src/md5sum.c

   1 /* Compute MD5 or SHA1 checksum of files or strings
   2    Copyright (C) 1995-2002 Free Software Foundation, Inc.
   3
   4    This program is free software; you can redistribute it and/or modify
   5    it under the terms of the GNU General Public License as published by
   6    the Free Software Foundation; either version 2, or (at your option)
   7    any later version.
   8
   9    This program is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12    GNU General Public License for more details.
  13
  14    You should have received a copy of the GNU General Public License
  15    along with this program; if not, write to the Free Software Foundation,
  16    Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  17
  18 /* Written by Ulrich Drepper <drepper@gnu.ai.mit.edu>.  */
  19
  20 #include <config.h>
  21
  22 #include <getopt.h>
  23 #include <stdio.h>
  24 #include <sys/types.h>
  25
  26 #include "system.h"
  27
  28 #include "md5.h"
  29 #include "sha.h"
  30 #include "checksum.h"
  31 #include "getline.h"
  32 #include "closeout.h"
  33 #include "error.h"
  34
  35 /* The official name of this program (e.g., no `g' prefix).  */
  36 #define PROGRAM_NAME (algorithm == ALG_MD5 ? "md5sum" : "shasum")
  37
  38 #define AUTHORS N_ ("Ulrich Drepper and Scott Miller")
  39
  40 /* Most systems do not distinguish between external and internal
  41    text representations.  */
  42 /* FIXME: This begs for an autoconf test.  */
  43 #if O_BINARY
  44 # define OPENOPTS(BINARY) ((BINARY) != 0 ? TEXT1TO1 : TEXTCNVT)
  45 # define TEXT1TO1 "rb"
  46 # define TEXTCNVT "r"
  47 #else
  48 # if defined VMS
  49 #  define OPENOPTS(BINARY) ((BINARY) != 0 ? TEXT1TO1 : TEXTCNVT)
  50 #  define TEXT1TO1 "rb", "ctx=stm"
  51 #  define TEXTCNVT "r", "ctx=stm"
  52 # else
  53 #  if UNIX || __UNIX__ || unix || __unix__ || _POSIX_VERSION
  54 #   define OPENOPTS(BINARY) "r"
  55 #  else
  56     /* The following line is intended to evoke an error.
  57        Using #error is not portable enough.  */
  58     "Cannot determine system type."
  59 #  endif
  60 # endif
  61 #endif
  62
  63
  64 #define DIGEST_TYPE_STRING(Alg) ((Alg) == ALG_MD5 ? "MD5" : "SHA1")
  65 #define DIGEST_STREAM(Alg) ((Alg) == ALG_MD5 ? md5_stream : sha_stream)
  66
  67 #define DIGEST_BITS(Alg) ((Alg) == ALG_MD5 ? 128 : 160)
  68 #define DIGEST_HEX_BYTES(Alg) (DIGEST_BITS (Alg) / 4)
  69 #define DIGEST_BIN_BYTES(Alg) (DIGEST_BITS (Alg) / 8)
  70
  71 #define MAX_DIGEST_BIN_BYTES MAX (DIGEST_BIN_BYTES (ALG_MD5), \
  72                                   DIGEST_BIN_BYTES (ALG_SHA1))
  73
  74 /* The minimum length of a valid digest line.  This length does
  75    not include any newline character at the end of a line.  */
  76 #define MIN_DIGEST_LINE_LENGTH(Alg) \
  77   (DIGEST_HEX_BYTES (Alg) /* length of hexadecimal message digest */ \
  78    + 2 /* blank and binary indicator */ \
  79    + 1 /* minimum filename length */ )
  80
  81 /* Nonzero if any of the files read were the standard input. */
  82 static int have_read_stdin;
  83
  84 /* The minimum length of a valid checksum line for the selected algorithm.  */
  85 static size_t min_digest_line_length;
  86
  87 /* Set to the length of a digest hex string for the selected algorithm.  */
  88 static size_t digest_hex_bytes;
  89
  90 /* With --check, don't generate any output.
  91    The exit code indicates success or failure.  */
  92 static int status_only = 0;
  93
  94 /* With --check, print a message to standard error warning about each
  95    improperly formatted checksum line.  */
  96 static int warn = 0;
  97
  98 /* Declared and set via one of the wrapper .c files.  */
  99 /* int algorithm = ALG_UNSPECIFIED; */
 100
 101 /* The name this program was run with.  */
 102 char *program_name;
 103
 104 static const struct option long_options[] =
 105 {
 106   { "binary", no_argument, 0, 'b' },
 107   { "check", no_argument, 0, 'c' },
 108   { "status", no_argument, 0, 2 },
 109   { "string", required_argument, 0, 1 },
 110   { "text", no_argument, 0, 't' },
 111   { "warn", no_argument, 0, 'w' },
 112   { GETOPT_HELP_OPTION_DECL },
 113   { GETOPT_VERSION_OPTION_DECL },
 114   { NULL, 0, NULL, 0 }
 115 };
 116
 117 void
 118 usage (int status)
 119 {
 120   if (status != 0)
 121     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 122              program_name);
 123   else
 124     {
 125       printf (_("\
 126 Usage: %s [OPTION] [FILE]...\n\
 127   or:  %s [OPTION] --check [FILE]\n\
 128 Print or check %s (%d-bit) checksums.\n\
 129 With no FILE, or when FILE is -, read standard input.\n\
 130 "),
 131               program_name, program_name,
 132               DIGEST_TYPE_STRING (algorithm),
 133               DIGEST_BITS (algorithm));
 134       printf (_("\
 135 \n\
 136   -b, --binary            read files in binary mode (default on DOS/Windows)\n\
 137   -c, --check             check %s sums against given list\n\
 138   -t, --text              read files in text mode (default)\n\
 139 \n\
 140 "),
 141               DIGEST_TYPE_STRING (algorithm));
 142       fputs (_("\
 143 The following two options are useful only when verifying checksums:\n\
 144       --status            don't output anything, status code shows success\n\
 145   -w, --warn              warn about improperly formated checksum lines\n\
 146 \n\
 147 "), stdout);
 148       fputs (HELP_OPTION_DESCRIPTION, stdout);
 149       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 150       printf (_("\
 151 \n\
 152 The sums are computed as described in %s.  When checking, the input\n\
 153 should be a former output of this program.  The default mode is to print\n\
 154 a line with checksum, a character indicating type (`*' for binary, ` ' for\n\
 155 text), and name for each FILE.\n"),
 156               (algorithm == ALG_MD5 ? "RFC 1321" : "FIPS-180-1"));
 157       printf (_("\nReport bugs to <%s>.\n"), PACKAGE_BUGREPORT);
 158     }
 159
 160   exit (status == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
 161 }
 162
 163 /* Split the string S (of length S_LEN) into three parts:
 164    a hexadecimal digest, binary flag, and the file name.
 165    S is modified.  */
 166
 167 static int
 168 split_3 (char *s, size_t s_len,
 169          unsigned char **hex_digest, int *binary, char **file_name)
 170 {
 171   size_t i;
 172   int escaped_filename = 0;
 173
 174 #define ISWHITE(c) ((c) == ' ' || (c) == '\t')
 175
 176   i = 0;
 177   while (ISWHITE (s[i]))
 178     ++i;
 179
 180   /* Ignore this line if it is too short.
 181      Each line must have at least `min_digest_line_length - 1' (or one more, if
 182      the first is a backslash) more characters to contain correct message digest
 183      information.  */
 184   if (s_len - i < min_digest_line_length + (s[i] == '\\'))
 185     return 1;
 186
 187   if (s[i] == '\\')
 188     {
 189       ++i;
 190       escaped_filename = 1;
 191     }
 192   *hex_digest = (unsigned char *) &s[i];
 193
 194   /* The first field has to be the n-character hexadecimal
 195      representation of the message digest.  If it is not followed
 196      immediately by a white space it's an error.  */
 197   i += digest_hex_bytes;
 198   if (!ISWHITE (s[i]))
 199     return 1;
 200
 201   s[i++] = '\0';
 202
 203   if (s[i] != ' ' && s[i] != '*')
 204     return 1;
 205   *binary = (s[i++] == '*');
 206
 207   /* All characters between the type indicator and end of line are
 208      significant -- that includes leading and trailing white space.  */
 209   *file_name = &s[i];
 210
 211   if (escaped_filename)
 212     {
 213       /* Translate each `\n' string in the file name to a NEWLINE,
 214          and each `\\' string to a backslash.  */
 215
 216       char *dst = &s[i];
 217
 218       while (i < s_len)
 219         {
 220           switch (s[i])
 221             {
 222             case '\\':
 223               if (i == s_len - 1)
 224                 {
 225                   /* A valid line does not end with a backslash.  */
 226                   return 1;
 227                 }
 228               ++i;
 229               switch (s[i++])
 230                 {
 231                 case 'n':
 232                   *dst++ = '\n';
 233                   break;
 234                 case '\\':
 235                   *dst++ = '\\';
 236                   break;
 237                 default:
 238                   /* Only `\' or `n' may follow a backslash.  */
 239                   return 1;
 240                 }
 241               break;
 242
 243             case '\0':
 244               /* The file name may not contain a NUL.  */
 245               return 1;
 246               break;
 247
 248             default:
 249               *dst++ = s[i++];
 250               break;
 251             }
 252         }
 253       *dst = '\0';
 254     }
 255   return 0;
 256 }
 257
 258 static int
 259 hex_digits (unsigned char const *s)
 260 {
 261   while (*s)
 262     {
 263       if (!ISXDIGIT (*s))
 264         return 0;
 265       ++s;
 266     }
 267   return 1;
 268 }
 269
 270 /* An interface to the function, DIGEST_STREAM,
 271    (either md5_stream or sha_stream).
 272    Operate on FILENAME (it may be "-") and put the result in *BIN_RESULT.
 273    Return non-zero upon failure, zero to indicate success.  */
 274
 275 static int
 276 digest_file (const char *filename, int binary, unsigned char *bin_result,
 277            int (*digest_stream)(FILE *, void *))
 278 {
 279   FILE *fp;
 280   int err;
 281
 282   if (STREQ (filename, "-"))
 283     {
 284       have_read_stdin = 1;
 285       fp = stdin;
 286 #if O_BINARY
 287       /* If we need binary reads from a pipe or redirected stdin, we need
 288          to switch it to BINARY mode here, since stdin is already open.  */
 289       if (binary)
 290         SET_BINARY (fileno (stdin));
 291 #endif
 292     }
 293   else
 294     {
 295       /* OPENOPTS is a macro.  It varies with the system.
 296          Some systems distinguish between internal and
 297          external text representations.  */
 298
 299       fp = fopen (filename, OPENOPTS (binary));
 300       if (fp == NULL)
 301         {
 302           error (0, errno, "%s", filename);
 303           return 1;
 304         }
 305     }
 306
 307   err = (*digest_stream) (fp, bin_result);
 308   if (err)
 309     {
 310       error (0, errno, "%s", filename);
 311       if (fp != stdin)
 312         fclose (fp);
 313       return 1;
 314     }
 315
 316   if (fp != stdin && fclose (fp) == EOF)
 317     {
 318       error (0, errno, "%s", filename);
 319       return 1;
 320     }
 321
 322   return 0;
 323 }
 324
 325 static int
 326 digest_check (const char *checkfile_name, int (*digest_stream)(FILE *, void *))
 327 {
 328   FILE *checkfile_stream;
 329   int n_properly_formated_lines = 0;
 330   int n_mismatched_checksums = 0;
 331   int n_open_or_read_failures = 0;
 332   unsigned char bin_buffer[MAX_DIGEST_BIN_BYTES];
 333   size_t line_number;
 334   char *line;
 335   size_t line_chars_allocated;
 336
 337   if (STREQ (checkfile_name, "-"))
 338     {
 339       have_read_stdin = 1;
 340       checkfile_name = _("standard input");
 341       checkfile_stream = stdin;
 342     }
 343   else
 344     {
 345       checkfile_stream = fopen (checkfile_name, "r");
 346       if (checkfile_stream == NULL)
 347         {
 348           error (0, errno, "%s", checkfile_name);
 349           return 1;
 350         }
 351     }
 352
 353   SET_MODE (fileno (checkfile_stream), O_TEXT);
 354   line_number = 0;
 355   line = NULL;
 356   line_chars_allocated = 0;
 357   do
 358     {
 359       char *filename;
 360       int binary;
 361       unsigned char *hex_digest;
 362       int err;
 363       int line_length;
 364
 365       ++line_number;
 366
 367       line_length = getline (&line, &line_chars_allocated, checkfile_stream);
 368       if (line_length <= 0)
 369         break;
 370
 371       /* Ignore comment lines, which begin with a '#' character.  */
 372       if (line[0] == '#')
 373         continue;
 374
 375       /* Remove any trailing newline.  */
 376       if (line[line_length - 1] == '\n')
 377         line[--line_length] = '\0';
 378
 379       err = split_3 (line, line_length, &hex_digest, &binary, &filename);
 380       if (err || !hex_digits (hex_digest))
 381         {
 382           if (warn)
 383             {
 384               error (0, 0,
 385                      _("%s: %lu: improperly formatted %s checksum line"),
 386                      checkfile_name, (unsigned long) line_number,
 387                      DIGEST_TYPE_STRING (algorithm));
 388             }
 389         }
 390       else
 391         {
 392           static const char bin2hex[] = { '0', '1', '2', '3',
 393                                           '4', '5', '6', '7',
 394                                           '8', '9', 'a', 'b',
 395                                           'c', 'd', 'e', 'f' };
 396           int fail;
 397
 398           ++n_properly_formated_lines;
 399
 400           fail = digest_file (filename, binary, bin_buffer, digest_stream);
 401
 402           if (fail)
 403             {
 404               ++n_open_or_read_failures;
 405               if (!status_only)
 406                 {
 407                   printf (_("%s: FAILED open or read\n"), filename);
 408                   fflush (stdout);
 409                 }
 410             }
 411           else
 412             {
 413               size_t digest_bin_bytes = digest_hex_bytes / 2;
 414               size_t cnt;
 415               /* Compare generated binary number with text representation
 416                  in check file.  Ignore case of hex digits.  */
 417               for (cnt = 0; cnt < digest_bin_bytes; ++cnt)
 418                 {
 419                   if (TOLOWER (hex_digest[2 * cnt])
 420                       != bin2hex[bin_buffer[cnt] >> 4]
 421                       || (TOLOWER (hex_digest[2 * cnt + 1])
 422                           != (bin2hex[bin_buffer[cnt] & 0xf])))
 423                     break;
 424                 }
 425               if (cnt != digest_bin_bytes)
 426                 ++n_mismatched_checksums;
 427
 428               if (!status_only)
 429                 {
 430                   printf ("%s: %s\n", filename,
 431                           (cnt != digest_bin_bytes ? _("FAILED") : _("OK")));
 432                   fflush (stdout);
 433                 }
 434             }
 435         }
 436     }
 437   while (!feof (checkfile_stream) && !ferror (checkfile_stream));
 438
 439   if (line)
 440     free (line);
 441
 442   if (ferror (checkfile_stream))
 443     {
 444       error (0, 0, _("%s: read error"), checkfile_name);
 445       return 1;
 446     }
 447
 448   if (checkfile_stream != stdin && fclose (checkfile_stream) == EOF)
 449     {
 450       error (0, errno, "%s", checkfile_name);
 451       return 1;
 452     }
 453
 454   if (n_properly_formated_lines == 0)
 455     {
 456       /* Warn if no tests are found.  */
 457       error (0, 0, _("%s: no properly formatted %s checksum lines found"),
 458              checkfile_name, DIGEST_TYPE_STRING (algorithm));
 459     }
 460   else
 461     {
 462       if (!status_only)
 463         {
 464           int n_computed_checkums = (n_properly_formated_lines
 465                                      - n_open_or_read_failures);
 466
 467           if (n_open_or_read_failures > 0)
 468             {
 469               error (0, 0,
 470                      _("WARNING: %d of %d listed %s could not be read"),
 471                      n_open_or_read_failures, n_properly_formated_lines,
 472                      (n_properly_formated_lines == 1
 473                       ? _("file") : _("files")));
 474             }
 475
 476           if (n_mismatched_checksums > 0)
 477             {
 478               error (0, 0,
 479                    _("WARNING: %d of %d computed %s did NOT match"),
 480                      n_mismatched_checksums, n_computed_checkums,
 481                      (n_computed_checkums == 1
 482                       ? _("checksum") : _("checksums")));
 483             }
 484         }
 485     }
 486
 487   return ((n_properly_formated_lines > 0 && n_mismatched_checksums == 0
 488            && n_open_or_read_failures == 0) ? 0 : 1);
 489 }
 490
 491 int
 492 main (int argc, char **argv)
 493 {
 494   unsigned char bin_buffer[MAX_DIGEST_BIN_BYTES];
 495   int do_check = 0;
 496   int opt;
 497   char **string = NULL;
 498   size_t n_strings = 0;
 499   size_t err = 0;
 500   int file_type_specified = 0;
 501
 502 #if O_BINARY
 503   /* Binary is default on MSDOS, so the actual file contents
 504      are used in computation.  */
 505   int binary = 1;
 506 #else
 507   /* Text is default of the Plumb/Lankester format.  */
 508   int binary = 0;
 509 #endif
 510
 511   /* Setting values of global variables.  */
 512   program_name = argv[0];
 513   setlocale (LC_ALL, "");
 514   bindtextdomain (PACKAGE, LOCALEDIR);
 515   textdomain (PACKAGE);
 516
 517   atexit (close_stdout);
 518
 519   while ((opt = getopt_long (argc, argv, "bctw", long_options, NULL)) != -1)
 520     switch (opt)
 521       {
 522       case 0:                   /* long option */
 523         break;
 524       case 1: /* --string */
 525         {
 526           if (string == NULL)
 527             string = (char **) xmalloc ((argc - 1) * sizeof (char *));
 528
 529           if (optarg == NULL)
 530             optarg = "";
 531           string[n_strings++] = optarg;
 532         }
 533         break;
 534       case 'b':
 535         file_type_specified = 1;
 536         binary = 1;
 537         break;
 538       case 'c':
 539         do_check = 1;
 540         break;
 541       case 2:
 542         status_only = 1;
 543         warn = 0;
 544         break;
 545       case 't':
 546         file_type_specified = 1;
 547         binary = 0;
 548         break;
 549       case 'w':
 550         status_only = 0;
 551         warn = 1;
 552         break;
 553       case_GETOPT_HELP_CHAR;
 554       case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 555       default:
 556         usage (EXIT_FAILURE);
 557       }
 558
 559   min_digest_line_length = MIN_DIGEST_LINE_LENGTH (algorithm);
 560   digest_hex_bytes = DIGEST_HEX_BYTES (algorithm);
 561
 562   if (file_type_specified && do_check)
 563     {
 564       error (0, 0, _("the --binary and --text options are meaningless when \
 565 verifying checksums"));
 566       usage (EXIT_FAILURE);
 567     }
 568
 569   if (n_strings > 0 && do_check)
 570     {
 571       error (0, 0,
 572              _("the --string and --check options are mutually exclusive"));
 573       usage (EXIT_FAILURE);
 574     }
 575
 576   if (status_only && !do_check)
 577     {
 578       error (0, 0,
 579        _("the --status option is meaningful only when verifying checksums"));
 580       usage (EXIT_FAILURE);
 581     }
 582
 583   if (warn && !do_check)
 584     {
 585       error (0, 0,
 586        _("the --warn option is meaningful only when verifying checksums"));
 587       usage (EXIT_FAILURE);
 588     }
 589
 590   if (n_strings > 0)
 591     {
 592       size_t i;
 593
 594       if (optind < argc)
 595         {
 596           error (0, 0, _("no files may be specified when using --string"));
 597           usage (EXIT_FAILURE);
 598         }
 599       for (i = 0; i < n_strings; ++i)
 600         {
 601           size_t cnt;
 602           if (algorithm == ALG_MD5)
 603             md5_buffer (string[i], strlen (string[i]), bin_buffer);
 604           else
 605             sha_buffer (string[i], strlen (string[i]), bin_buffer);
 606
 607           for (cnt = 0; cnt < (digest_hex_bytes / 2); ++cnt)
 608             printf ("%02x", bin_buffer[cnt]);
 609
 610           printf ("  \"%s\"\n", string[i]);
 611         }
 612     }
 613   else if (do_check)
 614     {
 615       if (optind + 1 < argc)
 616         {
 617           error (0, 0,
 618                  _("only one argument may be specified when using --check"));
 619           usage (EXIT_FAILURE);
 620         }
 621
 622       err = digest_check ((optind == argc) ? "-" : argv[optind],
 623                           DIGEST_STREAM (algorithm));
 624     }
 625   else
 626     {
 627       if (optind == argc)
 628         argv[argc++] = "-";
 629
 630       for (; optind < argc; ++optind)
 631         {
 632           int fail;
 633           char *file = argv[optind];
 634
 635           fail = digest_file (file, binary, bin_buffer,
 636                               DIGEST_STREAM (algorithm));
 637           err |= fail;
 638           if (!fail)
 639             {
 640               size_t i;
 641
 642               /* Output a leading backslash if the file name contains
 643                  a newline or backslash.  */
 644               if (strchr (file, '\n') || strchr (file, '\\'))
 645                 putchar ('\\');
 646
 647               for (i = 0; i < (digest_hex_bytes / 2); ++i)
 648                 printf ("%02x", bin_buffer[i]);
 649
 650               putchar (' ');
 651               if (binary)
 652                 putchar ('*');
 653               else
 654                 putchar (' ');
 655
 656               /* Translate each NEWLINE byte to the string, "\\n",
 657                  and each backslash to "\\\\".  */
 658               for (i = 0; i < strlen (file); ++i)
 659                 {
 660                   switch (file[i])
 661                     {
 662                     case '\n':
 663                       fputs ("\\n", stdout);
 664                       break;
 665
 666                     case '\\':
 667                       fputs ("\\\\", stdout);
 668                       break;
 669
 670                     default:
 671                       putchar (file[i]);
 672                       break;
 673                     }
 674                 }
 675               putchar ('\n');
 676             }
 677         }
 678     }
 679
 680   if (have_read_stdin && fclose (stdin) == EOF)
 681     error (EXIT_FAILURE, errno, _("standard input"));
 682
 683   exit (err == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
 684 }