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

   1 /* -*-c-*-
   2  *
   3  * Hash files using some secure hash function
   4  *
   5  * (c) 2000 Straylight/Edgeware
   6  */
   7
   8 /*----- Licensing notice --------------------------------------------------*
   9  *
  10  * This file is part of Catacomb.
  11  *
  12  * Catacomb is free software; you can redistribute it and/or modify
  13  * it under the terms of the GNU Library General Public License as
  14  * published by the Free Software Foundation; either version 2 of the
  15  * License, or (at your option) any later version.
  16  *
  17  * Catacomb is distributed in the hope that it will be useful,
  18  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  19  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  20  * GNU Library General Public License for more details.
  21  *
  22  * You should have received a copy of the GNU Library General Public
  23  * License along with Catacomb; if not, write to the Free
  24  * Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
  25  * MA 02111-1307, USA.
  26  */
  27
  28 /*----- Header files ------------------------------------------------------*/
  29
  30 #define _FILE_OFFSET_BITS 64
  31
  32 #include "config.h"
  33
  34 #include <assert.h>
  35 #include <errno.h>
  36 #include <stdio.h>
  37 #include <stdlib.h>
  38 #include <string.h>
  39
  40 #include <mLib/alloc.h>
  41 #include <mLib/dstr.h>
  42 #include <mLib/macros.h>
  43 #include <mLib/mdwopt.h>
  44 #include <mLib/quis.h>
  45 #include <mLib/report.h>
  46 #include <mLib/sub.h>
  47
  48 #include "ghash.h"
  49 #include "cc.h"
  50
  51 /*----- Static variables --------------------------------------------------*/
  52
  53 #define f_bogus 1u
  54 #define f_verbose 2u
  55 #define f_check 4u
  56 #define f_files 8u
  57 #define f_oddhash 16u
  58 #define f_escape 32u
  59 #define f_oddenc 64u
  60
  61 /*----- Guts --------------------------------------------------------------*/
  62
  63 static int checkjunk(const char *path, const struct stat *st, void *p)
  64 {
  65   const char *what;
  66   fhashstate *fh = p;
  67
  68   if (!st) {
  69     if (fh->f & f_verbose)
  70       fprintf(stderr, "JUNK (error %s) %s\n", strerror(errno), path);
  71     else
  72       moan("couldn't stat junk file `%s': %s", path, strerror(errno));
  73   } else {
  74     what = describefile(st);
  75     if (fh->f & f_verbose)
  76       fprintf(stderr, "JUNK %s %s\n", what, path);
  77     else
  78       moan("found junk %s `%s'", what, path);
  79   }
  80   return (0);
  81 }
  82
  83 static int warnjunk(const char *path, const struct stat *st, void *p)
  84 {
  85   if (st)
  86     moan("unexpected %s `%s'", describefile(st), path);
  87   else
  88     moan("couldn't stat unexpected file `%s': %s", path, strerror(errno));
  89   return (0);
  90 }
  91
  92 static int checkhash(fhashstate *fh, const char *file, const encodeops *e)
  93 {
  94   int rc = 0;
  95   hfpctx hfp;
  96   dstr dl = DSTR_INIT;
  97   dstr df = DSTR_INIT;
  98   unsigned long n = 0, nfail = 0;
  99   int hf;
 100
 101   if (!file || STRCMP(file, ==, "-"))
 102     hfp.fp = stdin;
 103   else if ((hfp.fp = fopen(file, fh->f & GSF_RAW ? "r" : "rb")) == 0) {
 104     moan("couldn't open `%s': %s", file, strerror(errno));
 105     return (EXIT_FAILURE);
 106   }
 107
 108   hfp.dline = &dl;
 109   hfp.dfile = &df;
 110   hfp.hbuf = xmalloc(2 * fh->gch->hashsz);
 111   hfp.gch = fh->gch;
 112   hfp.ee = e;
 113   hfp.f = fh->f;
 114
 115   while ((hf = hfparse(&hfp)) != HF_EOF) {
 116     switch (hf) {
 117       case HF_HASH:
 118         xfree(hfp.hbuf);
 119         hfp.hbuf = xmalloc(2 * hfp.gch->hashsz);
 120         fh->gch = hfp.gch;
 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)) {
 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, ==, "-"))
 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 -p, --progress          Show a progress indicator for large files.\n\
 250 -v, --verbose           Be verbose when checking digests.\n\
 251 \n\
 252 For a list of hashing algorithms and encodings, type `$ --list'.\n\
 253 ");
 254   if (gch)
 255     fprintf(fp, "The default message digest algorithm is %s.\n", gch->name);
 256 }
 257
 258 #define LISTS(LI)                                                       \
 259   LI("Lists", list, listtab[i].name, listtab[i].name)                   \
 260   LI("Hash functions", hash, ghashtab[i], ghashtab[i]->name)            \
 261   LI("Encodings", enc, encodingtab[i].name, encodingtab[i].name)
 262
 263 MAKELISTTAB(listtab, LISTS)
 264
 265 int main(int argc, char *argv[])
 266 {
 267   fhashstate fh;
 268   const encodeops *e = &encodingtab[ENC_HEX];
 269   int rc;
 270
 271   /* --- Initialization --- */
 272
 273   ego(argv[0]);
 274   sub_init();
 275   fhash_init(&fh, 0, 0);
 276
 277   /* --- Choose a hash function from the name --- */
 278
 279   {
 280     char *q = xstrdup(QUIS);
 281     size_t len = strlen(q);
 282     if (len > 3 && STRCMP(q + len - 3, ==, "sum")) {
 283       q[len - 3] = 0;
 284       fh.gch = gethash(q);
 285     }
 286     if (!fh.gch)
 287       fh.gch = gethash("md5");
 288     xfree(q);
 289   }
 290
 291   /* --- Read options --- */
 292
 293   for (;;) {
 294     static struct option opts[] = {
 295       { "help",         0,              0,      'h' },
 296       { "verbose",      0,              0,      'V' },
 297       { "usage",        0,              0,      'u' },
 298
 299       { "algorithm",    OPTF_ARGREQ,    0,      'a' },
 300       { "hash",         OPTF_ARGREQ,    0,      'a' },
 301       { "encoding",     OPTF_ARGREQ,    0,      'E' },
 302       { "list",         0,              0,      'l' },
 303
 304       { "files",        0,              0,      'f' },
 305       { "find",         0,              0,      'f' },
 306       { "null",         0,              0,      '0' },
 307
 308       { "escape",       0,              0,      'e' },
 309       { "check",        0,              0,      'c' },
 310       { "junk",         0,              0,      'j' },
 311       { "binary",       0,              0,      'b' },
 312       { "verbose",      0,              0,      'v' },
 313       { "progress",     0,              0,      'p' },
 314
 315       { 0,              0,              0,      0 }
 316     };
 317     int i = mdwopt(argc, argv, "hVu a:E:l f0 ecjbvp", opts, 0, 0, 0);
 318     if (i < 0)
 319       break;
 320
 321     switch (i) {
 322       case 'h':
 323         help(stdout, fh.gch);
 324         exit(0);
 325       case 'V':
 326         version(stdout);
 327         exit(0);
 328       case 'u':
 329         usage(stdout);
 330         exit(0);
 331       case 'l':
 332         exit(displaylists(listtab, argv + optind));
 333       case 'a':
 334         if ((fh.gch = gethash(optarg)) == 0)
 335           die(EXIT_FAILURE, "unknown hash algorithm `%s'", optarg);
 336         fh.f |= f_oddhash;
 337         break;
 338       case 'E':
 339         if ((e = getencoding(optarg)) == 0)
 340           die(EXIT_FAILURE, "unknown encoding `%s'", optarg);
 341         fh.f |= f_oddenc;
 342         break;
 343       case 'f':
 344         fh.f |= f_files;
 345         break;
 346       case '0':
 347         fh.f |= GSF_RAW;
 348         break;
 349       case 'e':
 350         fh.f |= f_escape;
 351         break;
 352       case 'c':
 353         fh.f |= f_check;
 354         break;
 355       case 'j':
 356         fh.f |= FHF_JUNK;
 357         break;
 358       case 'b':
 359         fh.f |= FHF_BINARY;
 360         break;
 361       case 'v':
 362         fh.f |= f_verbose;
 363         break;
 364       case 'p':
 365         fh.f |= FHF_PROGRESS;
 366         break;
 367       default:
 368         fh.f |= f_bogus;
 369         break;
 370     }
 371   }
 372
 373   if (fh.f & f_bogus) {
 374     usage(stderr);
 375     exit(EXIT_FAILURE);
 376   }
 377   argv += optind;
 378   argc -= optind;
 379
 380   /* --- Generate output --- */
 381
 382   if (!(fh.f & f_check) && (argc || (fh.f & f_files))) {
 383     if (fh.f & f_oddhash) printf("#hash %s\n", fh.gch->name);
 384     if (fh.f & f_oddenc) printf("#encoding %s\n", e->name);
 385     if (fh.f & f_escape) fputs("#escape\n", stdout);
 386   }
 387   if (!argc)
 388     rc = hashsum(&fh, 0, e);
 389   else {
 390     int i;
 391     int rrc;
 392
 393     rc = 0;
 394     for (i = 0; i < argc; i++) {
 395       if ((rrc = hashsum(&fh, argv[i], e)) != 0)
 396         rc = rrc;
 397     }
 398   }
 399
 400   if (fh.f & FHF_JUNK) {
 401     if (fh.f & f_check) {
 402       if (fhash_junk(&fh, checkjunk, &fh)) rc = EXIT_FAILURE;
 403     } else {
 404       if (fhash_junk(&fh, warnjunk, 0) < 0) rc = EXIT_FAILURE;
 405     }
 406   }
 407   fhash_free(&fh);
 408
 409   return (rc);
 410 }
 411
 412 /*----- That's all, folks -------------------------------------------------*/