mdw@git.distorted.org.uk Git - anag/blob - anag.c

   1 /* -*-c-*-
   2  *
   3  * $Id: anag.c,v 1.8 2004/04/08 01:36:19 mdw Exp $
   4  *
   5  * Main driver for anag
   6  *
   7  * (c) 2001 Mark Wooding
   8  */
   9
  10 /*----- Licensing notice --------------------------------------------------*
  11  *
  12  * This file is part of Anag: a simple wordgame helper.
  13  *
  14  * Anag is free software; you can redistribute it and/or modify
  15  * it under the terms of the GNU General Public License as published by
  16  * the Free Software Foundation; either version 2 of the License, or
  17  * (at your option) any later version.
  18  *
  19  * Anag 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 General Public License for more details.
  23  *
  24  * You should have received a copy of the GNU General Public License
  25  * along with Anag; if not, write to the Free Software Foundation,
  26  * Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  27  */
  28
  29 /*----- Header files ------------------------------------------------------*/
  30
  31 #include "anag.h"
  32
  33 /*----- Static variables --------------------------------------------------*/
  34
  35 static const char *file = DICTIONARY;
  36
  37 /*----- Help text functions -----------------------------------------------*/
  38
  39 static void usage(FILE *fp)
  40 {
  41   pquis(fp, "Usage: $ [-f file] expression\n");
  42 }
  43
  44 static void version(FILE *fp)
  45 {
  46   pquis(fp, "$, version " VERSION "\n");
  47 }
  48
  49 static void help(FILE *fp)
  50 {
  51   version(fp);
  52   fputc('\n', fp);
  53   usage(fp);
  54   fputs("\n\
  55 Searches a wordlist, printing all of the words which match an expression.\n\
  56 \n\
  57 Options supported are:\n\
  58 \n\
  59 -h, --help              display this help text\n\
  60 -v, --version           display the program's version number\n\
  61 -u, --usage             display a very brief usage message\n\
  62 -f, --file FILE         read wordlist from FILE, not `" DICTIONARY "'\n\
  63 \n\
  64 The basic tests in the expression are:\n\
  65 \n\
  66 -anagram WORD           matches a full-length anagram\n\
  67 -subgram WORD           matches words which only use letters in WORD\n\
  68 -wildcard PATTERN       matches with wildcards `*' and `?'\n\
  69 -trackword WORD         matches words which can be found in a trackword\n\
  70 -mono PATTERN           matches words isomorphic to the given PATTERN\n\
  71 "
  72 #ifdef HAVE_REGCOMP
  73 "\
  74 -regexp REGEXP          matches with an (extended) regular expression\n\
  75 "
  76 #endif
  77 #ifdef HAVE_PCRE
  78 "\
  79 -pcre REGEXP            matches with a Perl-like regular expression\n\
  80 "
  81 #endif
  82 "\
  83 \n\
  84 These simple tests can be combined using the operators `-a', `-o' and `-n'\n\
  85 (for `and', `or' and `not'; they may also be written `&', `|' and `!' if\n\
  86 you like), and grouped using parentheses `(' and `)'.\n\
  87 ", fp); /*"*/
  88 }
  89
  90 /*----- The options parser ------------------------------------------------*/
  91
  92 /* --- Options table structure --- */
  93
  94 struct opt {
  95   const char *name;
  96   unsigned nargs;
  97   unsigned f;
  98   unsigned tag;
  99 };
 100
 101 enum {
 102   O_HELP, O_VERSION, O_USAGE,
 103   O_FILE,
 104   O_AND, O_OR, O_NOT, O_LPAREN, O_RPAREN,
 105   O_ANAG, O_SUBG, O_WILD, O_TRACK, O_REGEXP, O_PCRE, O_MONO,
 106   O_EOF
 107 };
 108
 109 #define OF_SHORT 1u
 110
 111 static const struct opt opttab[] = {
 112
 113   /* --- Options -- don't form part of the language --- */
 114
 115   { "help",             0,      OF_SHORT,       O_HELP },
 116   { "version",          0,      OF_SHORT,       O_VERSION },
 117   { "usage",            0,      OF_SHORT,       O_USAGE },
 118   { "file",             1,      OF_SHORT,       O_FILE },
 119
 120   /* --- Operators -- provide the basic structure of the language --- *
 121    *
 122    * These are also given magical names by the parser.
 123    */
 124
 125   { "and",              0,      OF_SHORT,       O_AND },
 126   { "or",               0,      OF_SHORT,       O_OR },
 127   { "not",              0,      OF_SHORT,       O_NOT },
 128
 129   /* --- Actual matching oeprations -- do something useful --- */
 130
 131   { "anagram",          1,      0,              O_ANAG },
 132   { "subgram",          1,      0,              O_SUBG },
 133   { "wildcard",         1,      0,              O_WILD },
 134   { "trackword",        1,      0,              O_TRACK },
 135   { "mono",             1,      0,              O_MONO },
 136 #ifdef HAVE_REGCOMP
 137   { "regexp",           1,      0,              O_REGEXP },
 138 #endif
 139 #ifdef HAVE_PCRE
 140   { "pcre",             1,      0,              O_PCRE },
 141 #endif
 142
 143   /* --- End marker --- */
 144
 145   { 0,                  0,      0,              0 }
 146 };
 147
 148 static int ac;
 149 static const char *const *av;
 150 static int ai;
 151
 152 /* --- @nextopt@ --- *
 153  *
 154  * Arguments:   @const char ***arg@ = where to store the arg pointer
 155  *
 156  * Returns:     The tag of the next option.
 157  *
 158  * Use:         Scans the next option off the command line.  If the option
 159  *              doesn't form part of the language, it's processed internally,
 160  *              and you'll never see it from here.  On exit, the @arg@
 161  *              pointer is set to contain the address of the option scanned,
 162  *              followed by its arguments if any.  You're expected to know
 163  *              how many arguments there are for your option.
 164  */
 165
 166 static unsigned nextopt(const char *const **arg)
 167 {
 168   for (;;) {
 169     const struct opt *o, *oo;
 170     size_t sz;
 171     const char *p;
 172
 173     /* --- Pick the next option off the front --- */
 174
 175     *arg = av + ai;
 176     if (ai >= ac)
 177       return (O_EOF);
 178     p = av[ai++];
 179
 180     /* --- Cope with various forms of magic --- */
 181
 182     if (p[0] != '-') {
 183       if (!p[1]) switch (*p) {
 184         case '&': return (O_AND);
 185         case '|': return (O_OR);
 186         case '!': return (O_NOT);
 187         case '(': return (O_LPAREN);
 188         case ')': return (O_RPAREN);
 189       }
 190       goto bad;
 191     }
 192
 193     /* --- Now cope with other sorts of weirdies --- *
 194      *
 195      * By the end of this, a leading `-' or `--' will have been stripped.
 196      */
 197
 198     p++;
 199     if (!*p)
 200       goto bad;
 201     if (*p == '-')
 202       p++;
 203     if (!*p) {
 204       if (ai < ac)
 205         die("syntax error near `--': rubbish at end of line");
 206       return (O_EOF);
 207     }
 208
 209     /* --- Now look the word up in my table --- */
 210
 211     sz = strlen(p);
 212     oo = 0;
 213     for (o = opttab; o->name; o++) {
 214       if (strncmp(p, o->name, sz) == 0) {
 215         if (strlen(o->name) == sz || ((o->f & OF_SHORT) && sz == 1)) {
 216           oo = o;
 217           break;
 218         }
 219         if (oo) {
 220           die("ambiguous option name `-%s' (could match `-%s' or `-%s')",
 221               p, oo->name, o->name);
 222         }
 223         oo = o;
 224       }
 225     }
 226     if (!oo)
 227       die("unrecognized option name `-%s'", p);
 228
 229     /* --- Sort out the arguments --- */
 230
 231     if (ai + oo->nargs > ac)
 232       die("too few arguments for `-%s' (need %u)", oo->name, oo->nargs);
 233     ai += oo->nargs;
 234
 235     /* --- Now process the option --- */
 236
 237     switch (oo->tag) {
 238       case O_HELP:
 239         help(stdout);
 240         exit(0);
 241       case O_VERSION:
 242         version(stdout);
 243         exit(0);
 244       case O_USAGE:
 245         usage(stdout);
 246         exit(0);
 247       case O_FILE:
 248         file = (*arg)[1];
 249         break;
 250       default:
 251         return (oo->tag);
 252     }
 253     continue;
 254   bad:
 255     die("syntax error near `%s': unknown token type", av[ai - 1]);
 256   }
 257 }
 258
 259 /*----- Node types for operators ------------------------------------------*/
 260
 261 /* --- Node structures --- */
 262
 263 typedef struct node_bin {
 264   node n;
 265   node *left;
 266   node *right;
 267 } node_bin;
 268
 269 typedef struct node_un {
 270   node n;
 271   node *arg;
 272 } node_un;
 273
 274 /* --- Node functions --- */
 275
 276 static int n_or(node *nn, const char *p, size_t sz)
 277 {
 278   node_bin *n = (node_bin *)nn;
 279   return (n->left->func(n->left, p, sz) || n->right->func(n->right, p, sz));
 280 }
 281
 282 static int n_and(node *nn, const char *p, size_t sz)
 283 {
 284   node_bin *n = (node_bin *)nn;
 285   return (n->left->func(n->left, p, sz) && n->right->func(n->right, p, sz));
 286 }
 287
 288 static int n_not(node *nn, const char *p, size_t sz)
 289 {
 290   node_un *n = (node_un *)nn;
 291   return (!n->arg->func(n->arg, p, sz));
 292 }
 293
 294 /*----- Parser for the expression syntax ----------------------------------*/
 295
 296 /* --- A parser context --- */
 297
 298 typedef struct p_ctx {
 299   unsigned t;
 300   const char *const *a;
 301 } p_ctx;
 302
 303 /* --- Parser structure --- *
 304  *
 305  * This is a simple recursive descent parser.  The context retains
 306  * information about the current token.  Each function is passed the address
 307  * of a node pointer to fill in.  This simplifies the binary operator code
 308  * somewhat, relative to returning pointers to node trees.
 309  */
 310
 311 static void p_expr(p_ctx *p, node **/*nn*/);
 312
 313 static void p_next(p_ctx *p)
 314 {
 315   static const char *const eof[] = { "<end>", 0 };
 316   p->t = nextopt(&p->a);
 317   if (p->t == O_EOF)
 318     p->a = eof;
 319 }
 320
 321 static void p_factor(p_ctx *p, node **nn)
 322 {
 323   node_un *n;
 324   if (p->t == O_LPAREN) {
 325     p_next(p);
 326     p_expr(p, nn);
 327     if (p->t != O_RPAREN)
 328       die("syntax error near `%s': missing `)'", *p->a);
 329     p_next(p);
 330   } else if (p->t == O_NOT) {
 331     n = xmalloc(sizeof(node_un));
 332     n->n.func = n_not;
 333     *nn = &n->n;
 334     p_next(p);
 335     p_factor(p, &n->arg);
 336   } else {
 337     switch (p->t) {
 338       case O_ANAG: *nn = anagram(p->a + 1); break;
 339       case O_SUBG: *nn = subgram(p->a + 1); break;
 340       case O_WILD: *nn = wildcard(p->a + 1); break;
 341       case O_TRACK: *nn = trackword(p->a + 1); break;
 342 #ifdef HAVE_REGCOMP
 343       case O_REGEXP: *nn = regexp(p->a + 1); break;
 344 #endif
 345 #ifdef HAVE_PCRE
 346       case O_PCRE: *nn = pcrenode(p->a + 1); break;
 347 #endif
 348       case O_MONO: *nn = mono(p->a + 1); break;
 349       default: die("syntax error near `%s': unexpected token", *p->a);
 350     }
 351     p_next(p);
 352   }
 353 }
 354
 355 static void p_term(p_ctx *p, node **nn)
 356 {
 357   node_bin *n;
 358   for (;;) {
 359     p_factor(p, nn);
 360     switch (p->t) {
 361       case O_AND:
 362         p_next(p);
 363       default:
 364         break;
 365       case O_RPAREN:
 366       case O_OR:
 367       case O_EOF:
 368         return;
 369     }
 370     n = xmalloc(sizeof(node_bin));
 371     n->left = *nn;
 372     n->n.func = n_and;
 373     *nn = &n->n;
 374     nn = &n->right;
 375   }
 376 }
 377
 378 static void p_expr(p_ctx *p, node **nn)
 379 {
 380   node_bin *n;
 381   for (;;) {
 382     p_term(p, nn);
 383     if (p->t != O_OR)
 384       break;
 385     p_next(p);
 386     n = xmalloc(sizeof(node_bin));
 387     n->left = *nn;
 388     n->n.func = n_or;
 389     *nn = &n->n;
 390     nn = &n->right;
 391   }
 392 }
 393
 394 /* --- @p_argv@ --- *
 395  *
 396  * Arguments:   @int argc@ = number of command-line arguments
 397  *              @const char *const argv[]@ = vectoor of arguments
 398  *
 399  * Returns:     A compiled node, parsed from the arguments.
 400  *
 401  * Use:         Does the donkey-work of parsing a command-line.
 402  */
 403
 404 static node *p_argv(int argc, const char *const argv[])
 405 {
 406   p_ctx p;
 407   node *n;
 408
 409   av = argv;
 410   ac = argc;
 411   ai = 1;
 412   p_next(&p);
 413   if (p.t == O_EOF) {
 414     usage(stderr);
 415     pquis(stderr, "(Run `$ --help' for more detail.)\n");
 416     exit(EXIT_FAILURE);
 417   }
 418   p_expr(&p, &n);
 419   if (p.t != O_EOF) {
 420     die("syntax error near `%s': rubbish at end of line (too many `)'s?)",
 421         *p.a);
 422   }
 423   return (n);
 424 }
 425
 426 /*----- Main code ---------------------------------------------------------*/
 427
 428 /* --- @main@ --- *
 429  *
 430  * Arguments:   @int argc@ = number of command-line arguments
 431  *              @char *argv[]@ = vector of argument words
 432  *
 433  * Returns:     Zero on success, nonzero on failure.
 434  *
 435  * Use:         Picks entries from a word list which match particular
 436  *              expressions.  This might be of assistance to word-game types.
 437  */
 438
 439 int main(int argc, char *argv[])
 440 {
 441   node *n;
 442   FILE *fp;
 443   dstr d = DSTR_INIT;
 444   char *p, *q, *l;
 445
 446   ego(argv[0]);
 447   n = p_argv(argc, (const char *const *)argv);
 448
 449   if ((fp = fopen(file, "r")) == 0)
 450     die("error opening `%s': %s", file, strerror(errno));
 451   for (;;) {
 452     dstr_reset(&d);
 453     if (dstr_putline(&d, fp) < 0)
 454       break;
 455     l = d.buf + d.len;
 456     for (p = q = d.buf; p < l; p++) {
 457       if (!isalnum((unsigned char)*p))
 458         continue;
 459       *q++ = tolower((unsigned char)*p);
 460     }
 461     *q = 0;
 462     d.len = q - d.buf;
 463     if (n->func(n, d.buf, d.len)) {
 464       fwrite(d.buf, 1, d.len, stdout);
 465       fputc('\n', stdout);
 466     }
 467   }
 468   if (!feof(fp))
 469     die("error reading `%s': %s", file, strerror(errno));
 470   fclose(fp);
 471   return (0);
 472 }
 473
 474 /*----- That's all, folks -------------------------------------------------*/