util.c

   1 /* util.c -- utility functions for gzip support
   2
   3    Copyright (C) 1997-1999, 2001-2002, 2006, 2009-2024 Free Software
   4    Foundation, Inc.
   5    Copyright (C) 1992-1993 Jean-loup Gailly
   6
   7    This program is free software; you can redistribute it and/or modify
   8    it under the terms of the GNU General Public License as published by
   9    the Free Software Foundation; either version 3, or (at your option)
  10    any later version.
  11
  12    This program is distributed in the hope that it will be useful,
  13    but WITHOUT ANY WARRANTY; without even the implied warranty of
  14    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15    GNU General Public License for more details.
  16
  17    You should have received a copy of the GNU General Public License
  18    along with this program; if not, write to the Free Software Foundation,
  19    Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
  20
  21 #include <config.h>
  22 #include <ctype.h>
  23 #include <errno.h>
  24 #include <limits.h>
  25 #include <unistd.h>
  26 #include <fcntl.h>
  27 #include <stdlib.h>
  28 #include <errno.h>
  29
  30 #include "tailor.h"
  31 #include "gzip.h"
  32 #include <dirname.h>
  33 #include <xalloc.h>
  34
  35 #ifndef CHAR_BIT
  36 #  define CHAR_BIT 8
  37 #endif
  38
  39 #ifndef EPIPE
  40 # define EPIPE 0
  41 #endif
  42
  43 static int write_buffer (int, voidp, unsigned int);
  44
  45 /* ========================================================================
  46  * Table of CRC-32's of all single-byte values (made by makecrc.c)
  47  */
  48 static const ulg crc_32_tab[] = {
  49   0x00000000L, 0x77073096L, 0xee0e612cL, 0x990951baL, 0x076dc419L,
  50   0x706af48fL, 0xe963a535L, 0x9e6495a3L, 0x0edb8832L, 0x79dcb8a4L,
  51   0xe0d5e91eL, 0x97d2d988L, 0x09b64c2bL, 0x7eb17cbdL, 0xe7b82d07L,
  52   0x90bf1d91L, 0x1db71064L, 0x6ab020f2L, 0xf3b97148L, 0x84be41deL,
  53   0x1adad47dL, 0x6ddde4ebL, 0xf4d4b551L, 0x83d385c7L, 0x136c9856L,
  54   0x646ba8c0L, 0xfd62f97aL, 0x8a65c9ecL, 0x14015c4fL, 0x63066cd9L,
  55   0xfa0f3d63L, 0x8d080df5L, 0x3b6e20c8L, 0x4c69105eL, 0xd56041e4L,
  56   0xa2677172L, 0x3c03e4d1L, 0x4b04d447L, 0xd20d85fdL, 0xa50ab56bL,
  57   0x35b5a8faL, 0x42b2986cL, 0xdbbbc9d6L, 0xacbcf940L, 0x32d86ce3L,
  58   0x45df5c75L, 0xdcd60dcfL, 0xabd13d59L, 0x26d930acL, 0x51de003aL,
  59   0xc8d75180L, 0xbfd06116L, 0x21b4f4b5L, 0x56b3c423L, 0xcfba9599L,
  60   0xb8bda50fL, 0x2802b89eL, 0x5f058808L, 0xc60cd9b2L, 0xb10be924L,
  61   0x2f6f7c87L, 0x58684c11L, 0xc1611dabL, 0xb6662d3dL, 0x76dc4190L,
  62   0x01db7106L, 0x98d220bcL, 0xefd5102aL, 0x71b18589L, 0x06b6b51fL,
  63   0x9fbfe4a5L, 0xe8b8d433L, 0x7807c9a2L, 0x0f00f934L, 0x9609a88eL,
  64   0xe10e9818L, 0x7f6a0dbbL, 0x086d3d2dL, 0x91646c97L, 0xe6635c01L,
  65   0x6b6b51f4L, 0x1c6c6162L, 0x856530d8L, 0xf262004eL, 0x6c0695edL,
  66   0x1b01a57bL, 0x8208f4c1L, 0xf50fc457L, 0x65b0d9c6L, 0x12b7e950L,
  67   0x8bbeb8eaL, 0xfcb9887cL, 0x62dd1ddfL, 0x15da2d49L, 0x8cd37cf3L,
  68   0xfbd44c65L, 0x4db26158L, 0x3ab551ceL, 0xa3bc0074L, 0xd4bb30e2L,
  69   0x4adfa541L, 0x3dd895d7L, 0xa4d1c46dL, 0xd3d6f4fbL, 0x4369e96aL,
  70   0x346ed9fcL, 0xad678846L, 0xda60b8d0L, 0x44042d73L, 0x33031de5L,
  71   0xaa0a4c5fL, 0xdd0d7cc9L, 0x5005713cL, 0x270241aaL, 0xbe0b1010L,
  72   0xc90c2086L, 0x5768b525L, 0x206f85b3L, 0xb966d409L, 0xce61e49fL,
  73   0x5edef90eL, 0x29d9c998L, 0xb0d09822L, 0xc7d7a8b4L, 0x59b33d17L,
  74   0x2eb40d81L, 0xb7bd5c3bL, 0xc0ba6cadL, 0xedb88320L, 0x9abfb3b6L,
  75   0x03b6e20cL, 0x74b1d29aL, 0xead54739L, 0x9dd277afL, 0x04db2615L,
  76   0x73dc1683L, 0xe3630b12L, 0x94643b84L, 0x0d6d6a3eL, 0x7a6a5aa8L,
  77   0xe40ecf0bL, 0x9309ff9dL, 0x0a00ae27L, 0x7d079eb1L, 0xf00f9344L,
  78   0x8708a3d2L, 0x1e01f268L, 0x6906c2feL, 0xf762575dL, 0x806567cbL,
  79   0x196c3671L, 0x6e6b06e7L, 0xfed41b76L, 0x89d32be0L, 0x10da7a5aL,
  80   0x67dd4accL, 0xf9b9df6fL, 0x8ebeeff9L, 0x17b7be43L, 0x60b08ed5L,
  81   0xd6d6a3e8L, 0xa1d1937eL, 0x38d8c2c4L, 0x4fdff252L, 0xd1bb67f1L,
  82   0xa6bc5767L, 0x3fb506ddL, 0x48b2364bL, 0xd80d2bdaL, 0xaf0a1b4cL,
  83   0x36034af6L, 0x41047a60L, 0xdf60efc3L, 0xa867df55L, 0x316e8eefL,
  84   0x4669be79L, 0xcb61b38cL, 0xbc66831aL, 0x256fd2a0L, 0x5268e236L,
  85   0xcc0c7795L, 0xbb0b4703L, 0x220216b9L, 0x5505262fL, 0xc5ba3bbeL,
  86   0xb2bd0b28L, 0x2bb45a92L, 0x5cb36a04L, 0xc2d7ffa7L, 0xb5d0cf31L,
  87   0x2cd99e8bL, 0x5bdeae1dL, 0x9b64c2b0L, 0xec63f226L, 0x756aa39cL,
  88   0x026d930aL, 0x9c0906a9L, 0xeb0e363fL, 0x72076785L, 0x05005713L,
  89   0x95bf4a82L, 0xe2b87a14L, 0x7bb12baeL, 0x0cb61b38L, 0x92d28e9bL,
  90   0xe5d5be0dL, 0x7cdcefb7L, 0x0bdbdf21L, 0x86d3d2d4L, 0xf1d4e242L,
  91   0x68ddb3f8L, 0x1fda836eL, 0x81be16cdL, 0xf6b9265bL, 0x6fb077e1L,
  92   0x18b74777L, 0x88085ae6L, 0xff0f6a70L, 0x66063bcaL, 0x11010b5cL,
  93   0x8f659effL, 0xf862ae69L, 0x616bffd3L, 0x166ccf45L, 0xa00ae278L,
  94   0xd70dd2eeL, 0x4e048354L, 0x3903b3c2L, 0xa7672661L, 0xd06016f7L,
  95   0x4969474dL, 0x3e6e77dbL, 0xaed16a4aL, 0xd9d65adcL, 0x40df0b66L,
  96   0x37d83bf0L, 0xa9bcae53L, 0xdebb9ec5L, 0x47b2cf7fL, 0x30b5ffe9L,
  97   0xbdbdf21cL, 0xcabac28aL, 0x53b39330L, 0x24b4a3a6L, 0xbad03605L,
  98   0xcdd70693L, 0x54de5729L, 0x23d967bfL, 0xb3667a2eL, 0xc4614ab8L,
  99   0x5d681b02L, 0x2a6f2b94L, 0xb40bbe37L, 0xc30c8ea1L, 0x5a05df1bL,
 100   0x2d02ef8dL
 101 };
 102
 103 /* Shift register contents.  */
 104 static ulg crc = 0xffffffffL;
 105
 106 /* ===========================================================================
 107  * Copy input to output unchanged: zcat == cat with --force.
 108  * IN assertion: insize bytes have already been read in inbuf and inptr bytes
 109  * already processed or copied.
 110  * 'in' and 'out' are the input and output file descriptors.
 111  */
 112 int
 113 copy (int in, int out)
 114 {
 115     int got;
 116
 117     errno = 0;
 118     while (insize > inptr) {
 119         write_buf(out, (char*)inbuf + inptr, insize - inptr);
 120         got = read_buffer (in, (char *) inbuf, INBUFSIZ);
 121         if (got == -1)
 122             read_error();
 123         bytes_in += got;
 124         insize = (unsigned)got;
 125         inptr = 0;
 126     }
 127     return OK;
 128 }
 129
 130 /* ===========================================================================
 131  * Run a set of bytes through the crc shift register.  If s is a NULL
 132  * pointer, then initialize the crc shift register contents instead.
 133  * Return the current crc in either case.
 134  * S points to N bytes to pump through.
 135  */
 136 ulg
 137 updcrc (uch const *s, unsigned n)
 138 {
 139     register ulg c;         /* temporary variable */
 140
 141     if (s == NULL) {
 142         c = 0xffffffffL;
 143     } else {
 144         c = crc;
 145         if (n) do {
 146             c = crc_32_tab[((int)c ^ (*s++)) & 0xff] ^ (c >> 8);
 147         } while (--n);
 148     }
 149     crc = c;
 150     return c ^ 0xffffffffL;       /* (instead of ~c for 64-bit machines) */
 151 }
 152
 153 /* Return a current CRC value.  */
 154 ulg
 155 getcrc ()
 156 {
 157   return crc ^ 0xffffffffL;
 158 }
 159
 160 #ifdef IBM_Z_DFLTCC
 161 /* Set a new CRC value.  */
 162 void
 163 setcrc (ulg c)
 164 {
 165   crc = c ^ 0xffffffffL;
 166 }
 167 #endif
 168
 169 /* ===========================================================================
 170  * Clear input and output buffers
 171  */
 172 void clear_bufs()
 173 {
 174     outcnt = 0;
 175     insize = inptr = 0;
 176     bytes_in = bytes_out = 0L;
 177 }
 178
 179 /* ===========================================================================
 180  * Fill the input buffer. This is called only when the buffer is empty.
 181  * EOF_OK is set if EOF acceptable as a result.
 182  */
 183 int
 184 fill_inbuf (int eof_ok)
 185 {
 186     int len;
 187
 188     /* Read as much as possible */
 189     insize = 0;
 190     do {
 191         len = read_buffer (ifd, (char *) inbuf + insize, INBUFSIZ - insize);
 192         if (len == 0) break;
 193         if (len == -1) {
 194           read_error();
 195           break;
 196         }
 197         insize += len;
 198     } while (insize < INBUFSIZ);
 199
 200     if (insize == 0) {
 201         if (eof_ok) return EOF;
 202         flush_window();
 203         errno = 0;
 204         read_error();
 205     }
 206     bytes_in += (off_t)insize;
 207     inptr = 1;
 208     return inbuf[0];
 209 }
 210
 211 /* Like the standard read function, except do not attempt to read more
 212    than INT_MAX bytes at a time.  */
 213 int
 214 read_buffer (int fd, voidp buf, unsigned int cnt)
 215 {
 216   int len;
 217   if (INT_MAX < cnt)
 218     cnt = INT_MAX;
 219   len = read (fd, buf, cnt);
 220
 221 #if defined F_SETFL && O_NONBLOCK && defined EAGAIN
 222   /* Input files are opened O_NONBLOCK for security reasons.  On some
 223      file systems this can cause read to fail with errno == EAGAIN.  */
 224   if (len < 0 && errno == EAGAIN)
 225     {
 226       int flags = fcntl (fd, F_GETFL);
 227       if (0 <= flags)
 228         {
 229           if (! (flags & O_NONBLOCK))
 230             errno = EAGAIN;
 231           else if (fcntl (fd, F_SETFL, flags & ~O_NONBLOCK) != -1)
 232             len = read (fd, buf, cnt);
 233         }
 234     }
 235 #endif
 236
 237   return len;
 238 }
 239
 240 /* Likewise for 'write'.  */
 241 static int
 242 write_buffer (int fd, voidp buf, unsigned int cnt)
 243 {
 244   if (INT_MAX < cnt)
 245     cnt = INT_MAX;
 246   return write (fd, buf, cnt);
 247 }
 248
 249 /* ===========================================================================
 250  * Write the output buffer outbuf[0..outcnt-1] and update bytes_out.
 251  * (used for the compressed data only)
 252  */
 253 void flush_outbuf()
 254 {
 255     if (outcnt == 0) return;
 256
 257     write_buf (ofd, outbuf, outcnt);
 258     outcnt = 0;
 259 }
 260
 261 /* ===========================================================================
 262  * Write the output window window[0..outcnt-1] and update crc and bytes_out.
 263  * (Used for the decompressed data only.)
 264  */
 265 void flush_window()
 266 {
 267     if (outcnt == 0) return;
 268     updcrc(window, outcnt);
 269
 270     write_buf (ofd, window, outcnt);
 271     outcnt = 0;
 272 }
 273
 274 /* ===========================================================================
 275  * Update the count of output bytes.  If testing, do not do any
 276  * output.  Otherwise, write the buffer, checking for errors.
 277  */
 278 void
 279 write_buf (int fd, voidp buf, unsigned  cnt)
 280 {
 281     unsigned  n;
 282
 283     bytes_out += cnt;
 284     if (test)
 285       return;
 286
 287     while ((n = write_buffer (fd, buf, cnt)) != cnt) {
 288         if (n == (unsigned)(-1)) {
 289             write_error();
 290         }
 291         cnt -= n;
 292         buf = (voidp)((char*)buf+n);
 293     }
 294 }
 295
 296 /* ========================================================================
 297  * Put string s in lower case, return s.
 298  */
 299 char *
 300 strlwr (char *s)
 301 {
 302     char *t;
 303     for (t = s; *t; t++)
 304       *t = tolow ((unsigned char) *t);
 305     return s;
 306 }
 307
 308 /* ========================================================================
 309  * Return the base name of a file (remove any directory prefix and
 310  * any version suffix). For systems with file names that are not
 311  * case sensitive, force the base name to lower case.
 312  */
 313 char *
 314 gzip_base_name (char *fname)
 315 {
 316     fname = last_component (fname);
 317     if (casemap('A') == 'a') strlwr(fname);
 318     return fname;
 319 }
 320
 321 /* ========================================================================
 322  * Unlink a file, working around the unlink readonly bug (if present).
 323  */
 324 int
 325 xunlink (char *filename)
 326 {
 327   int r = unlink (filename);
 328
 329 #ifdef UNLINK_READONLY_BUG
 330   if (r != 0)
 331     {
 332       int e = errno;
 333       if (chmod (filename, S_IWUSR) != 0)
 334         {
 335           errno = e;
 336           return -1;
 337         }
 338
 339       r = unlink (filename);
 340     }
 341 #endif
 342
 343   return r;
 344 }
 345
 346 #ifdef NO_MULTIPLE_DOTS
 347 /* ========================================================================
 348  * Make a file name legal for file systems not allowing file names with
 349  * multiple dots or starting with a dot (such as MSDOS), by changing
 350  * all dots except the last one into underlines.  A target dependent
 351  * function can be used instead of this simple function by defining the macro
 352  * MAKE_LEGAL_NAME in tailor.h and providing the function in a target
 353  * dependent module.
 354  */
 355 void
 356 make_simple_name (char *name)
 357 {
 358     char *p = strrchr(name, '.');
 359     if (p == NULL) return;
 360     if (p == name) p++;
 361     do {
 362         if (*--p == '.') *p = '_';
 363     } while (p != name);
 364 }
 365 #endif
 366
 367 /* Convert the value of the environment variable ENVVAR_NAME
 368    to a newly allocated argument vector, and set *ARGCP and *ARGVP
 369    to the number of arguments and to the vector, respectively.
 370    Make the new vector's zeroth element equal to the old **ARGVP.
 371    Return a pointer to the newly allocated string storage.
 372
 373    If the vector would be empty, do not allocate storage,
 374    do not set *ARGCP and *ARGVP, and return NULL.  */
 375
 376 #define SEPARATOR       " \t"   /* separators in env variable */
 377
 378 char *add_envopt(
 379     int *argcp,          /* pointer to argc */
 380     char ***argvp,       /* pointer to argv */
 381     char const *envvar_name) /* name of environment variable */
 382 {
 383     char *p;             /* running pointer through env variable */
 384     char **oargv;        /* runs through old argv array */
 385     char **nargv;        /* runs through new argv array */
 386     int  nargc = 0;      /* number of arguments in env variable */
 387     char *env_val;
 388
 389     env_val = getenv(envvar_name);
 390     if (env_val == NULL) return NULL;
 391
 392     env_val = xstrdup (env_val);
 393
 394     for (p = env_val; *p; nargc++ ) {        /* move through env_val */
 395         p += strspn(p, SEPARATOR);           /* skip leading separators */
 396         if (*p == '\0') break;
 397
 398         p += strcspn(p, SEPARATOR);          /* find end of word */
 399         if (*p) *p++ = '\0';                 /* mark it */
 400     }
 401     if (nargc == 0) {
 402         free(env_val);
 403         return NULL;
 404     }
 405     *argcp = nargc + 1;
 406     /* Allocate the new argv array, with an extra element just in case
 407      * the original arg list did not end with a NULL.
 408      */
 409     nargv = xcalloc (*argcp + 1, sizeof (char *));
 410     oargv  = *argvp;
 411     *argvp = nargv;
 412
 413     /* Copy the program name first */
 414     *nargv++ = *oargv;
 415
 416     /* Then copy the environment args */
 417     for (p = env_val; nargc > 0; nargc--) {
 418         p += strspn(p, SEPARATOR);           /* skip separators */
 419         *(nargv++) = p;                      /* store start */
 420         while (*p++) ;                       /* skip over word */
 421     }
 422
 423     *nargv = NULL;
 424     return env_val;
 425 }
 426
 427 /* ========================================================================
 428  * Error handlers.
 429  */
 430 void
 431 gzip_error (char const *m)
 432 {
 433     fprintf (stderr, "\n%s: %s: %s\n", program_name, ifname, m);
 434     abort_gzip();
 435 }
 436
 437 void
 438 xalloc_die ()
 439 {
 440   fprintf (stderr, "\n%s: memory_exhausted\n", program_name);
 441   abort_gzip ();
 442 }
 443
 444 void warning (char const *m)
 445 {
 446     WARN ((stderr, "%s: %s: warning: %s\n", program_name, ifname, m));
 447 }
 448
 449 void read_error()
 450 {
 451     fprintf (stderr, "\n%s: %s: %s\n",
 452              program_name, ifname,
 453              errno ? strerror (errno) : "unexpected end of file");
 454     abort_gzip();
 455 }
 456
 457 void write_error()
 458 {
 459   int exitcode = errno == EPIPE ? WARNING : ERROR;
 460   if (! (exitcode == WARNING && quiet))
 461     fprintf (stderr, "\n%s: %s: %s\n", program_name, ofname, strerror (errno));
 462   finish_up_gzip (exitcode);
 463 }
 464
 465 /* ========================================================================
 466  * Display compression ratio on the given stream on 6 characters.
 467  */
 468 void
 469 display_ratio (off_t num, off_t den, FILE *file)
 470 {
 471     fprintf(file, "%5.1f%%", den == 0 ? 0 : 100.0 * num / den);
 472 }
 473
 474 /* ========================================================================
 475  * Print an off_t.  There's no completely portable way to use printf,
 476  * so we do it ourselves.
 477  */
 478 void
 479 fprint_off (FILE *file, off_t offset, int width)
 480 {
 481     char buf[CHAR_BIT * sizeof (off_t)];
 482     char *p = buf + sizeof buf;
 483
 484     /* Don't negate offset here; it might overflow.  */
 485     if (offset < 0) {
 486         do
 487           *--p = '0' - offset % 10;
 488         while ((offset /= 10) != 0);
 489
 490         *--p = '-';
 491     } else {
 492         do
 493           *--p = '0' + offset % 10;
 494         while ((offset /= 10) != 0);
 495     }
 496
 497     width -= buf + sizeof buf - p;
 498     while (0 < width--) {
 499         putc (' ', file);
 500     }
 501     for (;  p < buf + sizeof buf;  p++)
 502         putc (*p, file);
 503 }