5 * (c) 2001 Mark Wooding
8 /*----- Licensing notice --------------------------------------------------*
10 * This file is part of Anag: a simple wordgame helper.
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.
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.
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.
27 /*----- Header files ------------------------------------------------------*/
31 /*----- Static variables --------------------------------------------------*/
33 static const char *file = DICTIONARY;
35 /*----- Help text functions -----------------------------------------------*/
37 static void usage(FILE *fp)
38 { pquis(fp, "Usage: $ [-f file] expression\n"); }
40 static void version(FILE *fp)
41 { pquis(fp, "$, version " VERSION "\n"); }
43 static void help(FILE *fp)
49 Searches a wordlist, printing all of the words which match an expression.\n\
51 Options supported are:\n\
53 -h, --help display this help text\n\
54 -v, --version display the program's version number\n\
55 -u, --usage display a very brief usage message\n\
56 -f, --file FILE read wordlist from FILE, not `" DICTIONARY "'\n\
58 The basic tests in the expression are:\n\
60 -anagram WORD matches a full-length anagram\n\
61 -subgram WORD matches words which only use letters in WORD\n\
62 -wildcard PATTERN matches with wildcards `*' and `?'\n\
63 -trackword WORD matches words which can be found in a trackword\n\
64 -mono PATTERN matches words isomorphic to the given PATTERN\n\
68 -regexp REGEXP matches with an (extended) regular expression\n\
71 #if defined(HAVE_PCRE) || defined(HAVE_PCRE2)
73 -pcre REGEXP matches with a Perl-like regular expression\n\
77 -length [+|-]N matches if length is [at least|at most] N\n\
78 -longest output longest matches found here\n\
79 -shortest output shortest matches found here\n\
81 These simple tests can be combined using the operators `-a', `-o' and `-n'\n\
82 (for `and', `or' and `not'; they may also be written `&', `|' and `!' if\n\
83 you like), and grouped using parentheses `(' and `)'.\n\
87 /*----- The options parser ------------------------------------------------*/
89 /* --- Options table structure --- */
99 O_HELP, O_VERSION, O_USAGE,
101 O_AND, O_OR, O_NOT, O_LPAREN, O_RPAREN,
102 O_ANAG, O_SUBG, O_WILD, O_TRACK, O_REGEXP, O_PCRE, O_MONO, O_LENGTH,
103 O_LONGEST, O_SHORTEST,
109 static const struct opt opttab[] = {
111 /* --- Options -- don't form part of the language --- */
113 { "help", 0, OF_SHORT, O_HELP },
114 { "version", 0, OF_SHORT, O_VERSION },
115 { "usage", 0, OF_SHORT, O_USAGE },
116 { "file", 1, OF_SHORT, O_FILE },
118 /* --- Operators -- provide the basic structure of the language --- *
120 * These are also given magical names by the parser.
123 { "and", 0, OF_SHORT, O_AND },
124 { "or", 0, OF_SHORT, O_OR },
125 { "not", 0, OF_SHORT, O_NOT },
127 /* --- Actual matching operations -- do something useful --- */
129 { "anagram", 1, 0, O_ANAG },
130 { "subgram", 1, 0, O_SUBG },
131 { "wildcard", 1, 0, O_WILD },
132 { "trackword", 1, 0, O_TRACK },
133 { "mono", 1, 0, O_MONO },
135 { "regexp", 1, 0, O_REGEXP },
137 #if defined(HAVE_PCRE) || defined(HAVE_PCRE2)
138 { "pcre", 1, 0, O_PCRE },
140 { "length", 1, 0, O_LENGTH },
141 { "longest", 0, 0, O_LONGEST },
142 { "shortest", 0, 0, O_SHORTEST },
144 /* --- End marker --- */
150 static const char *const *av;
153 /* --- @nextopt@ --- *
155 * Arguments: @const char ***arg@ = where to store the arg pointer
157 * Returns: The tag of the next option.
159 * Use: Scans the next option off the command line. If the option
160 * doesn't form part of the language, it's processed internally,
161 * and you'll never see it from here. On exit, the @arg@
162 * pointer is set to contain the address of the option scanned,
163 * followed by its arguments if any. You're expected to know
164 * how many arguments there are for your option.
167 static unsigned nextopt(const char *const **arg)
170 const struct opt *o, *oo;
174 /* Pick the next option off the front. */
176 if (ai >= ac) return (O_EOF);
179 /* Cope with various forms of magic. */
181 if (!p[1]) switch (*p) {
182 case '&': return (O_AND);
183 case '|': return (O_OR);
184 case '!': return (O_NOT);
185 case '(': return (O_LPAREN);
186 case ')': return (O_RPAREN);
191 /* Now cope with other sorts of weirdies. By the end of this, a leading
192 * `-' or `--' will have been stripped.
198 if (ai < ac) die("syntax error near `--': rubbish at end of line");
202 /*Now look the word up in my table. */
205 for (o = opttab; o->name; o++) {
206 if (strncmp(p, o->name, sz) == 0) {
207 if (strlen(o->name) == sz || ((o->f & OF_SHORT) && sz == 1))
210 die("ambiguous option name `-%s' (could match `-%s' or `-%s')",
211 p, oo->name, o->name);
215 if (!oo) die("unrecognized option name `-%s'", p);
217 /* Sort out the arguments. */
218 if (ai + oo->nargs > ac)
219 die("too few arguments for `-%s' (need %u)", oo->name, oo->nargs);
222 /* Now process the option. */
224 case O_HELP: help(stdout); exit(0);
225 case O_VERSION: version(stdout); exit(0);
226 case O_USAGE: usage(stdout); exit(0);
227 case O_FILE: file = (*arg)[1]; break;
228 default: return (oo->tag);
232 die("syntax error near `%s': unknown token type", av[ai - 1]);
236 /*----- Node types for operators ------------------------------------------*/
238 /* --- Node structures --- */
240 typedef struct node_bin {
246 typedef struct node_un {
251 /* --- Node functions --- */
253 static int n_or(node *nn, const char *p, size_t sz)
255 node_bin *n = (node_bin *)nn;
256 return (n->left->func(n->left, p, sz) || n->right->func(n->right, p, sz));
259 static int n_and(node *nn, const char *p, size_t sz)
261 node_bin *n = (node_bin *)nn;
262 return (n->left->func(n->left, p, sz) && n->right->func(n->right, p, sz));
265 static int n_not(node *nn, const char *p, size_t sz)
267 node_un *n = (node_un *)nn;
268 return (!n->arg->func(n->arg, p, sz));
271 /*----- Other simple node types -------------------------------------------*/
273 enum { LESS = -1, EQUAL = 0, GREATER = 1 };
275 typedef struct node_numeric {
281 static void parse_numeric(const char *p, int *dir, int *i)
288 case '-': *dir = LESS; pp++; break;
289 case '+': *dir = GREATER; pp++; break;
290 default: *dir = EQUAL; break;
293 l = strtol(pp, &q, 0);
294 if (*q || errno || l < INT_MIN || l > INT_MAX)
295 die("bad numeric parameter `%s'", p);
299 static node *make_numeric(const char *const *av,
300 int (*func)(struct node *, const char *, size_t))
302 node_numeric *n = xmalloc(sizeof(*n));
303 parse_numeric(av[0], &n->dir, &n->i);
308 static int cmp_numeric(int x, int dir, int n)
311 case LESS: return (x <= n);
312 case EQUAL: return (x == n);
313 case GREATER: return (x >= n);
318 static int n_length(node *nn, const char *p, size_t sz)
320 node_numeric *n = (node_numeric *)nn;
321 return (cmp_numeric(sz, n->dir, n->i));
324 /*----- Parser for the expression syntax ----------------------------------*/
326 /* --- A parser context --- */
328 typedef struct p_ctx {
330 const char *const *a;
333 /* --- Parser structure --- *
335 * This is a simple recursive descent parser. The context retains
336 * information about the current token. Each function is passed the address
337 * of a node pointer to fill in. This simplifies the binary operator code
338 * somewhat, relative to returning pointers to node trees.
341 static void p_expr(p_ctx *p, node **/*nn*/);
343 static void p_next(p_ctx *p)
345 static const char *const eof[] = { "<end>", 0 };
346 p->t = nextopt(&p->a);
347 if (p->t == O_EOF) p->a = eof;
350 static void p_factor(p_ctx *p, node **nn)
353 if (p->t == O_LPAREN) {
356 if (p->t != O_RPAREN) die("syntax error near `%s': missing `)'", *p->a);
358 } else if (p->t == O_NOT) {
359 n = xmalloc(sizeof(node_un));
363 p_factor(p, &n->arg);
366 case O_ANAG: *nn = anagram(p->a + 1); break;
367 case O_SUBG: *nn = subgram(p->a + 1); break;
368 case O_WILD: *nn = wildcard(p->a + 1); break;
369 case O_TRACK: *nn = trackword(p->a + 1); break;
371 case O_REGEXP: *nn = regexp(p->a + 1); break;
373 #if defined(HAVE_PCRE) || defined(HAVE_PCRE2)
374 case O_PCRE: *nn = pcrenode(p->a + 1); break;
376 case O_MONO: *nn = mono(p->a + 1); break;
377 case O_LENGTH: *nn = make_numeric(p->a + 1, n_length); break;
378 case O_LONGEST: *nn = longest(p->a + 1); break;
379 case O_SHORTEST: *nn = shortest(p->a + 1); break;
380 default: die("syntax error near `%s': unexpected token", *p->a);
386 static void p_term(p_ctx *p, node **nn)
392 case O_AND: p_next(p); break;
394 case O_RPAREN: case O_OR: case O_EOF: return;
396 n = xmalloc(sizeof(node_bin));
404 static void p_expr(p_ctx *p, node **nn)
409 if (p->t != O_OR) break;
411 n = xmalloc(sizeof(node_bin));
419 /* --- @p_argv@ --- *
421 * Arguments: @int argc@ = number of command-line arguments
422 * @const char *const argv[]@ = vectoor of arguments
424 * Returns: A compiled node, parsed from the arguments.
426 * Use: Does the donkey-work of parsing a command-line.
429 static node *p_argv(int argc, const char *const argv[])
440 pquis(stderr, "(Run `$ --help' for more detail.)\n");
445 die("syntax error near `%s': rubbish at end of line (too many `)'s?)",
450 /*----- At-end stuff ------------------------------------------------------*/
452 /* --- @atend_register@ --- *
454 * Arguments: @int (*func)(void *)@ = function to call
455 * @void *p@ = handle to pass to it
459 * Use: Adds a function to the list of things to do at the end of the
460 * program. The function should return nonzero if it produced
464 typedef struct atend {
466 int (*func)(void */*p*/);
470 static atend *aa_head = 0, **aa_tail = &aa_head;
472 void atend_register(int (*func)(void */*p*/), void *p)
474 atend *a = xmalloc(sizeof(*a));
482 /*----- Main code ---------------------------------------------------------*/
486 * Arguments: @int argc@ = number of command-line arguments
487 * @char *argv[]@ = vector of argument words
489 * Returns: Zero on success, nonzero on failure.
491 * Use: Picks entries from a word list which match particular
492 * expressions. This might be of assistance to word-game types.
495 int main(int argc, char *argv[])
505 n = p_argv(argc, (const char *const *)argv);
507 if ((fp = fopen(file, "r")) == 0)
508 die("error opening `%s': %s", file, strerror(errno));
511 if (dstr_putline(&d, fp) < 0) break;
513 for (p = q = d.buf; p < l; p++) {
514 if (!isalnum((unsigned char)*p)) continue;
515 *q++ = tolower((unsigned char)*p);
519 if (n->func(n, d.buf, d.len)) {
520 fwrite(d.buf, 1, d.len, stdout);
525 if (ferror(fp) || fclose(fp))
526 die("error reading `%s': %s", file, strerror(errno));
527 for (a = aa_head; a; a = a->next)
528 if (a->func(a->p)) ok = 1;
529 if (fflush(stdout) || ferror(stdout) || fclose(stdout))
530 die("error writing output: %s", strerror(errno));
531 if (!ok) pquis(stderr, "$: no matches found\n");
532 return (ok ? EX_OK : EX_NONE);
535 /*----- That's all, folks -------------------------------------------------*/