Martin v. Löwis | d372aa8 | 2006-01-03 06:44:59 +0000 | [diff] [blame^] | 1 | /* gzjoin -- command to join gzip files into one gzip file |
| 2 | |
| 3 | Copyright (C) 2004 Mark Adler, all rights reserved |
| 4 | version 1.0, 11 Dec 2004 |
| 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 11 Dec 2004 - First version |
| 29 | * 1.1 12 Jun 2005 - Changed ssize_t to long for portability |
| 30 | */ |
| 31 | |
| 32 | /* |
| 33 | gzjoin takes one or more gzip files on the command line and writes out a |
| 34 | single gzip file that will uncompress to the concatenation of the |
| 35 | uncompressed data from the individual gzip files. gzjoin does this without |
| 36 | having to recompress any of the data and without having to calculate a new |
| 37 | crc32 for the concatenated uncompressed data. gzjoin does however have to |
| 38 | decompress all of the input data in order to find the bits in the compressed |
| 39 | data that need to be modified to concatenate the streams. |
| 40 | |
| 41 | gzjoin does not do an integrity check on the input gzip files other than |
| 42 | checking the gzip header and decompressing the compressed data. They are |
| 43 | otherwise assumed to be complete and correct. |
| 44 | |
| 45 | Each joint between gzip files removes at least 18 bytes of previous trailer |
| 46 | and subsequent header, and inserts an average of about three bytes to the |
| 47 | compressed data in order to connect the streams. The output gzip file |
| 48 | has a minimal ten-byte gzip header with no file name or modification time. |
| 49 | |
| 50 | This program was written to illustrate the use of the Z_BLOCK option of |
| 51 | inflate() and the crc32_combine() function. gzjoin will not compile with |
| 52 | versions of zlib earlier than 1.2.3. |
| 53 | */ |
| 54 | |
| 55 | #include <stdio.h> /* fputs(), fprintf(), fwrite(), putc() */ |
| 56 | #include <stdlib.h> /* exit(), malloc(), free() */ |
| 57 | #include <fcntl.h> /* open() */ |
| 58 | #include <unistd.h> /* close(), read(), lseek() */ |
| 59 | #include "zlib.h" |
| 60 | /* crc32(), crc32_combine(), inflateInit2(), inflate(), inflateEnd() */ |
| 61 | |
| 62 | #define local static |
| 63 | |
| 64 | /* exit with an error (return a value to allow use in an expression) */ |
| 65 | local int bail(char *why1, char *why2) |
| 66 | { |
| 67 | fprintf(stderr, "gzjoin error: %s%s, output incomplete\n", why1, why2); |
| 68 | exit(1); |
| 69 | return 0; |
| 70 | } |
| 71 | |
| 72 | /* -- simple buffered file input with access to the buffer -- */ |
| 73 | |
| 74 | #define CHUNK 32768 /* must be a power of two and fit in unsigned */ |
| 75 | |
| 76 | /* bin buffered input file type */ |
| 77 | typedef struct { |
| 78 | char *name; /* name of file for error messages */ |
| 79 | int fd; /* file descriptor */ |
| 80 | unsigned left; /* bytes remaining at next */ |
| 81 | unsigned char *next; /* next byte to read */ |
| 82 | unsigned char *buf; /* allocated buffer of length CHUNK */ |
| 83 | } bin; |
| 84 | |
| 85 | /* close a buffered file and free allocated memory */ |
| 86 | local void bclose(bin *in) |
| 87 | { |
| 88 | if (in != NULL) { |
| 89 | if (in->fd != -1) |
| 90 | close(in->fd); |
| 91 | if (in->buf != NULL) |
| 92 | free(in->buf); |
| 93 | free(in); |
| 94 | } |
| 95 | } |
| 96 | |
| 97 | /* open a buffered file for input, return a pointer to type bin, or NULL on |
| 98 | failure */ |
| 99 | local bin *bopen(char *name) |
| 100 | { |
| 101 | bin *in; |
| 102 | |
| 103 | in = malloc(sizeof(bin)); |
| 104 | if (in == NULL) |
| 105 | return NULL; |
| 106 | in->buf = malloc(CHUNK); |
| 107 | in->fd = open(name, O_RDONLY, 0); |
| 108 | if (in->buf == NULL || in->fd == -1) { |
| 109 | bclose(in); |
| 110 | return NULL; |
| 111 | } |
| 112 | in->left = 0; |
| 113 | in->next = in->buf; |
| 114 | in->name = name; |
| 115 | return in; |
| 116 | } |
| 117 | |
| 118 | /* load buffer from file, return -1 on read error, 0 or 1 on success, with |
| 119 | 1 indicating that end-of-file was reached */ |
| 120 | local int bload(bin *in) |
| 121 | { |
| 122 | long len; |
| 123 | |
| 124 | if (in == NULL) |
| 125 | return -1; |
| 126 | if (in->left != 0) |
| 127 | return 0; |
| 128 | in->next = in->buf; |
| 129 | do { |
| 130 | len = (long)read(in->fd, in->buf + in->left, CHUNK - in->left); |
| 131 | if (len < 0) |
| 132 | return -1; |
| 133 | in->left += (unsigned)len; |
| 134 | } while (len != 0 && in->left < CHUNK); |
| 135 | return len == 0 ? 1 : 0; |
| 136 | } |
| 137 | |
| 138 | /* get a byte from the file, bail if end of file */ |
| 139 | #define bget(in) (in->left ? 0 : bload(in), \ |
| 140 | in->left ? (in->left--, *(in->next)++) : \ |
| 141 | bail("unexpected end of file on ", in->name)) |
| 142 | |
| 143 | /* get a four-byte little-endian unsigned integer from file */ |
| 144 | local unsigned long bget4(bin *in) |
| 145 | { |
| 146 | unsigned long val; |
| 147 | |
| 148 | val = bget(in); |
| 149 | val += (unsigned long)(bget(in)) << 8; |
| 150 | val += (unsigned long)(bget(in)) << 16; |
| 151 | val += (unsigned long)(bget(in)) << 24; |
| 152 | return val; |
| 153 | } |
| 154 | |
| 155 | /* skip bytes in file */ |
| 156 | local void bskip(bin *in, unsigned skip) |
| 157 | { |
| 158 | /* check pointer */ |
| 159 | if (in == NULL) |
| 160 | return; |
| 161 | |
| 162 | /* easy case -- skip bytes in buffer */ |
| 163 | if (skip <= in->left) { |
| 164 | in->left -= skip; |
| 165 | in->next += skip; |
| 166 | return; |
| 167 | } |
| 168 | |
| 169 | /* skip what's in buffer, discard buffer contents */ |
| 170 | skip -= in->left; |
| 171 | in->left = 0; |
| 172 | |
| 173 | /* seek past multiples of CHUNK bytes */ |
| 174 | if (skip > CHUNK) { |
| 175 | unsigned left; |
| 176 | |
| 177 | left = skip & (CHUNK - 1); |
| 178 | if (left == 0) { |
| 179 | /* exact number of chunks: seek all the way minus one byte to check |
| 180 | for end-of-file with a read */ |
| 181 | lseek(in->fd, skip - 1, SEEK_CUR); |
| 182 | if (read(in->fd, in->buf, 1) != 1) |
| 183 | bail("unexpected end of file on ", in->name); |
| 184 | return; |
| 185 | } |
| 186 | |
| 187 | /* skip the integral chunks, update skip with remainder */ |
| 188 | lseek(in->fd, skip - left, SEEK_CUR); |
| 189 | skip = left; |
| 190 | } |
| 191 | |
| 192 | /* read more input and skip remainder */ |
| 193 | bload(in); |
| 194 | if (skip > in->left) |
| 195 | bail("unexpected end of file on ", in->name); |
| 196 | in->left -= skip; |
| 197 | in->next += skip; |
| 198 | } |
| 199 | |
| 200 | /* -- end of buffered input functions -- */ |
| 201 | |
| 202 | /* skip the gzip header from file in */ |
| 203 | local void gzhead(bin *in) |
| 204 | { |
| 205 | int flags; |
| 206 | |
| 207 | /* verify gzip magic header and compression method */ |
| 208 | if (bget(in) != 0x1f || bget(in) != 0x8b || bget(in) != 8) |
| 209 | bail(in->name, " is not a valid gzip file"); |
| 210 | |
| 211 | /* get and verify flags */ |
| 212 | flags = bget(in); |
| 213 | if ((flags & 0xe0) != 0) |
| 214 | bail("unknown reserved bits set in ", in->name); |
| 215 | |
| 216 | /* skip modification time, extra flags, and os */ |
| 217 | bskip(in, 6); |
| 218 | |
| 219 | /* skip extra field if present */ |
| 220 | if (flags & 4) { |
| 221 | unsigned len; |
| 222 | |
| 223 | len = bget(in); |
| 224 | len += (unsigned)(bget(in)) << 8; |
| 225 | bskip(in, len); |
| 226 | } |
| 227 | |
| 228 | /* skip file name if present */ |
| 229 | if (flags & 8) |
| 230 | while (bget(in) != 0) |
| 231 | ; |
| 232 | |
| 233 | /* skip comment if present */ |
| 234 | if (flags & 16) |
| 235 | while (bget(in) != 0) |
| 236 | ; |
| 237 | |
| 238 | /* skip header crc if present */ |
| 239 | if (flags & 2) |
| 240 | bskip(in, 2); |
| 241 | } |
| 242 | |
| 243 | /* write a four-byte little-endian unsigned integer to out */ |
| 244 | local void put4(unsigned long val, FILE *out) |
| 245 | { |
| 246 | putc(val & 0xff, out); |
| 247 | putc((val >> 8) & 0xff, out); |
| 248 | putc((val >> 16) & 0xff, out); |
| 249 | putc((val >> 24) & 0xff, out); |
| 250 | } |
| 251 | |
| 252 | /* Load up zlib stream from buffered input, bail if end of file */ |
| 253 | local void zpull(z_streamp strm, bin *in) |
| 254 | { |
| 255 | if (in->left == 0) |
| 256 | bload(in); |
| 257 | if (in->left == 0) |
| 258 | bail("unexpected end of file on ", in->name); |
| 259 | strm->avail_in = in->left; |
| 260 | strm->next_in = in->next; |
| 261 | } |
| 262 | |
| 263 | /* Write header for gzip file to out and initialize trailer. */ |
| 264 | local void gzinit(unsigned long *crc, unsigned long *tot, FILE *out) |
| 265 | { |
| 266 | fwrite("\x1f\x8b\x08\0\0\0\0\0\0\xff", 1, 10, out); |
| 267 | *crc = crc32(0L, Z_NULL, 0); |
| 268 | *tot = 0; |
| 269 | } |
| 270 | |
| 271 | /* Copy the compressed data from name, zeroing the last block bit of the last |
| 272 | block if clr is true, and adding empty blocks as needed to get to a byte |
| 273 | boundary. If clr is false, then the last block becomes the last block of |
| 274 | the output, and the gzip trailer is written. crc and tot maintains the |
| 275 | crc and length (modulo 2^32) of the output for the trailer. The resulting |
| 276 | gzip file is written to out. gzinit() must be called before the first call |
| 277 | of gzcopy() to write the gzip header and to initialize crc and tot. */ |
| 278 | local void gzcopy(char *name, int clr, unsigned long *crc, unsigned long *tot, |
| 279 | FILE *out) |
| 280 | { |
| 281 | int ret; /* return value from zlib functions */ |
| 282 | int pos; /* where the "last block" bit is in byte */ |
| 283 | int last; /* true if processing the last block */ |
| 284 | bin *in; /* buffered input file */ |
| 285 | unsigned char *start; /* start of compressed data in buffer */ |
| 286 | unsigned char *junk; /* buffer for uncompressed data -- discarded */ |
| 287 | z_off_t len; /* length of uncompressed data (support > 4 GB) */ |
| 288 | z_stream strm; /* zlib inflate stream */ |
| 289 | |
| 290 | /* open gzip file and skip header */ |
| 291 | in = bopen(name); |
| 292 | if (in == NULL) |
| 293 | bail("could not open ", name); |
| 294 | gzhead(in); |
| 295 | |
| 296 | /* allocate buffer for uncompressed data and initialize raw inflate |
| 297 | stream */ |
| 298 | junk = malloc(CHUNK); |
| 299 | strm.zalloc = Z_NULL; |
| 300 | strm.zfree = Z_NULL; |
| 301 | strm.opaque = Z_NULL; |
| 302 | strm.avail_in = 0; |
| 303 | strm.next_in = Z_NULL; |
| 304 | ret = inflateInit2(&strm, -15); |
| 305 | if (junk == NULL || ret != Z_OK) |
| 306 | bail("out of memory", ""); |
| 307 | |
| 308 | /* inflate and copy compressed data, clear last-block bit if requested */ |
| 309 | len = 0; |
| 310 | zpull(&strm, in); |
| 311 | start = strm.next_in; |
| 312 | last = start[0] & 1; |
| 313 | if (last && clr) |
| 314 | start[0] &= ~1; |
| 315 | strm.avail_out = 0; |
| 316 | for (;;) { |
| 317 | /* if input used and output done, write used input and get more */ |
| 318 | if (strm.avail_in == 0 && strm.avail_out != 0) { |
| 319 | fwrite(start, 1, strm.next_in - start, out); |
| 320 | start = in->buf; |
| 321 | in->left = 0; |
| 322 | zpull(&strm, in); |
| 323 | } |
| 324 | |
| 325 | /* decompress -- return early when end-of-block reached */ |
| 326 | strm.avail_out = CHUNK; |
| 327 | strm.next_out = junk; |
| 328 | ret = inflate(&strm, Z_BLOCK); |
| 329 | switch (ret) { |
| 330 | case Z_MEM_ERROR: |
| 331 | bail("out of memory", ""); |
| 332 | case Z_DATA_ERROR: |
| 333 | bail("invalid compressed data in ", in->name); |
| 334 | } |
| 335 | |
| 336 | /* update length of uncompressed data */ |
| 337 | len += CHUNK - strm.avail_out; |
| 338 | |
| 339 | /* check for block boundary (only get this when block copied out) */ |
| 340 | if (strm.data_type & 128) { |
| 341 | /* if that was the last block, then done */ |
| 342 | if (last) |
| 343 | break; |
| 344 | |
| 345 | /* number of unused bits in last byte */ |
| 346 | pos = strm.data_type & 7; |
| 347 | |
| 348 | /* find the next last-block bit */ |
| 349 | if (pos != 0) { |
| 350 | /* next last-block bit is in last used byte */ |
| 351 | pos = 0x100 >> pos; |
| 352 | last = strm.next_in[-1] & pos; |
| 353 | if (last && clr) |
| 354 | strm.next_in[-1] &= ~pos; |
| 355 | } |
| 356 | else { |
| 357 | /* next last-block bit is in next unused byte */ |
| 358 | if (strm.avail_in == 0) { |
| 359 | /* don't have that byte yet -- get it */ |
| 360 | fwrite(start, 1, strm.next_in - start, out); |
| 361 | start = in->buf; |
| 362 | in->left = 0; |
| 363 | zpull(&strm, in); |
| 364 | } |
| 365 | last = strm.next_in[0] & 1; |
| 366 | if (last && clr) |
| 367 | strm.next_in[0] &= ~1; |
| 368 | } |
| 369 | } |
| 370 | } |
| 371 | |
| 372 | /* update buffer with unused input */ |
| 373 | in->left = strm.avail_in; |
| 374 | in->next = strm.next_in; |
| 375 | |
| 376 | /* copy used input, write empty blocks to get to byte boundary */ |
| 377 | pos = strm.data_type & 7; |
| 378 | fwrite(start, 1, in->next - start - 1, out); |
| 379 | last = in->next[-1]; |
| 380 | if (pos == 0 || !clr) |
| 381 | /* already at byte boundary, or last file: write last byte */ |
| 382 | putc(last, out); |
| 383 | else { |
| 384 | /* append empty blocks to last byte */ |
| 385 | last &= ((0x100 >> pos) - 1); /* assure unused bits are zero */ |
| 386 | if (pos & 1) { |
| 387 | /* odd -- append an empty stored block */ |
| 388 | putc(last, out); |
| 389 | if (pos == 1) |
| 390 | putc(0, out); /* two more bits in block header */ |
| 391 | fwrite("\0\0\xff\xff", 1, 4, out); |
| 392 | } |
| 393 | else { |
| 394 | /* even -- append 1, 2, or 3 empty fixed blocks */ |
| 395 | switch (pos) { |
| 396 | case 6: |
| 397 | putc(last | 8, out); |
| 398 | last = 0; |
| 399 | case 4: |
| 400 | putc(last | 0x20, out); |
| 401 | last = 0; |
| 402 | case 2: |
| 403 | putc(last | 0x80, out); |
| 404 | putc(0, out); |
| 405 | } |
| 406 | } |
| 407 | } |
| 408 | |
| 409 | /* update crc and tot */ |
| 410 | *crc = crc32_combine(*crc, bget4(in), len); |
| 411 | *tot += (unsigned long)len; |
| 412 | |
| 413 | /* clean up */ |
| 414 | inflateEnd(&strm); |
| 415 | free(junk); |
| 416 | bclose(in); |
| 417 | |
| 418 | /* write trailer if this is the last gzip file */ |
| 419 | if (!clr) { |
| 420 | put4(*crc, out); |
| 421 | put4(*tot, out); |
| 422 | } |
| 423 | } |
| 424 | |
| 425 | /* join the gzip files on the command line, write result to stdout */ |
| 426 | int main(int argc, char **argv) |
| 427 | { |
| 428 | unsigned long crc, tot; /* running crc and total uncompressed length */ |
| 429 | |
| 430 | /* skip command name */ |
| 431 | argc--; |
| 432 | argv++; |
| 433 | |
| 434 | /* show usage if no arguments */ |
| 435 | if (argc == 0) { |
| 436 | fputs("gzjoin usage: gzjoin f1.gz [f2.gz [f3.gz ...]] > fjoin.gz\n", |
| 437 | stderr); |
| 438 | return 0; |
| 439 | } |
| 440 | |
| 441 | /* join gzip files on command line and write to stdout */ |
| 442 | gzinit(&crc, &tot, stdout); |
| 443 | while (argc--) |
| 444 | gzcopy(*argv++, argc, &crc, &tot, stdout); |
| 445 | |
| 446 | /* done */ |
| 447 | return 0; |
| 448 | } |