Martin v. Löwis | d372aa8 | 2006-01-03 06:44:59 +0000 | [diff] [blame^] | 1 | /* gun.c -- simple gunzip to give an example of the use of inflateBack() |
| 2 | * Copyright (C) 2003, 2005 Mark Adler |
| 3 | * For conditions of distribution and use, see copyright notice in zlib.h |
| 4 | Version 1.3 12 June 2005 Mark Adler */ |
| 5 | |
| 6 | /* Version history: |
| 7 | 1.0 16 Feb 2003 First version for testing of inflateBack() |
| 8 | 1.1 21 Feb 2005 Decompress concatenated gzip streams |
| 9 | Remove use of "this" variable (C++ keyword) |
| 10 | Fix return value for in() |
| 11 | Improve allocation failure checking |
| 12 | Add typecasting for void * structures |
| 13 | Add -h option for command version and usage |
| 14 | Add a bunch of comments |
| 15 | 1.2 20 Mar 2005 Add Unix compress (LZW) decompression |
| 16 | Copy file attributes from input file to output file |
| 17 | 1.3 12 Jun 2005 Add casts for error messages [Oberhumer] |
| 18 | */ |
| 19 | |
| 20 | /* |
| 21 | gun [ -t ] [ name ... ] |
| 22 | |
| 23 | decompresses the data in the named gzip files. If no arguments are given, |
| 24 | gun will decompress from stdin to stdout. The names must end in .gz, -gz, |
| 25 | .z, -z, _z, or .Z. The uncompressed data will be written to a file name |
| 26 | with the suffix stripped. On success, the original file is deleted. On |
| 27 | failure, the output file is deleted. For most failures, the command will |
| 28 | continue to process the remaining names on the command line. A memory |
| 29 | allocation failure will abort the command. If -t is specified, then the |
| 30 | listed files or stdin will be tested as gzip files for integrity (without |
| 31 | checking for a proper suffix), no output will be written, and no files |
| 32 | will be deleted. |
| 33 | |
| 34 | Like gzip, gun allows concatenated gzip streams and will decompress them, |
| 35 | writing all of the uncompressed data to the output. Unlike gzip, gun allows |
| 36 | an empty file on input, and will produce no error writing an empty output |
| 37 | file. |
| 38 | |
| 39 | gun will also decompress files made by Unix compress, which uses LZW |
| 40 | compression. These files are automatically detected by virtue of their |
| 41 | magic header bytes. Since the end of Unix compress stream is marked by the |
| 42 | end-of-file, they cannot be concantenated. If a Unix compress stream is |
| 43 | encountered in an input file, it is the last stream in that file. |
| 44 | |
| 45 | Like gunzip and uncompress, the file attributes of the orignal compressed |
| 46 | file are maintained in the final uncompressed file, to the extent that the |
| 47 | user permissions allow it. |
| 48 | |
| 49 | On my Mac OS X PowerPC G4, gun is almost twice as fast as gunzip (version |
| 50 | 1.2.4) is on the same file, when gun is linked with zlib 1.2.2. Also the |
| 51 | LZW decompression provided by gun is about twice as fast as the standard |
| 52 | Unix uncompress command. |
| 53 | */ |
| 54 | |
| 55 | /* external functions and related types and constants */ |
| 56 | #include <stdio.h> /* fprintf() */ |
| 57 | #include <stdlib.h> /* malloc(), free() */ |
| 58 | #include <string.h> /* strerror(), strcmp(), strlen(), memcpy() */ |
| 59 | #include <errno.h> /* errno */ |
| 60 | #include <fcntl.h> /* open() */ |
| 61 | #include <unistd.h> /* read(), write(), close(), chown(), unlink() */ |
| 62 | #include <sys/types.h> |
| 63 | #include <sys/stat.h> /* stat(), chmod() */ |
| 64 | #include <utime.h> /* utime() */ |
| 65 | #include "zlib.h" /* inflateBackInit(), inflateBack(), */ |
| 66 | /* inflateBackEnd(), crc32() */ |
| 67 | |
| 68 | /* function declaration */ |
| 69 | #define local static |
| 70 | |
| 71 | /* buffer constants */ |
| 72 | #define SIZE 32768U /* input and output buffer sizes */ |
| 73 | #define PIECE 16384 /* limits i/o chunks for 16-bit int case */ |
| 74 | |
| 75 | /* structure for infback() to pass to input function in() -- it maintains the |
| 76 | input file and a buffer of size SIZE */ |
| 77 | struct ind { |
| 78 | int infile; |
| 79 | unsigned char *inbuf; |
| 80 | }; |
| 81 | |
| 82 | /* Load input buffer, assumed to be empty, and return bytes loaded and a |
| 83 | pointer to them. read() is called until the buffer is full, or until it |
| 84 | returns end-of-file or error. Return 0 on error. */ |
| 85 | local unsigned in(void *in_desc, unsigned char **buf) |
| 86 | { |
| 87 | int ret; |
| 88 | unsigned len; |
| 89 | unsigned char *next; |
| 90 | struct ind *me = (struct ind *)in_desc; |
| 91 | |
| 92 | next = me->inbuf; |
| 93 | *buf = next; |
| 94 | len = 0; |
| 95 | do { |
| 96 | ret = PIECE; |
| 97 | if ((unsigned)ret > SIZE - len) |
| 98 | ret = (int)(SIZE - len); |
| 99 | ret = (int)read(me->infile, next, ret); |
| 100 | if (ret == -1) { |
| 101 | len = 0; |
| 102 | break; |
| 103 | } |
| 104 | next += ret; |
| 105 | len += ret; |
| 106 | } while (ret != 0 && len < SIZE); |
| 107 | return len; |
| 108 | } |
| 109 | |
| 110 | /* structure for infback() to pass to output function out() -- it maintains the |
| 111 | output file, a running CRC-32 check on the output and the total number of |
| 112 | bytes output, both for checking against the gzip trailer. (The length in |
| 113 | the gzip trailer is stored modulo 2^32, so it's ok if a long is 32 bits and |
| 114 | the output is greater than 4 GB.) */ |
| 115 | struct outd { |
| 116 | int outfile; |
| 117 | int check; /* true if checking crc and total */ |
| 118 | unsigned long crc; |
| 119 | unsigned long total; |
| 120 | }; |
| 121 | |
| 122 | /* Write output buffer and update the CRC-32 and total bytes written. write() |
| 123 | is called until all of the output is written or an error is encountered. |
| 124 | On success out() returns 0. For a write failure, out() returns 1. If the |
| 125 | output file descriptor is -1, then nothing is written. |
| 126 | */ |
| 127 | local int out(void *out_desc, unsigned char *buf, unsigned len) |
| 128 | { |
| 129 | int ret; |
| 130 | struct outd *me = (struct outd *)out_desc; |
| 131 | |
| 132 | if (me->check) { |
| 133 | me->crc = crc32(me->crc, buf, len); |
| 134 | me->total += len; |
| 135 | } |
| 136 | if (me->outfile != -1) |
| 137 | do { |
| 138 | ret = PIECE; |
| 139 | if ((unsigned)ret > len) |
| 140 | ret = (int)len; |
| 141 | ret = (int)write(me->outfile, buf, ret); |
| 142 | if (ret == -1) |
| 143 | return 1; |
| 144 | buf += ret; |
| 145 | len -= ret; |
| 146 | } while (len != 0); |
| 147 | return 0; |
| 148 | } |
| 149 | |
| 150 | /* next input byte macro for use inside lunpipe() and gunpipe() */ |
| 151 | #define NEXT() (have ? 0 : (have = in(indp, &next)), \ |
| 152 | last = have ? (have--, (int)(*next++)) : -1) |
| 153 | |
| 154 | /* memory for gunpipe() and lunpipe() -- |
| 155 | the first 256 entries of prefix[] and suffix[] are never used, could |
| 156 | have offset the index, but it's faster to waste the memory */ |
| 157 | unsigned char inbuf[SIZE]; /* input buffer */ |
| 158 | unsigned char outbuf[SIZE]; /* output buffer */ |
| 159 | unsigned short prefix[65536]; /* index to LZW prefix string */ |
| 160 | unsigned char suffix[65536]; /* one-character LZW suffix */ |
| 161 | unsigned char match[65280 + 2]; /* buffer for reversed match or gzip |
| 162 | 32K sliding window */ |
| 163 | |
| 164 | /* throw out what's left in the current bits byte buffer (this is a vestigial |
| 165 | aspect of the compressed data format derived from an implementation that |
| 166 | made use of a special VAX machine instruction!) */ |
| 167 | #define FLUSHCODE() \ |
| 168 | do { \ |
| 169 | left = 0; \ |
| 170 | rem = 0; \ |
| 171 | if (chunk > have) { \ |
| 172 | chunk -= have; \ |
| 173 | have = 0; \ |
| 174 | if (NEXT() == -1) \ |
| 175 | break; \ |
| 176 | chunk--; \ |
| 177 | if (chunk > have) { \ |
| 178 | chunk = have = 0; \ |
| 179 | break; \ |
| 180 | } \ |
| 181 | } \ |
| 182 | have -= chunk; \ |
| 183 | next += chunk; \ |
| 184 | chunk = 0; \ |
| 185 | } while (0) |
| 186 | |
| 187 | /* Decompress a compress (LZW) file from indp to outfile. The compress magic |
| 188 | header (two bytes) has already been read and verified. There are have bytes |
| 189 | of buffered input at next. strm is used for passing error information back |
| 190 | to gunpipe(). |
| 191 | |
| 192 | lunpipe() will return Z_OK on success, Z_BUF_ERROR for an unexpected end of |
| 193 | file, read error, or write error (a write error indicated by strm->next_in |
| 194 | not equal to Z_NULL), or Z_DATA_ERROR for invalid input. |
| 195 | */ |
| 196 | local int lunpipe(unsigned have, unsigned char *next, struct ind *indp, |
| 197 | int outfile, z_stream *strm) |
| 198 | { |
| 199 | int last; /* last byte read by NEXT(), or -1 if EOF */ |
| 200 | int chunk; /* bytes left in current chunk */ |
| 201 | int left; /* bits left in rem */ |
| 202 | unsigned rem; /* unused bits from input */ |
| 203 | int bits; /* current bits per code */ |
| 204 | unsigned code; /* code, table traversal index */ |
| 205 | unsigned mask; /* mask for current bits codes */ |
| 206 | int max; /* maximum bits per code for this stream */ |
| 207 | int flags; /* compress flags, then block compress flag */ |
| 208 | unsigned end; /* last valid entry in prefix/suffix tables */ |
| 209 | unsigned temp; /* current code */ |
| 210 | unsigned prev; /* previous code */ |
| 211 | unsigned final; /* last character written for previous code */ |
| 212 | unsigned stack; /* next position for reversed string */ |
| 213 | unsigned outcnt; /* bytes in output buffer */ |
| 214 | struct outd outd; /* output structure */ |
| 215 | |
| 216 | /* set up output */ |
| 217 | outd.outfile = outfile; |
| 218 | outd.check = 0; |
| 219 | |
| 220 | /* process remainder of compress header -- a flags byte */ |
| 221 | flags = NEXT(); |
| 222 | if (last == -1) |
| 223 | return Z_BUF_ERROR; |
| 224 | if (flags & 0x60) { |
| 225 | strm->msg = (char *)"unknown lzw flags set"; |
| 226 | return Z_DATA_ERROR; |
| 227 | } |
| 228 | max = flags & 0x1f; |
| 229 | if (max < 9 || max > 16) { |
| 230 | strm->msg = (char *)"lzw bits out of range"; |
| 231 | return Z_DATA_ERROR; |
| 232 | } |
| 233 | if (max == 9) /* 9 doesn't really mean 9 */ |
| 234 | max = 10; |
| 235 | flags &= 0x80; /* true if block compress */ |
| 236 | |
| 237 | /* clear table */ |
| 238 | bits = 9; |
| 239 | mask = 0x1ff; |
| 240 | end = flags ? 256 : 255; |
| 241 | |
| 242 | /* set up: get first 9-bit code, which is the first decompressed byte, but |
| 243 | don't create a table entry until the next code */ |
| 244 | if (NEXT() == -1) /* no compressed data is ok */ |
| 245 | return Z_OK; |
| 246 | final = prev = (unsigned)last; /* low 8 bits of code */ |
| 247 | if (NEXT() == -1) /* missing a bit */ |
| 248 | return Z_BUF_ERROR; |
| 249 | if (last & 1) { /* code must be < 256 */ |
| 250 | strm->msg = (char *)"invalid lzw code"; |
| 251 | return Z_DATA_ERROR; |
| 252 | } |
| 253 | rem = (unsigned)last >> 1; /* remaining 7 bits */ |
| 254 | left = 7; |
| 255 | chunk = bits - 2; /* 7 bytes left in this chunk */ |
| 256 | outbuf[0] = (unsigned char)final; /* write first decompressed byte */ |
| 257 | outcnt = 1; |
| 258 | |
| 259 | /* decode codes */ |
| 260 | stack = 0; |
| 261 | for (;;) { |
| 262 | /* if the table will be full after this, increment the code size */ |
| 263 | if (end >= mask && bits < max) { |
| 264 | FLUSHCODE(); |
| 265 | bits++; |
| 266 | mask <<= 1; |
| 267 | mask++; |
| 268 | } |
| 269 | |
| 270 | /* get a code of length bits */ |
| 271 | if (chunk == 0) /* decrement chunk modulo bits */ |
| 272 | chunk = bits; |
| 273 | code = rem; /* low bits of code */ |
| 274 | if (NEXT() == -1) { /* EOF is end of compressed data */ |
| 275 | /* write remaining buffered output */ |
| 276 | if (outcnt && out(&outd, outbuf, outcnt)) { |
| 277 | strm->next_in = outbuf; /* signal write error */ |
| 278 | return Z_BUF_ERROR; |
| 279 | } |
| 280 | return Z_OK; |
| 281 | } |
| 282 | code += (unsigned)last << left; /* middle (or high) bits of code */ |
| 283 | left += 8; |
| 284 | chunk--; |
| 285 | if (bits > left) { /* need more bits */ |
| 286 | if (NEXT() == -1) /* can't end in middle of code */ |
| 287 | return Z_BUF_ERROR; |
| 288 | code += (unsigned)last << left; /* high bits of code */ |
| 289 | left += 8; |
| 290 | chunk--; |
| 291 | } |
| 292 | code &= mask; /* mask to current code length */ |
| 293 | left -= bits; /* number of unused bits */ |
| 294 | rem = (unsigned)last >> (8 - left); /* unused bits from last byte */ |
| 295 | |
| 296 | /* process clear code (256) */ |
| 297 | if (code == 256 && flags) { |
| 298 | FLUSHCODE(); |
| 299 | bits = 9; /* initialize bits and mask */ |
| 300 | mask = 0x1ff; |
| 301 | end = 255; /* empty table */ |
| 302 | continue; /* get next code */ |
| 303 | } |
| 304 | |
| 305 | /* special code to reuse last match */ |
| 306 | temp = code; /* save the current code */ |
| 307 | if (code > end) { |
| 308 | /* Be picky on the allowed code here, and make sure that the code |
| 309 | we drop through (prev) will be a valid index so that random |
| 310 | input does not cause an exception. The code != end + 1 check is |
| 311 | empirically derived, and not checked in the original uncompress |
| 312 | code. If this ever causes a problem, that check could be safely |
| 313 | removed. Leaving this check in greatly improves gun's ability |
| 314 | to detect random or corrupted input after a compress header. |
| 315 | In any case, the prev > end check must be retained. */ |
| 316 | if (code != end + 1 || prev > end) { |
| 317 | strm->msg = (char *)"invalid lzw code"; |
| 318 | return Z_DATA_ERROR; |
| 319 | } |
| 320 | match[stack++] = (unsigned char)final; |
| 321 | code = prev; |
| 322 | } |
| 323 | |
| 324 | /* walk through linked list to generate output in reverse order */ |
| 325 | while (code >= 256) { |
| 326 | match[stack++] = suffix[code]; |
| 327 | code = prefix[code]; |
| 328 | } |
| 329 | match[stack++] = (unsigned char)code; |
| 330 | final = code; |
| 331 | |
| 332 | /* link new table entry */ |
| 333 | if (end < mask) { |
| 334 | end++; |
| 335 | prefix[end] = (unsigned short)prev; |
| 336 | suffix[end] = (unsigned char)final; |
| 337 | } |
| 338 | |
| 339 | /* set previous code for next iteration */ |
| 340 | prev = temp; |
| 341 | |
| 342 | /* write output in forward order */ |
| 343 | while (stack > SIZE - outcnt) { |
| 344 | while (outcnt < SIZE) |
| 345 | outbuf[outcnt++] = match[--stack]; |
| 346 | if (out(&outd, outbuf, outcnt)) { |
| 347 | strm->next_in = outbuf; /* signal write error */ |
| 348 | return Z_BUF_ERROR; |
| 349 | } |
| 350 | outcnt = 0; |
| 351 | } |
| 352 | do { |
| 353 | outbuf[outcnt++] = match[--stack]; |
| 354 | } while (stack); |
| 355 | |
| 356 | /* loop for next code with final and prev as the last match, rem and |
| 357 | left provide the first 0..7 bits of the next code, end is the last |
| 358 | valid table entry */ |
| 359 | } |
| 360 | } |
| 361 | |
| 362 | /* Decompress a gzip file from infile to outfile. strm is assumed to have been |
| 363 | successfully initialized with inflateBackInit(). The input file may consist |
| 364 | of a series of gzip streams, in which case all of them will be decompressed |
| 365 | to the output file. If outfile is -1, then the gzip stream(s) integrity is |
| 366 | checked and nothing is written. |
| 367 | |
| 368 | The return value is a zlib error code: Z_MEM_ERROR if out of memory, |
| 369 | Z_DATA_ERROR if the header or the compressed data is invalid, or if the |
| 370 | trailer CRC-32 check or length doesn't match, Z_BUF_ERROR if the input ends |
| 371 | prematurely or a write error occurs, or Z_ERRNO if junk (not a another gzip |
| 372 | stream) follows a valid gzip stream. |
| 373 | */ |
| 374 | local int gunpipe(z_stream *strm, int infile, int outfile) |
| 375 | { |
| 376 | int ret, first, last; |
| 377 | unsigned have, flags, len; |
| 378 | unsigned char *next; |
| 379 | struct ind ind, *indp; |
| 380 | struct outd outd; |
| 381 | |
| 382 | /* setup input buffer */ |
| 383 | ind.infile = infile; |
| 384 | ind.inbuf = inbuf; |
| 385 | indp = &ind; |
| 386 | |
| 387 | /* decompress concatenated gzip streams */ |
| 388 | have = 0; /* no input data read in yet */ |
| 389 | first = 1; /* looking for first gzip header */ |
| 390 | strm->next_in = Z_NULL; /* so Z_BUF_ERROR means EOF */ |
| 391 | for (;;) { |
| 392 | /* look for the two magic header bytes for a gzip stream */ |
| 393 | if (NEXT() == -1) { |
| 394 | ret = Z_OK; |
| 395 | break; /* empty gzip stream is ok */ |
| 396 | } |
| 397 | if (last != 31 || (NEXT() != 139 && last != 157)) { |
| 398 | strm->msg = (char *)"incorrect header check"; |
| 399 | ret = first ? Z_DATA_ERROR : Z_ERRNO; |
| 400 | break; /* not a gzip or compress header */ |
| 401 | } |
| 402 | first = 0; /* next non-header is junk */ |
| 403 | |
| 404 | /* process a compress (LZW) file -- can't be concatenated after this */ |
| 405 | if (last == 157) { |
| 406 | ret = lunpipe(have, next, indp, outfile, strm); |
| 407 | break; |
| 408 | } |
| 409 | |
| 410 | /* process remainder of gzip header */ |
| 411 | ret = Z_BUF_ERROR; |
| 412 | if (NEXT() != 8) { /* only deflate method allowed */ |
| 413 | if (last == -1) break; |
| 414 | strm->msg = (char *)"unknown compression method"; |
| 415 | ret = Z_DATA_ERROR; |
| 416 | break; |
| 417 | } |
| 418 | flags = NEXT(); /* header flags */ |
| 419 | NEXT(); /* discard mod time, xflgs, os */ |
| 420 | NEXT(); |
| 421 | NEXT(); |
| 422 | NEXT(); |
| 423 | NEXT(); |
| 424 | NEXT(); |
| 425 | if (last == -1) break; |
| 426 | if (flags & 0xe0) { |
| 427 | strm->msg = (char *)"unknown header flags set"; |
| 428 | ret = Z_DATA_ERROR; |
| 429 | break; |
| 430 | } |
| 431 | if (flags & 4) { /* extra field */ |
| 432 | len = NEXT(); |
| 433 | len += (unsigned)(NEXT()) << 8; |
| 434 | if (last == -1) break; |
| 435 | while (len > have) { |
| 436 | len -= have; |
| 437 | have = 0; |
| 438 | if (NEXT() == -1) break; |
| 439 | len--; |
| 440 | } |
| 441 | if (last == -1) break; |
| 442 | have -= len; |
| 443 | next += len; |
| 444 | } |
| 445 | if (flags & 8) /* file name */ |
| 446 | while (NEXT() != 0 && last != -1) |
| 447 | ; |
| 448 | if (flags & 16) /* comment */ |
| 449 | while (NEXT() != 0 && last != -1) |
| 450 | ; |
| 451 | if (flags & 2) { /* header crc */ |
| 452 | NEXT(); |
| 453 | NEXT(); |
| 454 | } |
| 455 | if (last == -1) break; |
| 456 | |
| 457 | /* set up output */ |
| 458 | outd.outfile = outfile; |
| 459 | outd.check = 1; |
| 460 | outd.crc = crc32(0L, Z_NULL, 0); |
| 461 | outd.total = 0; |
| 462 | |
| 463 | /* decompress data to output */ |
| 464 | strm->next_in = next; |
| 465 | strm->avail_in = have; |
| 466 | ret = inflateBack(strm, in, indp, out, &outd); |
| 467 | if (ret != Z_STREAM_END) break; |
| 468 | next = strm->next_in; |
| 469 | have = strm->avail_in; |
| 470 | strm->next_in = Z_NULL; /* so Z_BUF_ERROR means EOF */ |
| 471 | |
| 472 | /* check trailer */ |
| 473 | ret = Z_BUF_ERROR; |
| 474 | if (NEXT() != (outd.crc & 0xff) || |
| 475 | NEXT() != ((outd.crc >> 8) & 0xff) || |
| 476 | NEXT() != ((outd.crc >> 16) & 0xff) || |
| 477 | NEXT() != ((outd.crc >> 24) & 0xff)) { |
| 478 | /* crc error */ |
| 479 | if (last != -1) { |
| 480 | strm->msg = (char *)"incorrect data check"; |
| 481 | ret = Z_DATA_ERROR; |
| 482 | } |
| 483 | break; |
| 484 | } |
| 485 | if (NEXT() != (outd.total & 0xff) || |
| 486 | NEXT() != ((outd.total >> 8) & 0xff) || |
| 487 | NEXT() != ((outd.total >> 16) & 0xff) || |
| 488 | NEXT() != ((outd.total >> 24) & 0xff)) { |
| 489 | /* length error */ |
| 490 | if (last != -1) { |
| 491 | strm->msg = (char *)"incorrect length check"; |
| 492 | ret = Z_DATA_ERROR; |
| 493 | } |
| 494 | break; |
| 495 | } |
| 496 | |
| 497 | /* go back and look for another gzip stream */ |
| 498 | } |
| 499 | |
| 500 | /* clean up and return */ |
| 501 | return ret; |
| 502 | } |
| 503 | |
| 504 | /* Copy file attributes, from -> to, as best we can. This is best effort, so |
| 505 | no errors are reported. The mode bits, including suid, sgid, and the sticky |
| 506 | bit are copied (if allowed), the owner's user id and group id are copied |
| 507 | (again if allowed), and the access and modify times are copied. */ |
| 508 | local void copymeta(char *from, char *to) |
| 509 | { |
| 510 | struct stat was; |
| 511 | struct utimbuf when; |
| 512 | |
| 513 | /* get all of from's Unix meta data, return if not a regular file */ |
| 514 | if (stat(from, &was) != 0 || (was.st_mode & S_IFMT) != S_IFREG) |
| 515 | return; |
| 516 | |
| 517 | /* set to's mode bits, ignore errors */ |
| 518 | (void)chmod(to, was.st_mode & 07777); |
| 519 | |
| 520 | /* copy owner's user and group, ignore errors */ |
| 521 | (void)chown(to, was.st_uid, was.st_gid); |
| 522 | |
| 523 | /* copy access and modify times, ignore errors */ |
| 524 | when.actime = was.st_atime; |
| 525 | when.modtime = was.st_mtime; |
| 526 | (void)utime(to, &when); |
| 527 | } |
| 528 | |
| 529 | /* Decompress the file inname to the file outnname, of if test is true, just |
| 530 | decompress without writing and check the gzip trailer for integrity. If |
| 531 | inname is NULL or an empty string, read from stdin. If outname is NULL or |
| 532 | an empty string, write to stdout. strm is a pre-initialized inflateBack |
| 533 | structure. When appropriate, copy the file attributes from inname to |
| 534 | outname. |
| 535 | |
| 536 | gunzip() returns 1 if there is an out-of-memory error or an unexpected |
| 537 | return code from gunpipe(). Otherwise it returns 0. |
| 538 | */ |
| 539 | local int gunzip(z_stream *strm, char *inname, char *outname, int test) |
| 540 | { |
| 541 | int ret; |
| 542 | int infile, outfile; |
| 543 | |
| 544 | /* open files */ |
| 545 | if (inname == NULL || *inname == 0) { |
| 546 | inname = "-"; |
| 547 | infile = 0; /* stdin */ |
| 548 | } |
| 549 | else { |
| 550 | infile = open(inname, O_RDONLY, 0); |
| 551 | if (infile == -1) { |
| 552 | fprintf(stderr, "gun cannot open %s\n", inname); |
| 553 | return 0; |
| 554 | } |
| 555 | } |
| 556 | if (test) |
| 557 | outfile = -1; |
| 558 | else if (outname == NULL || *outname == 0) { |
| 559 | outname = "-"; |
| 560 | outfile = 1; /* stdout */ |
| 561 | } |
| 562 | else { |
| 563 | outfile = open(outname, O_CREAT | O_TRUNC | O_WRONLY, 0666); |
| 564 | if (outfile == -1) { |
| 565 | close(infile); |
| 566 | fprintf(stderr, "gun cannot create %s\n", outname); |
| 567 | return 0; |
| 568 | } |
| 569 | } |
| 570 | errno = 0; |
| 571 | |
| 572 | /* decompress */ |
| 573 | ret = gunpipe(strm, infile, outfile); |
| 574 | if (outfile > 2) close(outfile); |
| 575 | if (infile > 2) close(infile); |
| 576 | |
| 577 | /* interpret result */ |
| 578 | switch (ret) { |
| 579 | case Z_OK: |
| 580 | case Z_ERRNO: |
| 581 | if (infile > 2 && outfile > 2) { |
| 582 | copymeta(inname, outname); /* copy attributes */ |
| 583 | unlink(inname); |
| 584 | } |
| 585 | if (ret == Z_ERRNO) |
| 586 | fprintf(stderr, "gun warning: trailing garbage ignored in %s\n", |
| 587 | inname); |
| 588 | break; |
| 589 | case Z_DATA_ERROR: |
| 590 | if (outfile > 2) unlink(outname); |
| 591 | fprintf(stderr, "gun data error on %s: %s\n", inname, strm->msg); |
| 592 | break; |
| 593 | case Z_MEM_ERROR: |
| 594 | if (outfile > 2) unlink(outname); |
| 595 | fprintf(stderr, "gun out of memory error--aborting\n"); |
| 596 | return 1; |
| 597 | case Z_BUF_ERROR: |
| 598 | if (outfile > 2) unlink(outname); |
| 599 | if (strm->next_in != Z_NULL) { |
| 600 | fprintf(stderr, "gun write error on %s: %s\n", |
| 601 | outname, strerror(errno)); |
| 602 | } |
| 603 | else if (errno) { |
| 604 | fprintf(stderr, "gun read error on %s: %s\n", |
| 605 | inname, strerror(errno)); |
| 606 | } |
| 607 | else { |
| 608 | fprintf(stderr, "gun unexpected end of file on %s\n", |
| 609 | inname); |
| 610 | } |
| 611 | break; |
| 612 | default: |
| 613 | if (outfile > 2) unlink(outname); |
| 614 | fprintf(stderr, "gun internal error--aborting\n"); |
| 615 | return 1; |
| 616 | } |
| 617 | return 0; |
| 618 | } |
| 619 | |
| 620 | /* Process the gun command line arguments. See the command syntax near the |
| 621 | beginning of this source file. */ |
| 622 | int main(int argc, char **argv) |
| 623 | { |
| 624 | int ret, len, test; |
| 625 | char *outname; |
| 626 | unsigned char *window; |
| 627 | z_stream strm; |
| 628 | |
| 629 | /* initialize inflateBack state for repeated use */ |
| 630 | window = match; /* reuse LZW match buffer */ |
| 631 | strm.zalloc = Z_NULL; |
| 632 | strm.zfree = Z_NULL; |
| 633 | strm.opaque = Z_NULL; |
| 634 | ret = inflateBackInit(&strm, 15, window); |
| 635 | if (ret != Z_OK) { |
| 636 | fprintf(stderr, "gun out of memory error--aborting\n"); |
| 637 | return 1; |
| 638 | } |
| 639 | |
| 640 | /* decompress each file to the same name with the suffix removed */ |
| 641 | argc--; |
| 642 | argv++; |
| 643 | test = 0; |
| 644 | if (argc && strcmp(*argv, "-h") == 0) { |
| 645 | fprintf(stderr, "gun 1.3 (12 Jun 2005)\n"); |
| 646 | fprintf(stderr, "Copyright (c) 2005 Mark Adler\n"); |
| 647 | fprintf(stderr, "usage: gun [-t] [file1.gz [file2.Z ...]]\n"); |
| 648 | return 0; |
| 649 | } |
| 650 | if (argc && strcmp(*argv, "-t") == 0) { |
| 651 | test = 1; |
| 652 | argc--; |
| 653 | argv++; |
| 654 | } |
| 655 | if (argc) |
| 656 | do { |
| 657 | if (test) |
| 658 | outname = NULL; |
| 659 | else { |
| 660 | len = (int)strlen(*argv); |
| 661 | if (strcmp(*argv + len - 3, ".gz") == 0 || |
| 662 | strcmp(*argv + len - 3, "-gz") == 0) |
| 663 | len -= 3; |
| 664 | else if (strcmp(*argv + len - 2, ".z") == 0 || |
| 665 | strcmp(*argv + len - 2, "-z") == 0 || |
| 666 | strcmp(*argv + len - 2, "_z") == 0 || |
| 667 | strcmp(*argv + len - 2, ".Z") == 0) |
| 668 | len -= 2; |
| 669 | else { |
| 670 | fprintf(stderr, "gun error: no gz type on %s--skipping\n", |
| 671 | *argv); |
| 672 | continue; |
| 673 | } |
| 674 | outname = malloc(len + 1); |
| 675 | if (outname == NULL) { |
| 676 | fprintf(stderr, "gun out of memory error--aborting\n"); |
| 677 | ret = 1; |
| 678 | break; |
| 679 | } |
| 680 | memcpy(outname, *argv, len); |
| 681 | outname[len] = 0; |
| 682 | } |
| 683 | ret = gunzip(&strm, *argv, outname, test); |
| 684 | if (outname != NULL) free(outname); |
| 685 | if (ret) break; |
| 686 | } while (argv++, --argc); |
| 687 | else |
| 688 | ret = gunzip(&strm, NULL, NULL, test); |
| 689 | |
| 690 | /* clean up */ |
| 691 | inflateBackEnd(&strm); |
| 692 | return ret; |
| 693 | } |