git.distorted.org.uk Git - u/mdw/catacomb/blob - hashsum.c

   1 /* -*-c-*-
   2  *
   3  * $Id$
   4  *
   5  * Hash files using some secure hash function
   6  *
   7  * (c) 2000 Straylight/Edgeware
   8  */
   9
  10 /*----- Licensing notice --------------------------------------------------*
  11  *
  12  * This file is part of Catacomb.
  13  *
  14  * Catacomb is free software; you can redistribute it and/or modify
  15  * it under the terms of the GNU Library General Public License as
  16  * published by the Free Software Foundation; either version 2 of the
  17  * License, or (at your option) any later version.
  18  *
  19  * Catacomb is distributed in the hope that it will be useful,
  20  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  21  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  22  * GNU Library General Public License for more details.
  23  *
  24  * You should have received a copy of the GNU Library General Public
  25  * License along with Catacomb; if not, write to the Free
  26  * Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
  27  * MA 02111-1307, USA.
  28  */
  29
  30 /*----- Header files ------------------------------------------------------*/
  31
  32 #define _FILE_OFFSET_BITS 64
  33
  34 #include "config.h"
  35
  36 #include <assert.h>
  37 #include <errno.h>
  38 #include <stdio.h>
  39 #include <stdlib.h>
  40 #include <string.h>
  41
  42 #include <mLib/alloc.h>
  43 #include <mLib/dstr.h>
  44 #include <mLib/mdwopt.h>
  45 #include <mLib/quis.h>
  46 #include <mLib/report.h>
  47 #include <mLib/sub.h>
  48
  49 #include "ghash.h"
  50 #include "cc.h"
  51
  52 /*----- Static variables --------------------------------------------------*/
  53
  54 #define f_bogus 1u
  55 #define f_verbose 2u
  56 #define f_check 4u
  57 #define f_files 8u
  58 #define f_oddhash 16u
  59 #define f_escape 32u
  60 #define f_oddenc 64u
  61
  62 /*----- Guts --------------------------------------------------------------*/
  63
  64 static int checkjunk(const char *path, const struct stat *st, void *p)
  65 {
  66   const char *what;
  67   fhashstate *fh = p;
  68
  69   if (!st) {
  70     if (fh->f & f_verbose)
  71       fprintf(stderr, "JUNK (error %s) %s\n", strerror(errno), path);
  72     else
  73       moan("couldn't stat junk file `%s': %s", path, strerror(errno));
  74   } else {
  75     what = describefile(st);
  76     if (fh->f & f_verbose)
  77       fprintf(stderr, "JUNK %s %s\n", what, path);
  78     else
  79       moan("found junk %s `%s'", what, path);
  80   }
  81   return (0);
  82 }
  83
  84 static int warnjunk(const char *path, const struct stat *st, void *p)
  85 {
  86   if (st)
  87     moan("unexpected %s `%s'", describefile(st), path);
  88   else
  89     moan("couldn't stat unexpected file `%s': %s", path, strerror(errno));
  90   return (0);
  91 }
  92
  93 static int checkhash(fhashstate *fh, const char *file, const encodeops *e)
  94 {
  95   int rc;
  96   hfpctx hfp;
  97   dstr dl = DSTR_INIT;
  98   dstr df = DSTR_INIT;
  99   unsigned long n = 0, nfail = 0;
 100   int hf;
 101
 102   if (!file || strcmp(file, "-") == 0)
 103     hfp.fp = stdin;
 104   else if ((hfp.fp = fopen(file, fh->f & GSF_RAW ? "r" : "rb")) == 0) {
 105     moan("couldn't open `%s': %s", file, strerror(errno));
 106     return (EXIT_FAILURE);
 107   }
 108
 109   hfp.dline = &dl;
 110   hfp.dfile = &df;
 111   hfp.hbuf = xmalloc(2 * fh->gch->hashsz);
 112   hfp.gch = fh->gch;
 113   hfp.ee = e;
 114   hfp.f = fh->f;
 115
 116   while ((hf = hfparse(&hfp)) != HF_EOF) {
 117     switch (hf) {
 118       case HF_HASH:
 119         xfree(hfp.hbuf);
 120         hfp.hbuf = xmalloc(2 * hfp.gch->hashsz);
 121         break;
 122       case HF_FILE:
 123         if (fhash(fh, df.buf, hfp.hbuf + hfp.gch->hashsz)) {
 124           moan("couldn't read `%s': %s", df.buf, strerror(errno));
 125           rc = EXIT_FAILURE;
 126           continue;
 127         }
 128         if (memcmp(hfp.hbuf, hfp.hbuf + hfp.gch->hashsz,
 129                    hfp.gch->hashsz) != 0) {
 130           if (hfp.f & f_verbose)
 131             fprintf(stderr, "FAIL %s\n", df.buf);
 132           else
 133             moan("%s check failed for `%s'", hfp.gch->name, df.buf);
 134           nfail++;
 135           rc = EXIT_FAILURE;
 136         } else {
 137           if (hfp.f & f_verbose)
 138             fprintf(stderr, "OK %s\n", df.buf);
 139         }
 140         n++;
 141     }
 142   }
 143
 144   if (ferror(hfp.fp)) {
 145     moan("error reading input `%s': %s",
 146          file ? file : "<stdin>", strerror(errno));
 147     rc = EXIT_FAILURE;
 148   }
 149   dstr_destroy(&dl);
 150   dstr_destroy(&df);
 151   xfree(hfp.hbuf);
 152   if ((fh->f & f_verbose) && nfail)
 153     moan("%lu of %lu file(s) failed %s check", nfail, n, hfp.gch->name);
 154   else if (!n)
 155     moan("no files checked");
 156   return (rc);
 157 }
 158
 159 static int dohash(fhashstate *fh, const char *file, const encodeops *e)
 160 {
 161   int rc = 0;
 162   octet *p = xmalloc(fh->gch->hashsz);
 163
 164   if (fhash(fh, file, p)) {
 165     moan("couldn't read `%s': %s", file ? file : "<stdin>", strerror(errno));
 166     rc = EXIT_FAILURE;
 167   } else {
 168     e->put(p, fh->gch->hashsz, stdout);
 169     if (file) {
 170       fputc(' ', stdout);
 171       fputc(fh->f & FHF_BINARY ? '*' : ' ', stdout);
 172       if (fh->f & f_escape)
 173         putstring(stdout, file, 0);
 174       else
 175         fputs(file, stdout);
 176     }
 177     fputc('\n', stdout);
 178   }
 179
 180   xfree(p);
 181   return (rc);
 182 }
 183
 184 static int dofile(fhashstate *fh, const char *file, const encodeops *e)
 185   { return (fh->f & f_check ? checkhash : dohash)(fh, file, e); }
 186
 187 static int hashfiles(fhashstate *fh, const char *file, const encodeops *e)
 188 {
 189   FILE *fp;
 190   dstr d = DSTR_INIT;
 191   int rc = 0;
 192   int rrc;
 193
 194   if (!file || strcmp(file, "-") == 0)
 195     fp = stdin;
 196   else if ((fp = fopen(file, fh->f & GSF_RAW ? "r" : "rb")) == 0) {
 197     moan("couldn't open `%s': %s", file, strerror(errno));
 198     return (EXIT_FAILURE);
 199   }
 200
 201   for (;;) {
 202     DRESET(&d);
 203     if (getstring(fp, &d, GSF_FILE | fh->f))
 204       break;
 205     if ((rrc = dofile(fh, d.buf, e)) != 0)
 206       rc = rrc;
 207   }
 208
 209   return (rc);
 210 }
 211
 212 static int hashsum(fhashstate *fh, const char *file, const encodeops *e)
 213   { return (fh->f & f_files ? hashfiles : dofile)(fh, file, e); }
 214
 215 /*----- Main driver -------------------------------------------------------*/
 216
 217 void version(FILE *fp)
 218 {
 219   pquis(fp, "$, Catacomb version " VERSION "\n");
 220 }
 221
 222 static void usage(FILE *fp)
 223 {
 224   pquis(fp, "Usage: $ [-f0ebcv] [-a ALGORITHM] [-E ENC] [FILES...]\n");
 225 }
 226
 227 static void help(FILE *fp, const gchash *gch)
 228 {
 229   version(fp);
 230   fputc('\n', fp);
 231   usage(fp);
 232   pquis(fp, "\n\
 233 Generates or checks message digests on files.  Options available:\n\
 234 \n\
 235 -h, --help              Display this help message.\n\
 236 -V, --version           Display program's version number.\n\
 237 -u, --usage             Display a terse usage message.\n\
 238 -l, --list [ITEM...]    Show known hash functions and/or encodings.\n\
 239 \n\
 240 -a, --algorithm=ALG     Use the message digest algorithm ALG.\n\
 241 -E, --encoding=ENC      Represent hashes using encoding ENC.\n\
 242 \n\
 243 -f, --files             Read a list of file names from standard input.\n\
 244 -0, --null              File names are null terminated, not plain text.\n\
 245 \n\
 246 -e, --escape            Escape funny characters in filenames.\n\
 247 -c, --check             Check message digests rather than emitting them.\n\
 248 -b, --binary            When reading files, treat them as binary.\n\
 249 -v, --verbose           Be verbose when checking digests.\n\
 250 \n\
 251 For a list of hashing algorithms and encodings, type `$ --list'.\n\
 252 ");
 253   if (gch)
 254     fprintf(fp, "The default message digest algorithm is %s.\n", gch->name);
 255 }
 256
 257 #define LISTS(LI)                                                       \
 258   LI("Lists", list, listtab[i].name, listtab[i].name)                   \
 259   LI("Hash functions", hash, ghashtab[i], ghashtab[i]->name)            \
 260   LI("Encodings", enc, encodingtab[i].name, encodingtab[i].name)
 261
 262 MAKELISTTAB(listtab, LISTS)
 263
 264 int main(int argc, char *argv[])
 265 {
 266   fhashstate fh;
 267   const encodeops *e = &encodingtab[ENC_HEX];
 268   int rc;
 269
 270   /* --- Initialization --- */
 271
 272   ego(argv[0]);
 273   sub_init();
 274   fhash_init(&fh, 0, 0);
 275
 276   /* --- Choose a hash function from the name --- */
 277
 278   {
 279     char *q = xstrdup(QUIS);
 280     size_t len = strlen(q);
 281     if (len > 3 && strcmp(q + len - 3, "sum") == 0) {
 282       q[len - 3] = 0;
 283       fh.gch = gethash(q);
 284     }
 285     if (!fh.gch)
 286       fh.gch = gethash("md5");
 287     xfree(q);
 288   }
 289
 290   /* --- Read options --- */
 291
 292   for (;;) {
 293     static struct option opts[] = {
 294       { "help",         0,              0,      'h' },
 295       { "verbose",      0,              0,      'V' },
 296       { "usage",        0,              0,      'u' },
 297
 298       { "algorithm",    OPTF_ARGREQ,    0,      'a' },
 299       { "hash",         OPTF_ARGREQ,    0,      'a' },
 300       { "encoding",     OPTF_ARGREQ,    0,      'E' },
 301       { "list",         0,              0,      'l' },
 302
 303       { "files",        0,              0,      'f' },
 304       { "find",         0,              0,      'f' },
 305       { "null",         0,              0,      '0' },
 306
 307       { "escape",       0,              0,      'e' },
 308       { "check",        0,              0,      'c' },
 309       { "junk",         0,              0,      'j' },
 310       { "binary",       0,              0,      'b' },
 311       { "verbose",      0,              0,      'v' },
 312       { "progress",     0,              0,      'p' },
 313
 314       { 0,              0,              0,      0 }
 315     };
 316     int i = mdwopt(argc, argv, "hVu a:E:l f0 ecjbvp", opts, 0, 0, 0);
 317     if (i < 0)
 318       break;
 319
 320     switch (i) {
 321       case 'h':
 322         help(stdout, fh.gch);
 323         exit(0);
 324       case 'V':
 325         version(stdout);
 326         exit(0);
 327       case 'u':
 328         usage(stdout);
 329         exit(0);
 330       case 'l':
 331         exit(displaylists(listtab, argv + optind));
 332       case 'a':
 333         if ((fh.gch = gethash(optarg)) == 0)
 334           die(EXIT_FAILURE, "unknown hash algorithm `%s'", optarg);
 335         fh.f |= f_oddhash;
 336         break;
 337       case 'E':
 338         if ((e = getencoding(optarg)) == 0)
 339           die(EXIT_FAILURE, "unknown encoding `%s'", optarg);
 340         fh.f |= f_oddenc;
 341         break;
 342       case 'f':
 343         fh.f |= f_files;
 344         break;
 345       case '0':
 346         fh.f |= GSF_RAW;
 347         break;
 348       case 'e':
 349         fh.f |= f_escape;
 350         break;
 351       case 'c':
 352         fh.f |= f_check;
 353         break;
 354       case 'j':
 355         fh.f |= FHF_JUNK;
 356         break;
 357       case 'b':
 358         fh.f |= FHF_BINARY;
 359         break;
 360       case 'v':
 361         fh.f |= f_verbose;
 362         break;
 363       case 'p':
 364         fh.f |= FHF_PROGRESS;
 365         break;
 366       default:
 367         fh.f |= f_bogus;
 368         break;
 369     }
 370   }
 371
 372   if (fh.f & f_bogus) {
 373     usage(stderr);
 374     exit(EXIT_FAILURE);
 375   }
 376   argv += optind;
 377   argc -= optind;
 378
 379   /* --- Generate output --- */
 380
 381   if (!(fh.f & f_check) && (argc || (fh.f & f_files))) {
 382     if (fh.f & f_oddhash) printf("#hash %s\n", fh.gch->name);
 383     if (fh.f & f_oddenc) printf("#encoding %s\n", e->name);
 384     if (fh.f & f_escape) fputs("#escape\n", stdout);
 385   }
 386   if (!argc)
 387     rc = hashsum(&fh, 0, e);
 388   else {
 389     int i;
 390     int rrc;
 391
 392     rc = 0;
 393     for (i = 0; i < argc; i++) {
 394       if ((rrc = hashsum(&fh, argv[i], e)) != 0)
 395         rc = rrc;
 396     }
 397   }
 398
 399   if (fh.f & FHF_JUNK) {
 400     if (fh.f & f_check) {
 401       if (fhash_junk(&fh, checkjunk, &fh)) rc = EXIT_FAILURE;
 402     } else {
 403       if (fhash_junk(&fh, warnjunk, 0) < 0) rc = EXIT_FAILURE;
 404     }
 405   }
 406   fhash_free(&fh);
 407
 408   return (rc);
 409 }
 410
 411 /*----- That's all, folks -------------------------------------------------*/