src/md5sum.c

   1 /* Compute MD5 checksum of files or strings according to the definition
   2    of MD5 in RFC 1321 from April 1992.
   3    Copyright (C) 95, 96, 1997, 1998 Free Software Foundation, Inc.
   4
   5    This program is free software; you can redistribute it and/or modify
   6    it under the terms of the GNU General Public License as published by
   7    the Free Software Foundation; either version 2, or (at your option)
   8    any later version.
   9
  10    This program is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13    GNU General Public License for more details.
  14
  15    You should have received a copy of the GNU General Public License
  16    along with this program; if not, write to the Free Software Foundation,
  17    Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  18
  19 /* Written by Ulrich Drepper <drepper@gnu.ai.mit.edu>.  */
  20
  21 #ifdef HAVE_CONFIG_H
  22 # include <config.h>
  23 #endif
  24
  25 #include <getopt.h>
  26 #include <stdio.h>
  27 #include <sys/types.h>
  28
  29 #include "long-options.h"
  30 #include "md5.h"
  31 #include "getline.h"
  32 #include "system.h"
  33 #include "error.h"
  34
  35 /* Most systems do not distinguish between external and internal
  36    text representations.  */
  37 /* FIXME: This begs for an autoconf test.  */
  38 #if UNIX || __UNIX__ || unix || __unix__ || _POSIX_VERSION
  39 # define OPENOPTS(BINARY) "r"
  40 #else
  41 # define OPENOPTS(BINARY) ((BINARY) != 0 ? TEXT1TO1 : TEXTCNVT)
  42 # ifdef MSDOS
  43 #  define TEXT1TO1 "rb"
  44 #  define TEXTCNVT "r"
  45 # else
  46 #  if defined VMS
  47 #   define TEXT1TO1 "rb", "ctx=stm"
  48 #   define TEXTCNVT "r", "ctx=stm"
  49 #  else
  50     /* The following line is intended to evoke an error.
  51        Using #error is not portable enough.  */
  52     "Cannot determine system type."
  53 #  endif
  54 # endif
  55 #endif
  56
  57 #if _LIBC || STDC_HEADERS
  58 # define TOLOWER(c) tolower (c)
  59 #else
  60 # define TOLOWER(c) (ISUPPER (c) ? tolower (c) : (c))
  61 #endif
  62
  63 /* The minimum length of a valid digest line in a file produced
  64    by `md5sum FILE' and read by `md5sum --check'.  This length does
  65    not include any newline character at the end of a line.  */
  66 #define MIN_DIGEST_LINE_LENGTH (32 /* message digest length */ \
  67                                 + 2 /* blank and binary indicator */ \
  68                                 + 1 /* minimum filename length */ )
  69
  70 /* Nonzero if any of the files read were the standard input. */
  71 static int have_read_stdin;
  72
  73 /* With --check, don't generate any output.
  74    The exit code indicates success or failure.  */
  75 static int status_only = 0;
  76
  77 /* With --check, print a message to standard error warning about each
  78    improperly formatted MD5 checksum line.  */
  79 static int warn = 0;
  80
  81 /* The name this program was run with.  */
  82 char *program_name;
  83
  84 static const struct option long_options[] =
  85 {
  86   { "binary", no_argument, 0, 'b' },
  87   { "check", no_argument, 0, 'c' },
  88   { "status", no_argument, 0, 2 },
  89   { "string", required_argument, 0, 1 },
  90   { "text", no_argument, 0, 't' },
  91   { "warn", no_argument, 0, 'w' },
  92   { NULL, 0, NULL, 0 }
  93 };
  94
  95 static void
  96 usage (int status)
  97 {
  98   if (status != 0)
  99     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 100              program_name);
 101   else
 102     {
 103       printf (_("\
 104 Usage: %s [OPTION] [FILE]...\n\
 105   or:  %s [OPTION] --check [FILE]\n\
 106 Print or check MD5 checksums.\n\
 107 With no FILE, or when FILE is -, read standard input.\n\
 108 \n\
 109   -b, --binary            read files in binary mode\n\
 110   -c, --check             check MD5 sums against given list\n\
 111   -t, --text              read files in text mode (default)\n\
 112 \n\
 113 The following two options are useful only when verifying checksums:\n\
 114       --status            don't output anything, status code shows success\n\
 115   -w, --warn              warn about improperly formated MD5 checksum lines\n\
 116 \n\
 117       --help              display this help and exit\n\
 118       --version           output version information and exit\n\
 119 \n\
 120 The sums are computed as described in RFC 1321.  When checking, the input\n\
 121 should be a former output of this program.  The default mode is to print\n\
 122 a line with checksum, a character indicating type (`*' for binary, ` ' for\n\
 123 text), and name for each FILE.\n"),
 124               program_name, program_name, program_name);
 125       puts (_("\nReport bugs to <textutils-bugs@gnu.org>."));
 126     }
 127
 128   exit (status == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
 129 }
 130
 131 static int
 132 split_3 (char *s, size_t s_len, unsigned char **u, int *binary, char **w)
 133 {
 134   size_t i;
 135   int filename_has_newline = 0;
 136
 137 #define ISWHITE(c) ((c) == ' ' || (c) == '\t')
 138
 139   i = 0;
 140   while (ISWHITE (s[i]))
 141     ++i;
 142
 143   /* The line must have at least 35 (36 if the first is a backslash)
 144      more characters to contain correct message digest information.
 145      Ignore this line if it is too short.  */
 146   if (!(s_len - i >= MIN_DIGEST_LINE_LENGTH
 147         || (s[i] == '\\' && s_len - i >= 1 + MIN_DIGEST_LINE_LENGTH)))
 148     return 1;
 149
 150   if (s[i] == '\\')
 151     {
 152       ++i;
 153       filename_has_newline = 1;
 154     }
 155   *u = &s[i];
 156
 157   /* The first field has to be the 32-character hexadecimal
 158      representation of the message digest.  If it is not followed
 159      immediately by a white space it's an error.  */
 160   i += 32;
 161   if (!ISWHITE (s[i]))
 162     return 1;
 163
 164   s[i++] = '\0';
 165
 166   if (s[i] != ' ' && s[i] != '*')
 167     return 1;
 168   *binary = (s[i++] == '*');
 169
 170   /* All characters between the type indicator and end of line are
 171      significant -- that includes leading and trailing white space.  */
 172   *w = &s[i];
 173
 174   if (filename_has_newline)
 175     {
 176       /* Translate each `\n' string in the file name to a NEWLINE,
 177          and each `\\' string to a backslash.  */
 178
 179       char *dst = &s[i];
 180
 181       while (i < s_len)
 182         {
 183           switch (s[i])
 184             {
 185             case '\\':
 186               if (i == s_len - 1)
 187                 {
 188                   /* A valid line does not end with a backslash.  */
 189                   return 1;
 190                 }
 191               ++i;
 192               switch (s[i++])
 193                 {
 194                 case 'n':
 195                   *dst++ = '\n';
 196                   break;
 197                 case '\\':
 198                   *dst++ = '\\';
 199                   break;
 200                 default:
 201                   /* Only `\' or `n' may follow a backslash.  */
 202                   return 1;
 203                 }
 204               break;
 205
 206             case '\0':
 207               /* The file name may not contain a NUL.  */
 208               return 1;
 209               break;
 210
 211             default:
 212               *dst++ = s[i++];
 213               break;
 214             }
 215         }
 216       *dst = '\0';
 217     }
 218   return 0;
 219 }
 220
 221 static int
 222 hex_digits (unsigned char const *s)
 223 {
 224   while (*s)
 225     {
 226       if (!ISXDIGIT (*s))
 227         return 0;
 228       ++s;
 229     }
 230   return 1;
 231 }
 232
 233 /* An interface to md5_stream.  Operate on FILENAME (it may be "-") and
 234    put the result in *MD5_RESULT.  Return non-zero upon failure, zero
 235    to indicate success.  */
 236
 237 static int
 238 md5_file (const char *filename, int binary, unsigned char *md5_result)
 239 {
 240   FILE *fp;
 241   int err;
 242
 243   if (STREQ (filename, "-"))
 244     {
 245       have_read_stdin = 1;
 246       fp = stdin;
 247     }
 248   else
 249     {
 250       /* OPENOPTS is a macro.  It varies with the system.
 251          Some systems distinguish between internal and
 252          external text representations.  */
 253
 254       fp = fopen (filename, OPENOPTS (binary));
 255       if (fp == NULL)
 256         {
 257           error (0, errno, "%s", filename);
 258           return 1;
 259         }
 260     }
 261
 262   err = md5_stream (fp, md5_result);
 263   if (err)
 264     {
 265       error (0, errno, "%s", filename);
 266       if (fp != stdin)
 267         fclose (fp);
 268       return 1;
 269     }
 270
 271   if (fp != stdin && fclose (fp) == EOF)
 272     {
 273       error (0, errno, "%s", filename);
 274       return 1;
 275     }
 276
 277   return 0;
 278 }
 279
 280 static int
 281 md5_check (const char *checkfile_name)
 282 {
 283   FILE *checkfile_stream;
 284   int n_properly_formated_lines = 0;
 285   int n_mismatched_checksums = 0;
 286   int n_open_or_read_failures = 0;
 287   unsigned char md5buffer[16];
 288   size_t line_number;
 289   char *line;
 290   size_t line_chars_allocated;
 291
 292   if (STREQ (checkfile_name, "-"))
 293     {
 294       have_read_stdin = 1;
 295       checkfile_name = _("standard input");
 296       checkfile_stream = stdin;
 297     }
 298   else
 299     {
 300       checkfile_stream = fopen (checkfile_name, "r");
 301       if (checkfile_stream == NULL)
 302         {
 303           error (0, errno, "%s", checkfile_name);
 304           return 1;
 305         }
 306     }
 307
 308   line_number = 0;
 309   line = NULL;
 310   line_chars_allocated = 0;
 311   do
 312     {
 313       char *filename;
 314       int binary;
 315       unsigned char *md5num;
 316       int err;
 317       int line_length;
 318
 319       ++line_number;
 320
 321       line_length = getline (&line, &line_chars_allocated, checkfile_stream);
 322       if (line_length <= 0)
 323         break;
 324
 325       /* Ignore comment lines, which begin with a '#' character.  */
 326       if (line[0] == '#')
 327         continue;
 328
 329       /* Remove any trailing newline.  */
 330       if (line[line_length - 1] == '\n')
 331         line[--line_length] = '\0';
 332
 333       err = split_3 (line, line_length, &md5num, &binary, &filename);
 334       if (err || !hex_digits (md5num))
 335         {
 336           if (warn)
 337             {
 338               error (0, 0,
 339                      _("%s: %lu: improperly formatted MD5 checksum line"),
 340                      checkfile_name, (unsigned long) line_number);
 341             }
 342         }
 343       else
 344         {
 345           static const char bin2hex[] = { '0', '1', '2', '3',
 346                                           '4', '5', '6', '7',
 347                                           '8', '9', 'a', 'b',
 348                                           'c', 'd', 'e', 'f' };
 349           int fail;
 350
 351           ++n_properly_formated_lines;
 352
 353           fail = md5_file (filename, binary, md5buffer);
 354
 355           if (fail)
 356             {
 357               ++n_open_or_read_failures;
 358               if (!status_only)
 359                 {
 360                   printf (_("%s: FAILED open or read\n"), filename);
 361                   fflush (stdout);
 362                 }
 363             }
 364           else
 365             {
 366               size_t cnt;
 367               /* Compare generated binary number with text representation
 368                  in check file.  Ignore case of hex digits.  */
 369               for (cnt = 0; cnt < 16; ++cnt)
 370                 {
 371                   if (TOLOWER (md5num[2 * cnt]) != bin2hex[md5buffer[cnt] >> 4]
 372                       || (TOLOWER (md5num[2 * cnt + 1])
 373                           != (bin2hex[md5buffer[cnt] & 0xf])))
 374                     break;
 375                 }
 376               if (cnt != 16)
 377                 ++n_mismatched_checksums;
 378
 379               if (!status_only)
 380                 {
 381                   printf ("%s: %s\n", filename,
 382                           (cnt != 16 ? _("FAILED") : _("OK")));
 383                   fflush (stdout);
 384                 }
 385             }
 386         }
 387     }
 388   while (!feof (checkfile_stream) && !ferror (checkfile_stream));
 389
 390   if (line)
 391     free (line);
 392
 393   if (ferror (checkfile_stream))
 394     {
 395       error (0, 0, _("%s: read error"), checkfile_name);
 396       return 1;
 397     }
 398
 399   if (checkfile_stream != stdin && fclose (checkfile_stream) == EOF)
 400     {
 401       error (0, errno, "%s", checkfile_name);
 402       return 1;
 403     }
 404
 405   if (n_properly_formated_lines == 0)
 406     {
 407       /* Warn if no tests are found.  */
 408       error (0, 0, _("%s: no properly formatted MD5 checksum lines found"),
 409              checkfile_name);
 410     }
 411   else
 412     {
 413       if (!status_only)
 414         {
 415           int n_computed_checkums = (n_properly_formated_lines
 416                                      - n_open_or_read_failures);
 417
 418           if (n_open_or_read_failures > 0)
 419             {
 420               error (0, 0,
 421                    _("WARNING: %d of %d listed %s could not be read\n"),
 422                      n_open_or_read_failures, n_properly_formated_lines,
 423                      (n_properly_formated_lines == 1
 424                       ? _("file") : _("files")));
 425             }
 426
 427           if (n_mismatched_checksums > 0)
 428             {
 429               error (0, 0,
 430                    _("WARNING: %d of %d computed %s did NOT match"),
 431                      n_mismatched_checksums, n_computed_checkums,
 432                      (n_computed_checkums == 1
 433                       ? _("checksum") : _("checksums")));
 434             }
 435         }
 436     }
 437
 438   return ((n_properly_formated_lines > 0 && n_mismatched_checksums == 0
 439            && n_open_or_read_failures == 0) ? 0 : 1);
 440 }
 441
 442 int
 443 main (int argc, char **argv)
 444 {
 445   unsigned char md5buffer[16];
 446   int do_check = 0;
 447   int opt;
 448   char **string = NULL;
 449   size_t n_strings = 0;
 450   size_t err = 0;
 451   int file_type_specified = 0;
 452
 453   /* Text is default of the Plumb/Lankester format.  */
 454   int binary = 0;
 455
 456   /* Setting values of global variables.  */
 457   program_name = argv[0];
 458   setlocale (LC_ALL, "");
 459   bindtextdomain (PACKAGE, LOCALEDIR);
 460   textdomain (PACKAGE);
 461
 462   parse_long_options (argc, argv, "md5sum", GNU_PACKAGE, VERSION, usage);
 463
 464   while ((opt = getopt_long (argc, argv, "bctw", long_options, NULL)) != -1)
 465     switch (opt)
 466       {
 467       case 0:                   /* long option */
 468         break;
 469       case 1: /* --string */
 470         {
 471           if (string == NULL)
 472             string = (char **) xmalloc ((argc - 1) * sizeof (char *));
 473
 474           if (optarg == NULL)
 475             optarg = "";
 476           string[n_strings++] = optarg;
 477         }
 478         break;
 479       case 'b':
 480         file_type_specified = 1;
 481         binary = 1;
 482         break;
 483       case 'c':
 484         do_check = 1;
 485         break;
 486       case 2:
 487         status_only = 1;
 488         warn = 0;
 489         break;
 490       case 't':
 491         file_type_specified = 1;
 492         binary = 0;
 493         break;
 494       case 'w':
 495         status_only = 0;
 496         warn = 1;
 497         break;
 498       default:
 499         usage (EXIT_FAILURE);
 500       }
 501
 502   if (file_type_specified && do_check)
 503     {
 504       error (0, 0, _("the --binary and --text options are meaningless when \
 505 verifying checksums"));
 506       usage (EXIT_FAILURE);
 507     }
 508
 509   if (n_strings > 0 && do_check)
 510     {
 511       error (0, 0,
 512              _("the --string and --check options are mutually exclusive"));
 513       usage (EXIT_FAILURE);
 514     }
 515
 516   if (status_only && !do_check)
 517     {
 518       error (0, 0,
 519        _("the --status option is meaningful only when verifying checksums"));
 520       usage (EXIT_FAILURE);
 521     }
 522
 523   if (warn && !do_check)
 524     {
 525       error (0, 0,
 526        _("the --warn option is meaningful only when verifying checksums"));
 527       usage (EXIT_FAILURE);
 528     }
 529
 530   if (n_strings > 0)
 531     {
 532       size_t i;
 533
 534       if (optind < argc)
 535         {
 536           error (0, 0, _("no files may be specified when using --string"));
 537           usage (EXIT_FAILURE);
 538         }
 539       for (i = 0; i < n_strings; ++i)
 540         {
 541           size_t cnt;
 542           md5_buffer (string[i], strlen (string[i]), md5buffer);
 543
 544           for (cnt = 0; cnt < 16; ++cnt)
 545             printf ("%02x", md5buffer[cnt]);
 546
 547           printf ("  \"%s\"\n", string[i]);
 548         }
 549     }
 550   else if (do_check)
 551     {
 552       if (optind + 1 < argc)
 553         {
 554           error (0, 0,
 555                  _("only one argument may be specified when using --check"));
 556           usage (EXIT_FAILURE);
 557         }
 558
 559       err = md5_check ((optind == argc) ? "-" : argv[optind]);
 560     }
 561   else
 562     {
 563       if (optind == argc)
 564         argv[argc++] = "-";
 565
 566       for (; optind < argc; ++optind)
 567         {
 568           int fail;
 569           char *file = argv[optind];
 570
 571           fail = md5_file (file, binary, md5buffer);
 572           err |= fail;
 573           if (!fail)
 574             {
 575               size_t i;
 576
 577               /* Output a leading backslash if the file name contains
 578                  a newline.  */
 579               if (strchr (file, '\n'))
 580                 putchar ('\\');
 581
 582               for (i = 0; i < 16; ++i)
 583                 printf ("%02x", md5buffer[i]);
 584
 585               putchar (' ');
 586               if (binary)
 587                 putchar ('*');
 588               else
 589                 putchar (' ');
 590
 591               /* Translate each NEWLINE byte to the string, "\\n",
 592                  and each backslash to "\\\\".  */
 593               for (i = 0; i < strlen (file); ++i)
 594                 {
 595                   switch (file[i])
 596                     {
 597                     case '\n':
 598                       fputs ("\\n", stdout);
 599                       break;
 600
 601                     case '\\':
 602                       fputs ("\\\\", stdout);
 603                       break;
 604
 605                     default:
 606                       putchar (file[i]);
 607                       break;
 608                     }
 609                 }
 610               putchar ('\n');
 611             }
 612         }
 613     }
 614
 615   if (fclose (stdout) == EOF)
 616     error (EXIT_FAILURE, errno, _("write error"));
 617
 618   if (have_read_stdin && fclose (stdin) == EOF)
 619     error (EXIT_FAILURE, errno, _("standard input"));
 620
 621   exit (err == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
 622 }