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, 1996 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 #if UNIX || __UNIX__ || unix || __unix__ || _POSIX_VERSION
  38 # define OPENOPTS(BINARY) "r"
  39 #else
  40 # ifdef MSDOS
  41 #  define TEXT1TO1 "rb"
  42 #  define TEXTCNVT "r"
  43 # else
  44 #  if defined VMS
  45 #   define TEXT1TO1 "rb", "ctx=stm"
  46 #   define TEXTCNVT "r", "ctx=stm"
  47 #  else
  48     /* The following line is intended to evoke an error.
  49        Using #error is not portable enough.  */
  50     "Cannot determine system type."
  51 #  endif
  52 # endif
  53 # define OPENOPTS(BINARY) ((BINARY) != 0 ? TEXT1TO1 : TEXTCNVT)
  54 #endif
  55
  56 #if _LIBC || STDC_HEADERS
  57 # define TOLOWER(c) tolower (c)
  58 #else
  59 # define TOLOWER(c) (ISUPPER (c) ? tolower (c) : (c))
  60 #endif
  61
  62 /* The minimum length of a valid digest line in a file produced
  63    by `md5sum FILE' and read by `md5sum --check'.  This length does
  64    not include any newline character at the end of a line.  */
  65 #define MIN_DIGEST_LINE_LENGTH (32 /* message digest length */ \
  66                                 + 2 /* blank and binary indicator */ \
  67                                 + 1 /* minimum filename length */ )
  68
  69 /* Nonzero if any of the files read were the standard input. */
  70 static int have_read_stdin;
  71
  72 /* With --check, don't generate any output.
  73    The exit code indicates success or failure.  */
  74 static int status_only = 0;
  75
  76 /* With --check, print a message to standard error warning about each
  77    improperly formatted MD5 checksum line.  */
  78 static int warn = 0;
  79
  80 /* The name this program was run with.  */
  81 char *program_name;
  82
  83 static const struct option long_options[] =
  84 {
  85   { "binary", no_argument, 0, 'b' },
  86   { "check", no_argument, 0, 'c' },
  87   { "status", no_argument, 0, 2 },
  88   { "string", required_argument, 0, 1 },
  89   { "text", no_argument, 0, 't' },
  90   { "warn", no_argument, 0, 'w' },
  91   { NULL, 0, NULL, 0 }
  92 };
  93
  94 char *xmalloc ();
  95
  96 static void
  97 usage (int status)
  98 {
  99   if (status != 0)
 100     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 101              program_name);
 102   else
 103     {
 104       printf (_("\
 105 Usage: %s [OPTION] [FILE]...\n\
 106   or:  %s [OPTION] --check [FILE]\n\
 107 Print or check MD5 checksums.\n\
 108 With no FILE, or when FILE is -, read standard input.\n\
 109 \n\
 110   -b, --binary            read files in binary mode\n\
 111   -c, --check             check MD5 sums against given list\n\
 112   -t, --text              read files in text mode (default)\n\
 113 \n\
 114 The following two options are useful only when verifying checksums:\n\
 115       --status            don't output anything, status code shows success\n\
 116   -w, --warn              warn about improperly formated MD5 checksum lines\n\
 117 \n\
 118       --help              display this help and exit\n\
 119       --version           output version information and exit\n\
 120 \n\
 121 The sums are computed as described in RFC 1321.  When checking, the input\n\
 122 should be a former output of this program.  The default mode is to print\n\
 123 a line with checksum, a character indicating type (`*' for binary, ` ' for\n\
 124 text), and name for each FILE.\n"),
 125               program_name, program_name, program_name);
 126       puts (_("\nReport bugs to textutils-bugs@gnu.ai.mit.edu"));
 127     }
 128
 129   exit (status == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
 130 }
 131
 132 static int
 133 split_3 (char *s, size_t s_len, char **u, int *binary, char **w)
 134 {
 135   size_t i;
 136   int filename_has_newline = 0;
 137
 138 #define ISWHITE(c) ((c) == ' ' || (c) == '\t')
 139
 140   i = 0;
 141   while (ISWHITE (s[i]))
 142     ++i;
 143
 144   /* The line must have at least 35 (36 if the first is a backslash)
 145      more characters to contain correct message digest information.
 146      Ignore this line if it is too short.  */
 147   if (!(s_len - i >= MIN_DIGEST_LINE_LENGTH
 148         || (s[i] == '\\' && s_len - i >= 1 + MIN_DIGEST_LINE_LENGTH)))
 149     return 1;
 150
 151   if (s[i] == '\\')
 152     {
 153       ++i;
 154       filename_has_newline = 1;
 155     }
 156   *u = &s[i];
 157
 158   /* The first field has to be the 32-character hexadecimal
 159      representation of the message digest.  If it is not followed
 160      immediately by a white space it's an error.  */
 161   i += 32;
 162   if (!ISWHITE (s[i]))
 163     return 1;
 164
 165   s[i++] = '\0';
 166
 167   if (s[i] != ' ' && s[i] != '*')
 168     return 1;
 169   *binary = (s[i++] == '*');
 170
 171   /* All characters between the type indicator and end of line are
 172      significant -- that includes leading and trailing white space.  */
 173   *w = &s[i];
 174
 175   if (filename_has_newline)
 176     {
 177       /* Translate each `\n' string in the file name to a NEWLINE,
 178          and each `\\' string to a backslash.  */
 179
 180       char *dst = &s[i];
 181
 182       while (i < s_len)
 183         {
 184           switch (s[i])
 185             {
 186             case '\\':
 187               if (i == s_len - 1)
 188                 {
 189                   /* A valid line does not end with a backslash.  */
 190                   return 1;
 191                 }
 192               ++i;
 193               switch (s[i++])
 194                 {
 195                 case 'n':
 196                   *dst++ = '\n';
 197                   break;
 198                 case '\\':
 199                   *dst++ = '\\';
 200                   break;
 201                 default:
 202                   /* Only `\' or `n' may follow a backslash.  */
 203                   return 1;
 204                 }
 205               break;
 206
 207             case '\0':
 208               /* The file name may not contain a NUL.  */
 209               return 1;
 210               break;
 211
 212             default:
 213               *dst++ = s[i++];
 214               break;
 215             }
 216         }
 217       *dst = '\0';
 218     }
 219   return 0;
 220 }
 221
 222 static int
 223 hex_digits (const char *s)
 224 {
 225   while (*s)
 226     {
 227       if (!ISXDIGIT (*s))
 228         return 0;
 229       ++s;
 230     }
 231   return 1;
 232 }
 233
 234 /* An interface to md5_stream.  Operate on FILENAME (it may be "-") and
 235    put the result in *MD5_RESULT.  Return non-zero upon failure, zero
 236    to indicate success.  */
 237
 238 static int
 239 md5_file (const char *filename, int binary, unsigned char *md5_result)
 240 {
 241   FILE *fp;
 242   int err;
 243
 244   if (strcmp (filename, "-") == 0)
 245     {
 246       have_read_stdin = 1;
 247       fp = stdin;
 248     }
 249   else
 250     {
 251       /* OPENOPTS is a macro.  It varies with the system.
 252          Some systems distinguish between internal and
 253          external text representations.  */
 254
 255       fp = fopen (filename, OPENOPTS (binary));
 256       if (fp == NULL)
 257         {
 258           error (0, errno, "%s", filename);
 259           return 1;
 260         }
 261     }
 262
 263   err = md5_stream (fp, md5_result);
 264   if (err)
 265     {
 266       error (0, errno, "%s", filename);
 267       if (fp != stdin)
 268         fclose (fp);
 269       return 1;
 270     }
 271
 272   if (fp != stdin && fclose (fp) == EOF)
 273     {
 274       error (0, errno, "%s", filename);
 275       return 1;
 276     }
 277
 278   return 0;
 279 }
 280
 281 static int
 282 md5_check (const char *checkfile_name)
 283 {
 284   FILE *checkfile_stream;
 285   int n_properly_formated_lines = 0;
 286   int n_mismatched_checksums = 0;
 287   int n_open_or_read_failures = 0;
 288   unsigned char md5buffer[16];
 289   size_t line_number;
 290   char *line;
 291   size_t line_chars_allocated;
 292
 293   if (strcmp (checkfile_name, "-") == 0)
 294     {
 295       have_read_stdin = 1;
 296       checkfile_name = _("standard input");
 297       checkfile_stream = stdin;
 298     }
 299   else
 300     {
 301       checkfile_stream = fopen (checkfile_name, "r");
 302       if (checkfile_stream == NULL)
 303         {
 304           error (0, errno, "%s", checkfile_name);
 305           return 1;
 306         }
 307     }
 308
 309   line_number = 0;
 310   line = NULL;
 311   line_chars_allocated = 0;
 312   do
 313     {
 314       char *filename;
 315       int binary;
 316       char *md5num;
 317       int err;
 318       int line_length;
 319
 320       ++line_number;
 321
 322       line_length = getline (&line, &line_chars_allocated, checkfile_stream);
 323       if (line_length <= 0)
 324         break;
 325
 326       /* Ignore comment lines, which begin with a '#' character.  */
 327       if (line[0] == '#')
 328         continue;
 329
 330       /* Remove any trailing newline.  */
 331       if (line[line_length - 1] == '\n')
 332         line[--line_length] = '\0';
 333
 334       err = split_3 (line, line_length, &md5num, &binary, &filename);
 335       if (err || !hex_digits (md5num))
 336         {
 337           if (warn)
 338             {
 339               error (0, 0,
 340                      _("%s: %lu: improperly formatted MD5 checksum line"),
 341                      checkfile_name, (unsigned long) line_number);
 342             }
 343         }
 344       else
 345         {
 346           static const char bin2hex[] = { '0', '1', '2', '3',
 347                                           '4', '5', '6', '7',
 348                                           '8', '9', 'a', 'b',
 349                                           'c', 'd', 'e', 'f' };
 350           int fail;
 351
 352           ++n_properly_formated_lines;
 353
 354           fail = md5_file (filename, binary, md5buffer);
 355
 356           if (fail)
 357             {
 358               ++n_open_or_read_failures;
 359               if (!status_only)
 360                 {
 361                   printf (_("%s: FAILED open or read\n"), filename);
 362                   fflush (stdout);
 363                 }
 364             }
 365           else
 366             {
 367               size_t cnt;
 368               /* Compare generated binary number with text representation
 369                  in check file.  Ignore case of hex digits.  */
 370               for (cnt = 0; cnt < 16; ++cnt)
 371                 {
 372                   if (TOLOWER (md5num[2 * cnt]) != bin2hex[md5buffer[cnt] >> 4]
 373                       || (TOLOWER (md5num[2 * cnt + 1])
 374                           != (bin2hex[md5buffer[cnt] & 0xf])))
 375                     break;
 376                 }
 377               if (cnt != 16)
 378                 ++n_mismatched_checksums;
 379
 380               if (!status_only)
 381                 {
 382                   printf ("%s: %s\n", filename,
 383                           (cnt != 16 ? _("FAILED") : _("OK")));
 384                   fflush (stdout);
 385                 }
 386             }
 387         }
 388     }
 389   while (!feof (checkfile_stream) && !ferror (checkfile_stream));
 390
 391   if (line)
 392     free (line);
 393
 394   if (ferror (checkfile_stream))
 395     {
 396       error (0, 0, _("%s: read error"), checkfile_name);
 397       return 1;
 398     }
 399
 400   if (checkfile_stream != stdin && fclose (checkfile_stream) == EOF)
 401     {
 402       error (0, errno, "%s", checkfile_name);
 403       return 1;
 404     }
 405
 406   if (n_properly_formated_lines == 0)
 407     {
 408       /* Warn if no tests are found.  */
 409       error (0, 0, _("%s: no properly formatted MD5 checksum lines found"),
 410              checkfile_name);
 411     }
 412   else
 413     {
 414       if (!status_only)
 415         {
 416           int n_computed_checkums = (n_properly_formated_lines
 417                                      - n_open_or_read_failures);
 418
 419           if (n_open_or_read_failures > 0)
 420             {
 421               error (0, 0,
 422                    _("WARNING: %d of %d listed %s could not be read\n"),
 423                      n_open_or_read_failures, n_properly_formated_lines,
 424                      (n_properly_formated_lines == 1
 425                       ? _("file") : _("files")));
 426             }
 427
 428           if (n_mismatched_checksums > 0)
 429             {
 430               error (0, 0,
 431                    _("WARNING: %d of %d computed %s did NOT match"),
 432                      n_mismatched_checksums, n_computed_checkums,
 433                      (n_computed_checkums == 1
 434                       ? _("checksum") : _("checksums")));
 435             }
 436         }
 437     }
 438
 439   return ((n_properly_formated_lines > 0 && n_mismatched_checksums == 0
 440            && n_open_or_read_failures == 0) ? 0 : 1);
 441 }
 442
 443 int
 444 main (int argc, char **argv)
 445 {
 446   unsigned char md5buffer[16];
 447   int do_check = 0;
 448   int opt;
 449   char **string = NULL;
 450   size_t n_strings = 0;
 451   size_t i;
 452   size_t err = 0;
 453   int file_type_specified = 0;
 454
 455   /* Text is default of the Plumb/Lankester format.  */
 456   int binary = 0;
 457
 458   /* Setting values of global variables.  */
 459   program_name = argv[0];
 460   setlocale (LC_ALL, "");
 461   bindtextdomain (PACKAGE, LOCALEDIR);
 462   textdomain (PACKAGE);
 463
 464   parse_long_options (argc, argv, "md5sum", GNU_PACKAGE, VERSION, usage);
 465
 466   while ((opt = getopt_long (argc, argv, "bctw", long_options, NULL))
 467          != EOF)
 468     switch (opt)
 469       {
 470       case 0:                   /* long option */
 471         break;
 472       case 1: /* --string */
 473         {
 474           if (string == NULL)
 475             string = (char **) xmalloc ((argc - 1) * sizeof (char *));
 476
 477           if (optarg == NULL)
 478             optarg = "";
 479           string[n_strings++] = optarg;
 480         }
 481         break;
 482       case 'b':
 483         file_type_specified = 1;
 484         binary = 1;
 485         break;
 486       case 'c':
 487         do_check = 1;
 488         break;
 489       case 2:
 490         status_only = 1;
 491         warn = 0;
 492         break;
 493       case 't':
 494         file_type_specified = 1;
 495         binary = 0;
 496         break;
 497       case 'w':
 498         status_only = 0;
 499         warn = 1;
 500         break;
 501       default:
 502         usage (EXIT_FAILURE);
 503       }
 504
 505   if (file_type_specified && do_check)
 506     {
 507       error (0, 0, _("the --binary and --text options are meaningless when \
 508 verifying checksums"));
 509       usage (EXIT_FAILURE);
 510     }
 511
 512   if (n_strings > 0 && do_check)
 513     {
 514       error (0, 0,
 515              _("the --string and --check options are mutually exclusive"));
 516       usage (EXIT_FAILURE);
 517     }
 518
 519   if (status_only && !do_check)
 520     {
 521       error (0, 0,
 522        _("the --status option is meaningful only when verifying checksums"));
 523       usage (EXIT_FAILURE);
 524     }
 525
 526   if (warn && !do_check)
 527     {
 528       error (0, 0,
 529        _("the --warn option is meaningful only when verifying checksums"));
 530       usage (EXIT_FAILURE);
 531     }
 532
 533   if (n_strings > 0)
 534     {
 535       if (optind < argc)
 536         {
 537           error (0, 0, _("no files may be specified when using --string"));
 538           usage (EXIT_FAILURE);
 539         }
 540       for (i = 0; i < n_strings; ++i)
 541         {
 542           size_t cnt;
 543           md5_buffer (string[i], strlen (string[i]), md5buffer);
 544
 545           for (cnt = 0; cnt < 16; ++cnt)
 546             printf ("%02x", md5buffer[cnt]);
 547
 548           printf ("  \"%s\"\n", string[i]);
 549         }
 550     }
 551   else if (do_check)
 552     {
 553       if (optind + 1 < argc)
 554         {
 555           error (0, 0,
 556                  _("only one argument may be specified when using --check"));
 557           usage (EXIT_FAILURE);
 558         }
 559
 560       err = md5_check ((optind == argc) ? "-" : argv[optind]);
 561     }
 562   else
 563     {
 564       if (optind == argc)
 565         argv[argc++] = "-";
 566
 567       for (; optind < argc; ++optind)
 568         {
 569           int fail;
 570           char *file = argv[optind];
 571
 572           fail = md5_file (file, binary, md5buffer);
 573           err |= fail;
 574           if (!fail)
 575             {
 576               size_t i;
 577
 578               /* Output a leading backslash if the file name contains
 579                  a newline.  */
 580               if (strchr (file, '\n'))
 581                 putchar ('\\');
 582
 583               for (i = 0; i < 16; ++i)
 584                 printf ("%02x", md5buffer[i]);
 585
 586               putchar (' ');
 587               if (binary)
 588                 putchar ('*');
 589               else
 590                 putchar (' ');
 591
 592               /* Translate each NEWLINE byte to the string, "\\n",
 593                  and each backslash to "\\\\".  */
 594               for (i = 0; i < strlen (file); ++i)
 595                 {
 596                   switch (file[i])
 597                     {
 598                     case '\n':
 599                       fputs ("\\n", stdout);
 600                       break;
 601
 602                     case '\\':
 603                       fputs ("\\\\", stdout);
 604                       break;
 605
 606                     default:
 607                       putchar (file[i]);
 608                       break;
 609                     }
 610                 }
 611               putchar ('\n');
 612             }
 613         }
 614     }
 615
 616   if (fclose (stdout) == EOF)
 617     error (EXIT_FAILURE, errno, _("write error"));
 618
 619   if (have_read_stdin && fclose (stdin) == EOF)
 620     error (EXIT_FAILURE, errno, _("standard input"));
 621
 622   exit (err == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
 623 }