K by default) */
0116     uInt  w_bits;        /* log2(w_size)  (8..16) */
0117     uInt  w_mask;        /* w_size - 1 */
0118 
0119     Bytef *window;
0120     /* Sliding window. Input bytes are read into the second half of the window,
0121      * and move to the first half later to keep a dictionary of at least wSize
0122      * bytes. With this organization, matches are limited to a distance of
0123      * wSize-MAX_MATCH bytes, but this ensures that IO is always
0124      * performed with a length multiple of the block size. Also, it limits
0125      * the window size to 64K, which is quite useful on MSDOS.
0126      * To do: use the user input buffer as sliding window.
0127      */
0128 
0129     ulg window_size;
0130     /* Actual size of window: 2*wSize, except when the user input buffer
0131      * is directly used as sliding window.
0132      */
0133 
0134     Posf *prev;
0135     /* Link to older string with same hash index. To limit the size of this
0136      * array to 64K, this link is maintained only for the last 32K strings.
0137      * An index in this array is thus a window index modulo 32K.
0138      */
0139 
0140     Posf *head; /* Heads of the hash chains or NIL. */
0141 
0142     uInt  ins_h;          /* hash index of string to be inserted */
0143     uInt  hash_size;      /* number of elements in hash table */
0144     uInt  hash_bits;      /* log2(hash_size) */
0145     uInt  hash_mask;      /* hash_size-1 */
0146 
0147     uInt  hash_shift;
0148     /* Number of bits by which ins_h must be shifted at each input
0149      * step. It must be such that after MIN_MATCH steps, the oldest
0150      * byte no longer takes part in the hash key, that is:
0151      *   hash_shift * MIN_MATCH >= hash_bits
0152      */
0153 
0154     long block_start;
0155     /* Window position at the beginning of the current output block. Gets
0156      * negative when the window is moved backwards.
0157      */
0158 
0159     uInt match_length;           /* length of best match */
0160     IPos prev_match;             /* previous match */
0161     int match_available;         /* set if previous match exists */
0162     uInt strstart;               /* start of string to insert */
0163     uInt match_start;            /* start of matching string */
0164     uInt lookahead;              /* number of valid bytes ahead in window */
0165 
0166     uInt prev_length;
0167     /* Length of the best match at previous step. Matches not greater than this
0168      * are discarded. This is used in the lazy match evaluation.
0169      */
0170 
0171     uInt max_chain_length;
0172     /* To speed up deflation, hash chains are never searched beyond this
0173      * length.  A higher limit improves compression ratio but degrades the
0174      * speed.
0175      */
0176 
0177     uInt max_lazy_match;
0178     /* Attempt to find a better match only when the current match is strictly
0179      * smaller than this value. This mechanism is used only for compression
0180      * levels >= 4.
0181      */
0182 #   define max_insert_length  max_lazy_match
0183     /* Insert new strings in the hash table only if the match length is not
0184      * greater than this length. This saves time but degrades compression.
0185      * max_insert_length is used only for compression levels <= 3.
0186      */
0187 
0188     int level;    /* compression level (1..9) */
0189     int strategy; /* favor or force Huffman coding*/
0190 
0191     uInt good_match;
0192     /* Use a faster search when the previous match is longer than this */
0193 
0194     int nice_match; /* Stop searching when current match exceeds this */
0195 
0196                 /* used by trees.c: */
0197     /* Didn't use ct_data typedef below to suppress compiler warning */
0198     struct ct_data_s dyn_ltree[HEAP_SIZE];   /* literal and length tree */
0199     struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */
0200     struct ct_data_s bl_tree[2*BL_CODES+1];  /* Huffman tree for bit lengths */
0201 
0202     struct tree_desc_s l_desc;               /* desc. for literal tree */
0203     struct tree_desc_s d_desc;               /* desc. for distance tree */
0204     struct tree_desc_s bl_desc;              /* desc. for bit length tree */
0205 
0206     ush bl_count[MAX_BITS+1];
0207     /* number of codes at each bit length for an optimal tree */
0208 
0209     int heap[2*L_CODES+1];      /* heap used to build the Huffman trees */
0210     int heap_len;               /* number of elements in the heap */
0211     int heap_max;               /* element of largest frequency */
0212     /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
0213      * The same heap array is used to build all trees.
0214      */
0215 
0216     uch depth[2*L_CODES+1];
0217     /* Depth of each subtree used as tie breaker for trees of equal frequency
0218      */
0219 
0220     uchf *sym_buf;        /* buffer for distances and literals/lengths */
0221 
0222     uInt  lit_bufsize;
0223     /* Size of match buffer for literals/lengths.  There are 4 reasons for
0224      * limiting lit_bufsize to 64K:
0225      *   - frequencies can be kept in 16 bit counters
0226      *   - if compression is not successful for the first block, all input
0227      *     data is still in the window so we can still emit a stored block even
0228      *     when input comes from standard input.  (This can also be done for
0229      *     all blocks if lit_bufsize is not greater than 32K.)
0230      *   - if compression is not successful for a file smaller than 64K, we can
0231      *     even emit a stored file instead of a stored block (saving 5 bytes).
0232      *     This is applicable only for zip (not gzip or zlib).
0233      *   - creating new Huffman trees less frequently may not provide fast
0234      *     adaptation to changes in the input data statistics. (Take for
0235      *     example a binary file with poorly compressible code followed by
0236      *     a highly compressible string table.) Smaller buffer sizes give
0237      *     fast adaptation but have of course the overhead of transmitting
0238      *     trees more frequently.
0239      *   - I can't count above 4
0240      */
0241 
0242     uInt sym_next;      /* running index in sym_buf */
0243     uInt sym_end;       /* symbol table full when sym_next reaches this */
0244 
0245     ulg opt_len;        /* bit length of current block with optimal trees */
0246     ulg static_len;     /* bit length of current block with static trees */
0247     uInt matches;       /* number of string matches in current block */
0248     uInt insert;        /* bytes at end of window left to insert */
0249 
0250 #ifdef ZLIB_DEBUG
0251     ulg compressed_len; /* total bit length of compressed file mod 2^32 */
0252     ulg bits_sent;      /* bit length of compressed data sent mod 2^32 */
0253 #endif
0254 
0255     ush bi_buf;
0256     /* Output buffer. bits are inserted starting at the bottom (least
0257      * significant bits).
0258      */
0259     int bi_valid;
0260     /* Number of valid bits in bi_buf.  All bits above the last valid bit
0261      * are always zero.
0262      */
0263 
0264     ulg high_water;
0265     /* High water mark offset in window for initialized bytes -- bytes above
0266      * this are set to zero in order to avoid memory check warnings when
0267      * longest match routines access bytes past the input.  This is then
0268      * updated to the new high water mark.
0269      */
0270 
0271 } FAR deflate_state;
0272 
0273 /* Output a byte on the stream.
0274  * IN assertion: there is enough room in pending_buf.
0275  */
0276 #define put_byte(s, c) {s->pending_buf[s->pending++] = (Bytef)(c);}
0277 
0278 
0279 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
0280 /* Minimum amount of lookahead, except at the end of the input file.
0281  * See deflate.c for comments about the MIN_MATCH+1.
0282  */
0283 
0284 #define MAX_DIST(s)  ((s)->w_size-MIN_LOOKAHEAD)
0285 /* In order to simplify the code, particularly on 16 bit machines, match
0286  * distances are limited to MAX_DIST instead of WSIZE.
0287  */
0288 
0289 #define WIN_INIT MAX_MATCH
0290 /* Number of bytes after end of data in window to initialize in order to avoid
0291    memory checker errors from longest match routines */
0292 
0293         /* in trees.c */
0294 void ZLIB_INTERNAL _tr_init OF((deflate_state *s));
0295 int ZLIB_INTERNAL _tr_tally OF((deflate_state *s, unsigned dist, unsigned lc));
0296 void ZLIB_INTERNAL _tr_flush_block OF((deflate_state *s, charf *buf,
0297                         ulg stored_len, int last));
0298 void ZLIB_INTERNAL _tr_flush_bits OF((deflate_state *s));
0299 void ZLIB_INTERNAL _tr_align OF((deflate_state *s));
0300 void ZLIB_INTERNAL _tr_stored_block OF((deflate_state *s, charf *buf,
0301                         ulg stored_len, int last));
0302 
0303 #define d_code(dist) \
0304    ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)])
0305 /* Mapping from a distance to a distance code. dist is the distance - 1 and
0306  * must not have side effects. _dist_code[256] and _dist_code[257] are never
0307  * used.
0308  */
0309 
0310 #ifndef ZLIB_DEBUG
0311 /* Inline versions of _tr_tally for speed: */
0312 
0313 #if defined(GEN_TREES_H) || !defined(STDC)
0314   extern uch ZLIB_INTERNAL _length_code[];
0315   extern uch ZLIB_INTERNAL _dist_code[];
0316 #else
0317   extern const uch ZLIB_INTERNAL _length_code[];
0318   extern const uch ZLIB_INTERNAL _dist_code[];
0319 #endif
0320 
0321 # define _tr_tally_lit(s, c, flush) \
0322   { uch cc = (c); \
0323     s->sym_buf[s->sym_next++] = 0; \
0324     s->sym_buf[s->sym_next++] = 0; \
0325     s->sym_buf[s->sym_next++] = cc; \
0326     s->dyn_ltree[cc].Freq++; \
0327     flush = (s->sym_next == s->sym_end); \
0328    }
0329 # define _tr_tally_dist(s, distance, length, flush) \
0330   { uch len = (uch)(length); \
0331     ush dist = (ush)(distance); \
0332     s->sym_buf[s->sym_next++] = (uch)dist; \
0333     s->sym_buf[s->sym_next++] = (uch)(dist >> 8); \
0334     s->sym_buf[s->sym_next++] = len; \
0335     dist--; \
0336     s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \
0337     s->dyn_dtree[d_code(dist)].Freq++; \
0338     flush = (s->sym_next == s->sym_end); \
0339   }
0340 #else
0341 # define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c)
0342 # define _tr_tally_dist(s, distance, length, flush) \
0343               flush = _tr_tally(s, distance, length)
0344 #endif
0345 
0346 #endif /* DEFLATE_H */