1 /* unlzw.c -- decompress files in LZW format.
2 * The code in this file is directly derived from the public domain 'compress'
3 * written by Spencer Thomas, Joe Orost, James Woods, Jim McKie, Steve Davies,
4 * Ken Turkowski, Dave Mack and Peter Jannesen.
6 * This is a temporary version which will be rewritten in some future version
7 * to accommodate in-memory decompression.
19 typedef unsigned char char_type;
20 typedef long code_int;
21 typedef unsigned long count_int;
22 typedef unsigned short count_short;
23 typedef unsigned long cmp_code_int;
25 #define MAXCODE(n) (1L << (n))
28 # define BYTEORDER 0000
58 #if BYTEORDER == 4321 && NOALLIGN == 1
59 # define input(b,o,c,n,m){ \
60 (c) = (*(long *)(&(b)[(o)>>3])>>((o)&0x7))&(m); \
64 # define input(b,o,c,n,m){ \
65 char_type *p = &(b)[(o)>>3]; \
66 (c) = ((((long)(p[0]))|((long)(p[1])<<8)| \
67 ((long)(p[2])<<16))>>((o)&0x7))&(m); \
73 /* DECLARE(ush, tab_prefix, (1<<BITS)); -- prefix code */
74 # define tab_prefixof(i) tab_prefix[i]
75 # define clear_tab_prefixof() memzero(tab_prefix, 256);
77 /* DECLARE(ush, tab_prefix0, (1<<(BITS-1)); -- prefix for even codes */
78 /* DECLARE(ush, tab_prefix1, (1<<(BITS-1)); -- prefix for odd codes */
80 # define tab_prefixof(i) tab_prefix[(i)&1][(i)>>1]
81 # define clear_tab_prefixof() \
82 memzero(tab_prefix0, 128), \
83 memzero(tab_prefix1, 128);
85 #define de_stack ((char_type *)(&d_buf[DIST_BUFSIZE-1]))
86 #define tab_suffixof(i) tab_suffix[i]
88 int block_mode = BLOCK_MODE; /* block compress mode -C compatible with 2.0 */
90 /* ============================================================================
91 * Decompress in to out. This routine adapts to the codes in the
92 * file building the "string" table on-the-fly; requiring no table to
93 * be stored in the compressed file.
94 * IN assertions: the buffer inbuf contains already the beginning of
95 * the compressed data, from offsets iptr to insize-1 included.
96 * The magic header has already been checked and skipped.
97 * bytes_in and bytes_out have been initialized.
100 int in, out; /* input and output file descriptors */
110 /* int insize; (global) */
119 tab_prefix[0] = tab_prefix0;
120 tab_prefix[1] = tab_prefix1;
122 maxbits = get_byte();
123 block_mode = maxbits & BLOCK_MODE;
124 if ((maxbits & LZW_RESERVED) != 0) {
125 WARN((stderr, "\n%s: %s: warning, unknown flags 0x%x\n",
126 program_name, ifname, (unsigned int) maxbits & LZW_RESERVED));
129 maxmaxcode = MAXCODE(maxbits);
131 if (maxbits > BITS) {
133 "\n%s: %s: compressed with %d bits, can only handle %d bits\n",
134 program_name, ifname, maxbits, BITS);
139 maxcode = MAXCODE(n_bits = INIT_BITS)-1;
140 bitmask = (1<<n_bits)-1;
146 free_ent = ((block_mode) ? FIRST : 256);
148 clear_tab_prefixof(); /* Initialize the first 256 entries in the table. */
150 for (code = 255 ; code >= 0 ; --code) {
151 tab_suffixof(code) = (char_type)code;
160 e = o <= insize ? insize - o : 0;
162 for (i = 0 ; i < e ; ++i) {
163 inbuf[i] = inbuf[i+o];
168 if (insize < INBUF_EXTRA) {
169 rsize = read_buffer (in, (char *) inbuf + insize, INBUFSIZ);
174 bytes_in += (off_t)rsize;
176 inbits = ((rsize != 0) ? ((long)insize - insize%n_bits)<<3 :
177 ((long)insize<<3)-(n_bits-1));
179 while (inbits > posbits) {
180 if (free_ent > maxcode) {
181 posbits = ((posbits-1) +
182 ((n_bits<<3)-(posbits-1+(n_bits<<3))%(n_bits<<3)));
184 if (n_bits == maxbits) {
185 maxcode = maxmaxcode;
187 maxcode = MAXCODE(n_bits)-1;
189 bitmask = (1<<n_bits)-1;
192 input(inbuf,posbits,code,n_bits,bitmask);
193 Tracev((stderr, "%ld ", code));
197 gzip_error ("corrupt input.");
198 outbuf[outpos++] = (char_type)(finchar = (int)(oldcode=code));
201 if (code == CLEAR && block_mode) {
202 clear_tab_prefixof();
203 free_ent = FIRST - 1;
204 posbits = ((posbits-1) +
205 ((n_bits<<3)-(posbits-1+(n_bits<<3))%(n_bits<<3)));
206 maxcode = MAXCODE(n_bits = INIT_BITS)-1;
207 bitmask = (1<<n_bits)-1;
213 if (code >= free_ent) { /* Special case for KwKwK string. */
214 if (code > free_ent) {
219 p = &inbuf[posbits>>3];
221 "code:%ld free_ent:%ld n_bits:%d insize:%u\n",
222 code, free_ent, n_bits, insize);
224 "posbits:%ld inbuf:%02X %02X %02X %02X %02X\n",
225 posbits, p[-1],p[0],p[1],p[2],p[3]);
227 if (!test && outpos > 0) {
228 write_buf(out, (char*)outbuf, outpos);
229 bytes_out += (off_t)outpos;
231 gzip_error (to_stdout
233 : "corrupt input. Use zcat to recover some data.");
235 *--stackp = (char_type)finchar;
239 while ((cmp_code_int)code >= (cmp_code_int)256) {
240 /* Generate output characters in reverse order */
241 *--stackp = tab_suffixof(code);
242 code = tab_prefixof(code);
244 *--stackp = (char_type)(finchar = tab_suffixof(code));
246 /* And put them out in forward order */
250 if (outpos+(i = (de_stack-stackp)) >= OUTBUFSIZ) {
252 if (i > OUTBUFSIZ-outpos) i = OUTBUFSIZ-outpos;
255 memcpy(outbuf+outpos, stackp, i);
258 if (outpos >= OUTBUFSIZ) {
260 write_buf(out, (char*)outbuf, outpos);
261 bytes_out += (off_t)outpos;
266 } while ((i = (de_stack-stackp)) > 0);
268 memcpy(outbuf+outpos, stackp, i);
273 if ((code = free_ent) < maxmaxcode) { /* Generate the new entry. */
275 tab_prefixof(code) = (unsigned short)oldcode;
276 tab_suffixof(code) = (char_type)finchar;
279 oldcode = incode; /* Remember previous code. */
281 } while (rsize != 0);
283 if (!test && outpos > 0) {
284 write_buf(out, (char*)outbuf, outpos);
285 bytes_out += (off_t)outpos;