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
)
39 pquis(fp
, "Usage: $ [-f file] expression\n");
42 static void version(FILE *fp
)
44 pquis(fp
, "$, version " VERSION
"\n");
47 static void help(FILE *fp
)
53 Searches a wordlist, printing all of the words which match an expression.\n\
55 Options supported are:\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\
62 The basic tests in the expression are:\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\
68 -mono PATTERN matches words isomorphic to the given PATTERN\n\
72 -regexp REGEXP matches with an (extended) regular expression\n\
77 -pcre REGEXP matches with a Perl-like regular expression\n\
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\
85 These 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\
87 you like), and grouped using parentheses `(' and `)'.\n\
91 /*----- The options parser ------------------------------------------------*/
93 /* --- Options table structure --- */
103 O_HELP
, O_VERSION
, O_USAGE
,
105 O_AND
, O_OR
, O_NOT
, O_LPAREN
, O_RPAREN
,
106 O_ANAG
, O_SUBG
, O_WILD
, O_TRACK
, O_REGEXP
, O_PCRE
, O_MONO
, O_LENGTH
,
107 O_LONGEST
, O_SHORTEST
,
113 static const struct opt opttab
[] = {
115 /* --- Options -- don't form part of the language --- */
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
},
122 /* --- Operators -- provide the basic structure of the language --- *
124 * These are also given magical names by the parser.
127 { "and", 0, OF_SHORT
, O_AND
},
128 { "or", 0, OF_SHORT
, O_OR
},
129 { "not", 0, OF_SHORT
, O_NOT
},
131 /* --- Actual matching operations -- do something useful --- */
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 { "mono", 1, 0, O_MONO
},
139 { "regexp", 1, 0, O_REGEXP
},
142 { "pcre", 1, 0, O_PCRE
},
144 { "length", 1, 0, O_LENGTH
},
145 { "longest", 0, 0, O_LONGEST
},
146 { "shortest", 0, 0, O_SHORTEST
},
148 /* --- End marker --- */
154 static const char *const *av
;
157 /* --- @nextopt@ --- *
159 * Arguments: @const char ***arg@ = where to store the arg pointer
161 * Returns: The tag of the next option.
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.
171 static unsigned nextopt(const char *const **arg
)
174 const struct opt
*o
, *oo
;
178 /* --- Pick the next option off the front --- */
185 /* --- Cope with various forms of magic --- */
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
);
198 /* --- Now cope with other sorts of weirdies --- *
200 * By the end of this, a leading `-' or `--' will have been stripped.
210 die("syntax error near `--': rubbish at end of line");
214 /* --- Now look the word up in my table --- */
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)) {
225 die("ambiguous option name `-%s' (could match `-%s' or `-%s')",
226 p
, oo
->name
, o
->name
);
232 die("unrecognized option name `-%s'", p
);
234 /* --- Sort out the arguments --- */
236 if (ai
+ oo
->nargs
> ac
)
237 die("too few arguments for `-%s' (need %u)", oo
->name
, oo
->nargs
);
240 /* --- Now process the option --- */
260 die("syntax error near `%s': unknown token type", av
[ai
- 1]);
264 /*----- Node types for operators ------------------------------------------*/
266 /* --- Node structures --- */
268 typedef struct node_bin
{
274 typedef struct node_un
{
279 /* --- Node functions --- */
281 static int n_or(node
*nn
, const char *p
, size_t sz
)
283 node_bin
*n
= (node_bin
*)nn
;
284 return (n
->left
->func(n
->left
, p
, sz
) || n
->right
->func(n
->right
, p
, sz
));
287 static int n_and(node
*nn
, const char *p
, size_t sz
)
289 node_bin
*n
= (node_bin
*)nn
;
290 return (n
->left
->func(n
->left
, p
, sz
) && n
->right
->func(n
->right
, p
, sz
));
293 static int n_not(node
*nn
, const char *p
, size_t sz
)
295 node_un
*n
= (node_un
*)nn
;
296 return (!n
->arg
->func(n
->arg
, p
, sz
));
299 /*----- Other simple node types -------------------------------------------*/
301 enum { LESS
= -1, EQUAL
= 0, GREATER
= 1 };
303 typedef struct node_numeric
{
309 static void parse_numeric(const char *p
, int *dir
, int *i
)
316 case '-': *dir
= LESS
; pp
++; break;
317 case '+': *dir
= GREATER
; pp
++; break;
318 default: *dir
= EQUAL
; break;
321 l
= strtol(pp
, &q
, 0);
322 if (*q
|| errno
|| l
< INT_MIN
|| l
> INT_MAX
)
323 die("bad numeric parameter `%s'", p
);
327 static node
*make_numeric(const char *const *av
,
328 int (*func
)(struct node
*, const char *, size_t))
330 node_numeric
*n
= xmalloc(sizeof(*n
));
331 parse_numeric(av
[0], &n
->dir
, &n
->i
);
336 static int cmp_numeric(int x
, int dir
, int n
)
339 case LESS
: return (x
<= n
);
340 case EQUAL
: return (x
== n
);
341 case GREATER
: return (x
>= n
);
346 static int n_length(node
*nn
, const char *p
, size_t sz
)
348 node_numeric
*n
= (node_numeric
*)nn
;
349 return (cmp_numeric(sz
, n
->dir
, n
->i
));
352 /*----- Parser for the expression syntax ----------------------------------*/
354 /* --- A parser context --- */
356 typedef struct p_ctx
{
358 const char *const *a
;
361 /* --- Parser structure --- *
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.
369 static void p_expr(p_ctx
*p
, node
**/
*nn*/
);
371 static void p_next(p_ctx
*p
)
373 static const char *const eof
[] = { "<end>", 0 };
374 p
->t
= nextopt(&p
->a
);
379 static void p_factor(p_ctx
*p
, node
**nn
)
382 if (p
->t
== O_LPAREN
) {
385 if (p
->t
!= O_RPAREN
)
386 die("syntax error near `%s': missing `)'", *p
->a
);
388 } else if (p
->t
== O_NOT
) {
389 n
= xmalloc(sizeof(node_un
));
393 p_factor(p
, &n
->arg
);
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;
401 case O_REGEXP
: *nn
= regexp(p
->a
+ 1); break;
404 case O_PCRE
: *nn
= pcrenode(p
->a
+ 1); break;
406 case O_MONO
: *nn
= mono(p
->a
+ 1); break;
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;
410 default: die("syntax error near `%s': unexpected token", *p
->a
);
416 static void p_term(p_ctx
*p
, node
**nn
)
431 n
= xmalloc(sizeof(node_bin
));
439 static void p_expr(p_ctx
*p
, node
**nn
)
447 n
= xmalloc(sizeof(node_bin
));
455 /* --- @p_argv@ --- *
457 * Arguments: @int argc@ = number of command-line arguments
458 * @const char *const argv[]@ = vectoor of arguments
460 * Returns: A compiled node, parsed from the arguments.
462 * Use: Does the donkey-work of parsing a command-line.
465 static node
*p_argv(int argc
, const char *const argv
[])
476 pquis(stderr
, "(Run `$ --help' for more detail.)\n");
481 die("syntax error near `%s': rubbish at end of line (too many `)'s?)",
487 /*----- At-end stuff ------------------------------------------------------*/
489 /* --- @atend_register@ --- *
491 * Arguments: @int (*func)(void *)@ = function to call
492 * @void *p@ = handle to pass to it
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
501 typedef struct atend
{
503 int (*func
)(void */
*p*/
);
507 static atend
*aa_head
= 0, **aa_tail
= &aa_head
;
509 void atend_register(int (*func
)(void */
*p*/
), void *p
)
511 atend
*a
= xmalloc(sizeof(*a
));
519 /*----- Main code ---------------------------------------------------------*/
523 * Arguments: @int argc@ = number of command-line arguments
524 * @char *argv[]@ = vector of argument words
526 * Returns: Zero on success, nonzero on failure.
528 * Use: Picks entries from a word list which match particular
529 * expressions. This might be of assistance to word-game types.
532 int main(int argc
, char *argv
[])
542 n
= p_argv(argc
, (const char *const *)argv
);
544 if ((fp
= fopen(file
, "r")) == 0)
545 die("error opening `%s': %s", file
, strerror(errno
));
548 if (dstr_putline(&d
, fp
) < 0)
551 for (p
= q
= d
.buf
; p
< l
; p
++) {
552 if (!isalnum((unsigned char)*p
))
554 *q
++ = tolower((unsigned char)*p
);
558 if (n
->func(n
, d
.buf
, d
.len
)) {
559 fwrite(d
.buf
, 1, d
.len
, stdout
);
564 if (ferror(fp
) || fclose(fp
))
565 die("error reading `%s': %s", file
, strerror(errno
));
566 for (a
= aa_head
; a
; a
= a
->next
) {
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
);
576 /*----- That's all, folks -------------------------------------------------*/