maint: update all copyright year number ranges
[gzip.git] / unpack.c
blob1483e364f55cdf662cc26b6230d5b1e6b8a3c332
1 /* unpack.c -- decompress files in pack format.
3 Copyright (C) 1997, 1999, 2006, 2009-2013 Free Software Foundation, Inc.
4 Copyright (C) 1992-1993 Jean-loup Gailly
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 3, or (at your option)
9 any later version.
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with this program; if not, write to the Free Software Foundation,
18 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
20 #include <config.h>
21 #include "tailor.h"
22 #include "gzip.h"
24 #define MIN(a,b) ((a) <= (b) ? (a) : (b))
25 /* The arguments must not have side effects. */
27 #define MAX_BITLEN 25
28 /* Maximum length of Huffman codes. (Minor modifications to the code
29 * would be needed to support 32 bits codes, but pack never generates
30 * more than 24 bits anyway.)
33 #define LITERALS 256
34 /* Number of literals, excluding the End of Block (EOB) code */
36 #define MAX_PEEK 12
37 /* Maximum number of 'peek' bits used to optimize traversal of the
38 * Huffman tree.
41 local ulg orig_len; /* original uncompressed length */
42 local int max_len; /* maximum bit length of Huffman codes */
44 local uch literal[LITERALS];
45 /* The literal bytes present in the Huffman tree. The EOB code is not
46 * represented.
49 local int lit_base[MAX_BITLEN+1];
50 /* All literals of a given bit length are contiguous in literal[] and
51 * have contiguous codes. literal[code+lit_base[len]] is the literal
52 * for a code of len bits.
55 local int leaves [MAX_BITLEN+1]; /* Number of leaves for each bit length */
56 local int parents[MAX_BITLEN+1]; /* Number of parents for each bit length */
58 local int peek_bits; /* Number of peek bits currently used */
60 /* local uch prefix_len[1 << MAX_PEEK]; */
61 #define prefix_len outbuf
62 /* For each bit pattern b of peek_bits bits, prefix_len[b] is the length
63 * of the Huffman code starting with a prefix of b (upper bits), or 0
64 * if all codes of prefix b have more than peek_bits bits. It is not
65 * necessary to have a huge table (large MAX_PEEK) because most of the
66 * codes encountered in the input stream are short codes (by construction).
67 * So for most codes a single lookup will be necessary.
69 #if (1<<MAX_PEEK) > OUTBUFSIZ
70 error cannot overlay prefix_len and outbuf
71 #endif
73 local ulg bitbuf;
74 /* Bits are added on the low part of bitbuf and read from the high part. */
76 local int valid; /* number of valid bits in bitbuf */
77 /* all bits above the last valid bit are always zero */
79 /* Read an input byte, reporting an error at EOF. */
80 static unsigned char
81 read_byte (void)
83 int b = get_byte ();
84 if (b < 0)
85 gzip_error ("invalid compressed data -- unexpected end of file");
86 return b;
89 /* Set code to the next 'bits' input bits without skipping them. code
90 * must be the name of a simple variable and bits must not have side effects.
91 * IN assertions: bits <= 25 (so that we still have room for an extra byte
92 * when valid is only 24), and mask = (1<<bits)-1.
94 #define look_bits(code,bits,mask) \
95 { \
96 while (valid < (bits)) bitbuf = (bitbuf<<8) | read_byte(), valid += 8; \
97 code = (bitbuf >> (valid-(bits))) & (mask); \
100 /* Skip the given number of bits (after having peeked at them): */
101 #define skip_bits(bits) (valid -= (bits))
103 #define clear_bitbuf() (valid = 0, bitbuf = 0)
105 /* Local functions */
107 local void read_tree (void);
108 local void build_tree (void);
110 /* ===========================================================================
111 * Read the Huffman tree.
113 local void read_tree()
115 int len; /* bit length */
116 int base; /* base offset for a sequence of leaves */
117 int n;
118 int max_leaves = 1;
120 /* Read the original input size, MSB first */
121 orig_len = 0;
122 for (n = 1; n <= 4; n++)
123 orig_len = (orig_len << 8) | read_byte ();
125 /* Read the maximum bit length of Huffman codes. */
126 max_len = read_byte ();
127 if (! (0 < max_len && max_len <= MAX_BITLEN))
128 gzip_error ("invalid compressed data -- "
129 "Huffman code bit length out of range");
131 /* Get the number of leaves at each bit length */
132 n = 0;
133 for (len = 1; len <= max_len; len++) {
134 leaves[len] = read_byte ();
135 if (max_leaves - (len == max_len) < leaves[len])
136 gzip_error ("too many leaves in Huffman tree");
137 max_leaves = (max_leaves - leaves[len] + 1) * 2 - 1;
138 n += leaves[len];
140 if (LITERALS <= n) {
141 gzip_error ("too many leaves in Huffman tree");
143 Trace((stderr, "orig_len %lu, max_len %d, leaves %d\n",
144 orig_len, max_len, n));
145 /* There are at least 2 and at most 256 leaves of length max_len.
146 * (Pack arbitrarily rejects empty files and files consisting of
147 * a single byte even repeated.) To fit the last leaf count in a
148 * byte, it is offset by 2. However, the last literal is the EOB
149 * code, and is not transmitted explicitly in the tree, so we must
150 * adjust here by one only.
152 leaves[max_len]++;
154 /* Now read the leaves themselves */
155 base = 0;
156 for (len = 1; len <= max_len; len++) {
157 /* Remember where the literals of this length start in literal[] : */
158 lit_base[len] = base;
159 /* And read the literals: */
160 for (n = leaves[len]; n > 0; n--) {
161 literal[base++] = read_byte ();
164 leaves[max_len]++; /* Now include the EOB code in the Huffman tree */
167 /* ===========================================================================
168 * Build the Huffman tree and the prefix table.
170 local void build_tree()
172 int nodes = 0; /* number of nodes (parents+leaves) at current bit length */
173 int len; /* current bit length */
174 uch *prefixp; /* pointer in prefix_len */
176 for (len = max_len; len >= 1; len--) {
177 /* The number of parent nodes at this level is half the total
178 * number of nodes at parent level:
180 nodes >>= 1;
181 parents[len] = nodes;
182 /* Update lit_base by the appropriate bias to skip the parent nodes
183 * (which are not represented in the literal array):
185 lit_base[len] -= nodes;
186 /* Restore nodes to be parents+leaves: */
187 nodes += leaves[len];
189 /* Construct the prefix table, from shortest leaves to longest ones.
190 * The shortest code is all ones, so we start at the end of the table.
192 peek_bits = MIN(max_len, MAX_PEEK);
193 prefixp = &prefix_len[1<<peek_bits];
194 for (len = 1; len <= peek_bits; len++) {
195 int prefixes = leaves[len] << (peek_bits-len); /* may be 0 */
196 while (prefixes--) *--prefixp = (uch)len;
198 /* The length of all other codes is unknown: */
199 while (prefixp > prefix_len) *--prefixp = 0;
202 /* ===========================================================================
203 * Unpack in to out. This routine does not support the old pack format
204 * with magic header \037\037.
206 * IN assertions: the buffer inbuf contains already the beginning of
207 * the compressed data, from offsets inptr to insize-1 included.
208 * The magic header has already been checked. The output buffer is cleared.
210 int unpack(in, out)
211 int in, out; /* input and output file descriptors */
213 int len; /* Bit length of current code */
214 unsigned eob; /* End Of Block code */
215 register unsigned peek; /* lookahead bits */
216 unsigned peek_mask; /* Mask for peek_bits bits */
218 ifd = in;
219 ofd = out;
221 read_tree(); /* Read the Huffman tree */
222 build_tree(); /* Build the prefix table */
223 clear_bitbuf(); /* Initialize bit input */
224 peek_mask = (1<<peek_bits)-1;
226 /* The eob code is the largest code among all leaves of maximal length: */
227 eob = leaves[max_len]-1;
228 Trace((stderr, "eob %d %x\n", max_len, eob));
230 /* Decode the input data: */
231 for (;;) {
232 /* Since eob is the longest code and not shorter than max_len,
233 * we can peek at max_len bits without having the risk of reading
234 * beyond the end of file.
236 look_bits(peek, peek_bits, peek_mask);
237 len = prefix_len[peek];
238 if (len > 0) {
239 peek >>= peek_bits - len; /* discard the extra bits */
240 } else {
241 /* Code of more than peek_bits bits, we must traverse the tree */
242 ulg mask = peek_mask;
243 len = peek_bits;
245 /* Loop as long as peek is a parent node. */
246 while (peek < parents[len])
248 len++, mask = (mask<<1)+1;
249 look_bits(peek, len, mask);
252 /* At this point, peek is the next complete code, of len bits */
253 if (peek == eob)
254 break; /* End of file. */
255 if (eob < peek)
256 gzip_error ("invalid compressed data--code out of range");
257 put_ubyte(literal[peek+lit_base[len]]);
258 Tracev((stderr,"%02d %04x %c\n", len, peek,
259 literal[peek+lit_base[len]]));
260 skip_bits(len);
261 } /* for (;;) */
263 flush_window();
264 if (orig_len != (ulg)(bytes_out & 0xffffffff)) {
265 gzip_error ("invalid compressed data--length error");
267 return OK;