build: update gnulib submodule to latest
[coreutils/ericb.git] / src / md5sum.c
blob6f6e6373d5586418c513b1972d352378e97cd4d9
1 /* Compute checksums of files or strings.
2 Copyright (C) 1995-2011 Free Software Foundation, Inc.
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 3 of the License, or
7 (at your option) any later version.
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.
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>. */
17 /* Written by Ulrich Drepper <drepper@gnu.ai.mit.edu>. */
19 #include <config.h>
21 #include <getopt.h>
22 #include <sys/types.h>
24 #include "system.h"
26 #if HASH_ALGO_MD5
27 # include "md5.h"
28 #endif
29 #if HASH_ALGO_SHA1
30 # include "sha1.h"
31 #endif
32 #if HASH_ALGO_SHA256 || HASH_ALGO_SHA224
33 # include "sha256.h"
34 #endif
35 #if HASH_ALGO_SHA512 || HASH_ALGO_SHA384
36 # include "sha512.h"
37 #endif
38 #include "error.h"
39 #include "fadvise.h"
40 #include "stdio--.h"
41 #include "xfreopen.h"
43 /* The official name of this program (e.g., no `g' prefix). */
44 #if HASH_ALGO_MD5
45 # define PROGRAM_NAME "md5sum"
46 # define DIGEST_TYPE_STRING "MD5"
47 # define DIGEST_STREAM md5_stream
48 # define DIGEST_BITS 128
49 # define DIGEST_REFERENCE "RFC 1321"
50 # define DIGEST_ALIGN 4
51 #elif HASH_ALGO_SHA1
52 # define PROGRAM_NAME "sha1sum"
53 # define DIGEST_TYPE_STRING "SHA1"
54 # define DIGEST_STREAM sha1_stream
55 # define DIGEST_BITS 160
56 # define DIGEST_REFERENCE "FIPS-180-1"
57 # define DIGEST_ALIGN 4
58 #elif HASH_ALGO_SHA256
59 # define PROGRAM_NAME "sha256sum"
60 # define DIGEST_TYPE_STRING "SHA256"
61 # define DIGEST_STREAM sha256_stream
62 # define DIGEST_BITS 256
63 # define DIGEST_REFERENCE "FIPS-180-2"
64 # define DIGEST_ALIGN 4
65 #elif HASH_ALGO_SHA224
66 # define PROGRAM_NAME "sha224sum"
67 # define DIGEST_TYPE_STRING "SHA224"
68 # define DIGEST_STREAM sha224_stream
69 # define DIGEST_BITS 224
70 # define DIGEST_REFERENCE "RFC 3874"
71 # define DIGEST_ALIGN 4
72 #elif HASH_ALGO_SHA512
73 # define PROGRAM_NAME "sha512sum"
74 # define DIGEST_TYPE_STRING "SHA512"
75 # define DIGEST_STREAM sha512_stream
76 # define DIGEST_BITS 512
77 # define DIGEST_REFERENCE "FIPS-180-2"
78 # define DIGEST_ALIGN 8
79 #elif HASH_ALGO_SHA384
80 # define PROGRAM_NAME "sha384sum"
81 # define DIGEST_TYPE_STRING "SHA384"
82 # define DIGEST_STREAM sha384_stream
83 # define DIGEST_BITS 384
84 # define DIGEST_REFERENCE "FIPS-180-2"
85 # define DIGEST_ALIGN 8
86 #else
87 # error "Can't decide which hash algorithm to compile."
88 #endif
90 #define DIGEST_HEX_BYTES (DIGEST_BITS / 4)
91 #define DIGEST_BIN_BYTES (DIGEST_BITS / 8)
93 #define AUTHORS \
94 proper_name ("Ulrich Drepper"), \
95 proper_name ("Scott Miller"), \
96 proper_name ("David Madore")
98 /* The minimum length of a valid digest line. This length does
99 not include any newline character at the end of a line. */
100 #define MIN_DIGEST_LINE_LENGTH \
101 (DIGEST_HEX_BYTES /* length of hexadecimal message digest */ \
102 + 2 /* blank and binary indicator */ \
103 + 1 /* minimum filename length */ )
105 /* True if any of the files read were the standard input. */
106 static bool have_read_stdin;
108 /* The minimum length of a valid checksum line for the selected algorithm. */
109 static size_t min_digest_line_length;
111 /* Set to the length of a digest hex string for the selected algorithm. */
112 static size_t digest_hex_bytes;
114 /* With --check, don't generate any output.
115 The exit code indicates success or failure. */
116 static bool status_only = false;
118 /* With --check, print a message to standard error warning about each
119 improperly formatted checksum line. */
120 static bool warn = false;
122 /* With --check, suppress the "OK" printed for each verified file. */
123 static bool quiet = false;
125 /* For long options that have no equivalent short option, use a
126 non-character as a pseudo short option, starting with CHAR_MAX + 1. */
127 enum
129 STATUS_OPTION = CHAR_MAX + 1,
130 QUIET_OPTION
133 static struct option const long_options[] =
135 { "binary", no_argument, NULL, 'b' },
136 { "check", no_argument, NULL, 'c' },
137 { "quiet", no_argument, NULL, QUIET_OPTION },
138 { "status", no_argument, NULL, STATUS_OPTION },
139 { "text", no_argument, NULL, 't' },
140 { "warn", no_argument, NULL, 'w' },
141 { GETOPT_HELP_OPTION_DECL },
142 { GETOPT_VERSION_OPTION_DECL },
143 { NULL, 0, NULL, 0 }
146 void
147 usage (int status)
149 if (status != EXIT_SUCCESS)
150 fprintf (stderr, _("Try `%s --help' for more information.\n"),
151 program_name);
152 else
154 printf (_("\
155 Usage: %s [OPTION]... [FILE]...\n\
156 Print or check %s (%d-bit) checksums.\n\
157 With no FILE, or when FILE is -, read standard input.\n\
160 program_name,
161 DIGEST_TYPE_STRING,
162 DIGEST_BITS);
163 if (O_BINARY)
164 fputs (_("\
165 -b, --binary read in binary mode (default unless reading tty stdin)\n\
166 "), stdout);
167 else
168 fputs (_("\
169 -b, --binary read in binary mode\n\
170 "), stdout);
171 printf (_("\
172 -c, --check read %s sums from the FILEs and check them\n"),
173 DIGEST_TYPE_STRING);
174 if (O_BINARY)
175 fputs (_("\
176 -t, --text read in text mode (default if reading tty stdin)\n\
177 "), stdout);
178 else
179 fputs (_("\
180 -t, --text read in text mode (default)\n\
181 "), stdout);
182 fputs (_("\
184 The following three options are useful only when verifying checksums:\n\
185 --quiet don't print OK for each successfully verified file\n\
186 --status don't output anything, status code shows success\n\
187 -w, --warn warn about improperly formatted checksum lines\n\
189 "), stdout);
190 fputs (HELP_OPTION_DESCRIPTION, stdout);
191 fputs (VERSION_OPTION_DESCRIPTION, stdout);
192 printf (_("\
194 The sums are computed as described in %s. When checking, the input\n\
195 should be a former output of this program. The default mode is to print\n\
196 a line with checksum, a character indicating type (`*' for binary, ` ' for\n\
197 text), and name for each FILE.\n"),
198 DIGEST_REFERENCE);
199 emit_ancillary_info ();
202 exit (status);
205 #define ISWHITE(c) ((c) == ' ' || (c) == '\t')
207 /* Split the checksum string S (of length S_LEN) from a BSD 'md5' or
208 'sha1' command into two parts: a hexadecimal digest, and the file
209 name. S is modified. Return true if successful. */
211 static bool
212 bsd_split_3 (char *s, size_t s_len, unsigned char **hex_digest,
213 char **file_name)
215 size_t i;
217 if (s_len == 0)
218 return false;
220 *file_name = s;
222 /* Find end of filename. The BSD 'md5' and 'sha1' commands do not escape
223 filenames, so search backwards for the last ')'. */
224 i = s_len - 1;
225 while (i && s[i] != ')')
226 i--;
228 if (s[i] != ')')
229 return false;
231 s[i++] = '\0';
233 while (ISWHITE (s[i]))
234 i++;
236 if (s[i] != '=')
237 return false;
239 i++;
241 while (ISWHITE (s[i]))
242 i++;
244 *hex_digest = (unsigned char *) &s[i];
245 return true;
248 /* Split the string S (of length S_LEN) into three parts:
249 a hexadecimal digest, binary flag, and the file name.
250 S is modified. Return true if successful. */
252 static bool
253 split_3 (char *s, size_t s_len,
254 unsigned char **hex_digest, int *binary, char **file_name)
256 bool escaped_filename = false;
257 size_t algo_name_len;
259 size_t i = 0;
260 while (ISWHITE (s[i]))
261 ++i;
263 /* Check for BSD-style checksum line. */
264 algo_name_len = strlen (DIGEST_TYPE_STRING);
265 if (STREQ_LEN (s + i, DIGEST_TYPE_STRING, algo_name_len))
267 if (s[i + algo_name_len] == ' ')
268 ++i;
269 if (s[i + algo_name_len] == '(')
271 *binary = 0;
272 return bsd_split_3 (s + i + algo_name_len + 1,
273 s_len - (i + algo_name_len + 1),
274 hex_digest, file_name);
278 /* Ignore this line if it is too short.
279 Each line must have at least `min_digest_line_length - 1' (or one more, if
280 the first is a backslash) more characters to contain correct message digest
281 information. */
282 if (s_len - i < min_digest_line_length + (s[i] == '\\'))
283 return false;
285 if (s[i] == '\\')
287 ++i;
288 escaped_filename = true;
290 *hex_digest = (unsigned char *) &s[i];
292 /* The first field has to be the n-character hexadecimal
293 representation of the message digest. If it is not followed
294 immediately by a white space it's an error. */
295 i += digest_hex_bytes;
296 if (!ISWHITE (s[i]))
297 return false;
299 s[i++] = '\0';
301 if (s[i] != ' ' && s[i] != '*')
302 return false;
303 *binary = (s[i++] == '*');
305 /* All characters between the type indicator and end of line are
306 significant -- that includes leading and trailing white space. */
307 *file_name = &s[i];
309 if (escaped_filename)
311 /* Translate each `\n' string in the file name to a NEWLINE,
312 and each `\\' string to a backslash. */
314 char *dst = &s[i];
316 while (i < s_len)
318 switch (s[i])
320 case '\\':
321 if (i == s_len - 1)
323 /* A valid line does not end with a backslash. */
324 return false;
326 ++i;
327 switch (s[i++])
329 case 'n':
330 *dst++ = '\n';
331 break;
332 case '\\':
333 *dst++ = '\\';
334 break;
335 default:
336 /* Only `\' or `n' may follow a backslash. */
337 return false;
339 break;
341 case '\0':
342 /* The file name may not contain a NUL. */
343 return false;
344 break;
346 default:
347 *dst++ = s[i++];
348 break;
351 *dst = '\0';
353 return true;
356 /* Return true if S is a NUL-terminated string of DIGEST_HEX_BYTES hex digits.
357 Otherwise, return false. */
358 static bool
359 hex_digits (unsigned char const *s)
361 unsigned int i;
362 for (i = 0; i < digest_hex_bytes; i++)
364 if (!isxdigit (*s))
365 return false;
366 ++s;
368 return *s == '\0';
371 /* An interface to the function, DIGEST_STREAM.
372 Operate on FILENAME (it may be "-").
374 *BINARY indicates whether the file is binary. BINARY < 0 means it
375 depends on whether binary mode makes any difference and the file is
376 a terminal; in that case, clear *BINARY if the file was treated as
377 text because it was a terminal.
379 Put the checksum in *BIN_RESULT, which must be properly aligned.
380 Return true if successful. */
382 static bool
383 digest_file (const char *filename, int *binary, unsigned char *bin_result)
385 FILE *fp;
386 int err;
387 bool is_stdin = STREQ (filename, "-");
389 if (is_stdin)
391 have_read_stdin = true;
392 fp = stdin;
393 if (O_BINARY && *binary)
395 if (*binary < 0)
396 *binary = ! isatty (STDIN_FILENO);
397 if (*binary)
398 xfreopen (NULL, "rb", stdin);
401 else
403 fp = fopen (filename, (O_BINARY && *binary ? "rb" : "r"));
404 if (fp == NULL)
406 error (0, errno, "%s", filename);
407 return false;
411 fadvise (fp, FADVISE_SEQUENTIAL);
413 err = DIGEST_STREAM (fp, bin_result);
414 if (err)
416 error (0, errno, "%s", filename);
417 if (fp != stdin)
418 fclose (fp);
419 return false;
422 if (!is_stdin && fclose (fp) != 0)
424 error (0, errno, "%s", filename);
425 return false;
428 return true;
431 static bool
432 digest_check (const char *checkfile_name)
434 FILE *checkfile_stream;
435 uintmax_t n_misformatted_lines = 0;
436 uintmax_t n_properly_formatted_lines = 0;
437 uintmax_t n_mismatched_checksums = 0;
438 uintmax_t n_open_or_read_failures = 0;
439 unsigned char bin_buffer_unaligned[DIGEST_BIN_BYTES + DIGEST_ALIGN];
440 /* Make sure bin_buffer is properly aligned. */
441 unsigned char *bin_buffer = ptr_align (bin_buffer_unaligned, DIGEST_ALIGN);
442 uintmax_t line_number;
443 char *line;
444 size_t line_chars_allocated;
445 bool is_stdin = STREQ (checkfile_name, "-");
447 if (is_stdin)
449 have_read_stdin = true;
450 checkfile_name = _("standard input");
451 checkfile_stream = stdin;
453 else
455 checkfile_stream = fopen (checkfile_name, "r");
456 if (checkfile_stream == NULL)
458 error (0, errno, "%s", checkfile_name);
459 return false;
463 line_number = 0;
464 line = NULL;
465 line_chars_allocated = 0;
468 char *filename IF_LINT ( = NULL);
469 int binary;
470 unsigned char *hex_digest IF_LINT ( = NULL);
471 ssize_t line_length;
473 ++line_number;
474 if (line_number == 0)
475 error (EXIT_FAILURE, 0, _("%s: too many checksum lines"),
476 checkfile_name);
478 line_length = getline (&line, &line_chars_allocated, checkfile_stream);
479 if (line_length <= 0)
480 break;
482 /* Ignore comment lines, which begin with a '#' character. */
483 if (line[0] == '#')
484 continue;
486 /* Remove any trailing newline. */
487 if (line[line_length - 1] == '\n')
488 line[--line_length] = '\0';
490 if (! (split_3 (line, line_length, &hex_digest, &binary, &filename)
491 && ! (is_stdin && STREQ (filename, "-"))
492 && hex_digits (hex_digest)))
494 ++n_misformatted_lines;
496 if (warn)
498 error (0, 0,
499 _("%s: %" PRIuMAX
500 ": improperly formatted %s checksum line"),
501 checkfile_name, line_number,
502 DIGEST_TYPE_STRING);
505 else
507 static const char bin2hex[] = { '0', '1', '2', '3',
508 '4', '5', '6', '7',
509 '8', '9', 'a', 'b',
510 'c', 'd', 'e', 'f' };
511 bool ok;
513 ++n_properly_formatted_lines;
515 ok = digest_file (filename, &binary, bin_buffer);
517 if (!ok)
519 ++n_open_or_read_failures;
520 if (!status_only)
522 printf (_("%s: FAILED open or read\n"), filename);
525 else
527 size_t digest_bin_bytes = digest_hex_bytes / 2;
528 size_t cnt;
529 /* Compare generated binary number with text representation
530 in check file. Ignore case of hex digits. */
531 for (cnt = 0; cnt < digest_bin_bytes; ++cnt)
533 if (tolower (hex_digest[2 * cnt])
534 != bin2hex[bin_buffer[cnt] >> 4]
535 || (tolower (hex_digest[2 * cnt + 1])
536 != (bin2hex[bin_buffer[cnt] & 0xf])))
537 break;
539 if (cnt != digest_bin_bytes)
540 ++n_mismatched_checksums;
542 if (!status_only)
544 if (cnt != digest_bin_bytes)
545 printf ("%s: %s\n", filename, _("FAILED"));
546 else if (!quiet)
547 printf ("%s: %s\n", filename, _("OK"));
552 while (!feof (checkfile_stream) && !ferror (checkfile_stream));
554 free (line);
556 if (ferror (checkfile_stream))
558 error (0, 0, _("%s: read error"), checkfile_name);
559 return false;
562 if (!is_stdin && fclose (checkfile_stream) != 0)
564 error (0, errno, "%s", checkfile_name);
565 return false;
568 if (n_properly_formatted_lines == 0)
570 /* Warn if no tests are found. */
571 error (0, 0, _("%s: no properly formatted %s checksum lines found"),
572 checkfile_name, DIGEST_TYPE_STRING);
574 else
576 if (!status_only)
578 if (n_misformatted_lines != 0)
579 error (0, 0,
580 (ngettext
581 ("WARNING: %" PRIuMAX " line is improperly formatted",
582 "WARNING: %" PRIuMAX " lines are improperly formatted",
583 select_plural (n_misformatted_lines))),
584 n_misformatted_lines);
586 if (n_open_or_read_failures != 0)
587 error (0, 0,
588 (ngettext
589 ("WARNING: %" PRIuMAX " listed file could not be read",
590 "WARNING: %" PRIuMAX " listed files could not be read",
591 select_plural (n_open_or_read_failures))),
592 n_open_or_read_failures);
594 if (n_mismatched_checksums != 0)
595 error (0, 0,
596 (ngettext
597 ("WARNING: %" PRIuMAX " computed checksum did NOT match",
598 "WARNING: %" PRIuMAX " computed checksums did NOT match",
599 select_plural (n_mismatched_checksums))),
600 n_mismatched_checksums);
604 return (n_properly_formatted_lines != 0
605 && n_mismatched_checksums == 0
606 && n_open_or_read_failures == 0);
610 main (int argc, char **argv)
612 unsigned char bin_buffer_unaligned[DIGEST_BIN_BYTES + DIGEST_ALIGN];
613 /* Make sure bin_buffer is properly aligned. */
614 unsigned char *bin_buffer = ptr_align (bin_buffer_unaligned, DIGEST_ALIGN);
615 bool do_check = false;
616 int opt;
617 bool ok = true;
618 int binary = -1;
620 /* Setting values of global variables. */
621 initialize_main (&argc, &argv);
622 set_program_name (argv[0]);
623 setlocale (LC_ALL, "");
624 bindtextdomain (PACKAGE, LOCALEDIR);
625 textdomain (PACKAGE);
627 atexit (close_stdout);
629 /* Line buffer stdout to ensure lines are written atomically and immediately
630 so that processes running in parallel do not intersperse their output. */
631 setvbuf (stdout, NULL, _IOLBF, 0);
633 while ((opt = getopt_long (argc, argv, "bctw", long_options, NULL)) != -1)
634 switch (opt)
636 case 'b':
637 binary = 1;
638 break;
639 case 'c':
640 do_check = true;
641 break;
642 case STATUS_OPTION:
643 status_only = true;
644 warn = false;
645 quiet = false;
646 break;
647 case 't':
648 binary = 0;
649 break;
650 case 'w':
651 status_only = false;
652 warn = true;
653 quiet = false;
654 break;
655 case QUIET_OPTION:
656 status_only = false;
657 warn = false;
658 quiet = true;
659 break;
660 case_GETOPT_HELP_CHAR;
661 case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
662 default:
663 usage (EXIT_FAILURE);
666 min_digest_line_length = MIN_DIGEST_LINE_LENGTH;
667 digest_hex_bytes = DIGEST_HEX_BYTES;
669 if (0 <= binary && do_check)
671 error (0, 0, _("the --binary and --text options are meaningless when "
672 "verifying checksums"));
673 usage (EXIT_FAILURE);
676 if (status_only && !do_check)
678 error (0, 0,
679 _("the --status option is meaningful only when verifying checksums"));
680 usage (EXIT_FAILURE);
683 if (warn && !do_check)
685 error (0, 0,
686 _("the --warn option is meaningful only when verifying checksums"));
687 usage (EXIT_FAILURE);
690 if (quiet && !do_check)
692 error (0, 0,
693 _("the --quiet option is meaningful only when verifying checksums"));
694 usage (EXIT_FAILURE);
697 if (!O_BINARY && binary < 0)
698 binary = 0;
700 if (optind == argc)
701 argv[argc++] = bad_cast ("-");
703 for (; optind < argc; ++optind)
705 char *file = argv[optind];
707 if (do_check)
708 ok &= digest_check (file);
709 else
711 int file_is_binary = binary;
713 if (! digest_file (file, &file_is_binary, bin_buffer))
714 ok = false;
715 else
717 size_t i;
719 /* Output a leading backslash if the file name contains
720 a newline or backslash. */
721 if (strchr (file, '\n') || strchr (file, '\\'))
722 putchar ('\\');
724 for (i = 0; i < (digest_hex_bytes / 2); ++i)
725 printf ("%02x", bin_buffer[i]);
727 putchar (' ');
728 if (file_is_binary)
729 putchar ('*');
730 else
731 putchar (' ');
733 /* Translate each NEWLINE byte to the string, "\\n",
734 and each backslash to "\\\\". */
735 for (i = 0; i < strlen (file); ++i)
737 switch (file[i])
739 case '\n':
740 fputs ("\\n", stdout);
741 break;
743 case '\\':
744 fputs ("\\\\", stdout);
745 break;
747 default:
748 putchar (file[i]);
749 break;
752 putchar ('\n');
757 if (have_read_stdin && fclose (stdin) == EOF)
758 error (EXIT_FAILURE, errno, _("standard input"));
760 exit (ok ? EXIT_SUCCESS : EXIT_FAILURE);