cannam@89
|
1 /* gun.c -- simple gunzip to give an example of the use of inflateBack()
|
cannam@89
|
2 * Copyright (C) 2003, 2005, 2008, 2010 Mark Adler
|
cannam@89
|
3 * For conditions of distribution and use, see copyright notice in zlib.h
|
cannam@89
|
4 Version 1.6 17 January 2010 Mark Adler */
|
cannam@89
|
5
|
cannam@89
|
6 /* Version history:
|
cannam@89
|
7 1.0 16 Feb 2003 First version for testing of inflateBack()
|
cannam@89
|
8 1.1 21 Feb 2005 Decompress concatenated gzip streams
|
cannam@89
|
9 Remove use of "this" variable (C++ keyword)
|
cannam@89
|
10 Fix return value for in()
|
cannam@89
|
11 Improve allocation failure checking
|
cannam@89
|
12 Add typecasting for void * structures
|
cannam@89
|
13 Add -h option for command version and usage
|
cannam@89
|
14 Add a bunch of comments
|
cannam@89
|
15 1.2 20 Mar 2005 Add Unix compress (LZW) decompression
|
cannam@89
|
16 Copy file attributes from input file to output file
|
cannam@89
|
17 1.3 12 Jun 2005 Add casts for error messages [Oberhumer]
|
cannam@89
|
18 1.4 8 Dec 2006 LZW decompression speed improvements
|
cannam@89
|
19 1.5 9 Feb 2008 Avoid warning in latest version of gcc
|
cannam@89
|
20 1.6 17 Jan 2010 Avoid signed/unsigned comparison warnings
|
cannam@89
|
21 */
|
cannam@89
|
22
|
cannam@89
|
23 /*
|
cannam@89
|
24 gun [ -t ] [ name ... ]
|
cannam@89
|
25
|
cannam@89
|
26 decompresses the data in the named gzip files. If no arguments are given,
|
cannam@89
|
27 gun will decompress from stdin to stdout. The names must end in .gz, -gz,
|
cannam@89
|
28 .z, -z, _z, or .Z. The uncompressed data will be written to a file name
|
cannam@89
|
29 with the suffix stripped. On success, the original file is deleted. On
|
cannam@89
|
30 failure, the output file is deleted. For most failures, the command will
|
cannam@89
|
31 continue to process the remaining names on the command line. A memory
|
cannam@89
|
32 allocation failure will abort the command. If -t is specified, then the
|
cannam@89
|
33 listed files or stdin will be tested as gzip files for integrity (without
|
cannam@89
|
34 checking for a proper suffix), no output will be written, and no files
|
cannam@89
|
35 will be deleted.
|
cannam@89
|
36
|
cannam@89
|
37 Like gzip, gun allows concatenated gzip streams and will decompress them,
|
cannam@89
|
38 writing all of the uncompressed data to the output. Unlike gzip, gun allows
|
cannam@89
|
39 an empty file on input, and will produce no error writing an empty output
|
cannam@89
|
40 file.
|
cannam@89
|
41
|
cannam@89
|
42 gun will also decompress files made by Unix compress, which uses LZW
|
cannam@89
|
43 compression. These files are automatically detected by virtue of their
|
cannam@89
|
44 magic header bytes. Since the end of Unix compress stream is marked by the
|
cannam@89
|
45 end-of-file, they cannot be concantenated. If a Unix compress stream is
|
cannam@89
|
46 encountered in an input file, it is the last stream in that file.
|
cannam@89
|
47
|
cannam@89
|
48 Like gunzip and uncompress, the file attributes of the orignal compressed
|
cannam@89
|
49 file are maintained in the final uncompressed file, to the extent that the
|
cannam@89
|
50 user permissions allow it.
|
cannam@89
|
51
|
cannam@89
|
52 On my Mac OS X PowerPC G4, gun is almost twice as fast as gunzip (version
|
cannam@89
|
53 1.2.4) is on the same file, when gun is linked with zlib 1.2.2. Also the
|
cannam@89
|
54 LZW decompression provided by gun is about twice as fast as the standard
|
cannam@89
|
55 Unix uncompress command.
|
cannam@89
|
56 */
|
cannam@89
|
57
|
cannam@89
|
58 /* external functions and related types and constants */
|
cannam@89
|
59 #include <stdio.h> /* fprintf() */
|
cannam@89
|
60 #include <stdlib.h> /* malloc(), free() */
|
cannam@89
|
61 #include <string.h> /* strerror(), strcmp(), strlen(), memcpy() */
|
cannam@89
|
62 #include <errno.h> /* errno */
|
cannam@89
|
63 #include <fcntl.h> /* open() */
|
cannam@89
|
64 #include <unistd.h> /* read(), write(), close(), chown(), unlink() */
|
cannam@89
|
65 #include <sys/types.h>
|
cannam@89
|
66 #include <sys/stat.h> /* stat(), chmod() */
|
cannam@89
|
67 #include <utime.h> /* utime() */
|
cannam@89
|
68 #include "zlib.h" /* inflateBackInit(), inflateBack(), */
|
cannam@89
|
69 /* inflateBackEnd(), crc32() */
|
cannam@89
|
70
|
cannam@89
|
71 /* function declaration */
|
cannam@89
|
72 #define local static
|
cannam@89
|
73
|
cannam@89
|
74 /* buffer constants */
|
cannam@89
|
75 #define SIZE 32768U /* input and output buffer sizes */
|
cannam@89
|
76 #define PIECE 16384 /* limits i/o chunks for 16-bit int case */
|
cannam@89
|
77
|
cannam@89
|
78 /* structure for infback() to pass to input function in() -- it maintains the
|
cannam@89
|
79 input file and a buffer of size SIZE */
|
cannam@89
|
80 struct ind {
|
cannam@89
|
81 int infile;
|
cannam@89
|
82 unsigned char *inbuf;
|
cannam@89
|
83 };
|
cannam@89
|
84
|
cannam@89
|
85 /* Load input buffer, assumed to be empty, and return bytes loaded and a
|
cannam@89
|
86 pointer to them. read() is called until the buffer is full, or until it
|
cannam@89
|
87 returns end-of-file or error. Return 0 on error. */
|
cannam@89
|
88 local unsigned in(void *in_desc, unsigned char **buf)
|
cannam@89
|
89 {
|
cannam@89
|
90 int ret;
|
cannam@89
|
91 unsigned len;
|
cannam@89
|
92 unsigned char *next;
|
cannam@89
|
93 struct ind *me = (struct ind *)in_desc;
|
cannam@89
|
94
|
cannam@89
|
95 next = me->inbuf;
|
cannam@89
|
96 *buf = next;
|
cannam@89
|
97 len = 0;
|
cannam@89
|
98 do {
|
cannam@89
|
99 ret = PIECE;
|
cannam@89
|
100 if ((unsigned)ret > SIZE - len)
|
cannam@89
|
101 ret = (int)(SIZE - len);
|
cannam@89
|
102 ret = (int)read(me->infile, next, ret);
|
cannam@89
|
103 if (ret == -1) {
|
cannam@89
|
104 len = 0;
|
cannam@89
|
105 break;
|
cannam@89
|
106 }
|
cannam@89
|
107 next += ret;
|
cannam@89
|
108 len += ret;
|
cannam@89
|
109 } while (ret != 0 && len < SIZE);
|
cannam@89
|
110 return len;
|
cannam@89
|
111 }
|
cannam@89
|
112
|
cannam@89
|
113 /* structure for infback() to pass to output function out() -- it maintains the
|
cannam@89
|
114 output file, a running CRC-32 check on the output and the total number of
|
cannam@89
|
115 bytes output, both for checking against the gzip trailer. (The length in
|
cannam@89
|
116 the gzip trailer is stored modulo 2^32, so it's ok if a long is 32 bits and
|
cannam@89
|
117 the output is greater than 4 GB.) */
|
cannam@89
|
118 struct outd {
|
cannam@89
|
119 int outfile;
|
cannam@89
|
120 int check; /* true if checking crc and total */
|
cannam@89
|
121 unsigned long crc;
|
cannam@89
|
122 unsigned long total;
|
cannam@89
|
123 };
|
cannam@89
|
124
|
cannam@89
|
125 /* Write output buffer and update the CRC-32 and total bytes written. write()
|
cannam@89
|
126 is called until all of the output is written or an error is encountered.
|
cannam@89
|
127 On success out() returns 0. For a write failure, out() returns 1. If the
|
cannam@89
|
128 output file descriptor is -1, then nothing is written.
|
cannam@89
|
129 */
|
cannam@89
|
130 local int out(void *out_desc, unsigned char *buf, unsigned len)
|
cannam@89
|
131 {
|
cannam@89
|
132 int ret;
|
cannam@89
|
133 struct outd *me = (struct outd *)out_desc;
|
cannam@89
|
134
|
cannam@89
|
135 if (me->check) {
|
cannam@89
|
136 me->crc = crc32(me->crc, buf, len);
|
cannam@89
|
137 me->total += len;
|
cannam@89
|
138 }
|
cannam@89
|
139 if (me->outfile != -1)
|
cannam@89
|
140 do {
|
cannam@89
|
141 ret = PIECE;
|
cannam@89
|
142 if ((unsigned)ret > len)
|
cannam@89
|
143 ret = (int)len;
|
cannam@89
|
144 ret = (int)write(me->outfile, buf, ret);
|
cannam@89
|
145 if (ret == -1)
|
cannam@89
|
146 return 1;
|
cannam@89
|
147 buf += ret;
|
cannam@89
|
148 len -= ret;
|
cannam@89
|
149 } while (len != 0);
|
cannam@89
|
150 return 0;
|
cannam@89
|
151 }
|
cannam@89
|
152
|
cannam@89
|
153 /* next input byte macro for use inside lunpipe() and gunpipe() */
|
cannam@89
|
154 #define NEXT() (have ? 0 : (have = in(indp, &next)), \
|
cannam@89
|
155 last = have ? (have--, (int)(*next++)) : -1)
|
cannam@89
|
156
|
cannam@89
|
157 /* memory for gunpipe() and lunpipe() --
|
cannam@89
|
158 the first 256 entries of prefix[] and suffix[] are never used, could
|
cannam@89
|
159 have offset the index, but it's faster to waste the memory */
|
cannam@89
|
160 unsigned char inbuf[SIZE]; /* input buffer */
|
cannam@89
|
161 unsigned char outbuf[SIZE]; /* output buffer */
|
cannam@89
|
162 unsigned short prefix[65536]; /* index to LZW prefix string */
|
cannam@89
|
163 unsigned char suffix[65536]; /* one-character LZW suffix */
|
cannam@89
|
164 unsigned char match[65280 + 2]; /* buffer for reversed match or gzip
|
cannam@89
|
165 32K sliding window */
|
cannam@89
|
166
|
cannam@89
|
167 /* throw out what's left in the current bits byte buffer (this is a vestigial
|
cannam@89
|
168 aspect of the compressed data format derived from an implementation that
|
cannam@89
|
169 made use of a special VAX machine instruction!) */
|
cannam@89
|
170 #define FLUSHCODE() \
|
cannam@89
|
171 do { \
|
cannam@89
|
172 left = 0; \
|
cannam@89
|
173 rem = 0; \
|
cannam@89
|
174 if (chunk > have) { \
|
cannam@89
|
175 chunk -= have; \
|
cannam@89
|
176 have = 0; \
|
cannam@89
|
177 if (NEXT() == -1) \
|
cannam@89
|
178 break; \
|
cannam@89
|
179 chunk--; \
|
cannam@89
|
180 if (chunk > have) { \
|
cannam@89
|
181 chunk = have = 0; \
|
cannam@89
|
182 break; \
|
cannam@89
|
183 } \
|
cannam@89
|
184 } \
|
cannam@89
|
185 have -= chunk; \
|
cannam@89
|
186 next += chunk; \
|
cannam@89
|
187 chunk = 0; \
|
cannam@89
|
188 } while (0)
|
cannam@89
|
189
|
cannam@89
|
190 /* Decompress a compress (LZW) file from indp to outfile. The compress magic
|
cannam@89
|
191 header (two bytes) has already been read and verified. There are have bytes
|
cannam@89
|
192 of buffered input at next. strm is used for passing error information back
|
cannam@89
|
193 to gunpipe().
|
cannam@89
|
194
|
cannam@89
|
195 lunpipe() will return Z_OK on success, Z_BUF_ERROR for an unexpected end of
|
cannam@89
|
196 file, read error, or write error (a write error indicated by strm->next_in
|
cannam@89
|
197 not equal to Z_NULL), or Z_DATA_ERROR for invalid input.
|
cannam@89
|
198 */
|
cannam@89
|
199 local int lunpipe(unsigned have, unsigned char *next, struct ind *indp,
|
cannam@89
|
200 int outfile, z_stream *strm)
|
cannam@89
|
201 {
|
cannam@89
|
202 int last; /* last byte read by NEXT(), or -1 if EOF */
|
cannam@89
|
203 unsigned chunk; /* bytes left in current chunk */
|
cannam@89
|
204 int left; /* bits left in rem */
|
cannam@89
|
205 unsigned rem; /* unused bits from input */
|
cannam@89
|
206 int bits; /* current bits per code */
|
cannam@89
|
207 unsigned code; /* code, table traversal index */
|
cannam@89
|
208 unsigned mask; /* mask for current bits codes */
|
cannam@89
|
209 int max; /* maximum bits per code for this stream */
|
cannam@89
|
210 unsigned flags; /* compress flags, then block compress flag */
|
cannam@89
|
211 unsigned end; /* last valid entry in prefix/suffix tables */
|
cannam@89
|
212 unsigned temp; /* current code */
|
cannam@89
|
213 unsigned prev; /* previous code */
|
cannam@89
|
214 unsigned final; /* last character written for previous code */
|
cannam@89
|
215 unsigned stack; /* next position for reversed string */
|
cannam@89
|
216 unsigned outcnt; /* bytes in output buffer */
|
cannam@89
|
217 struct outd outd; /* output structure */
|
cannam@89
|
218 unsigned char *p;
|
cannam@89
|
219
|
cannam@89
|
220 /* set up output */
|
cannam@89
|
221 outd.outfile = outfile;
|
cannam@89
|
222 outd.check = 0;
|
cannam@89
|
223
|
cannam@89
|
224 /* process remainder of compress header -- a flags byte */
|
cannam@89
|
225 flags = NEXT();
|
cannam@89
|
226 if (last == -1)
|
cannam@89
|
227 return Z_BUF_ERROR;
|
cannam@89
|
228 if (flags & 0x60) {
|
cannam@89
|
229 strm->msg = (char *)"unknown lzw flags set";
|
cannam@89
|
230 return Z_DATA_ERROR;
|
cannam@89
|
231 }
|
cannam@89
|
232 max = flags & 0x1f;
|
cannam@89
|
233 if (max < 9 || max > 16) {
|
cannam@89
|
234 strm->msg = (char *)"lzw bits out of range";
|
cannam@89
|
235 return Z_DATA_ERROR;
|
cannam@89
|
236 }
|
cannam@89
|
237 if (max == 9) /* 9 doesn't really mean 9 */
|
cannam@89
|
238 max = 10;
|
cannam@89
|
239 flags &= 0x80; /* true if block compress */
|
cannam@89
|
240
|
cannam@89
|
241 /* clear table */
|
cannam@89
|
242 bits = 9;
|
cannam@89
|
243 mask = 0x1ff;
|
cannam@89
|
244 end = flags ? 256 : 255;
|
cannam@89
|
245
|
cannam@89
|
246 /* set up: get first 9-bit code, which is the first decompressed byte, but
|
cannam@89
|
247 don't create a table entry until the next code */
|
cannam@89
|
248 if (NEXT() == -1) /* no compressed data is ok */
|
cannam@89
|
249 return Z_OK;
|
cannam@89
|
250 final = prev = (unsigned)last; /* low 8 bits of code */
|
cannam@89
|
251 if (NEXT() == -1) /* missing a bit */
|
cannam@89
|
252 return Z_BUF_ERROR;
|
cannam@89
|
253 if (last & 1) { /* code must be < 256 */
|
cannam@89
|
254 strm->msg = (char *)"invalid lzw code";
|
cannam@89
|
255 return Z_DATA_ERROR;
|
cannam@89
|
256 }
|
cannam@89
|
257 rem = (unsigned)last >> 1; /* remaining 7 bits */
|
cannam@89
|
258 left = 7;
|
cannam@89
|
259 chunk = bits - 2; /* 7 bytes left in this chunk */
|
cannam@89
|
260 outbuf[0] = (unsigned char)final; /* write first decompressed byte */
|
cannam@89
|
261 outcnt = 1;
|
cannam@89
|
262
|
cannam@89
|
263 /* decode codes */
|
cannam@89
|
264 stack = 0;
|
cannam@89
|
265 for (;;) {
|
cannam@89
|
266 /* if the table will be full after this, increment the code size */
|
cannam@89
|
267 if (end >= mask && bits < max) {
|
cannam@89
|
268 FLUSHCODE();
|
cannam@89
|
269 bits++;
|
cannam@89
|
270 mask <<= 1;
|
cannam@89
|
271 mask++;
|
cannam@89
|
272 }
|
cannam@89
|
273
|
cannam@89
|
274 /* get a code of length bits */
|
cannam@89
|
275 if (chunk == 0) /* decrement chunk modulo bits */
|
cannam@89
|
276 chunk = bits;
|
cannam@89
|
277 code = rem; /* low bits of code */
|
cannam@89
|
278 if (NEXT() == -1) { /* EOF is end of compressed data */
|
cannam@89
|
279 /* write remaining buffered output */
|
cannam@89
|
280 if (outcnt && out(&outd, outbuf, outcnt)) {
|
cannam@89
|
281 strm->next_in = outbuf; /* signal write error */
|
cannam@89
|
282 return Z_BUF_ERROR;
|
cannam@89
|
283 }
|
cannam@89
|
284 return Z_OK;
|
cannam@89
|
285 }
|
cannam@89
|
286 code += (unsigned)last << left; /* middle (or high) bits of code */
|
cannam@89
|
287 left += 8;
|
cannam@89
|
288 chunk--;
|
cannam@89
|
289 if (bits > left) { /* need more bits */
|
cannam@89
|
290 if (NEXT() == -1) /* can't end in middle of code */
|
cannam@89
|
291 return Z_BUF_ERROR;
|
cannam@89
|
292 code += (unsigned)last << left; /* high bits of code */
|
cannam@89
|
293 left += 8;
|
cannam@89
|
294 chunk--;
|
cannam@89
|
295 }
|
cannam@89
|
296 code &= mask; /* mask to current code length */
|
cannam@89
|
297 left -= bits; /* number of unused bits */
|
cannam@89
|
298 rem = (unsigned)last >> (8 - left); /* unused bits from last byte */
|
cannam@89
|
299
|
cannam@89
|
300 /* process clear code (256) */
|
cannam@89
|
301 if (code == 256 && flags) {
|
cannam@89
|
302 FLUSHCODE();
|
cannam@89
|
303 bits = 9; /* initialize bits and mask */
|
cannam@89
|
304 mask = 0x1ff;
|
cannam@89
|
305 end = 255; /* empty table */
|
cannam@89
|
306 continue; /* get next code */
|
cannam@89
|
307 }
|
cannam@89
|
308
|
cannam@89
|
309 /* special code to reuse last match */
|
cannam@89
|
310 temp = code; /* save the current code */
|
cannam@89
|
311 if (code > end) {
|
cannam@89
|
312 /* Be picky on the allowed code here, and make sure that the code
|
cannam@89
|
313 we drop through (prev) will be a valid index so that random
|
cannam@89
|
314 input does not cause an exception. The code != end + 1 check is
|
cannam@89
|
315 empirically derived, and not checked in the original uncompress
|
cannam@89
|
316 code. If this ever causes a problem, that check could be safely
|
cannam@89
|
317 removed. Leaving this check in greatly improves gun's ability
|
cannam@89
|
318 to detect random or corrupted input after a compress header.
|
cannam@89
|
319 In any case, the prev > end check must be retained. */
|
cannam@89
|
320 if (code != end + 1 || prev > end) {
|
cannam@89
|
321 strm->msg = (char *)"invalid lzw code";
|
cannam@89
|
322 return Z_DATA_ERROR;
|
cannam@89
|
323 }
|
cannam@89
|
324 match[stack++] = (unsigned char)final;
|
cannam@89
|
325 code = prev;
|
cannam@89
|
326 }
|
cannam@89
|
327
|
cannam@89
|
328 /* walk through linked list to generate output in reverse order */
|
cannam@89
|
329 p = match + stack;
|
cannam@89
|
330 while (code >= 256) {
|
cannam@89
|
331 *p++ = suffix[code];
|
cannam@89
|
332 code = prefix[code];
|
cannam@89
|
333 }
|
cannam@89
|
334 stack = p - match;
|
cannam@89
|
335 match[stack++] = (unsigned char)code;
|
cannam@89
|
336 final = code;
|
cannam@89
|
337
|
cannam@89
|
338 /* link new table entry */
|
cannam@89
|
339 if (end < mask) {
|
cannam@89
|
340 end++;
|
cannam@89
|
341 prefix[end] = (unsigned short)prev;
|
cannam@89
|
342 suffix[end] = (unsigned char)final;
|
cannam@89
|
343 }
|
cannam@89
|
344
|
cannam@89
|
345 /* set previous code for next iteration */
|
cannam@89
|
346 prev = temp;
|
cannam@89
|
347
|
cannam@89
|
348 /* write output in forward order */
|
cannam@89
|
349 while (stack > SIZE - outcnt) {
|
cannam@89
|
350 while (outcnt < SIZE)
|
cannam@89
|
351 outbuf[outcnt++] = match[--stack];
|
cannam@89
|
352 if (out(&outd, outbuf, outcnt)) {
|
cannam@89
|
353 strm->next_in = outbuf; /* signal write error */
|
cannam@89
|
354 return Z_BUF_ERROR;
|
cannam@89
|
355 }
|
cannam@89
|
356 outcnt = 0;
|
cannam@89
|
357 }
|
cannam@89
|
358 p = match + stack;
|
cannam@89
|
359 do {
|
cannam@89
|
360 outbuf[outcnt++] = *--p;
|
cannam@89
|
361 } while (p > match);
|
cannam@89
|
362 stack = 0;
|
cannam@89
|
363
|
cannam@89
|
364 /* loop for next code with final and prev as the last match, rem and
|
cannam@89
|
365 left provide the first 0..7 bits of the next code, end is the last
|
cannam@89
|
366 valid table entry */
|
cannam@89
|
367 }
|
cannam@89
|
368 }
|
cannam@89
|
369
|
cannam@89
|
370 /* Decompress a gzip file from infile to outfile. strm is assumed to have been
|
cannam@89
|
371 successfully initialized with inflateBackInit(). The input file may consist
|
cannam@89
|
372 of a series of gzip streams, in which case all of them will be decompressed
|
cannam@89
|
373 to the output file. If outfile is -1, then the gzip stream(s) integrity is
|
cannam@89
|
374 checked and nothing is written.
|
cannam@89
|
375
|
cannam@89
|
376 The return value is a zlib error code: Z_MEM_ERROR if out of memory,
|
cannam@89
|
377 Z_DATA_ERROR if the header or the compressed data is invalid, or if the
|
cannam@89
|
378 trailer CRC-32 check or length doesn't match, Z_BUF_ERROR if the input ends
|
cannam@89
|
379 prematurely or a write error occurs, or Z_ERRNO if junk (not a another gzip
|
cannam@89
|
380 stream) follows a valid gzip stream.
|
cannam@89
|
381 */
|
cannam@89
|
382 local int gunpipe(z_stream *strm, int infile, int outfile)
|
cannam@89
|
383 {
|
cannam@89
|
384 int ret, first, last;
|
cannam@89
|
385 unsigned have, flags, len;
|
cannam@89
|
386 unsigned char *next = NULL;
|
cannam@89
|
387 struct ind ind, *indp;
|
cannam@89
|
388 struct outd outd;
|
cannam@89
|
389
|
cannam@89
|
390 /* setup input buffer */
|
cannam@89
|
391 ind.infile = infile;
|
cannam@89
|
392 ind.inbuf = inbuf;
|
cannam@89
|
393 indp = &ind;
|
cannam@89
|
394
|
cannam@89
|
395 /* decompress concatenated gzip streams */
|
cannam@89
|
396 have = 0; /* no input data read in yet */
|
cannam@89
|
397 first = 1; /* looking for first gzip header */
|
cannam@89
|
398 strm->next_in = Z_NULL; /* so Z_BUF_ERROR means EOF */
|
cannam@89
|
399 for (;;) {
|
cannam@89
|
400 /* look for the two magic header bytes for a gzip stream */
|
cannam@89
|
401 if (NEXT() == -1) {
|
cannam@89
|
402 ret = Z_OK;
|
cannam@89
|
403 break; /* empty gzip stream is ok */
|
cannam@89
|
404 }
|
cannam@89
|
405 if (last != 31 || (NEXT() != 139 && last != 157)) {
|
cannam@89
|
406 strm->msg = (char *)"incorrect header check";
|
cannam@89
|
407 ret = first ? Z_DATA_ERROR : Z_ERRNO;
|
cannam@89
|
408 break; /* not a gzip or compress header */
|
cannam@89
|
409 }
|
cannam@89
|
410 first = 0; /* next non-header is junk */
|
cannam@89
|
411
|
cannam@89
|
412 /* process a compress (LZW) file -- can't be concatenated after this */
|
cannam@89
|
413 if (last == 157) {
|
cannam@89
|
414 ret = lunpipe(have, next, indp, outfile, strm);
|
cannam@89
|
415 break;
|
cannam@89
|
416 }
|
cannam@89
|
417
|
cannam@89
|
418 /* process remainder of gzip header */
|
cannam@89
|
419 ret = Z_BUF_ERROR;
|
cannam@89
|
420 if (NEXT() != 8) { /* only deflate method allowed */
|
cannam@89
|
421 if (last == -1) break;
|
cannam@89
|
422 strm->msg = (char *)"unknown compression method";
|
cannam@89
|
423 ret = Z_DATA_ERROR;
|
cannam@89
|
424 break;
|
cannam@89
|
425 }
|
cannam@89
|
426 flags = NEXT(); /* header flags */
|
cannam@89
|
427 NEXT(); /* discard mod time, xflgs, os */
|
cannam@89
|
428 NEXT();
|
cannam@89
|
429 NEXT();
|
cannam@89
|
430 NEXT();
|
cannam@89
|
431 NEXT();
|
cannam@89
|
432 NEXT();
|
cannam@89
|
433 if (last == -1) break;
|
cannam@89
|
434 if (flags & 0xe0) {
|
cannam@89
|
435 strm->msg = (char *)"unknown header flags set";
|
cannam@89
|
436 ret = Z_DATA_ERROR;
|
cannam@89
|
437 break;
|
cannam@89
|
438 }
|
cannam@89
|
439 if (flags & 4) { /* extra field */
|
cannam@89
|
440 len = NEXT();
|
cannam@89
|
441 len += (unsigned)(NEXT()) << 8;
|
cannam@89
|
442 if (last == -1) break;
|
cannam@89
|
443 while (len > have) {
|
cannam@89
|
444 len -= have;
|
cannam@89
|
445 have = 0;
|
cannam@89
|
446 if (NEXT() == -1) break;
|
cannam@89
|
447 len--;
|
cannam@89
|
448 }
|
cannam@89
|
449 if (last == -1) break;
|
cannam@89
|
450 have -= len;
|
cannam@89
|
451 next += len;
|
cannam@89
|
452 }
|
cannam@89
|
453 if (flags & 8) /* file name */
|
cannam@89
|
454 while (NEXT() != 0 && last != -1)
|
cannam@89
|
455 ;
|
cannam@89
|
456 if (flags & 16) /* comment */
|
cannam@89
|
457 while (NEXT() != 0 && last != -1)
|
cannam@89
|
458 ;
|
cannam@89
|
459 if (flags & 2) { /* header crc */
|
cannam@89
|
460 NEXT();
|
cannam@89
|
461 NEXT();
|
cannam@89
|
462 }
|
cannam@89
|
463 if (last == -1) break;
|
cannam@89
|
464
|
cannam@89
|
465 /* set up output */
|
cannam@89
|
466 outd.outfile = outfile;
|
cannam@89
|
467 outd.check = 1;
|
cannam@89
|
468 outd.crc = crc32(0L, Z_NULL, 0);
|
cannam@89
|
469 outd.total = 0;
|
cannam@89
|
470
|
cannam@89
|
471 /* decompress data to output */
|
cannam@89
|
472 strm->next_in = next;
|
cannam@89
|
473 strm->avail_in = have;
|
cannam@89
|
474 ret = inflateBack(strm, in, indp, out, &outd);
|
cannam@89
|
475 if (ret != Z_STREAM_END) break;
|
cannam@89
|
476 next = strm->next_in;
|
cannam@89
|
477 have = strm->avail_in;
|
cannam@89
|
478 strm->next_in = Z_NULL; /* so Z_BUF_ERROR means EOF */
|
cannam@89
|
479
|
cannam@89
|
480 /* check trailer */
|
cannam@89
|
481 ret = Z_BUF_ERROR;
|
cannam@89
|
482 if (NEXT() != (int)(outd.crc & 0xff) ||
|
cannam@89
|
483 NEXT() != (int)((outd.crc >> 8) & 0xff) ||
|
cannam@89
|
484 NEXT() != (int)((outd.crc >> 16) & 0xff) ||
|
cannam@89
|
485 NEXT() != (int)((outd.crc >> 24) & 0xff)) {
|
cannam@89
|
486 /* crc error */
|
cannam@89
|
487 if (last != -1) {
|
cannam@89
|
488 strm->msg = (char *)"incorrect data check";
|
cannam@89
|
489 ret = Z_DATA_ERROR;
|
cannam@89
|
490 }
|
cannam@89
|
491 break;
|
cannam@89
|
492 }
|
cannam@89
|
493 if (NEXT() != (int)(outd.total & 0xff) ||
|
cannam@89
|
494 NEXT() != (int)((outd.total >> 8) & 0xff) ||
|
cannam@89
|
495 NEXT() != (int)((outd.total >> 16) & 0xff) ||
|
cannam@89
|
496 NEXT() != (int)((outd.total >> 24) & 0xff)) {
|
cannam@89
|
497 /* length error */
|
cannam@89
|
498 if (last != -1) {
|
cannam@89
|
499 strm->msg = (char *)"incorrect length check";
|
cannam@89
|
500 ret = Z_DATA_ERROR;
|
cannam@89
|
501 }
|
cannam@89
|
502 break;
|
cannam@89
|
503 }
|
cannam@89
|
504
|
cannam@89
|
505 /* go back and look for another gzip stream */
|
cannam@89
|
506 }
|
cannam@89
|
507
|
cannam@89
|
508 /* clean up and return */
|
cannam@89
|
509 return ret;
|
cannam@89
|
510 }
|
cannam@89
|
511
|
cannam@89
|
512 /* Copy file attributes, from -> to, as best we can. This is best effort, so
|
cannam@89
|
513 no errors are reported. The mode bits, including suid, sgid, and the sticky
|
cannam@89
|
514 bit are copied (if allowed), the owner's user id and group id are copied
|
cannam@89
|
515 (again if allowed), and the access and modify times are copied. */
|
cannam@89
|
516 local void copymeta(char *from, char *to)
|
cannam@89
|
517 {
|
cannam@89
|
518 struct stat was;
|
cannam@89
|
519 struct utimbuf when;
|
cannam@89
|
520
|
cannam@89
|
521 /* get all of from's Unix meta data, return if not a regular file */
|
cannam@89
|
522 if (stat(from, &was) != 0 || (was.st_mode & S_IFMT) != S_IFREG)
|
cannam@89
|
523 return;
|
cannam@89
|
524
|
cannam@89
|
525 /* set to's mode bits, ignore errors */
|
cannam@89
|
526 (void)chmod(to, was.st_mode & 07777);
|
cannam@89
|
527
|
cannam@89
|
528 /* copy owner's user and group, ignore errors */
|
cannam@89
|
529 (void)chown(to, was.st_uid, was.st_gid);
|
cannam@89
|
530
|
cannam@89
|
531 /* copy access and modify times, ignore errors */
|
cannam@89
|
532 when.actime = was.st_atime;
|
cannam@89
|
533 when.modtime = was.st_mtime;
|
cannam@89
|
534 (void)utime(to, &when);
|
cannam@89
|
535 }
|
cannam@89
|
536
|
cannam@89
|
537 /* Decompress the file inname to the file outnname, of if test is true, just
|
cannam@89
|
538 decompress without writing and check the gzip trailer for integrity. If
|
cannam@89
|
539 inname is NULL or an empty string, read from stdin. If outname is NULL or
|
cannam@89
|
540 an empty string, write to stdout. strm is a pre-initialized inflateBack
|
cannam@89
|
541 structure. When appropriate, copy the file attributes from inname to
|
cannam@89
|
542 outname.
|
cannam@89
|
543
|
cannam@89
|
544 gunzip() returns 1 if there is an out-of-memory error or an unexpected
|
cannam@89
|
545 return code from gunpipe(). Otherwise it returns 0.
|
cannam@89
|
546 */
|
cannam@89
|
547 local int gunzip(z_stream *strm, char *inname, char *outname, int test)
|
cannam@89
|
548 {
|
cannam@89
|
549 int ret;
|
cannam@89
|
550 int infile, outfile;
|
cannam@89
|
551
|
cannam@89
|
552 /* open files */
|
cannam@89
|
553 if (inname == NULL || *inname == 0) {
|
cannam@89
|
554 inname = "-";
|
cannam@89
|
555 infile = 0; /* stdin */
|
cannam@89
|
556 }
|
cannam@89
|
557 else {
|
cannam@89
|
558 infile = open(inname, O_RDONLY, 0);
|
cannam@89
|
559 if (infile == -1) {
|
cannam@89
|
560 fprintf(stderr, "gun cannot open %s\n", inname);
|
cannam@89
|
561 return 0;
|
cannam@89
|
562 }
|
cannam@89
|
563 }
|
cannam@89
|
564 if (test)
|
cannam@89
|
565 outfile = -1;
|
cannam@89
|
566 else if (outname == NULL || *outname == 0) {
|
cannam@89
|
567 outname = "-";
|
cannam@89
|
568 outfile = 1; /* stdout */
|
cannam@89
|
569 }
|
cannam@89
|
570 else {
|
cannam@89
|
571 outfile = open(outname, O_CREAT | O_TRUNC | O_WRONLY, 0666);
|
cannam@89
|
572 if (outfile == -1) {
|
cannam@89
|
573 close(infile);
|
cannam@89
|
574 fprintf(stderr, "gun cannot create %s\n", outname);
|
cannam@89
|
575 return 0;
|
cannam@89
|
576 }
|
cannam@89
|
577 }
|
cannam@89
|
578 errno = 0;
|
cannam@89
|
579
|
cannam@89
|
580 /* decompress */
|
cannam@89
|
581 ret = gunpipe(strm, infile, outfile);
|
cannam@89
|
582 if (outfile > 2) close(outfile);
|
cannam@89
|
583 if (infile > 2) close(infile);
|
cannam@89
|
584
|
cannam@89
|
585 /* interpret result */
|
cannam@89
|
586 switch (ret) {
|
cannam@89
|
587 case Z_OK:
|
cannam@89
|
588 case Z_ERRNO:
|
cannam@89
|
589 if (infile > 2 && outfile > 2) {
|
cannam@89
|
590 copymeta(inname, outname); /* copy attributes */
|
cannam@89
|
591 unlink(inname);
|
cannam@89
|
592 }
|
cannam@89
|
593 if (ret == Z_ERRNO)
|
cannam@89
|
594 fprintf(stderr, "gun warning: trailing garbage ignored in %s\n",
|
cannam@89
|
595 inname);
|
cannam@89
|
596 break;
|
cannam@89
|
597 case Z_DATA_ERROR:
|
cannam@89
|
598 if (outfile > 2) unlink(outname);
|
cannam@89
|
599 fprintf(stderr, "gun data error on %s: %s\n", inname, strm->msg);
|
cannam@89
|
600 break;
|
cannam@89
|
601 case Z_MEM_ERROR:
|
cannam@89
|
602 if (outfile > 2) unlink(outname);
|
cannam@89
|
603 fprintf(stderr, "gun out of memory error--aborting\n");
|
cannam@89
|
604 return 1;
|
cannam@89
|
605 case Z_BUF_ERROR:
|
cannam@89
|
606 if (outfile > 2) unlink(outname);
|
cannam@89
|
607 if (strm->next_in != Z_NULL) {
|
cannam@89
|
608 fprintf(stderr, "gun write error on %s: %s\n",
|
cannam@89
|
609 outname, strerror(errno));
|
cannam@89
|
610 }
|
cannam@89
|
611 else if (errno) {
|
cannam@89
|
612 fprintf(stderr, "gun read error on %s: %s\n",
|
cannam@89
|
613 inname, strerror(errno));
|
cannam@89
|
614 }
|
cannam@89
|
615 else {
|
cannam@89
|
616 fprintf(stderr, "gun unexpected end of file on %s\n",
|
cannam@89
|
617 inname);
|
cannam@89
|
618 }
|
cannam@89
|
619 break;
|
cannam@89
|
620 default:
|
cannam@89
|
621 if (outfile > 2) unlink(outname);
|
cannam@89
|
622 fprintf(stderr, "gun internal error--aborting\n");
|
cannam@89
|
623 return 1;
|
cannam@89
|
624 }
|
cannam@89
|
625 return 0;
|
cannam@89
|
626 }
|
cannam@89
|
627
|
cannam@89
|
628 /* Process the gun command line arguments. See the command syntax near the
|
cannam@89
|
629 beginning of this source file. */
|
cannam@89
|
630 int main(int argc, char **argv)
|
cannam@89
|
631 {
|
cannam@89
|
632 int ret, len, test;
|
cannam@89
|
633 char *outname;
|
cannam@89
|
634 unsigned char *window;
|
cannam@89
|
635 z_stream strm;
|
cannam@89
|
636
|
cannam@89
|
637 /* initialize inflateBack state for repeated use */
|
cannam@89
|
638 window = match; /* reuse LZW match buffer */
|
cannam@89
|
639 strm.zalloc = Z_NULL;
|
cannam@89
|
640 strm.zfree = Z_NULL;
|
cannam@89
|
641 strm.opaque = Z_NULL;
|
cannam@89
|
642 ret = inflateBackInit(&strm, 15, window);
|
cannam@89
|
643 if (ret != Z_OK) {
|
cannam@89
|
644 fprintf(stderr, "gun out of memory error--aborting\n");
|
cannam@89
|
645 return 1;
|
cannam@89
|
646 }
|
cannam@89
|
647
|
cannam@89
|
648 /* decompress each file to the same name with the suffix removed */
|
cannam@89
|
649 argc--;
|
cannam@89
|
650 argv++;
|
cannam@89
|
651 test = 0;
|
cannam@89
|
652 if (argc && strcmp(*argv, "-h") == 0) {
|
cannam@89
|
653 fprintf(stderr, "gun 1.6 (17 Jan 2010)\n");
|
cannam@89
|
654 fprintf(stderr, "Copyright (C) 2003-2010 Mark Adler\n");
|
cannam@89
|
655 fprintf(stderr, "usage: gun [-t] [file1.gz [file2.Z ...]]\n");
|
cannam@89
|
656 return 0;
|
cannam@89
|
657 }
|
cannam@89
|
658 if (argc && strcmp(*argv, "-t") == 0) {
|
cannam@89
|
659 test = 1;
|
cannam@89
|
660 argc--;
|
cannam@89
|
661 argv++;
|
cannam@89
|
662 }
|
cannam@89
|
663 if (argc)
|
cannam@89
|
664 do {
|
cannam@89
|
665 if (test)
|
cannam@89
|
666 outname = NULL;
|
cannam@89
|
667 else {
|
cannam@89
|
668 len = (int)strlen(*argv);
|
cannam@89
|
669 if (strcmp(*argv + len - 3, ".gz") == 0 ||
|
cannam@89
|
670 strcmp(*argv + len - 3, "-gz") == 0)
|
cannam@89
|
671 len -= 3;
|
cannam@89
|
672 else if (strcmp(*argv + len - 2, ".z") == 0 ||
|
cannam@89
|
673 strcmp(*argv + len - 2, "-z") == 0 ||
|
cannam@89
|
674 strcmp(*argv + len - 2, "_z") == 0 ||
|
cannam@89
|
675 strcmp(*argv + len - 2, ".Z") == 0)
|
cannam@89
|
676 len -= 2;
|
cannam@89
|
677 else {
|
cannam@89
|
678 fprintf(stderr, "gun error: no gz type on %s--skipping\n",
|
cannam@89
|
679 *argv);
|
cannam@89
|
680 continue;
|
cannam@89
|
681 }
|
cannam@89
|
682 outname = malloc(len + 1);
|
cannam@89
|
683 if (outname == NULL) {
|
cannam@89
|
684 fprintf(stderr, "gun out of memory error--aborting\n");
|
cannam@89
|
685 ret = 1;
|
cannam@89
|
686 break;
|
cannam@89
|
687 }
|
cannam@89
|
688 memcpy(outname, *argv, len);
|
cannam@89
|
689 outname[len] = 0;
|
cannam@89
|
690 }
|
cannam@89
|
691 ret = gunzip(&strm, *argv, outname, test);
|
cannam@89
|
692 if (outname != NULL) free(outname);
|
cannam@89
|
693 if (ret) break;
|
cannam@89
|
694 } while (argv++, --argc);
|
cannam@89
|
695 else
|
cannam@89
|
696 ret = gunzip(&strm, NULL, NULL, test);
|
cannam@89
|
697
|
cannam@89
|
698 /* clean up */
|
cannam@89
|
699 inflateBackEnd(&strm);
|
cannam@89
|
700 return ret;
|
cannam@89
|
701 }
|