chiark / gitweb /
Quick lick of paint before we really get started.
[anag] / anag.c
CommitLineData
6e403221 1/* -*-c-*-
2 *
6e403221 3 * Main driver for anag
4 *
5 * (c) 2001 Mark Wooding
6 */
7
0279756e 8/*----- Licensing notice --------------------------------------------------*
6e403221 9 *
10 * This file is part of Anag: a simple wordgame helper.
11 *
12 * Anag is free software; you can redistribute it and/or modify
13 * it under the terms of the GNU General Public License as published by
14 * the Free Software Foundation; either version 2 of the License, or
15 * (at your option) any later version.
0279756e 16 *
6e403221 17 * Anag 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 General Public License for more details.
0279756e 21 *
6e403221 22 * You should have received a copy of the GNU General Public License
23 * along with Anag; if not, write to the Free Software Foundation,
24 * Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
25 */
26
6e403221 27/*----- Header files ------------------------------------------------------*/
28
29#include "anag.h"
30
31/*----- Static variables --------------------------------------------------*/
32
33static const char *file = DICTIONARY;
34
35/*----- Help text functions -----------------------------------------------*/
36
37static void usage(FILE *fp)
38{
39 pquis(fp, "Usage: $ [-f file] expression\n");
40}
41
42static void version(FILE *fp)
43{
44 pquis(fp, "$, version " VERSION "\n");
45}
46
47static void help(FILE *fp)
48{
49 version(fp);
50 fputc('\n', fp);
51 usage(fp);
52 fputs("\n\
53Searches a wordlist, printing all of the words which match an expression.\n\
2668675c 54\n\
55Options supported are:\n\
56\n\
57-h, --help display this help text\n\
58-v, --version display the program's version number\n\
59-u, --usage display a very brief usage message\n\
60-f, --file FILE read wordlist from FILE, not `" DICTIONARY "'\n\
61\n\
6e403221 62The basic tests in the expression are:\n\
63\n\
64-anagram WORD matches a full-length anagram\n\
65-subgram WORD matches words which only use letters in WORD\n\
66-wildcard PATTERN matches with wildcards `*' and `?'\n\
67-trackword WORD matches words which can be found in a trackword\n\
fe9969ff 68-mono PATTERN matches words isomorphic to the given PATTERN\n\
a10122de 69"
70#ifdef HAVE_REGCOMP
71"\
72-regexp REGEXP matches with an (extended) regular expression\n\
73"
74#endif
d9af4a2b 75#ifdef HAVE_PCRE
76"\
77-pcre REGEXP matches with a Perl-like regular expression\n\
78"
79#endif
a10122de 80"\
94ed9b30 81-length [+|-]N matches if length is [at least|at most] N\n\
82-longest output longest matches found here\n\
83-shortest output shortest matches found here\n\
6e403221 84\n\
85These simple tests can be combined using the operators `-a', `-o' and `-n'\n\
86(for `and', `or' and `not'; they may also be written `&', `|' and `!' if\n\
87you like), and grouped using parentheses `(' and `)'.\n\
d9af4a2b 88", fp); /*"*/
6e403221 89}
90
91/*----- The options parser ------------------------------------------------*/
92
93/* --- Options table structure --- */
94
95struct opt {
96 const char *name;
97 unsigned nargs;
98 unsigned f;
99 unsigned tag;
100};
101
102enum {
103 O_HELP, O_VERSION, O_USAGE,
104 O_FILE,
105 O_AND, O_OR, O_NOT, O_LPAREN, O_RPAREN,
94ed9b30 106 O_ANAG, O_SUBG, O_WILD, O_TRACK, O_REGEXP, O_PCRE, O_MONO, O_LENGTH,
107 O_LONGEST, O_SHORTEST,
6e403221 108 O_EOF
109};
110
111#define OF_SHORT 1u
112
113static 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
94ed9b30 131 /* --- Actual matching operations -- do something useful --- */
6e403221 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 },
fe9969ff 137 { "mono", 1, 0, O_MONO },
a10122de 138#ifdef HAVE_REGCOMP
139 { "regexp", 1, 0, O_REGEXP },
140#endif
d9af4a2b 141#ifdef HAVE_PCRE
142 { "pcre", 1, 0, O_PCRE },
143#endif
0279756e
MW
144 { "length", 1, 0, O_LENGTH },
145 { "longest", 0, 0, O_LONGEST },
146 { "shortest", 0, 0, O_SHORTEST },
6e403221 147
148 /* --- End marker --- */
149
150 { 0, 0, 0, 0 }
151};
152
153static int ac;
154static const char *const *av;
155static int ai;
156
157/* --- @nextopt@ --- *
158 *
159 * Arguments: @const char ***arg@ = where to store the arg pointer
160 *
161 * Returns: The tag of the next option.
162 *
163 * Use: Scans the next option off the command line. If the option
164 * doesn't form part of the language, it's processed internally,
165 * and you'll never see it from here. On exit, the @arg@
166 * pointer is set to contain the address of the option scanned,
167 * followed by its arguments if any. You're expected to know
168 * how many arguments there are for your option.
169 */
170
171static unsigned nextopt(const char *const **arg)
172{
173 for (;;) {
174 const struct opt *o, *oo;
175 size_t sz;
176 const char *p;
177
178 /* --- Pick the next option off the front --- */
179
180 *arg = av + ai;
181 if (ai >= ac)
182 return (O_EOF);
183 p = av[ai++];
184
185 /* --- Cope with various forms of magic --- */
186
187 if (p[0] != '-') {
188 if (!p[1]) switch (*p) {
189 case '&': return (O_AND);
190 case '|': return (O_OR);
191 case '!': return (O_NOT);
192 case '(': return (O_LPAREN);
193 case ')': return (O_RPAREN);
194 }
195 goto bad;
196 }
197
198 /* --- Now cope with other sorts of weirdies --- *
199 *
200 * By the end of this, a leading `-' or `--' will have been stripped.
201 */
202
203 p++;
204 if (!*p)
205 goto bad;
206 if (*p == '-')
207 p++;
208 if (!*p) {
209 if (ai < ac)
210 die("syntax error near `--': rubbish at end of line");
211 return (O_EOF);
212 }
213
214 /* --- Now look the word up in my table --- */
215
216 sz = strlen(p);
217 oo = 0;
218 for (o = opttab; o->name; o++) {
219 if (strncmp(p, o->name, sz) == 0) {
220 if (strlen(o->name) == sz || ((o->f & OF_SHORT) && sz == 1)) {
221 oo = o;
222 break;
223 }
224 if (oo) {
225 die("ambiguous option name `-%s' (could match `-%s' or `-%s')",
226 p, oo->name, o->name);
227 }
228 oo = o;
229 }
230 }
231 if (!oo)
232 die("unrecognized option name `-%s'", p);
233
234 /* --- Sort out the arguments --- */
235
236 if (ai + oo->nargs > ac)
237 die("too few arguments for `-%s' (need %u)", oo->name, oo->nargs);
238 ai += oo->nargs;
239
240 /* --- Now process the option --- */
241
242 switch (oo->tag) {
243 case O_HELP:
244 help(stdout);
245 exit(0);
246 case O_VERSION:
247 version(stdout);
248 exit(0);
249 case O_USAGE:
250 usage(stdout);
251 exit(0);
252 case O_FILE:
253 file = (*arg)[1];
254 break;
255 default:
256 return (oo->tag);
257 }
60dffc01 258 continue;
6e403221 259 bad:
260 die("syntax error near `%s': unknown token type", av[ai - 1]);
261 }
262}
263
264/*----- Node types for operators ------------------------------------------*/
265
266/* --- Node structures --- */
267
268typedef struct node_bin {
269 node n;
270 node *left;
271 node *right;
272} node_bin;
273
274typedef struct node_un {
275 node n;
276 node *arg;
277} node_un;
278
279/* --- Node functions --- */
280
281static int n_or(node *nn, const char *p, size_t sz)
282{
283 node_bin *n = (node_bin *)nn;
284 return (n->left->func(n->left, p, sz) || n->right->func(n->right, p, sz));
285}
286
287static int n_and(node *nn, const char *p, size_t sz)
288{
289 node_bin *n = (node_bin *)nn;
290 return (n->left->func(n->left, p, sz) && n->right->func(n->right, p, sz));
291}
292
293static int n_not(node *nn, const char *p, size_t sz)
294{
295 node_un *n = (node_un *)nn;
296 return (!n->arg->func(n->arg, p, sz));
297}
298
94ed9b30 299/*----- Other simple node types -------------------------------------------*/
300
301enum { LESS = -1, EQUAL = 0, GREATER = 1 };
302
303typedef struct node_numeric {
304 node n;
305 int dir;
306 int i;
307} node_numeric;
308
309static void parse_numeric(const char *p, int *dir, int *i)
310{
311 long l;
312 const char *pp = p;
313 char *q;
314
315 switch (*pp) {
316 case '-': *dir = LESS; pp++; break;
317 case '+': *dir = GREATER; pp++; break;
318 default: *dir = EQUAL; break;
319 }
320 errno = 0;
321 l = strtol(pp, &q, 0);
322 if (*q || errno || l < INT_MIN || l > INT_MAX)
323 die("bad numeric parameter `%s'", p);
324 *i = l;
325}
326
327static node *make_numeric(const char *const *av,
328 int (*func)(struct node *, const char *, size_t))
329{
330 node_numeric *n = xmalloc(sizeof(*n));
331 parse_numeric(av[0], &n->dir, &n->i);
332 n->n.func = func;
333 return (&n->n);
334}
335
336static int cmp_numeric(int x, int dir, int n)
337{
338 switch (dir) {
339 case LESS: return (x <= n);
340 case EQUAL: return (x == n);
341 case GREATER: return (x >= n);
342 }
343 abort();
344}
345
346static int n_length(node *nn, const char *p, size_t sz)
347{
348 node_numeric *n = (node_numeric *)nn;
349 return (cmp_numeric(sz, n->dir, n->i));
350}
351
6e403221 352/*----- Parser for the expression syntax ----------------------------------*/
353
354/* --- A parser context --- */
355
356typedef struct p_ctx {
357 unsigned t;
358 const char *const *a;
359} p_ctx;
360
361/* --- Parser structure --- *
362 *
363 * This is a simple recursive descent parser. The context retains
364 * information about the current token. Each function is passed the address
365 * of a node pointer to fill in. This simplifies the binary operator code
366 * somewhat, relative to returning pointers to node trees.
367 */
368
369static void p_expr(p_ctx *p, node **/*nn*/);
370
371static void p_next(p_ctx *p)
372{
373 static const char *const eof[] = { "<end>", 0 };
374 p->t = nextopt(&p->a);
375 if (p->t == O_EOF)
376 p->a = eof;
377}
378
379static void p_factor(p_ctx *p, node **nn)
380{
381 node_un *n;
382 if (p->t == O_LPAREN) {
383 p_next(p);
384 p_expr(p, nn);
385 if (p->t != O_RPAREN)
1d2d1062 386 die("syntax error near `%s': missing `)'", *p->a);
6e403221 387 p_next(p);
388 } else if (p->t == O_NOT) {
389 n = xmalloc(sizeof(node_un));
390 n->n.func = n_not;
391 *nn = &n->n;
392 p_next(p);
393 p_factor(p, &n->arg);
394 } else {
395 switch (p->t) {
396 case O_ANAG: *nn = anagram(p->a + 1); break;
397 case O_SUBG: *nn = subgram(p->a + 1); break;
398 case O_WILD: *nn = wildcard(p->a + 1); break;
399 case O_TRACK: *nn = trackword(p->a + 1); break;
a10122de 400#ifdef HAVE_REGCOMP
401 case O_REGEXP: *nn = regexp(p->a + 1); break;
402#endif
d9af4a2b 403#ifdef HAVE_PCRE
404 case O_PCRE: *nn = pcrenode(p->a + 1); break;
405#endif
fe9969ff 406 case O_MONO: *nn = mono(p->a + 1); break;
94ed9b30 407 case O_LENGTH: *nn = make_numeric(p->a + 1, n_length); break;
408 case O_LONGEST: *nn = longest(p->a + 1); break;
409 case O_SHORTEST: *nn = shortest(p->a + 1); break;
6e403221 410 default: die("syntax error near `%s': unexpected token", *p->a);
411 }
412 p_next(p);
413 }
414}
415
416static void p_term(p_ctx *p, node **nn)
417{
418 node_bin *n;
419 for (;;) {
420 p_factor(p, nn);
421 switch (p->t) {
422 case O_AND:
423 p_next(p);
424 default:
425 break;
6e403221 426 case O_RPAREN:
427 case O_OR:
428 case O_EOF:
429 return;
430 }
431 n = xmalloc(sizeof(node_bin));
432 n->left = *nn;
433 n->n.func = n_and;
434 *nn = &n->n;
435 nn = &n->right;
436 }
437}
438
439static void p_expr(p_ctx *p, node **nn)
440{
441 node_bin *n;
442 for (;;) {
443 p_term(p, nn);
444 if (p->t != O_OR)
445 break;
446 p_next(p);
447 n = xmalloc(sizeof(node_bin));
448 n->left = *nn;
449 n->n.func = n_or;
450 *nn = &n->n;
451 nn = &n->right;
452 }
453}
454
455/* --- @p_argv@ --- *
456 *
457 * Arguments: @int argc@ = number of command-line arguments
458 * @const char *const argv[]@ = vectoor of arguments
459 *
460 * Returns: A compiled node, parsed from the arguments.
461 *
462 * Use: Does the donkey-work of parsing a command-line.
463 */
464
465static node *p_argv(int argc, const char *const argv[])
466{
467 p_ctx p;
468 node *n;
469
470 av = argv;
471 ac = argc;
472 ai = 1;
473 p_next(&p);
2668675c 474 if (p.t == O_EOF) {
475 usage(stderr);
476 pquis(stderr, "(Run `$ --help' for more detail.)\n");
477 exit(EXIT_FAILURE);
478 }
6e403221 479 p_expr(&p, &n);
480 if (p.t != O_EOF) {
481 die("syntax error near `%s': rubbish at end of line (too many `)'s?)",
482 *p.a);
483 }
484 return (n);
485}
486
94ed9b30 487/*----- At-end stuff ------------------------------------------------------*/
488
489/* --- @atend_register@ --- *
490 *
491 * Arguments: @int (*func)(void *)@ = function to call
492 * @void *p@ = handle to pass to it
493 *
494 * Returns: ---
495 *
496 * Use: Adds a function to the list of things to do at the end of the
497 * program. The function should return nonzero if it produced
498 * any output.
499 */
500
501typedef struct atend {
502 struct atend *next;
503 int (*func)(void */*p*/);
504 void *p;
505} atend;
506
507static atend *aa_head = 0, **aa_tail = &aa_head;
508
509void atend_register(int (*func)(void */*p*/), void *p)
510{
511 atend *a = xmalloc(sizeof(*a));
512 a->next = 0;
513 a->func = func;
514 a->p = p;
515 *aa_tail = a;
516 aa_tail = &a->next;
517}
518
6e403221 519/*----- Main code ---------------------------------------------------------*/
520
521/* --- @main@ --- *
522 *
523 * Arguments: @int argc@ = number of command-line arguments
524 * @char *argv[]@ = vector of argument words
525 *
526 * Returns: Zero on success, nonzero on failure.
527 *
528 * Use: Picks entries from a word list which match particular
529 * expressions. This might be of assistance to word-game types.
530 */
531
532int main(int argc, char *argv[])
533{
534 node *n;
535 FILE *fp;
536 dstr d = DSTR_INIT;
94ed9b30 537 int ok = 0;
6e403221 538 char *p, *q, *l;
94ed9b30 539 atend *a;
6e403221 540
541 ego(argv[0]);
542 n = p_argv(argc, (const char *const *)argv);
543
544 if ((fp = fopen(file, "r")) == 0)
545 die("error opening `%s': %s", file, strerror(errno));
546 for (;;) {
547 dstr_reset(&d);
548 if (dstr_putline(&d, fp) < 0)
549 break;
550 l = d.buf + d.len;
551 for (p = q = d.buf; p < l; p++) {
552 if (!isalnum((unsigned char)*p))
553 continue;
554 *q++ = tolower((unsigned char)*p);
555 }
556 *q = 0;
557 d.len = q - d.buf;
558 if (n->func(n, d.buf, d.len)) {
559 fwrite(d.buf, 1, d.len, stdout);
560 fputc('\n', stdout);
94ed9b30 561 ok = 1;
6e403221 562 }
563 }
94ed9b30 564 if (ferror(fp) || fclose(fp))
6e403221 565 die("error reading `%s': %s", file, strerror(errno));
94ed9b30 566 for (a = aa_head; a; a = a->next) {
567 if (a->func(a->p))
568 ok = 1;
569 }
570 if (fflush(stdout) || ferror(stdout) || fclose(stdout))
571 die("error writing output: %s", strerror(errno));
572 if (!ok) pquis(stderr, "$: no matches found\n");
573 return (ok ? EX_OK : EX_NONE);
6e403221 574}
575
576/*----- That's all, folks -------------------------------------------------*/