• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /* inftree9.c -- generate Huffman trees for efficient decoding
2  * Copyright (C) 1995-2005 Mark Adler
3  * For conditions of distribution and use, see copyright notice in zlib.h
4  */
5 
6 #include "zutil.h"
7 #include "inftree9.h"
8 
9 #define MAXBITS 15
10 
11 const char inflate9_copyright[] =
12    " inflate9 1.2.3 Copyright 1995-2005 Mark Adler ";
13 /*
14   If you use the zlib library in a product, an acknowledgment is welcome
15   in the documentation of your product. If for some reason you cannot
16   include such an acknowledgment, I would appreciate that you keep this
17   copyright string in the executable of your product.
18  */
19 
20 /*
21    Build a set of tables to decode the provided canonical Huffman code.
22    The code lengths are lens[0..codes-1].  The result starts at *table,
23    whose indices are 0..2^bits-1.  work is a writable array of at least
24    lens shorts, which is used as a work area.  type is the type of code
25    to be generated, CODES, LENS, or DISTS.  On return, zero is success,
26    -1 is an invalid code, and +1 means that ENOUGH isn't enough.  table
27    on return points to the next available entry's address.  bits is the
28    requested root table index bits, and on return it is the actual root
29    table index bits.  It will differ if the request is greater than the
30    longest code or if it is less than the shortest code.
31  */
inflate_table9(type,lens,codes,table,bits,work)32 int inflate_table9(type, lens, codes, table, bits, work)
33 codetype type;
34 unsigned short FAR *lens;
35 unsigned codes;
36 code FAR * FAR *table;
37 unsigned FAR *bits;
38 unsigned short FAR *work;
39 {
40     unsigned len;               /* a code's length in bits */
41     unsigned sym;               /* index of code symbols */
42     unsigned min, max;          /* minimum and maximum code lengths */
43     unsigned root;              /* number of index bits for root table */
44     unsigned curr;              /* number of index bits for current table */
45     unsigned drop;              /* code bits to drop for sub-table */
46     int left;                   /* number of prefix codes available */
47     unsigned used;              /* code entries in table used */
48     unsigned huff;              /* Huffman code */
49     unsigned incr;              /* for incrementing code, index */
50     unsigned fill;              /* index for replicating entries */
51     unsigned low;               /* low bits for current root entry */
52     unsigned mask;              /* mask for low root bits */
53     code this;                  /* table entry for duplication */
54     code FAR *next;             /* next available space in table */
55     const unsigned short FAR *base;     /* base value table to use */
56     const unsigned short FAR *extra;    /* extra bits table to use */
57     int end;                    /* use base and extra for symbol > end */
58     unsigned short count[MAXBITS+1];    /* number of codes of each length */
59     unsigned short offs[MAXBITS+1];     /* offsets in table for each length */
60     static const unsigned short lbase[31] = { /* Length codes 257..285 base */
61         3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17,
62         19, 23, 27, 31, 35, 43, 51, 59, 67, 83, 99, 115,
63         131, 163, 195, 227, 3, 0, 0};
64     static const unsigned short lext[31] = { /* Length codes 257..285 extra */
65         128, 128, 128, 128, 128, 128, 128, 128, 129, 129, 129, 129,
66         130, 130, 130, 130, 131, 131, 131, 131, 132, 132, 132, 132,
67         133, 133, 133, 133, 144, 201, 196};
68     static const unsigned short dbase[32] = { /* Distance codes 0..31 base */
69         1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49,
70         65, 97, 129, 193, 257, 385, 513, 769, 1025, 1537, 2049, 3073,
71         4097, 6145, 8193, 12289, 16385, 24577, 32769, 49153};
72     static const unsigned short dext[32] = { /* Distance codes 0..31 extra */
73         128, 128, 128, 128, 129, 129, 130, 130, 131, 131, 132, 132,
74         133, 133, 134, 134, 135, 135, 136, 136, 137, 137, 138, 138,
75         139, 139, 140, 140, 141, 141, 142, 142};
76 
77     /*
78        Process a set of code lengths to create a canonical Huffman code.  The
79        code lengths are lens[0..codes-1].  Each length corresponds to the
80        symbols 0..codes-1.  The Huffman code is generated by first sorting the
81        symbols by length from short to long, and retaining the symbol order
82        for codes with equal lengths.  Then the code starts with all zero bits
83        for the first code of the shortest length, and the codes are integer
84        increments for the same length, and zeros are appended as the length
85        increases.  For the deflate format, these bits are stored backwards
86        from their more natural integer increment ordering, and so when the
87        decoding tables are built in the large loop below, the integer codes
88        are incremented backwards.
89 
90        This routine assumes, but does not check, that all of the entries in
91        lens[] are in the range 0..MAXBITS.  The caller must assure this.
92        1..MAXBITS is interpreted as that code length.  zero means that that
93        symbol does not occur in this code.
94 
95        The codes are sorted by computing a count of codes for each length,
96        creating from that a table of starting indices for each length in the
97        sorted table, and then entering the symbols in order in the sorted
98        table.  The sorted table is work[], with that space being provided by
99        the caller.
100 
101        The length counts are used for other purposes as well, i.e. finding
102        the minimum and maximum length codes, determining if there are any
103        codes at all, checking for a valid set of lengths, and looking ahead
104        at length counts to determine sub-table sizes when building the
105        decoding tables.
106      */
107 
108     /* accumulate lengths for codes (assumes lens[] all in 0..MAXBITS) */
109     for (len = 0; len <= MAXBITS; len++)
110         count[len] = 0;
111     for (sym = 0; sym < codes; sym++)
112         count[lens[sym]]++;
113 
114     /* bound code lengths, force root to be within code lengths */
115     root = *bits;
116     for (max = MAXBITS; max >= 1; max--)
117         if (count[max] != 0) break;
118     if (root > max) root = max;
119     if (max == 0) return -1;            /* no codes! */
120     for (min = 1; min <= MAXBITS; min++)
121         if (count[min] != 0) break;
122     if (root < min) root = min;
123 
124     /* check for an over-subscribed or incomplete set of lengths */
125     left = 1;
126     for (len = 1; len <= MAXBITS; len++) {
127         left <<= 1;
128         left -= count[len];
129         if (left < 0) return -1;        /* over-subscribed */
130     }
131     if (left > 0 && (type == CODES || max != 1))
132         return -1;                      /* incomplete set */
133 
134     /* generate offsets into symbol table for each length for sorting */
135     offs[1] = 0;
136     for (len = 1; len < MAXBITS; len++)
137         offs[len + 1] = offs[len] + count[len];
138 
139     /* sort symbols by length, by symbol order within each length */
140     for (sym = 0; sym < codes; sym++)
141         if (lens[sym] != 0) work[offs[lens[sym]]++] = (unsigned short)sym;
142 
143     /*
144        Create and fill in decoding tables.  In this loop, the table being
145        filled is at next and has curr index bits.  The code being used is huff
146        with length len.  That code is converted to an index by dropping drop
147        bits off of the bottom.  For codes where len is less than drop + curr,
148        those top drop + curr - len bits are incremented through all values to
149        fill the table with replicated entries.
150 
151        root is the number of index bits for the root table.  When len exceeds
152        root, sub-tables are created pointed to by the root entry with an index
153        of the low root bits of huff.  This is saved in low to check for when a
154        new sub-table should be started.  drop is zero when the root table is
155        being filled, and drop is root when sub-tables are being filled.
156 
157        When a new sub-table is needed, it is necessary to look ahead in the
158        code lengths to determine what size sub-table is needed.  The length
159        counts are used for this, and so count[] is decremented as codes are
160        entered in the tables.
161 
162        used keeps track of how many table entries have been allocated from the
163        provided *table space.  It is checked when a LENS table is being made
164        against the space in *table, ENOUGH, minus the maximum space needed by
165        the worst case distance code, MAXD.  This should never happen, but the
166        sufficiency of ENOUGH has not been proven exhaustively, hence the check.
167        This assumes that when type == LENS, bits == 9.
168 
169        sym increments through all symbols, and the loop terminates when
170        all codes of length max, i.e. all codes, have been processed.  This
171        routine permits incomplete codes, so another loop after this one fills
172        in the rest of the decoding tables with invalid code markers.
173      */
174 
175     /* set up for code type */
176     switch (type) {
177     case CODES:
178         base = extra = work;    /* dummy value--not used */
179         end = 19;
180         break;
181     case LENS:
182         base = lbase;
183         base -= 257;
184         extra = lext;
185         extra -= 257;
186         end = 256;
187         break;
188     default:            /* DISTS */
189         base = dbase;
190         extra = dext;
191         end = -1;
192     }
193 
194     /* initialize state for loop */
195     huff = 0;                   /* starting code */
196     sym = 0;                    /* starting code symbol */
197     len = min;                  /* starting code length */
198     next = *table;              /* current table to fill in */
199     curr = root;                /* current table index bits */
200     drop = 0;                   /* current bits to drop from code for index */
201     low = (unsigned)(-1);       /* trigger new sub-table when len > root */
202     used = 1U << root;          /* use root table entries */
203     mask = used - 1;            /* mask for comparing low */
204 
205     /* check available table space */
206     if (type == LENS && used >= ENOUGH - MAXD)
207         return 1;
208 
209     /* process all codes and make table entries */
210     for (;;) {
211         /* create table entry */
212         this.bits = (unsigned char)(len - drop);
213         if ((int)(work[sym]) < end) {
214             this.op = (unsigned char)0;
215             this.val = work[sym];
216         }
217         else if ((int)(work[sym]) > end) {
218             this.op = (unsigned char)(extra[work[sym]]);
219             this.val = base[work[sym]];
220         }
221         else {
222             this.op = (unsigned char)(32 + 64);         /* end of block */
223             this.val = 0;
224         }
225 
226         /* replicate for those indices with low len bits equal to huff */
227         incr = 1U << (len - drop);
228         fill = 1U << curr;
229         do {
230             fill -= incr;
231             next[(huff >> drop) + fill] = this;
232         } while (fill != 0);
233 
234         /* backwards increment the len-bit code huff */
235         incr = 1U << (len - 1);
236         while (huff & incr)
237             incr >>= 1;
238         if (incr != 0) {
239             huff &= incr - 1;
240             huff += incr;
241         }
242         else
243             huff = 0;
244 
245         /* go to next symbol, update count, len */
246         sym++;
247         if (--(count[len]) == 0) {
248             if (len == max) break;
249             len = lens[work[sym]];
250         }
251 
252         /* create new sub-table if needed */
253         if (len > root && (huff & mask) != low) {
254             /* if first time, transition to sub-tables */
255             if (drop == 0)
256                 drop = root;
257 
258             /* increment past last table */
259             next += 1U << curr;
260 
261             /* determine length of next table */
262             curr = len - drop;
263             left = (int)(1 << curr);
264             while (curr + drop < max) {
265                 left -= count[curr + drop];
266                 if (left <= 0) break;
267                 curr++;
268                 left <<= 1;
269             }
270 
271             /* check for enough space */
272             used += 1U << curr;
273             if (type == LENS && used >= ENOUGH - MAXD)
274                 return 1;
275 
276             /* point entry in root table to sub-table */
277             low = huff & mask;
278             (*table)[low].op = (unsigned char)curr;
279             (*table)[low].bits = (unsigned char)root;
280             (*table)[low].val = (unsigned short)(next - *table);
281         }
282     }
283 
284     /*
285        Fill in rest of table for incomplete codes.  This loop is similar to the
286        loop above in incrementing huff for table indices.  It is assumed that
287        len is equal to curr + drop, so there is no loop needed to increment
288        through high index bits.  When the current sub-table is filled, the loop
289        drops back to the root table to fill in any remaining entries there.
290      */
291     this.op = (unsigned char)64;                /* invalid code marker */
292     this.bits = (unsigned char)(len - drop);
293     this.val = (unsigned short)0;
294     while (huff != 0) {
295         /* when done with sub-table, drop back to root table */
296         if (drop != 0 && (huff & mask) != low) {
297             drop = 0;
298             len = root;
299             next = *table;
300             curr = root;
301             this.bits = (unsigned char)len;
302         }
303 
304         /* put invalid code marker in table */
305         next[huff >> drop] = this;
306 
307         /* backwards increment the len-bit code huff */
308         incr = 1U << (len - 1);
309         while (huff & incr)
310             incr >>= 1;
311         if (incr != 0) {
312             huff &= incr - 1;
313             huff += incr;
314         }
315         else
316             huff = 0;
317     }
318 
319     /* set return parameters */
320     *table += used;
321     *bits = root;
322     return 0;
323 }
324