xref: /illumos-gate/usr/src/contrib/zlib/deflate.h (revision 148fd93e57d3d5813d90f1291e6bd30de45c7723)
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 #ifndef DEFLATE_H
12 #define DEFLATE_H
13 
14 #include "zutil.h"
15 
16 /* define NO_GZIP when compiling if you want to disable gzip header and
17    trailer creation by deflate().  NO_GZIP would be used to avoid linking in
18    the crc code when it is not needed.  For shared libraries, gzip encoding
19    should be left enabled. */
20 #ifndef NO_GZIP
21 #  define GZIP
22 #endif
23 
24 /* ===========================================================================
25  * Internal compression state.
26  */
27 
28 #define LENGTH_CODES 29
29 /* number of length codes, not counting the special END_BLOCK code */
30 
31 #define LITERALS  256
32 /* number of literal bytes 0..255 */
33 
34 #define L_CODES (LITERALS+1+LENGTH_CODES)
35 /* number of Literal or Length codes, including the END_BLOCK code */
36 
37 #define D_CODES   30
38 /* number of distance codes */
39 
40 #define BL_CODES  19
41 /* number of codes used to transfer the bit lengths */
42 
43 #define HEAP_SIZE (2*L_CODES+1)
44 /* maximum heap size */
45 
46 #define MAX_BITS 15
47 /* All codes must not exceed MAX_BITS bits */
48 
49 #define Buf_size 16
50 /* size of bit buffer in bi_buf */
51 
52 #define INIT_STATE    42    /* zlib header -> BUSY_STATE */
53 #ifdef GZIP
54 #  define GZIP_STATE  57    /* gzip header -> BUSY_STATE | EXTRA_STATE */
55 #endif
56 #define EXTRA_STATE   69    /* gzip extra block -> NAME_STATE */
57 #define NAME_STATE    73    /* gzip file name -> COMMENT_STATE */
58 #define COMMENT_STATE 91    /* gzip comment -> HCRC_STATE */
59 #define HCRC_STATE   103    /* gzip header CRC -> BUSY_STATE */
60 #define BUSY_STATE   113    /* deflate -> FINISH_STATE */
61 #define FINISH_STATE 666    /* stream complete */
62 /* Stream status */
63 
64 
65 /* Data structure describing a single value and its code string. */
66 typedef struct ct_data_s {
67     union {
68         ush  freq;       /* frequency count */
69         ush  code;       /* bit string */
70     } fc;
71     union {
72         ush  dad;        /* father node in Huffman tree */
73         ush  len;        /* length of bit string */
74     } dl;
75 } FAR ct_data;
76 
77 #define Freq fc.freq
78 #define Code fc.code
79 #define Dad  dl.dad
80 #define Len  dl.len
81 
82 typedef struct static_tree_desc_s  static_tree_desc;
83 
84 typedef struct tree_desc_s {
85     ct_data *dyn_tree;           /* the dynamic tree */
86     int     max_code;            /* largest code with non zero frequency */
87     const static_tree_desc *stat_desc;  /* the corresponding static tree */
88 } FAR tree_desc;
89 
90 typedef ush Pos;
91 typedef Pos FAR Posf;
92 typedef unsigned IPos;
93 
94 /* A Pos is an index in the character window. We use short instead of int to
95  * save space in the various tables. IPos is used only for parameter passing.
96  */
97 
98 typedef struct internal_state {
99     z_streamp strm;      /* pointer back to this zlib stream */
100     int   status;        /* as the name implies */
101     Bytef *pending_buf;  /* output still pending */
102     ulg   pending_buf_size; /* size of pending_buf */
103     Bytef *pending_out;  /* next pending byte to output to the stream */
104     ulg   pending;       /* nb of bytes in the pending buffer */
105     int   wrap;          /* bit 0 true for zlib, bit 1 true for gzip */
106     gz_headerp  gzhead;  /* gzip header information to write */
107     ulg   gzindex;       /* where in extra, name, or comment */
108     Byte  method;        /* can only be DEFLATED */
109     int   last_flush;    /* value of flush param for previous deflate call */
110 
111                 /* used by deflate.c: */
112 
113     uInt  w_size;        /* LZ77 window size (32K by default) */
114     uInt  w_bits;        /* log2(w_size)  (8..16) */
115     uInt  w_mask;        /* w_size - 1 */
116 
117     Bytef *window;
118     /* Sliding window. Input bytes are read into the second half of the window,
119      * and move to the first half later to keep a dictionary of at least wSize
120      * bytes. With this organization, matches are limited to a distance of
121      * wSize-MAX_MATCH bytes, but this ensures that IO is always
122      * performed with a length multiple of the block size. Also, it limits
123      * the window size to 64K, which is quite useful on MSDOS.
124      * To do: use the user input buffer as sliding window.
125      */
126 
127     ulg window_size;
128     /* Actual size of window: 2*wSize, except when the user input buffer
129      * is directly used as sliding window.
130      */
131 
132     Posf *prev;
133     /* Link to older string with same hash index. To limit the size of this
134      * array to 64K, this link is maintained only for the last 32K strings.
135      * An index in this array is thus a window index modulo 32K.
136      */
137 
138     Posf *head; /* Heads of the hash chains or NIL. */
139 
140     uInt  ins_h;          /* hash index of string to be inserted */
141     uInt  hash_size;      /* number of elements in hash table */
142     uInt  hash_bits;      /* log2(hash_size) */
143     uInt  hash_mask;      /* hash_size-1 */
144 
145     uInt  hash_shift;
146     /* Number of bits by which ins_h must be shifted at each input
147      * step. It must be such that after MIN_MATCH steps, the oldest
148      * byte no longer takes part in the hash key, that is:
149      *   hash_shift * MIN_MATCH >= hash_bits
150      */
151 
152     long block_start;
153     /* Window position at the beginning of the current output block. Gets
154      * negative when the window is moved backwards.
155      */
156 
157     uInt match_length;           /* length of best match */
158     IPos prev_match;             /* previous match */
159     int match_available;         /* set if previous match exists */
160     uInt strstart;               /* start of string to insert */
161     uInt match_start;            /* start of matching string */
162     uInt lookahead;              /* number of valid bytes ahead in window */
163 
164     uInt prev_length;
165     /* Length of the best match at previous step. Matches not greater than this
166      * are discarded. This is used in the lazy match evaluation.
167      */
168 
169     uInt max_chain_length;
170     /* To speed up deflation, hash chains are never searched beyond this
171      * length.  A higher limit improves compression ratio but degrades the
172      * speed.
173      */
174 
175     uInt max_lazy_match;
176     /* Attempt to find a better match only when the current match is strictly
177      * smaller than this value. This mechanism is used only for compression
178      * levels >= 4.
179      */
180 #   define max_insert_length  max_lazy_match
181     /* Insert new strings in the hash table only if the match length is not
182      * greater than this length. This saves time but degrades compression.
183      * max_insert_length is used only for compression levels <= 3.
184      */
185 
186     int level;    /* compression level (1..9) */
187     int strategy; /* favor or force Huffman coding*/
188 
189     uInt good_match;
190     /* Use a faster search when the previous match is longer than this */
191 
192     int nice_match; /* Stop searching when current match exceeds this */
193 
194                 /* used by trees.c: */
195     /* Didn't use ct_data typedef below to suppress compiler warning */
196     struct ct_data_s dyn_ltree[HEAP_SIZE];   /* literal and length tree */
197     struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */
198     struct ct_data_s bl_tree[2*BL_CODES+1];  /* Huffman tree for bit lengths */
199 
200     struct tree_desc_s l_desc;               /* desc. for literal tree */
201     struct tree_desc_s d_desc;               /* desc. for distance tree */
202     struct tree_desc_s bl_desc;              /* desc. for bit length tree */
203 
204     ush bl_count[MAX_BITS+1];
205     /* number of codes at each bit length for an optimal tree */
206 
207     int heap[2*L_CODES+1];      /* heap used to build the Huffman trees */
208     int heap_len;               /* number of elements in the heap */
209     int heap_max;               /* element of largest frequency */
210     /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
211      * The same heap array is used to build all trees.
212      */
213 
214     uch depth[2*L_CODES+1];
215     /* Depth of each subtree used as tie breaker for trees of equal frequency
216      */
217 
218     uchf *sym_buf;        /* buffer for distances and literals/lengths */
219 
220     uInt  lit_bufsize;
221     /* Size of match buffer for literals/lengths.  There are 4 reasons for
222      * limiting lit_bufsize to 64K:
223      *   - frequencies can be kept in 16 bit counters
224      *   - if compression is not successful for the first block, all input
225      *     data is still in the window so we can still emit a stored block even
226      *     when input comes from standard input.  (This can also be done for
227      *     all blocks if lit_bufsize is not greater than 32K.)
228      *   - if compression is not successful for a file smaller than 64K, we can
229      *     even emit a stored file instead of a stored block (saving 5 bytes).
230      *     This is applicable only for zip (not gzip or zlib).
231      *   - creating new Huffman trees less frequently may not provide fast
232      *     adaptation to changes in the input data statistics. (Take for
233      *     example a binary file with poorly compressible code followed by
234      *     a highly compressible string table.) Smaller buffer sizes give
235      *     fast adaptation but have of course the overhead of transmitting
236      *     trees more frequently.
237      *   - I can't count above 4
238      */
239 
240     uInt sym_next;      /* running index in sym_buf */
241     uInt sym_end;       /* symbol table full when sym_next reaches this */
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     uInt insert;        /* bytes at end of window left to insert */
247 
248 #ifdef ZLIB_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     ulg high_water;
263     /* High water mark offset in window for initialized bytes -- bytes above
264      * this are set to zero in order to avoid memory check warnings when
265      * longest match routines access bytes past the input.  This is then
266      * updated to the new high water mark.
267      */
268 
269 } FAR deflate_state;
270 
271 /* Output a byte on the stream.
272  * IN assertion: there is enough room in pending_buf.
273  */
274 #define put_byte(s, c) {s->pending_buf[s->pending++] = (Bytef)(c);}
275 
276 
277 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
278 /* Minimum amount of lookahead, except at the end of the input file.
279  * See deflate.c for comments about the MIN_MATCH+1.
280  */
281 
282 #define MAX_DIST(s)  ((s)->w_size-MIN_LOOKAHEAD)
283 /* In order to simplify the code, particularly on 16 bit machines, match
284  * distances are limited to MAX_DIST instead of WSIZE.
285  */
286 
287 #define WIN_INIT MAX_MATCH
288 /* Number of bytes after end of data in window to initialize in order to avoid
289    memory checker errors from longest match routines */
290 
291         /* in trees.c */
292 void ZLIB_INTERNAL _tr_init OF((deflate_state *s));
293 int ZLIB_INTERNAL _tr_tally OF((deflate_state *s, unsigned dist, unsigned lc));
294 void ZLIB_INTERNAL _tr_flush_block OF((deflate_state *s, charf *buf,
295                         ulg stored_len, int last));
296 void ZLIB_INTERNAL _tr_flush_bits OF((deflate_state *s));
297 void ZLIB_INTERNAL _tr_align OF((deflate_state *s));
298 void ZLIB_INTERNAL _tr_stored_block OF((deflate_state *s, charf *buf,
299                         ulg stored_len, int last));
300 
301 #define d_code(dist) \
302    ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)])
303 /* Mapping from a distance to a distance code. dist is the distance - 1 and
304  * must not have side effects. _dist_code[256] and _dist_code[257] are never
305  * used.
306  */
307 
308 #ifndef ZLIB_DEBUG
309 /* Inline versions of _tr_tally for speed: */
310 
311 #if defined(GEN_TREES_H) || !defined(STDC)
312   extern uch ZLIB_INTERNAL _length_code[];
313   extern uch ZLIB_INTERNAL _dist_code[];
314 #else
315   extern const uch ZLIB_INTERNAL _length_code[];
316   extern const uch ZLIB_INTERNAL _dist_code[];
317 #endif
318 
319 # define _tr_tally_lit(s, c, flush) \
320   { uch cc = (c); \
321     s->sym_buf[s->sym_next++] = 0; \
322     s->sym_buf[s->sym_next++] = 0; \
323     s->sym_buf[s->sym_next++] = cc; \
324     s->dyn_ltree[cc].Freq++; \
325     flush = (s->sym_next == s->sym_end); \
326    }
327 # define _tr_tally_dist(s, distance, length, flush) \
328   { uch len = (uch)(length); \
329     ush dist = (ush)(distance); \
330     s->sym_buf[s->sym_next++] = dist; \
331     s->sym_buf[s->sym_next++] = dist >> 8; \
332     s->sym_buf[s->sym_next++] = len; \
333     dist--; \
334     s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \
335     s->dyn_dtree[d_code(dist)].Freq++; \
336     flush = (s->sym_next == s->sym_end); \
337   }
338 #else
339 # define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c)
340 # define _tr_tally_dist(s, distance, length, flush) \
341               flush = _tr_tally(s, distance, length)
342 #endif
343 
344 #endif /* DEFLATE_H */
345