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

   1 /* -*-c-*-
   2  *
   3  * $Id: anag.c,v 1.3 2001/02/16 21:45:19 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 /*----- Revision history --------------------------------------------------*
  30  *
  31  * $Log: anag.c,v $
  32  * Revision 1.3  2001/02/16 21:45:19  mdw
  33  * Be more helpful.  Improve full help message.  Special-case error for
  34  * empty command strings.
  35  *
  36  * Revision 1.2  2001/02/07 09:09:11  mdw
  37  * Fix spurious error when `-file' is used.
  38  *
  39  * Revision 1.1  2001/02/04 17:14:42  mdw
  40  * Initial checkin
  41  *
  42  */
  43
  44 /*----- Header files ------------------------------------------------------*/
  45
  46 #include "anag.h"
  47
  48 /*----- Static variables --------------------------------------------------*/
  49
  50 static const char *file = DICTIONARY;
  51
  52 /*----- Help text functions -----------------------------------------------*/
  53
  54 static void usage(FILE *fp)
  55 {
  56   pquis(fp, "Usage: $ [-f file] expression\n");
  57 }
  58
  59 static void version(FILE *fp)
  60 {
  61   pquis(fp, "$, version " VERSION "\n");
  62 }
  63
  64 static void help(FILE *fp)
  65 {
  66   version(fp);
  67   fputc('\n', fp);
  68   usage(fp);
  69   fputs("\n\
  70 Searches a wordlist, printing all of the words which match an expression.\n\
  71 \n\
  72 Options supported are:\n\
  73 \n\
  74 -h, --help              display this help text\n\
  75 -v, --version           display the program's version number\n\
  76 -u, --usage             display a very brief usage message\n\
  77 -f, --file FILE         read wordlist from FILE, not `" DICTIONARY "'\n\
  78 \n\
  79 The basic tests in the expression are:\n\
  80 \n\
  81 -anagram WORD           matches a full-length anagram\n\
  82 -subgram WORD           matches words which only use letters in WORD\n\
  83 -wildcard PATTERN       matches with wildcards `*' and `?'\n\
  84 -trackword WORD         matches words which can be found in a trackword\n\
  85 \n\
  86 These simple tests can be combined using the operators `-a', `-o' and `-n'\n\
  87 (for `and', `or' and `not'; they may also be written `&', `|' and `!' if\n\
  88 you like), and grouped using parentheses `(' and `)'.\n\
  89 ", fp);
  90 }
  91
  92 /*----- The options parser ------------------------------------------------*/
  93
  94 /* --- Options table structure --- */
  95
  96 struct opt {
  97   const char *name;
  98   unsigned nargs;
  99   unsigned f;
 100   unsigned tag;
 101 };
 102
 103 enum {
 104   O_HELP, O_VERSION, O_USAGE,
 105   O_FILE,
 106   O_AND, O_OR, O_NOT, O_LPAREN, O_RPAREN,
 107   O_ANAG, O_SUBG, O_WILD, O_TRACK,
 108   O_EOF
 109 };
 110
 111 #define OF_SHORT 1u
 112
 113 static const struct opt opttab[] = {
 114
 115   /* --- Options -- don't form part of the language --- */
 116
 117   { "help",             0,      OF_SHORT,       O_HELP },
 118   { "version",          0,      OF_SHORT,       O_VERSION },
 119   { "usage",            0,      OF_SHORT,       O_USAGE },
 120   { "file",             1,      OF_SHORT,       O_FILE },
 121
 122   /* --- Operators -- provide the basic structure of the language --- *
 123    *
 124    * These are also given magical names by the parser.
 125    */
 126
 127   { "and",              0,      OF_SHORT,       O_AND },
 128   { "or",               0,      OF_SHORT,       O_OR },
 129   { "not",              0,      OF_SHORT,       O_NOT },
 130
 131   /* --- Actual matching oeprations -- do something useful --- */
 132
 133   { "anagram",          1,      0,              O_ANAG },
 134   { "subgram",          1,      0,              O_SUBG },
 135   { "wildcard",         1,      0,              O_WILD },
 136   { "trackword",        1,      0,              O_TRACK },
 137
 138   /* --- End marker --- */
 139
 140   { 0,                  0,      0,              0 }
 141 };
 142
 143 static int ac;
 144 static const char *const *av;
 145 static int ai;
 146
 147 /* --- @nextopt@ --- *
 148  *
 149  * Arguments:   @const char ***arg@ = where to store the arg pointer
 150  *
 151  * Returns:     The tag of the next option.
 152  *
 153  * Use:         Scans the next option off the command line.  If the option
 154  *              doesn't form part of the language, it's processed internally,
 155  *              and you'll never see it from here.  On exit, the @arg@
 156  *              pointer is set to contain the address of the option scanned,
 157  *              followed by its arguments if any.  You're expected to know
 158  *              how many arguments there are for your option.
 159  */
 160
 161 static unsigned nextopt(const char *const **arg)
 162 {
 163   for (;;) {
 164     const struct opt *o, *oo;
 165     size_t sz;
 166     const char *p;
 167
 168     /* --- Pick the next option off the front --- */
 169
 170     *arg = av + ai;
 171     if (ai >= ac)
 172       return (O_EOF);
 173     p = av[ai++];
 174
 175     /* --- Cope with various forms of magic --- */
 176
 177     if (p[0] != '-') {
 178       if (!p[1]) switch (*p) {
 179         case '&': return (O_AND);
 180         case '|': return (O_OR);
 181         case '!': return (O_NOT);
 182         case '(': return (O_LPAREN);
 183         case ')': return (O_RPAREN);
 184       }
 185       goto bad;
 186     }
 187
 188     /* --- Now cope with other sorts of weirdies --- *
 189      *
 190      * By the end of this, a leading `-' or `--' will have been stripped.
 191      */
 192
 193     p++;
 194     if (!*p)
 195       goto bad;
 196     if (*p == '-')
 197       p++;
 198     if (!*p) {
 199       if (ai < ac)
 200         die("syntax error near `--': rubbish at end of line");
 201       return (O_EOF);
 202     }
 203
 204     /* --- Now look the word up in my table --- */
 205
 206     sz = strlen(p);
 207     oo = 0;
 208     for (o = opttab; o->name; o++) {
 209       if (strncmp(p, o->name, sz) == 0) {
 210         if (strlen(o->name) == sz || ((o->f & OF_SHORT) && sz == 1)) {
 211           oo = o;
 212           break;
 213         }
 214         if (oo) {
 215           die("ambiguous option name `-%s' (could match `-%s' or `-%s')",
 216               p, oo->name, o->name);
 217         }
 218         oo = o;
 219       }
 220     }
 221     if (!oo)
 222       die("unrecognized option name `-%s'", p);
 223
 224     /* --- Sort out the arguments --- */
 225
 226     if (ai + oo->nargs > ac)
 227       die("too few arguments for `-%s' (need %u)", oo->name, oo->nargs);
 228     ai += oo->nargs;
 229
 230     /* --- Now process the option --- */
 231
 232     switch (oo->tag) {
 233       case O_HELP:
 234         help(stdout);
 235         exit(0);
 236       case O_VERSION:
 237         version(stdout);
 238         exit(0);
 239       case O_USAGE:
 240         usage(stdout);
 241         exit(0);
 242       case O_FILE:
 243         file = (*arg)[1];
 244         break;
 245       default:
 246         return (oo->tag);
 247     }
 248     continue;
 249   bad:
 250     die("syntax error near `%s': unknown token type", av[ai - 1]);
 251   }
 252 }
 253
 254 /*----- Node types for operators ------------------------------------------*/
 255
 256 /* --- Node structures --- */
 257
 258 typedef struct node_bin {
 259   node n;
 260   node *left;
 261   node *right;
 262 } node_bin;
 263
 264 typedef struct node_un {
 265   node n;
 266   node *arg;
 267 } node_un;
 268
 269 /* --- Node functions --- */
 270
 271 static int n_or(node *nn, const char *p, size_t sz)
 272 {
 273   node_bin *n = (node_bin *)nn;
 274   return (n->left->func(n->left, p, sz) || n->right->func(n->right, p, sz));
 275 }
 276
 277 static int n_and(node *nn, const char *p, size_t sz)
 278 {
 279   node_bin *n = (node_bin *)nn;
 280   return (n->left->func(n->left, p, sz) && n->right->func(n->right, p, sz));
 281 }
 282
 283 static int n_not(node *nn, const char *p, size_t sz)
 284 {
 285   node_un *n = (node_un *)nn;
 286   return (!n->arg->func(n->arg, p, sz));
 287 }
 288
 289 /*----- Parser for the expression syntax ----------------------------------*/
 290
 291 /* --- A parser context --- */
 292
 293 typedef struct p_ctx {
 294   unsigned t;
 295   const char *const *a;
 296 } p_ctx;
 297
 298 /* --- Parser structure --- *
 299  *
 300  * This is a simple recursive descent parser.  The context retains
 301  * information about the current token.  Each function is passed the address
 302  * of a node pointer to fill in.  This simplifies the binary operator code
 303  * somewhat, relative to returning pointers to node trees.
 304  */
 305
 306 static void p_expr(p_ctx *p, node **/*nn*/);
 307
 308 static void p_next(p_ctx *p)
 309 {
 310   static const char *const eof[] = { "<end>", 0 };
 311   p->t = nextopt(&p->a);
 312   if (p->t == O_EOF)
 313     p->a = eof;
 314 }
 315
 316 static void p_factor(p_ctx *p, node **nn)
 317 {
 318   node_un *n;
 319   if (p->t == O_LPAREN) {
 320     p_next(p);
 321     p_expr(p, nn);
 322     if (p->t != O_RPAREN)
 323       die("syntax error near `%s': missing `('", *p->a);
 324     p_next(p);
 325   } else if (p->t == O_NOT) {
 326     n = xmalloc(sizeof(node_un));
 327     n->n.func = n_not;
 328     *nn = &n->n;
 329     p_next(p);
 330     p_factor(p, &n->arg);
 331   } else {
 332     switch (p->t) {
 333       case O_ANAG: *nn = anagram(p->a + 1); break;
 334       case O_SUBG: *nn = subgram(p->a + 1); break;
 335       case O_WILD: *nn = wildcard(p->a + 1); break;
 336       case O_TRACK: *nn = trackword(p->a + 1); break;
 337       default: die("syntax error near `%s': unexpected token", *p->a);
 338     }
 339     p_next(p);
 340   }
 341 }
 342
 343 static void p_term(p_ctx *p, node **nn)
 344 {
 345   node_bin *n;
 346   for (;;) {
 347     p_factor(p, nn);
 348     switch (p->t) {
 349       case O_AND:
 350         p_next(p);
 351       default:
 352         break;
 353       case O_LPAREN:
 354       case O_RPAREN:
 355       case O_OR:
 356       case O_EOF:
 357         return;
 358     }
 359     n = xmalloc(sizeof(node_bin));
 360     n->left = *nn;
 361     n->n.func = n_and;
 362     *nn = &n->n;
 363     nn = &n->right;
 364   }
 365 }
 366
 367 static void p_expr(p_ctx *p, node **nn)
 368 {
 369   node_bin *n;
 370   for (;;) {
 371     p_term(p, nn);
 372     if (p->t != O_OR)
 373       break;
 374     p_next(p);
 375     n = xmalloc(sizeof(node_bin));
 376     n->left = *nn;
 377     n->n.func = n_or;
 378     *nn = &n->n;
 379     nn = &n->right;
 380   }
 381 }
 382
 383 /* --- @p_argv@ --- *
 384  *
 385  * Arguments:   @int argc@ = number of command-line arguments
 386  *              @const char *const argv[]@ = vectoor of arguments
 387  *
 388  * Returns:     A compiled node, parsed from the arguments.
 389  *
 390  * Use:         Does the donkey-work of parsing a command-line.
 391  */
 392
 393 static node *p_argv(int argc, const char *const argv[])
 394 {
 395   p_ctx p;
 396   node *n;
 397
 398   av = argv;
 399   ac = argc;
 400   ai = 1;
 401   p_next(&p);
 402   if (p.t == O_EOF) {
 403     usage(stderr);
 404     pquis(stderr, "(Run `$ --help' for more detail.)\n");
 405     exit(EXIT_FAILURE);
 406   }
 407   p_expr(&p, &n);
 408   if (p.t != O_EOF) {
 409     die("syntax error near `%s': rubbish at end of line (too many `)'s?)",
 410         *p.a);
 411   }
 412   return (n);
 413 }
 414
 415 /*----- Main code ---------------------------------------------------------*/
 416
 417 /* --- @main@ --- *
 418  *
 419  * Arguments:   @int argc@ = number of command-line arguments
 420  *              @char *argv[]@ = vector of argument words
 421  *
 422  * Returns:     Zero on success, nonzero on failure.
 423  *
 424  * Use:         Picks entries from a word list which match particular
 425  *              expressions.  This might be of assistance to word-game types.
 426  */
 427
 428 int main(int argc, char *argv[])
 429 {
 430   node *n;
 431   FILE *fp;
 432   dstr d = DSTR_INIT;
 433   char *p, *q, *l;
 434
 435   ego(argv[0]);
 436   n = p_argv(argc, (const char *const *)argv);
 437
 438   if ((fp = fopen(file, "r")) == 0)
 439     die("error opening `%s': %s", file, strerror(errno));
 440   for (;;) {
 441     dstr_reset(&d);
 442     if (dstr_putline(&d, fp) < 0)
 443       break;
 444     l = d.buf + d.len;
 445     for (p = q = d.buf; p < l; p++) {
 446       if (!isalnum((unsigned char)*p))
 447         continue;
 448       *q++ = tolower((unsigned char)*p);
 449     }
 450     *q = 0;
 451     d.len = q - d.buf;
 452     if (n->func(n, d.buf, d.len)) {
 453       fwrite(d.buf, 1, d.len, stdout);
 454       fputc('\n', stdout);
 455     }
 456   }
 457   if (!feof(fp))
 458     die("error reading `%s': %s", file, strerror(errno));
 459   fclose(fp);
 460   return (0);
 461 }
 462
 463 /*----- That's all, folks -------------------------------------------------*/