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) 1995-1999 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 "md5.h"
  30 #include "getline.h"
  31 #include "system.h"
  32 #include "error.h"
  33
  34 /* The official name of this program (e.g., no `g' prefix).  */
  35 #define PROGRAM_NAME "md5sum"
  36
  37 #define AUTHORS "Ulrich Drepper"
  38
  39 /* Most systems do not distinguish between external and internal
  40    text representations.  */
  41 /* FIXME: This begs for an autoconf test.  */
  42 #if O_BINARY
  43 # define OPENOPTS(BINARY) ((BINARY) != 0 ? TEXT1TO1 : TEXTCNVT)
  44 # define TEXT1TO1 "rb"
  45 # define TEXTCNVT "r"
  46 #else
  47 # if defined VMS
  48 #  define OPENOPTS(BINARY) ((BINARY) != 0 ? TEXT1TO1 : TEXTCNVT)
  49 #  define TEXT1TO1 "rb", "ctx=stm"
  50 #  define TEXTCNVT "r", "ctx=stm"
  51 # else
  52 #  if UNIX || __UNIX__ || unix || __unix__ || _POSIX_VERSION
  53 #   define OPENOPTS(BINARY) "r"
  54 #  else
  55     /* The following line is intended to evoke an error.
  56        Using #error is not portable enough.  */
  57     "Cannot determine system type."
  58 #  endif
  59 # endif
  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   { GETOPT_HELP_OPTION_DECL },
  92   { GETOPT_VERSION_OPTION_DECL },
  93   { NULL, 0, NULL, 0 }
  94 };
  95
  96 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 (default on DOS/Windows)\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);
 126       puts (_("\nReport bugs to <bug-textutils@gnu.org>."));
 127     }
 128
 129   exit (status == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
 130 }
 131
 132 static int
 133 split_3 (char *s, size_t s_len, unsigned char **u, int *binary, char **w)
 134 {
 135   size_t i;
 136   int escaped_filename = 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       escaped_filename = 1;
 155     }
 156   *u = (unsigned char *) &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 (escaped_filename)
 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 (unsigned char const *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 (STREQ (filename, "-"))
 245     {
 246       have_read_stdin = 1;
 247       fp = stdin;
 248 #if O_BINARY
 249       /* If we need binary reads from a pipe or redirected stdin, we need
 250          to switch it to BINARY mode here, since stdin is already open.  */
 251       if (binary)
 252         SET_BINARY (fileno (stdin));
 253 #endif
 254     }
 255   else
 256     {
 257       /* OPENOPTS is a macro.  It varies with the system.
 258          Some systems distinguish between internal and
 259          external text representations.  */
 260
 261       fp = fopen (filename, OPENOPTS (binary));
 262       if (fp == NULL)
 263         {
 264           error (0, errno, "%s", filename);
 265           return 1;
 266         }
 267     }
 268
 269   err = md5_stream (fp, md5_result);
 270   if (err)
 271     {
 272       error (0, errno, "%s", filename);
 273       if (fp != stdin)
 274         fclose (fp);
 275       return 1;
 276     }
 277
 278   if (fp != stdin && fclose (fp) == EOF)
 279     {
 280       error (0, errno, "%s", filename);
 281       return 1;
 282     }
 283
 284   return 0;
 285 }
 286
 287 static int
 288 md5_check (const char *checkfile_name)
 289 {
 290   FILE *checkfile_stream;
 291   int n_properly_formated_lines = 0;
 292   int n_mismatched_checksums = 0;
 293   int n_open_or_read_failures = 0;
 294   unsigned char md5buffer[16];
 295   size_t line_number;
 296   char *line;
 297   size_t line_chars_allocated;
 298
 299   if (STREQ (checkfile_name, "-"))
 300     {
 301       have_read_stdin = 1;
 302       checkfile_name = _("standard input");
 303       checkfile_stream = stdin;
 304     }
 305   else
 306     {
 307       checkfile_stream = fopen (checkfile_name, "r");
 308       if (checkfile_stream == NULL)
 309         {
 310           error (0, errno, "%s", checkfile_name);
 311           return 1;
 312         }
 313     }
 314
 315   line_number = 0;
 316   line = NULL;
 317   line_chars_allocated = 0;
 318   do
 319     {
 320       char *filename;
 321       int binary;
 322       unsigned char *md5num;
 323       int err;
 324       int line_length;
 325
 326       ++line_number;
 327
 328       line_length = getline (&line, &line_chars_allocated, checkfile_stream);
 329       if (line_length <= 0)
 330         break;
 331
 332       /* Ignore comment lines, which begin with a '#' character.  */
 333       if (line[0] == '#')
 334         continue;
 335
 336       /* Remove any trailing newline.  */
 337       if (line[line_length - 1] == '\n')
 338         line[--line_length] = '\0';
 339
 340       err = split_3 (line, line_length, &md5num, &binary, &filename);
 341       if (err || !hex_digits (md5num))
 342         {
 343           if (warn)
 344             {
 345               error (0, 0,
 346                      _("%s: %lu: improperly formatted MD5 checksum line"),
 347                      checkfile_name, (unsigned long) line_number);
 348             }
 349         }
 350       else
 351         {
 352           static const char bin2hex[] = { '0', '1', '2', '3',
 353                                           '4', '5', '6', '7',
 354                                           '8', '9', 'a', 'b',
 355                                           'c', 'd', 'e', 'f' };
 356           int fail;
 357
 358           ++n_properly_formated_lines;
 359
 360           fail = md5_file (filename, binary, md5buffer);
 361
 362           if (fail)
 363             {
 364               ++n_open_or_read_failures;
 365               if (!status_only)
 366                 {
 367                   printf (_("%s: FAILED open or read\n"), filename);
 368                   fflush (stdout);
 369                 }
 370             }
 371           else
 372             {
 373               size_t cnt;
 374               /* Compare generated binary number with text representation
 375                  in check file.  Ignore case of hex digits.  */
 376               for (cnt = 0; cnt < 16; ++cnt)
 377                 {
 378                   if (TOLOWER (md5num[2 * cnt]) != bin2hex[md5buffer[cnt] >> 4]
 379                       || (TOLOWER (md5num[2 * cnt + 1])
 380                           != (bin2hex[md5buffer[cnt] & 0xf])))
 381                     break;
 382                 }
 383               if (cnt != 16)
 384                 ++n_mismatched_checksums;
 385
 386               if (!status_only)
 387                 {
 388                   printf ("%s: %s\n", filename,
 389                           (cnt != 16 ? _("FAILED") : _("OK")));
 390                   fflush (stdout);
 391                 }
 392             }
 393         }
 394     }
 395   while (!feof (checkfile_stream) && !ferror (checkfile_stream));
 396
 397   if (line)
 398     free (line);
 399
 400   if (ferror (checkfile_stream))
 401     {
 402       error (0, 0, _("%s: read error"), checkfile_name);
 403       return 1;
 404     }
 405
 406   if (checkfile_stream != stdin && fclose (checkfile_stream) == EOF)
 407     {
 408       error (0, errno, "%s", checkfile_name);
 409       return 1;
 410     }
 411
 412   if (n_properly_formated_lines == 0)
 413     {
 414       /* Warn if no tests are found.  */
 415       error (0, 0, _("%s: no properly formatted MD5 checksum lines found"),
 416              checkfile_name);
 417     }
 418   else
 419     {
 420       if (!status_only)
 421         {
 422           int n_computed_checkums = (n_properly_formated_lines
 423                                      - n_open_or_read_failures);
 424
 425           if (n_open_or_read_failures > 0)
 426             {
 427               error (0, 0,
 428                    _("WARNING: %d of %d listed %s could not be read\n"),
 429                      n_open_or_read_failures, n_properly_formated_lines,
 430                      (n_properly_formated_lines == 1
 431                       ? _("file") : _("files")));
 432             }
 433
 434           if (n_mismatched_checksums > 0)
 435             {
 436               error (0, 0,
 437                    _("WARNING: %d of %d computed %s did NOT match"),
 438                      n_mismatched_checksums, n_computed_checkums,
 439                      (n_computed_checkums == 1
 440                       ? _("checksum") : _("checksums")));
 441             }
 442         }
 443     }
 444
 445   return ((n_properly_formated_lines > 0 && n_mismatched_checksums == 0
 446            && n_open_or_read_failures == 0) ? 0 : 1);
 447 }
 448
 449 int
 450 main (int argc, char **argv)
 451 {
 452   unsigned char md5buffer[16];
 453   int do_check = 0;
 454   int opt;
 455   char **string = NULL;
 456   size_t n_strings = 0;
 457   size_t err = 0;
 458   int file_type_specified = 0;
 459
 460 #if O_BINARY
 461   /* Binary is default on MSDOS, so the actual file contents
 462      are used in computation.  */
 463   int binary = 1;
 464 #else
 465   /* Text is default of the Plumb/Lankester format.  */
 466   int binary = 0;
 467 #endif
 468
 469   /* Setting values of global variables.  */
 470   program_name = argv[0];
 471   setlocale (LC_ALL, "");
 472   bindtextdomain (PACKAGE, LOCALEDIR);
 473   textdomain (PACKAGE);
 474
 475   while ((opt = getopt_long (argc, argv, "bctw", long_options, NULL)) != -1)
 476     switch (opt)
 477       {
 478       case 0:                   /* long option */
 479         break;
 480       case 1: /* --string */
 481         {
 482           if (string == NULL)
 483             string = (char **) xmalloc ((argc - 1) * sizeof (char *));
 484
 485           if (optarg == NULL)
 486             optarg = "";
 487           string[n_strings++] = optarg;
 488         }
 489         break;
 490       case 'b':
 491         file_type_specified = 1;
 492         binary = 1;
 493         break;
 494       case 'c':
 495         do_check = 1;
 496         break;
 497       case 2:
 498         status_only = 1;
 499         warn = 0;
 500         break;
 501       case 't':
 502         file_type_specified = 1;
 503         binary = 0;
 504         break;
 505       case 'w':
 506         status_only = 0;
 507         warn = 1;
 508         break;
 509       case_GETOPT_HELP_CHAR;
 510       case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 511       default:
 512         usage (EXIT_FAILURE);
 513       }
 514
 515   if (file_type_specified && do_check)
 516     {
 517       error (0, 0, _("the --binary and --text options are meaningless when \
 518 verifying checksums"));
 519       usage (EXIT_FAILURE);
 520     }
 521
 522   if (n_strings > 0 && do_check)
 523     {
 524       error (0, 0,
 525              _("the --string and --check options are mutually exclusive"));
 526       usage (EXIT_FAILURE);
 527     }
 528
 529   if (status_only && !do_check)
 530     {
 531       error (0, 0,
 532        _("the --status option is meaningful only when verifying checksums"));
 533       usage (EXIT_FAILURE);
 534     }
 535
 536   if (warn && !do_check)
 537     {
 538       error (0, 0,
 539        _("the --warn option is meaningful only when verifying checksums"));
 540       usage (EXIT_FAILURE);
 541     }
 542
 543   if (n_strings > 0)
 544     {
 545       size_t i;
 546
 547       if (optind < argc)
 548         {
 549           error (0, 0, _("no files may be specified when using --string"));
 550           usage (EXIT_FAILURE);
 551         }
 552       for (i = 0; i < n_strings; ++i)
 553         {
 554           size_t cnt;
 555           md5_buffer (string[i], strlen (string[i]), md5buffer);
 556
 557           for (cnt = 0; cnt < 16; ++cnt)
 558             printf ("%02x", md5buffer[cnt]);
 559
 560           printf ("  \"%s\"\n", string[i]);
 561         }
 562     }
 563   else if (do_check)
 564     {
 565       if (optind + 1 < argc)
 566         {
 567           error (0, 0,
 568                  _("only one argument may be specified when using --check"));
 569           usage (EXIT_FAILURE);
 570         }
 571
 572       err = md5_check ((optind == argc) ? "-" : argv[optind]);
 573     }
 574   else
 575     {
 576       if (optind == argc)
 577         argv[argc++] = "-";
 578
 579       for (; optind < argc; ++optind)
 580         {
 581           int fail;
 582           char *file = argv[optind];
 583
 584           fail = md5_file (file, binary, md5buffer);
 585           err |= fail;
 586           if (!fail)
 587             {
 588               size_t i;
 589
 590               /* Output a leading backslash if the file name contains
 591                  a newline or backslash.  */
 592               if (strchr (file, '\n') || strchr (file, '\\'))
 593                 putchar ('\\');
 594
 595               for (i = 0; i < 16; ++i)
 596                 printf ("%02x", md5buffer[i]);
 597
 598               putchar (' ');
 599               if (binary)
 600                 putchar ('*');
 601               else
 602                 putchar (' ');
 603
 604               /* Translate each NEWLINE byte to the string, "\\n",
 605                  and each backslash to "\\\\".  */
 606               for (i = 0; i < strlen (file); ++i)
 607                 {
 608                   switch (file[i])
 609                     {
 610                     case '\n':
 611                       fputs ("\\n", stdout);
 612                       break;
 613
 614                     case '\\':
 615                       fputs ("\\\\", stdout);
 616                       break;
 617
 618                     default:
 619                       putchar (file[i]);
 620                       break;
 621                     }
 622                 }
 623               putchar ('\n');
 624             }
 625         }
 626     }
 627
 628   if (fclose (stdout) == EOF)
 629     error (EXIT_FAILURE, errno, _("write error"));
 630
 631   if (have_read_stdin && fclose (stdin) == EOF)
 632     error (EXIT_FAILURE, errno, _("standard input"));
 633
 634   exit (err == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
 635 }