1 /*- 2 * Copyright (c) 2003-2007 Tim Kientzle 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR 15 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 16 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 17 * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, 18 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 19 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 20 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 21 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 22 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 23 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 24 */ 25 26 #include "bsdtar_platform.h" 27 __FBSDID("$FreeBSD: src/usr.bin/tar/util.c,v 1.23 2008/12/15 06:00:25 kientzle Exp $"); 28 29 #ifdef HAVE_SYS_STAT_H 30 #include <sys/stat.h> 31 #endif 32 #ifdef HAVE_SYS_TYPES_H 33 #include <sys/types.h> /* Linux doesn't define mode_t, etc. in sys/stat.h. */ 34 #endif 35 #include <ctype.h> 36 #ifdef HAVE_ERRNO_H 37 #include <errno.h> 38 #endif 39 #ifdef HAVE_IO_H 40 #include <io.h> 41 #endif 42 #ifdef HAVE_STDARG_H 43 #include <stdarg.h> 44 #endif 45 #ifdef HAVE_STDINT_H 46 #include <stdint.h> 47 #endif 48 #include <stdio.h> 49 #ifdef HAVE_STDLIB_H 50 #include <stdlib.h> 51 #endif 52 #ifdef HAVE_STRING_H 53 #include <string.h> 54 #endif 55 #ifdef HAVE_WCTYPE_H 56 #include <wctype.h> 57 #else 58 /* If we don't have wctype, we need to hack up some version of iswprint(). */ 59 #define iswprint isprint 60 #endif 61 62 #include "bsdtar.h" 63 #include "err.h" 64 65 static size_t bsdtar_expand_char(char *, size_t, char); 66 static const char *strip_components(const char *path, int elements); 67 68 #if defined(_WIN32) && !defined(__CYGWIN__) 69 #define read _read 70 #endif 71 72 /* TODO: Hack up a version of mbtowc for platforms with no wide 73 * character support at all. I think the following might suffice, 74 * but it needs careful testing. 75 * #if !HAVE_MBTOWC 76 * #define mbtowc(wcp, p, n) ((*wcp = *p), 1) 77 * #endif 78 */ 79 80 /* 81 * Print a string, taking care with any non-printable characters. 82 * 83 * Note that we use a stack-allocated buffer to receive the formatted 84 * string if we can. This is partly performance (avoiding a call to 85 * malloc()), partly out of expedience (we have to call vsnprintf() 86 * before malloc() anyway to find out how big a buffer we need; we may 87 * as well point that first call at a small local buffer in case it 88 * works), but mostly for safety (so we can use this to print messages 89 * about out-of-memory conditions). 90 */ 91 92 void 93 safe_fprintf(FILE *f, const char *fmt, ...) 94 { 95 char fmtbuff_stack[256]; /* Place to format the printf() string. */ 96 char outbuff[256]; /* Buffer for outgoing characters. */ 97 char *fmtbuff_heap; /* If fmtbuff_stack is too small, we use malloc */ 98 char *fmtbuff; /* Pointer to fmtbuff_stack or fmtbuff_heap. */ 99 int fmtbuff_length; 100 int length, n; 101 va_list ap; 102 const char *p; 103 unsigned i; 104 wchar_t wc; 105 char try_wc; 106 107 /* Use a stack-allocated buffer if we can, for speed and safety. */ 108 fmtbuff_heap = NULL; 109 fmtbuff_length = sizeof(fmtbuff_stack); 110 fmtbuff = fmtbuff_stack; 111 112 /* Try formatting into the stack buffer. */ 113 va_start(ap, fmt); 114 length = vsnprintf(fmtbuff, fmtbuff_length, fmt, ap); 115 va_end(ap); 116 117 /* If the result was too large, allocate a buffer on the heap. */ 118 while (length < 0 || length >= fmtbuff_length) { 119 if (length >= fmtbuff_length) 120 fmtbuff_length = length+1; 121 else if (fmtbuff_length < 8192) 122 fmtbuff_length *= 2; 123 else if (fmtbuff_length < 1000000) 124 fmtbuff_length += fmtbuff_length / 4; 125 else { 126 length = fmtbuff_length; 127 fmtbuff_heap[length-1] = '\0'; 128 break; 129 } 130 free(fmtbuff_heap); 131 fmtbuff_heap = malloc(fmtbuff_length); 132 133 /* Reformat the result into the heap buffer if we can. */ 134 if (fmtbuff_heap != NULL) { 135 fmtbuff = fmtbuff_heap; 136 va_start(ap, fmt); 137 length = vsnprintf(fmtbuff, fmtbuff_length, fmt, ap); 138 va_end(ap); 139 } else { 140 /* Leave fmtbuff pointing to the truncated 141 * string in fmtbuff_stack. */ 142 length = sizeof(fmtbuff_stack) - 1; 143 break; 144 } 145 } 146 147 /* Note: mbrtowc() has a cleaner API, but mbtowc() seems a bit 148 * more portable, so we use that here instead. */ 149 if (mbtowc(NULL, NULL, 1) == -1) { /* Reset the shift state. */ 150 /* mbtowc() should never fail in practice, but 151 * handle the theoretical error anyway. */ 152 free(fmtbuff_heap); 153 return; 154 } 155 156 /* Write data, expanding unprintable characters. */ 157 p = fmtbuff; 158 i = 0; 159 try_wc = 1; 160 while (*p != '\0') { 161 162 /* Convert to wide char, test if the wide 163 * char is printable in the current locale. */ 164 if (try_wc && (n = mbtowc(&wc, p, length)) != -1) { 165 length -= n; 166 if (iswprint(wc) && wc != L'\\') { 167 /* Printable, copy the bytes through. */ 168 while (n-- > 0) 169 outbuff[i++] = *p++; 170 } else { 171 /* Not printable, format the bytes. */ 172 while (n-- > 0) 173 i += (unsigned)bsdtar_expand_char( 174 outbuff, i, *p++); 175 } 176 } else { 177 /* After any conversion failure, don't bother 178 * trying to convert the rest. */ 179 i += (unsigned)bsdtar_expand_char(outbuff, i, *p++); 180 try_wc = 0; 181 } 182 183 /* If our output buffer is full, dump it and keep going. */ 184 if (i > (sizeof(outbuff) - 20)) { 185 outbuff[i] = '\0'; 186 fprintf(f, "%s", outbuff); 187 i = 0; 188 } 189 } 190 outbuff[i] = '\0'; 191 fprintf(f, "%s", outbuff); 192 193 /* If we allocated a heap-based formatting buffer, free it now. */ 194 free(fmtbuff_heap); 195 } 196 197 /* 198 * Render an arbitrary sequence of bytes into printable ASCII characters. 199 */ 200 static size_t 201 bsdtar_expand_char(char *buff, size_t offset, char c) 202 { 203 size_t i = offset; 204 205 if (isprint((unsigned char)c) && c != '\\') 206 buff[i++] = c; 207 else { 208 buff[i++] = '\\'; 209 switch (c) { 210 case '\a': buff[i++] = 'a'; break; 211 case '\b': buff[i++] = 'b'; break; 212 case '\f': buff[i++] = 'f'; break; 213 case '\n': buff[i++] = 'n'; break; 214 #if '\r' != '\n' 215 /* On some platforms, \n and \r are the same. */ 216 case '\r': buff[i++] = 'r'; break; 217 #endif 218 case '\t': buff[i++] = 't'; break; 219 case '\v': buff[i++] = 'v'; break; 220 case '\\': buff[i++] = '\\'; break; 221 default: 222 sprintf(buff + i, "%03o", 0xFF & (int)c); 223 i += 3; 224 } 225 } 226 227 return (i - offset); 228 } 229 230 int 231 yes(const char *fmt, ...) 232 { 233 char buff[32]; 234 char *p; 235 ssize_t l; 236 237 va_list ap; 238 va_start(ap, fmt); 239 vfprintf(stderr, fmt, ap); 240 va_end(ap); 241 fprintf(stderr, " (y/N)? "); 242 fflush(stderr); 243 244 l = read(2, buff, sizeof(buff) - 1); 245 if (l < 0) { 246 fprintf(stderr, "Keyboard read failed\n"); 247 exit(1); 248 } 249 if (l == 0) 250 return (0); 251 buff[l] = 0; 252 253 for (p = buff; *p != '\0'; p++) { 254 if (isspace((unsigned char)*p)) 255 continue; 256 switch(*p) { 257 case 'y': case 'Y': 258 return (1); 259 case 'n': case 'N': 260 return (0); 261 default: 262 return (0); 263 } 264 } 265 266 return (0); 267 } 268 269 /*- 270 * The logic here for -C <dir> attempts to avoid 271 * chdir() as long as possible. For example: 272 * "-C /foo -C /bar file" needs chdir("/bar") but not chdir("/foo") 273 * "-C /foo -C bar file" needs chdir("/foo/bar") 274 * "-C /foo -C bar /file1" does not need chdir() 275 * "-C /foo -C bar /file1 file2" needs chdir("/foo/bar") before file2 276 * 277 * The only correct way to handle this is to record a "pending" chdir 278 * request and combine multiple requests intelligently until we 279 * need to process a non-absolute file. set_chdir() adds the new dir 280 * to the pending list; do_chdir() actually executes any pending chdir. 281 * 282 * This way, programs that build tar command lines don't have to worry 283 * about -C with non-existent directories; such requests will only 284 * fail if the directory must be accessed. 285 * 286 */ 287 void 288 set_chdir(struct bsdtar *bsdtar, const char *newdir) 289 { 290 #if defined(_WIN32) && !defined(__CYGWIN__) 291 if (newdir[0] == '/' || newdir[0] == '\\' || 292 /* Detect this type, for example, "C:\" or "C:/" */ 293 (((newdir[0] >= 'a' && newdir[0] <= 'z') || 294 (newdir[0] >= 'A' && newdir[0] <= 'Z')) && 295 newdir[1] == ':' && (newdir[2] == '/' || newdir[2] == '\\'))) { 296 #else 297 if (newdir[0] == '/') { 298 #endif 299 /* The -C /foo -C /bar case; dump first one. */ 300 free(bsdtar->pending_chdir); 301 bsdtar->pending_chdir = NULL; 302 } 303 if (bsdtar->pending_chdir == NULL) 304 /* Easy case: no previously-saved dir. */ 305 bsdtar->pending_chdir = strdup(newdir); 306 else { 307 /* The -C /foo -C bar case; concatenate */ 308 char *old_pending = bsdtar->pending_chdir; 309 size_t old_len = strlen(old_pending); 310 bsdtar->pending_chdir = malloc(old_len + strlen(newdir) + 2); 311 if (old_pending[old_len - 1] == '/') 312 old_pending[old_len - 1] = '\0'; 313 if (bsdtar->pending_chdir != NULL) 314 sprintf(bsdtar->pending_chdir, "%s/%s", 315 old_pending, newdir); 316 free(old_pending); 317 } 318 if (bsdtar->pending_chdir == NULL) 319 lafe_errc(1, errno, "No memory"); 320 } 321 322 void 323 do_chdir(struct bsdtar *bsdtar) 324 { 325 if (bsdtar->pending_chdir == NULL) 326 return; 327 328 if (chdir(bsdtar->pending_chdir) != 0) { 329 lafe_errc(1, 0, "could not chdir to '%s'\n", 330 bsdtar->pending_chdir); 331 } 332 free(bsdtar->pending_chdir); 333 bsdtar->pending_chdir = NULL; 334 } 335 336 static const char * 337 strip_components(const char *p, int elements) 338 { 339 /* Skip as many elements as necessary. */ 340 while (elements > 0) { 341 switch (*p++) { 342 case '/': 343 #if defined(_WIN32) && !defined(__CYGWIN__) 344 case '\\': /* Support \ path sep on Windows ONLY. */ 345 #endif 346 elements--; 347 break; 348 case '\0': 349 /* Path is too short, skip it. */ 350 return (NULL); 351 } 352 } 353 354 /* Skip any / characters. This handles short paths that have 355 * additional / termination. This also handles the case where 356 * the logic above stops in the middle of a duplicate // 357 * sequence (which would otherwise get converted to an 358 * absolute path). */ 359 for (;;) { 360 switch (*p) { 361 case '/': 362 #if defined(_WIN32) && !defined(__CYGWIN__) 363 case '\\': /* Support \ path sep on Windows ONLY. */ 364 #endif 365 ++p; 366 break; 367 case '\0': 368 return (NULL); 369 default: 370 return (p); 371 } 372 } 373 } 374 375 /* 376 * Handle --strip-components and any future path-rewriting options. 377 * Returns non-zero if the pathname should not be extracted. 378 * 379 * TODO: Support pax-style regex path rewrites. 380 */ 381 int 382 edit_pathname(struct bsdtar *bsdtar, struct archive_entry *entry) 383 { 384 const char *name = archive_entry_pathname(entry); 385 #if defined(HAVE_REGEX_H) || defined(HAVE_PCREPOSIX_H) 386 char *subst_name; 387 int r; 388 389 r = apply_substitution(bsdtar, name, &subst_name, 0, 0); 390 if (r == -1) { 391 lafe_warnc(0, "Invalid substitution, skipping entry"); 392 return 1; 393 } 394 if (r == 1) { 395 archive_entry_copy_pathname(entry, subst_name); 396 if (*subst_name == '\0') { 397 free(subst_name); 398 return -1; 399 } else 400 free(subst_name); 401 name = archive_entry_pathname(entry); 402 } 403 404 if (archive_entry_hardlink(entry)) { 405 r = apply_substitution(bsdtar, archive_entry_hardlink(entry), &subst_name, 0, 1); 406 if (r == -1) { 407 lafe_warnc(0, "Invalid substitution, skipping entry"); 408 return 1; 409 } 410 if (r == 1) { 411 archive_entry_copy_hardlink(entry, subst_name); 412 free(subst_name); 413 } 414 } 415 if (archive_entry_symlink(entry) != NULL) { 416 r = apply_substitution(bsdtar, archive_entry_symlink(entry), &subst_name, 1, 0); 417 if (r == -1) { 418 lafe_warnc(0, "Invalid substitution, skipping entry"); 419 return 1; 420 } 421 if (r == 1) { 422 archive_entry_copy_symlink(entry, subst_name); 423 free(subst_name); 424 } 425 } 426 #endif 427 428 /* Strip leading dir names as per --strip-components option. */ 429 if (bsdtar->strip_components > 0) { 430 const char *linkname = archive_entry_hardlink(entry); 431 432 name = strip_components(name, bsdtar->strip_components); 433 if (name == NULL) 434 return (1); 435 436 if (linkname != NULL) { 437 linkname = strip_components(linkname, 438 bsdtar->strip_components); 439 if (linkname == NULL) 440 return (1); 441 archive_entry_copy_hardlink(entry, linkname); 442 } 443 } 444 445 /* By default, don't write or restore absolute pathnames. */ 446 if (!bsdtar->option_absolute_paths) { 447 const char *rp, *p = name; 448 int slashonly = 1; 449 450 /* Remove leading "//./" or "//?/" or "//?/UNC/" 451 * (absolute path prefixes used by Windows API) */ 452 if ((p[0] == '/' || p[0] == '\\') && 453 (p[1] == '/' || p[1] == '\\') && 454 (p[2] == '.' || p[2] == '?') && 455 (p[3] == '/' || p[3] == '\\')) 456 { 457 if (p[2] == '?' && 458 (p[4] == 'U' || p[4] == 'u') && 459 (p[5] == 'N' || p[5] == 'n') && 460 (p[6] == 'C' || p[6] == 'c') && 461 (p[7] == '/' || p[7] == '\\')) 462 p += 8; 463 else 464 p += 4; 465 slashonly = 0; 466 } 467 do { 468 rp = p; 469 /* Remove leading drive letter from archives created 470 * on Windows. */ 471 if (((p[0] >= 'a' && p[0] <= 'z') || 472 (p[0] >= 'A' && p[0] <= 'Z')) && 473 p[1] == ':') { 474 p += 2; 475 slashonly = 0; 476 } 477 /* Remove leading "/../", "//", etc. */ 478 while (p[0] == '/' || p[0] == '\\') { 479 if (p[1] == '.' && p[2] == '.' && 480 (p[3] == '/' || p[3] == '\\')) { 481 p += 3; /* Remove "/..", leave "/" 482 * for next pass. */ 483 slashonly = 0; 484 } else 485 p += 1; /* Remove "/". */ 486 } 487 } while (rp != p); 488 489 if (p != name && !bsdtar->warned_lead_slash) { 490 /* Generate a warning the first time this happens. */ 491 if (slashonly) 492 lafe_warnc(0, 493 "Removing leading '%c' from member names", 494 name[0]); 495 else 496 lafe_warnc(0, 497 "Removing leading drive letter from " 498 "member names"); 499 bsdtar->warned_lead_slash = 1; 500 } 501 502 /* Special case: Stripping everything yields ".". */ 503 if (*p == '\0') 504 name = "."; 505 else 506 name = p; 507 } else { 508 /* Strip redundant leading '/' characters. */ 509 while (name[0] == '/' && name[1] == '/') 510 name++; 511 } 512 513 /* Safely replace name in archive_entry. */ 514 if (name != archive_entry_pathname(entry)) { 515 char *q = strdup(name); 516 archive_entry_copy_pathname(entry, q); 517 free(q); 518 } 519 return (0); 520 } 521 522 /* 523 * It would be nice to just use printf() for formatting large numbers, 524 * but the compatibility problems are quite a headache. Hence the 525 * following simple utility function. 526 */ 527 const char * 528 tar_i64toa(int64_t n0) 529 { 530 static char buff[24]; 531 uint64_t n = n0 < 0 ? -n0 : n0; 532 char *p = buff + sizeof(buff); 533 534 *--p = '\0'; 535 do { 536 *--p = '0' + (int)(n % 10); 537 } while (n /= 10); 538 if (n0 < 0) 539 *--p = '-'; 540 return p; 541 } 542 543 /* 544 * Like strcmp(), but try to be a little more aware of the fact that 545 * we're comparing two paths. Right now, it just handles leading 546 * "./" and trailing '/' specially, so that "a/b/" == "./a/b" 547 * 548 * TODO: Make this better, so that "./a//b/./c/" == "a/b/c" 549 * TODO: After this works, push it down into libarchive. 550 * TODO: Publish the path normalization routines in libarchive so 551 * that bsdtar can normalize paths and use fast strcmp() instead 552 * of this. 553 * 554 * Note: This is currently only used within write.c, so should 555 * not handle \ path separators. 556 */ 557 558 int 559 pathcmp(const char *a, const char *b) 560 { 561 /* Skip leading './' */ 562 if (a[0] == '.' && a[1] == '/' && a[2] != '\0') 563 a += 2; 564 if (b[0] == '.' && b[1] == '/' && b[2] != '\0') 565 b += 2; 566 /* Find the first difference, or return (0) if none. */ 567 while (*a == *b) { 568 if (*a == '\0') 569 return (0); 570 a++; 571 b++; 572 } 573 /* 574 * If one ends in '/' and the other one doesn't, 575 * they're the same. 576 */ 577 if (a[0] == '/' && a[1] == '\0' && b[0] == '\0') 578 return (0); 579 if (a[0] == '\0' && b[0] == '/' && b[1] == '\0') 580 return (0); 581 /* They're really different, return the correct sign. */ 582 return (*(const unsigned char *)a - *(const unsigned char *)b); 583 } 584