Martin v. Löwis | d372aa8 | 2006-01-03 06:44:59 +0000 | [diff] [blame^] | 1 | /* gzappend -- command to append to a gzip file |
| 2 | |
| 3 | Copyright (C) 2003 Mark Adler, all rights reserved |
| 4 | version 1.1, 4 Nov 2003 |
| 5 | |
| 6 | This software is provided 'as-is', without any express or implied |
| 7 | warranty. In no event will the author be held liable for any damages |
| 8 | arising from the use of this software. |
| 9 | |
| 10 | Permission is granted to anyone to use this software for any purpose, |
| 11 | including commercial applications, and to alter it and redistribute it |
| 12 | freely, subject to the following restrictions: |
| 13 | |
| 14 | 1. The origin of this software must not be misrepresented; you must not |
| 15 | claim that you wrote the original software. If you use this software |
| 16 | in a product, an acknowledgment in the product documentation would be |
| 17 | appreciated but is not required. |
| 18 | 2. Altered source versions must be plainly marked as such, and must not be |
| 19 | misrepresented as being the original software. |
| 20 | 3. This notice may not be removed or altered from any source distribution. |
| 21 | |
| 22 | Mark Adler madler@alumni.caltech.edu |
| 23 | */ |
| 24 | |
| 25 | /* |
| 26 | * Change history: |
| 27 | * |
| 28 | * 1.0 19 Oct 2003 - First version |
| 29 | * 1.1 4 Nov 2003 - Expand and clarify some comments and notes |
| 30 | * - Add version and copyright to help |
| 31 | * - Send help to stdout instead of stderr |
| 32 | * - Add some preemptive typecasts |
| 33 | * - Add L to constants in lseek() calls |
| 34 | * - Remove some debugging information in error messages |
| 35 | * - Use new data_type definition for zlib 1.2.1 |
| 36 | * - Simplfy and unify file operations |
| 37 | * - Finish off gzip file in gztack() |
| 38 | * - Use deflatePrime() instead of adding empty blocks |
| 39 | * - Keep gzip file clean on appended file read errors |
| 40 | * - Use in-place rotate instead of auxiliary buffer |
| 41 | * (Why you ask? Because it was fun to write!) |
| 42 | */ |
| 43 | |
| 44 | /* |
| 45 | gzappend takes a gzip file and appends to it, compressing files from the |
| 46 | command line or data from stdin. The gzip file is written to directly, to |
| 47 | avoid copying that file, in case it's large. Note that this results in the |
| 48 | unfriendly behavior that if gzappend fails, the gzip file is corrupted. |
| 49 | |
| 50 | This program was written to illustrate the use of the new Z_BLOCK option of |
| 51 | zlib 1.2.x's inflate() function. This option returns from inflate() at each |
| 52 | block boundary to facilitate locating and modifying the last block bit at |
| 53 | the start of the final deflate block. Also whether using Z_BLOCK or not, |
| 54 | another required feature of zlib 1.2.x is that inflate() now provides the |
| 55 | number of unusued bits in the last input byte used. gzappend will not work |
| 56 | with versions of zlib earlier than 1.2.1. |
| 57 | |
| 58 | gzappend first decompresses the gzip file internally, discarding all but |
| 59 | the last 32K of uncompressed data, and noting the location of the last block |
| 60 | bit and the number of unused bits in the last byte of the compressed data. |
| 61 | The gzip trailer containing the CRC-32 and length of the uncompressed data |
| 62 | is verified. This trailer will be later overwritten. |
| 63 | |
| 64 | Then the last block bit is cleared by seeking back in the file and rewriting |
| 65 | the byte that contains it. Seeking forward, the last byte of the compressed |
| 66 | data is saved along with the number of unused bits to initialize deflate. |
| 67 | |
| 68 | A deflate process is initialized, using the last 32K of the uncompressed |
| 69 | data from the gzip file to initialize the dictionary. If the total |
| 70 | uncompressed data was less than 32K, then all of it is used to initialize |
| 71 | the dictionary. The deflate output bit buffer is also initialized with the |
| 72 | last bits from the original deflate stream. From here on, the data to |
| 73 | append is simply compressed using deflate, and written to the gzip file. |
| 74 | When that is complete, the new CRC-32 and uncompressed length are written |
| 75 | as the trailer of the gzip file. |
| 76 | */ |
| 77 | |
| 78 | #include <stdio.h> |
| 79 | #include <stdlib.h> |
| 80 | #include <string.h> |
| 81 | #include <fcntl.h> |
| 82 | #include <unistd.h> |
| 83 | #include "zlib.h" |
| 84 | |
| 85 | #define local static |
| 86 | #define LGCHUNK 14 |
| 87 | #define CHUNK (1U << LGCHUNK) |
| 88 | #define DSIZE 32768U |
| 89 | |
| 90 | /* print an error message and terminate with extreme prejudice */ |
| 91 | local void bye(char *msg1, char *msg2) |
| 92 | { |
| 93 | fprintf(stderr, "gzappend error: %s%s\n", msg1, msg2); |
| 94 | exit(1); |
| 95 | } |
| 96 | |
| 97 | /* return the greatest common divisor of a and b using Euclid's algorithm, |
| 98 | modified to be fast when one argument much greater than the other, and |
| 99 | coded to avoid unnecessary swapping */ |
| 100 | local unsigned gcd(unsigned a, unsigned b) |
| 101 | { |
| 102 | unsigned c; |
| 103 | |
| 104 | while (a && b) |
| 105 | if (a > b) { |
| 106 | c = b; |
| 107 | while (a - c >= c) |
| 108 | c <<= 1; |
| 109 | a -= c; |
| 110 | } |
| 111 | else { |
| 112 | c = a; |
| 113 | while (b - c >= c) |
| 114 | c <<= 1; |
| 115 | b -= c; |
| 116 | } |
| 117 | return a + b; |
| 118 | } |
| 119 | |
| 120 | /* rotate list[0..len-1] left by rot positions, in place */ |
| 121 | local void rotate(unsigned char *list, unsigned len, unsigned rot) |
| 122 | { |
| 123 | unsigned char tmp; |
| 124 | unsigned cycles; |
| 125 | unsigned char *start, *last, *to, *from; |
| 126 | |
| 127 | /* normalize rot and handle degenerate cases */ |
| 128 | if (len < 2) return; |
| 129 | if (rot >= len) rot %= len; |
| 130 | if (rot == 0) return; |
| 131 | |
| 132 | /* pointer to last entry in list */ |
| 133 | last = list + (len - 1); |
| 134 | |
| 135 | /* do simple left shift by one */ |
| 136 | if (rot == 1) { |
| 137 | tmp = *list; |
| 138 | memcpy(list, list + 1, len - 1); |
| 139 | *last = tmp; |
| 140 | return; |
| 141 | } |
| 142 | |
| 143 | /* do simple right shift by one */ |
| 144 | if (rot == len - 1) { |
| 145 | tmp = *last; |
| 146 | memmove(list + 1, list, len - 1); |
| 147 | *list = tmp; |
| 148 | return; |
| 149 | } |
| 150 | |
| 151 | /* otherwise do rotate as a set of cycles in place */ |
| 152 | cycles = gcd(len, rot); /* number of cycles */ |
| 153 | do { |
| 154 | start = from = list + cycles; /* start index is arbitrary */ |
| 155 | tmp = *from; /* save entry to be overwritten */ |
| 156 | for (;;) { |
| 157 | to = from; /* next step in cycle */ |
| 158 | from += rot; /* go right rot positions */ |
| 159 | if (from > last) from -= len; /* (pointer better not wrap) */ |
| 160 | if (from == start) break; /* all but one shifted */ |
| 161 | *to = *from; /* shift left */ |
| 162 | } |
| 163 | *to = tmp; /* complete the circle */ |
| 164 | } while (--cycles); |
| 165 | } |
| 166 | |
| 167 | /* structure for gzip file read operations */ |
| 168 | typedef struct { |
| 169 | int fd; /* file descriptor */ |
| 170 | int size; /* 1 << size is bytes in buf */ |
| 171 | unsigned left; /* bytes available at next */ |
| 172 | unsigned char *buf; /* buffer */ |
| 173 | unsigned char *next; /* next byte in buffer */ |
| 174 | char *name; /* file name for error messages */ |
| 175 | } file; |
| 176 | |
| 177 | /* reload buffer */ |
| 178 | local int readin(file *in) |
| 179 | { |
| 180 | int len; |
| 181 | |
| 182 | len = read(in->fd, in->buf, 1 << in->size); |
| 183 | if (len == -1) bye("error reading ", in->name); |
| 184 | in->left = (unsigned)len; |
| 185 | in->next = in->buf; |
| 186 | return len; |
| 187 | } |
| 188 | |
| 189 | /* read from file in, exit if end-of-file */ |
| 190 | local int readmore(file *in) |
| 191 | { |
| 192 | if (readin(in) == 0) bye("unexpected end of ", in->name); |
| 193 | return 0; |
| 194 | } |
| 195 | |
| 196 | #define read1(in) (in->left == 0 ? readmore(in) : 0, \ |
| 197 | in->left--, *(in->next)++) |
| 198 | |
| 199 | /* skip over n bytes of in */ |
| 200 | local void skip(file *in, unsigned n) |
| 201 | { |
| 202 | unsigned bypass; |
| 203 | |
| 204 | if (n > in->left) { |
| 205 | n -= in->left; |
| 206 | bypass = n & ~((1U << in->size) - 1); |
| 207 | if (bypass) { |
| 208 | if (lseek(in->fd, (off_t)bypass, SEEK_CUR) == -1) |
| 209 | bye("seeking ", in->name); |
| 210 | n -= bypass; |
| 211 | } |
| 212 | readmore(in); |
| 213 | if (n > in->left) |
| 214 | bye("unexpected end of ", in->name); |
| 215 | } |
| 216 | in->left -= n; |
| 217 | in->next += n; |
| 218 | } |
| 219 | |
| 220 | /* read a four-byte unsigned integer, little-endian, from in */ |
| 221 | unsigned long read4(file *in) |
| 222 | { |
| 223 | unsigned long val; |
| 224 | |
| 225 | val = read1(in); |
| 226 | val += (unsigned)read1(in) << 8; |
| 227 | val += (unsigned long)read1(in) << 16; |
| 228 | val += (unsigned long)read1(in) << 24; |
| 229 | return val; |
| 230 | } |
| 231 | |
| 232 | /* skip over gzip header */ |
| 233 | local void gzheader(file *in) |
| 234 | { |
| 235 | int flags; |
| 236 | unsigned n; |
| 237 | |
| 238 | if (read1(in) != 31 || read1(in) != 139) bye(in->name, " not a gzip file"); |
| 239 | if (read1(in) != 8) bye("unknown compression method in", in->name); |
| 240 | flags = read1(in); |
| 241 | if (flags & 0xe0) bye("unknown header flags set in", in->name); |
| 242 | skip(in, 6); |
| 243 | if (flags & 4) { |
| 244 | n = read1(in); |
| 245 | n += (unsigned)(read1(in)) << 8; |
| 246 | skip(in, n); |
| 247 | } |
| 248 | if (flags & 8) while (read1(in) != 0) ; |
| 249 | if (flags & 16) while (read1(in) != 0) ; |
| 250 | if (flags & 2) skip(in, 2); |
| 251 | } |
| 252 | |
| 253 | /* decompress gzip file "name", return strm with a deflate stream ready to |
| 254 | continue compression of the data in the gzip file, and return a file |
| 255 | descriptor pointing to where to write the compressed data -- the deflate |
| 256 | stream is initialized to compress using level "level" */ |
| 257 | local int gzscan(char *name, z_stream *strm, int level) |
| 258 | { |
| 259 | int ret, lastbit, left, full; |
| 260 | unsigned have; |
| 261 | unsigned long crc, tot; |
| 262 | unsigned char *window; |
| 263 | off_t lastoff, end; |
| 264 | file gz; |
| 265 | |
| 266 | /* open gzip file */ |
| 267 | gz.name = name; |
| 268 | gz.fd = open(name, O_RDWR, 0); |
| 269 | if (gz.fd == -1) bye("cannot open ", name); |
| 270 | gz.buf = malloc(CHUNK); |
| 271 | if (gz.buf == NULL) bye("out of memory", ""); |
| 272 | gz.size = LGCHUNK; |
| 273 | gz.left = 0; |
| 274 | |
| 275 | /* skip gzip header */ |
| 276 | gzheader(&gz); |
| 277 | |
| 278 | /* prepare to decompress */ |
| 279 | window = malloc(DSIZE); |
| 280 | if (window == NULL) bye("out of memory", ""); |
| 281 | strm->zalloc = Z_NULL; |
| 282 | strm->zfree = Z_NULL; |
| 283 | strm->opaque = Z_NULL; |
| 284 | ret = inflateInit2(strm, -15); |
| 285 | if (ret != Z_OK) bye("out of memory", " or library mismatch"); |
| 286 | |
| 287 | /* decompress the deflate stream, saving append information */ |
| 288 | lastbit = 0; |
| 289 | lastoff = lseek(gz.fd, 0L, SEEK_CUR) - gz.left; |
| 290 | left = 0; |
| 291 | strm->avail_in = gz.left; |
| 292 | strm->next_in = gz.next; |
| 293 | crc = crc32(0L, Z_NULL, 0); |
| 294 | have = full = 0; |
| 295 | do { |
| 296 | /* if needed, get more input */ |
| 297 | if (strm->avail_in == 0) { |
| 298 | readmore(&gz); |
| 299 | strm->avail_in = gz.left; |
| 300 | strm->next_in = gz.next; |
| 301 | } |
| 302 | |
| 303 | /* set up output to next available section of sliding window */ |
| 304 | strm->avail_out = DSIZE - have; |
| 305 | strm->next_out = window + have; |
| 306 | |
| 307 | /* inflate and check for errors */ |
| 308 | ret = inflate(strm, Z_BLOCK); |
| 309 | if (ret == Z_STREAM_ERROR) bye("internal stream error!", ""); |
| 310 | if (ret == Z_MEM_ERROR) bye("out of memory", ""); |
| 311 | if (ret == Z_DATA_ERROR) |
| 312 | bye("invalid compressed data--format violated in", name); |
| 313 | |
| 314 | /* update crc and sliding window pointer */ |
| 315 | crc = crc32(crc, window + have, DSIZE - have - strm->avail_out); |
| 316 | if (strm->avail_out) |
| 317 | have = DSIZE - strm->avail_out; |
| 318 | else { |
| 319 | have = 0; |
| 320 | full = 1; |
| 321 | } |
| 322 | |
| 323 | /* process end of block */ |
| 324 | if (strm->data_type & 128) { |
| 325 | if (strm->data_type & 64) |
| 326 | left = strm->data_type & 0x1f; |
| 327 | else { |
| 328 | lastbit = strm->data_type & 0x1f; |
| 329 | lastoff = lseek(gz.fd, 0L, SEEK_CUR) - strm->avail_in; |
| 330 | } |
| 331 | } |
| 332 | } while (ret != Z_STREAM_END); |
| 333 | inflateEnd(strm); |
| 334 | gz.left = strm->avail_in; |
| 335 | gz.next = strm->next_in; |
| 336 | |
| 337 | /* save the location of the end of the compressed data */ |
| 338 | end = lseek(gz.fd, 0L, SEEK_CUR) - gz.left; |
| 339 | |
| 340 | /* check gzip trailer and save total for deflate */ |
| 341 | if (crc != read4(&gz)) |
| 342 | bye("invalid compressed data--crc mismatch in ", name); |
| 343 | tot = strm->total_out; |
| 344 | if ((tot & 0xffffffffUL) != read4(&gz)) |
| 345 | bye("invalid compressed data--length mismatch in", name); |
| 346 | |
| 347 | /* if not at end of file, warn */ |
| 348 | if (gz.left || readin(&gz)) |
| 349 | fprintf(stderr, |
| 350 | "gzappend warning: junk at end of gzip file overwritten\n"); |
| 351 | |
| 352 | /* clear last block bit */ |
| 353 | lseek(gz.fd, lastoff - (lastbit != 0), SEEK_SET); |
| 354 | if (read(gz.fd, gz.buf, 1) != 1) bye("reading after seek on ", name); |
| 355 | *gz.buf = (unsigned char)(*gz.buf ^ (1 << ((8 - lastbit) & 7))); |
| 356 | lseek(gz.fd, -1L, SEEK_CUR); |
| 357 | if (write(gz.fd, gz.buf, 1) != 1) bye("writing after seek to ", name); |
| 358 | |
| 359 | /* if window wrapped, build dictionary from window by rotating */ |
| 360 | if (full) { |
| 361 | rotate(window, DSIZE, have); |
| 362 | have = DSIZE; |
| 363 | } |
| 364 | |
| 365 | /* set up deflate stream with window, crc, total_in, and leftover bits */ |
| 366 | ret = deflateInit2(strm, level, Z_DEFLATED, -15, 8, Z_DEFAULT_STRATEGY); |
| 367 | if (ret != Z_OK) bye("out of memory", ""); |
| 368 | deflateSetDictionary(strm, window, have); |
| 369 | strm->adler = crc; |
| 370 | strm->total_in = tot; |
| 371 | if (left) { |
| 372 | lseek(gz.fd, --end, SEEK_SET); |
| 373 | if (read(gz.fd, gz.buf, 1) != 1) bye("reading after seek on ", name); |
| 374 | deflatePrime(strm, 8 - left, *gz.buf); |
| 375 | } |
| 376 | lseek(gz.fd, end, SEEK_SET); |
| 377 | |
| 378 | /* clean up and return */ |
| 379 | free(window); |
| 380 | free(gz.buf); |
| 381 | return gz.fd; |
| 382 | } |
| 383 | |
| 384 | /* append file "name" to gzip file gd using deflate stream strm -- if last |
| 385 | is true, then finish off the deflate stream at the end */ |
| 386 | local void gztack(char *name, int gd, z_stream *strm, int last) |
| 387 | { |
| 388 | int fd, len, ret; |
| 389 | unsigned left; |
| 390 | unsigned char *in, *out; |
| 391 | |
| 392 | /* open file to compress and append */ |
| 393 | fd = 0; |
| 394 | if (name != NULL) { |
| 395 | fd = open(name, O_RDONLY, 0); |
| 396 | if (fd == -1) |
| 397 | fprintf(stderr, "gzappend warning: %s not found, skipping ...\n", |
| 398 | name); |
| 399 | } |
| 400 | |
| 401 | /* allocate buffers */ |
| 402 | in = fd == -1 ? NULL : malloc(CHUNK); |
| 403 | out = malloc(CHUNK); |
| 404 | if (out == NULL) bye("out of memory", ""); |
| 405 | |
| 406 | /* compress input file and append to gzip file */ |
| 407 | do { |
| 408 | /* get more input */ |
| 409 | len = fd == -1 ? 0 : read(fd, in, CHUNK); |
| 410 | if (len == -1) { |
| 411 | fprintf(stderr, |
| 412 | "gzappend warning: error reading %s, skipping rest ...\n", |
| 413 | name); |
| 414 | len = 0; |
| 415 | } |
| 416 | strm->avail_in = (unsigned)len; |
| 417 | strm->next_in = in; |
| 418 | if (len) strm->adler = crc32(strm->adler, in, (unsigned)len); |
| 419 | |
| 420 | /* compress and write all available output */ |
| 421 | do { |
| 422 | strm->avail_out = CHUNK; |
| 423 | strm->next_out = out; |
| 424 | ret = deflate(strm, last && len == 0 ? Z_FINISH : Z_NO_FLUSH); |
| 425 | left = CHUNK - strm->avail_out; |
| 426 | while (left) { |
| 427 | len = write(gd, out + CHUNK - strm->avail_out - left, left); |
| 428 | if (len == -1) bye("writing gzip file", ""); |
| 429 | left -= (unsigned)len; |
| 430 | } |
| 431 | } while (strm->avail_out == 0 && ret != Z_STREAM_END); |
| 432 | } while (len != 0); |
| 433 | |
| 434 | /* write trailer after last entry */ |
| 435 | if (last) { |
| 436 | deflateEnd(strm); |
| 437 | out[0] = (unsigned char)(strm->adler); |
| 438 | out[1] = (unsigned char)(strm->adler >> 8); |
| 439 | out[2] = (unsigned char)(strm->adler >> 16); |
| 440 | out[3] = (unsigned char)(strm->adler >> 24); |
| 441 | out[4] = (unsigned char)(strm->total_in); |
| 442 | out[5] = (unsigned char)(strm->total_in >> 8); |
| 443 | out[6] = (unsigned char)(strm->total_in >> 16); |
| 444 | out[7] = (unsigned char)(strm->total_in >> 24); |
| 445 | len = 8; |
| 446 | do { |
| 447 | ret = write(gd, out + 8 - len, len); |
| 448 | if (ret == -1) bye("writing gzip file", ""); |
| 449 | len -= ret; |
| 450 | } while (len); |
| 451 | close(gd); |
| 452 | } |
| 453 | |
| 454 | /* clean up and return */ |
| 455 | free(out); |
| 456 | if (in != NULL) free(in); |
| 457 | if (fd > 0) close(fd); |
| 458 | } |
| 459 | |
| 460 | /* process the compression level option if present, scan the gzip file, and |
| 461 | append the specified files, or append the data from stdin if no other file |
| 462 | names are provided on the command line -- the gzip file must be writable |
| 463 | and seekable */ |
| 464 | int main(int argc, char **argv) |
| 465 | { |
| 466 | int gd, level; |
| 467 | z_stream strm; |
| 468 | |
| 469 | /* ignore command name */ |
| 470 | argv++; |
| 471 | |
| 472 | /* provide usage if no arguments */ |
| 473 | if (*argv == NULL) { |
| 474 | printf("gzappend 1.1 (4 Nov 2003) Copyright (C) 2003 Mark Adler\n"); |
| 475 | printf( |
| 476 | "usage: gzappend [-level] file.gz [ addthis [ andthis ... ]]\n"); |
| 477 | return 0; |
| 478 | } |
| 479 | |
| 480 | /* set compression level */ |
| 481 | level = Z_DEFAULT_COMPRESSION; |
| 482 | if (argv[0][0] == '-') { |
| 483 | if (argv[0][1] < '0' || argv[0][1] > '9' || argv[0][2] != 0) |
| 484 | bye("invalid compression level", ""); |
| 485 | level = argv[0][1] - '0'; |
| 486 | if (*++argv == NULL) bye("no gzip file name after options", ""); |
| 487 | } |
| 488 | |
| 489 | /* prepare to append to gzip file */ |
| 490 | gd = gzscan(*argv++, &strm, level); |
| 491 | |
| 492 | /* append files on command line, or from stdin if none */ |
| 493 | if (*argv == NULL) |
| 494 | gztack(NULL, gd, &strm, 1); |
| 495 | else |
| 496 | do { |
| 497 | gztack(*argv, gd, &strm, argv[1] == NULL); |
| 498 | } while (*++argv != NULL); |
| 499 | return 0; |
| 500 | } |