xref: /freebsd/usr.bin/locale/locale.c (revision b52f49a9a0f22207ad5130ad8faba08de3ed23d8)
1 /*-
2  * Copyright (c) 2002, 2003 Alexey Zelkin <phantom@FreeBSD.org>
3  * All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions
7  * are met:
8  * 1. Redistributions of source code must retain the above copyright
9  *    notice, this list of conditions and the following disclaimer.
10  * 2. Redistributions in binary form must reproduce the above copyright
11  *    notice, this list of conditions and the following disclaimer in the
12  *    documentation and/or other materials provided with the distribution.
13  *
14  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
15  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
18  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
24  * SUCH DAMAGE.
25  *
26  * $FreeBSD$
27  */
28 
29 /*
30  * XXX: implement missing int_* (LC_MONETARY) (require libc modification) and
31  *	era_* (LC_TIME) keywords (require libc & nl_langinfo(3) extensions)
32  *
33  * XXX: correctly handle reserved 'charmap' keyword and '-m' option (require
34  *      localedef(1) implementation).  Currently it's handled via
35  *	nl_langinfo(CODESET).
36  */
37 
38 #include <sys/types.h>
39 #include <dirent.h>
40 #include <err.h>
41 #include <locale.h>
42 #include <langinfo.h>
43 #include <rune.h>		/* for _PATH_LOCALE */
44 #include <stdio.h>
45 #include <stdlib.h>
46 #include <string.h>
47 #include <stringlist.h>
48 #include <unistd.h>
49 
50 /* Local prototypes */
51 void	init_locales_list(void);
52 void	list_locales(void);
53 const char *lookup_localecat(int);
54 char	*kwval_lconv(int);
55 int	kwval_lookup(char *, char **, int *, int *);
56 void	showdetails(char *);
57 void	showkeywordslist(void);
58 void	showlocale(void);
59 void	usage(void);
60 
61 /* Global variables */
62 static StringList *locales = NULL;
63 
64 int	all_locales = 0;
65 int	all_charmaps = 0;
66 int	prt_categories = 0;
67 int	prt_keywords = 0;
68 int	more_params = 0;
69 
70 struct _lcinfo {
71 	const char	*name;
72 	int		id;
73 } lcinfo [] = {
74 	{ "LC_CTYPE",		LC_CTYPE },
75 	{ "LC_COLLATE",		LC_COLLATE },
76 	{ "LC_TIME",		LC_TIME },
77 	{ "LC_NUMERIC",		LC_NUMERIC },
78 	{ "LC_MONETARY",	LC_MONETARY },
79 	{ "LC_MESSAGES",	LC_MESSAGES }
80 };
81 #define NLCINFO (sizeof(lcinfo)/sizeof(lcinfo[0]))
82 
83 /* ids for values not referenced by nl_langinfo() */
84 #define	KW_ZERO			10000
85 #define	KW_GROUPING		(KW_ZERO+1)
86 #define KW_INT_CURR_SYMBOL 	(KW_ZERO+2)
87 #define KW_CURRENCY_SYMBOL 	(KW_ZERO+3)
88 #define KW_MON_DECIMAL_POINT 	(KW_ZERO+4)
89 #define KW_MON_THOUSANDS_SEP 	(KW_ZERO+5)
90 #define KW_MON_GROUPING 	(KW_ZERO+6)
91 #define KW_POSITIVE_SIGN 	(KW_ZERO+7)
92 #define KW_NEGATIVE_SIGN 	(KW_ZERO+8)
93 #define KW_INT_FRAC_DIGITS 	(KW_ZERO+9)
94 #define KW_FRAC_DIGITS 		(KW_ZERO+10)
95 #define KW_P_CS_PRECEDES 	(KW_ZERO+11)
96 #define KW_P_SEP_BY_SPACE 	(KW_ZERO+12)
97 #define KW_N_CS_PRECEDES 	(KW_ZERO+13)
98 #define KW_N_SEP_BY_SPACE 	(KW_ZERO+14)
99 #define KW_P_SIGN_POSN 		(KW_ZERO+15)
100 #define KW_N_SIGN_POSN 		(KW_ZERO+16)
101 
102 struct _kwinfo {
103 	const char	*name;
104 	int		isstr;		/* true - string, false - number */
105 	int		catid;		/* LC_* */
106 	int		value_ref;
107 	const char	*comment;
108 } kwinfo [] = {
109 	{ "charmap",		1, LC_CTYPE,	CODESET, "" },	/* hack */
110 
111 	{ "decimal_point",	1, LC_NUMERIC,	RADIXCHAR, "" },
112 	{ "thousands_sep",	1, LC_NUMERIC,	THOUSEP, "" },
113 	{ "grouping",		1, LC_NUMERIC,	KW_GROUPING, "" },
114 	{ "radixchar",		1, LC_NUMERIC,	RADIXCHAR,
115 	  "Same as decimal_point (FreeBSD only)" },		/* compat */
116 	{ "thousep",		1, LC_NUMERIC,	THOUSEP,
117 	  "Same as thousands_sep (FreeBSD only)" },		/* compat */
118 
119 	{ "int_curr_symbol",	1, LC_MONETARY,	KW_INT_CURR_SYMBOL, "" },
120 	{ "currency_symbol",	1, LC_MONETARY,	KW_CURRENCY_SYMBOL, "" },
121 	{ "mon_decimal_point",	1, LC_MONETARY,	KW_MON_DECIMAL_POINT, "" },
122 	{ "mon_thousands_sep",	1, LC_MONETARY,	KW_MON_THOUSANDS_SEP, "" },
123 	{ "mon_grouping",	1, LC_MONETARY,	KW_MON_GROUPING, "" },
124 	{ "positive_sign",	1, LC_MONETARY,	KW_POSITIVE_SIGN, "" },
125 	{ "negative_sign",	1, LC_MONETARY,	KW_NEGATIVE_SIGN, "" },
126 
127 	{ "int_frac_digits",	0, LC_MONETARY,	KW_INT_FRAC_DIGITS, "" },
128 	{ "frac_digits",	0, LC_MONETARY,	KW_FRAC_DIGITS, "" },
129 	{ "p_cs_precedes",	0, LC_MONETARY,	KW_P_CS_PRECEDES, "" },
130 	{ "p_sep_by_space",	0, LC_MONETARY,	KW_P_SEP_BY_SPACE, "" },
131 	{ "n_cs_precedes",	0, LC_MONETARY,	KW_N_CS_PRECEDES, "" },
132 	{ "n_sep_by_space",	0, LC_MONETARY,	KW_N_SEP_BY_SPACE, "" },
133 	{ "p_sign_posn",	0, LC_MONETARY,	KW_P_SIGN_POSN, "" },
134 	{ "n_sign_posn",	0, LC_MONETARY,	KW_N_SIGN_POSN, "" },
135 
136 	{ "d_t_fmt",		1, LC_TIME,	D_T_FMT, "" },
137 	{ "d_fmt",		1, LC_TIME,	D_FMT, "" },
138 	{ "t_fmt",		1, LC_TIME,	T_FMT, "" },
139 	{ "am_str",		1, LC_TIME,	AM_STR, "" },
140 	{ "pm_str",		1, LC_TIME,	PM_STR, "" },
141 	{ "t_fmt_ampm",		1, LC_TIME,	T_FMT_AMPM, "" },
142 	{ "day_1",		1, LC_TIME,	DAY_1, "" },
143 	{ "day_2",		1, LC_TIME,	DAY_2, "" },
144 	{ "day_3",		1, LC_TIME,	DAY_3, "" },
145 	{ "day_4",		1, LC_TIME,	DAY_4, "" },
146 	{ "day_5",		1, LC_TIME,	DAY_5, "" },
147 	{ "day_6",		1, LC_TIME,	DAY_6, "" },
148 	{ "day_7",		1, LC_TIME,	DAY_7, "" },
149 	{ "abday_1",		1, LC_TIME,	ABDAY_1, "" },
150 	{ "abday_2",		1, LC_TIME,	ABDAY_2, "" },
151 	{ "abday_3",		1, LC_TIME,	ABDAY_3, "" },
152 	{ "abday_4",		1, LC_TIME,	ABDAY_4, "" },
153 	{ "abday_5",		1, LC_TIME,	ABDAY_5, "" },
154 	{ "abday_6",		1, LC_TIME,	ABDAY_6, "" },
155 	{ "abday_7",		1, LC_TIME,	ABDAY_7, "" },
156 	{ "mon_1",		1, LC_TIME,	MON_1, "" },
157 	{ "mon_2",		1, LC_TIME,	MON_2, "" },
158 	{ "mon_3",		1, LC_TIME,	MON_3, "" },
159 	{ "mon_4",		1, LC_TIME,	MON_4, "" },
160 	{ "mon_5",		1, LC_TIME,	MON_5, "" },
161 	{ "mon_6",		1, LC_TIME,	MON_6, "" },
162 	{ "mon_7",		1, LC_TIME,	MON_7, "" },
163 	{ "mon_8",		1, LC_TIME,	MON_8, "" },
164 	{ "mon_9",		1, LC_TIME,	MON_9, "" },
165 	{ "mon_10",		1, LC_TIME,	MON_10, "" },
166 	{ "mon_11",		1, LC_TIME,	MON_11, "" },
167 	{ "mon_12",		1, LC_TIME,	MON_12, "" },
168 	{ "abmon_1",		1, LC_TIME,	ABMON_1, "" },
169 	{ "abmon_2",		1, LC_TIME,	ABMON_2, "" },
170 	{ "abmon_3",		1, LC_TIME,	ABMON_3, "" },
171 	{ "abmon_4",		1, LC_TIME,	ABMON_4, "" },
172 	{ "abmon_5",		1, LC_TIME,	ABMON_5, "" },
173 	{ "abmon_6",		1, LC_TIME,	ABMON_6, "" },
174 	{ "abmon_7",		1, LC_TIME,	ABMON_7, "" },
175 	{ "abmon_8",		1, LC_TIME,	ABMON_8, "" },
176 	{ "abmon_9",		1, LC_TIME,	ABMON_9, "" },
177 	{ "abmon_10",		1, LC_TIME,	ABMON_10, "" },
178 	{ "abmon_11",		1, LC_TIME,	ABMON_11, "" },
179 	{ "abmon_12",		1, LC_TIME,	ABMON_12, "" },
180 	{ "era",		1, LC_TIME,	ERA, "(unavailable)" },
181 	{ "era_d_fmt",		1, LC_TIME,	ERA_D_FMT, "(unavailable)" },
182 	{ "era_d_t_fmt",	1, LC_TIME,	ERA_D_T_FMT, "(unavailable)" },
183 	{ "era_t_fmt",		1, LC_TIME,	ERA_T_FMT, "(unavailable)" },
184 	{ "alt_digits",		1, LC_TIME,	ALT_DIGITS, "" },
185 	{ "d_md_order",		1, LC_TIME,	D_MD_ORDER,
186 	  "(FreeBSD only)"				},	/* local */
187 
188 	{ "yesexpr",		1, LC_MESSAGES, YESEXPR, "" },
189 	{ "noexpr",		1, LC_MESSAGES, NOEXPR, "" },
190 	{ "yesstr",		1, LC_MESSAGES, YESSTR,
191 	  "(POSIX legacy)" },					/* compat */
192 	{ "nostr",		1, LC_MESSAGES, NOSTR,
193 	  "(POSIX legacy)" }					/* compat */
194 
195 };
196 #define NKWINFO (sizeof(kwinfo)/sizeof(kwinfo[0]))
197 
198 int
199 main(int argc, char *argv[])
200 {
201 	char	ch;
202 	int	tmp;
203 
204 	while ((ch = getopt(argc, argv, "ackm")) != -1) {
205 		switch (ch) {
206 		case 'a':
207 			all_locales = 1;
208 			break;
209 		case 'c':
210 			prt_categories = 1;
211 			break;
212 		case 'k':
213 			prt_keywords = 1;
214 			break;
215 		case 'm':
216 			all_charmaps = 1;
217 			break;
218 		default:
219 			usage();
220 		}
221 	}
222 	argc -= optind;
223 	argv += optind;
224 
225 	/* validate arguments */
226 	if (all_locales && all_charmaps)
227 		usage();
228 	if ((all_locales || all_charmaps) && argc > 0)
229 		usage();
230 	if ((all_locales || all_charmaps) && (prt_categories || prt_keywords))
231 		usage();
232 	if ((prt_categories || prt_keywords) && argc <= 0)
233 		usage();
234 
235 	/* process '-a' */
236 	if (all_locales) {
237 		list_locales();
238 		exit(0);
239 	}
240 
241 	/* process '-m' */
242 	if (all_charmaps) {
243 		/*
244 		 * XXX: charmaps are not supported by FreeBSD now.  It
245 		 * need to be implemented as soon as localedef(1) implemented.
246 		 */
247 		exit(1);
248 	}
249 
250 	/* check for special case '-k list' */
251 	tmp = 0;
252 	if (prt_keywords && argc > 0)
253 		while (tmp < argc)
254 			if (strcasecmp(argv[tmp++], "list") == 0) {
255 				showkeywordslist();
256 				exit(0);
257 			}
258 
259 	/* process '-c' and/or '-k' */
260 	if (prt_categories || prt_keywords || argc > 0) {
261 		setlocale(LC_ALL, "");
262 		while (argc > 0) {
263 			showdetails(*argv);
264 			argv++;
265 			argc--;
266 		}
267 		exit(0);
268 	}
269 
270 	/* no arguments, show current locale state */
271 	showlocale();
272 
273 	return (0);
274 }
275 
276 void
277 usage(void)
278 {
279 	printf("Usage: locale [ -a | -m ]\n"
280                "       locale [ -ck ] name ...\n");
281 	exit(1);
282 }
283 
284 /*
285  * Output information about all available locales
286  *
287  * XXX actually output of this function does not guarantee that locale
288  *     is really available to application, since it can be broken or
289  *     inconsistent thus setlocale() will fail.  Maybe add '-V' function to
290  *     also validate these locales?
291  */
292 void
293 list_locales(void)
294 {
295 	size_t i;
296 
297 	init_locales_list();
298 	for (i = 0; i < locales->sl_cur; i++) {
299 		printf("%s\n", locales->sl_str[i]);
300 	}
301 }
302 
303 
304 /*
305  * qsort() helper function
306  */
307 static int
308 scmp(const void *s1, const void *s2)
309 {
310 	return strcmp(*(const char **)s1, *(const char **)s2);
311 }
312 
313 
314 /*
315  * Retrieve sorted list of system locales (or user locales, if PATH_LOCALE
316  * environment variable is set)
317  */
318 void
319 init_locales_list(void)
320 {
321 	DIR *dirp;
322 	struct dirent *dp;
323 	const char *dirname;
324 
325 	/* why call this function twice ? */
326 	if (locales != NULL)
327 		return;
328 
329 	/* initialize StringList */
330 	locales = sl_init();
331 	if (locales == NULL)
332 		err(1, "could not allocate memory");
333 
334 	/* get actual locales directory name */
335 	dirname = getenv("PATH_LOCALE");
336 	if (dirname == NULL)
337 		dirname = _PATH_LOCALE;
338 
339 	/* open locales directory */
340 	dirp = opendir(dirname);
341 	if (dirp == NULL)
342 		err(1, "could not open directory '%s'", dirname);
343 
344 	/* scan directory and store its contents except "." and ".." */
345 	while ((dp = readdir(dirp)) != NULL) {
346 		if (*(dp->d_name) == '.')
347 			continue;		/* exclude "." and ".." */
348 		sl_add(locales, strdup(dp->d_name));
349 	}
350 	closedir(dirp);
351 
352         /* make sure that 'POSIX' and 'C' locales are present in the list.
353 	 * POSIX 1003.1-2001 requires presence of 'POSIX' name only here, but
354          * we also list 'C' for constistency
355          */
356 	if (sl_find(locales, "POSIX") == NULL)
357 		sl_add(locales, "POSIX");
358 
359 	if (sl_find(locales, "C") == NULL)
360 		sl_add(locales, "C");
361 
362 	/* make output nicer, sort the list */
363 	qsort(locales->sl_str, locales->sl_cur, sizeof(char *), scmp);
364 }
365 
366 /*
367  * Show current locale status, depending on environment variables
368  */
369 void
370 showlocale(void)
371 {
372 	size_t	i;
373 	const char *lang, *vval, *eval;
374 
375 	setlocale(LC_ALL, "");
376 
377 	lang = getenv("LANG");
378 	if (lang == NULL) {
379 		lang = "";
380 	}
381 	printf("LANG=%s\n", lang);
382 	/* XXX: if LANG is null, then set it to "C" to get implied values? */
383 
384 	for (i = 0; i < NLCINFO; i++) {
385 		vval = setlocale(lcinfo[i].id, NULL);
386 		eval = getenv(lcinfo[i].name);
387 		if (eval != NULL && !strcmp(eval, vval)
388 				&& strcmp(lang, vval)) {
389 			/*
390 			 * Appropriate environment variable set, its value
391 			 * is valid and not overriden by LC_ALL
392 			 *
393 			 * XXX: possible side effect: if both LANG and
394 			 * overriden environment variable are set into same
395 			 * value, then it'll be assumed as 'implied'
396 			 */
397 			printf("%s=%s\n", lcinfo[i].name, vval);
398 		} else {
399 			printf("%s=\"%s\"\n", lcinfo[i].name, vval);
400 		}
401 	}
402 
403 	vval = getenv("LC_ALL");
404 	if (vval == NULL) {
405 		vval = "";
406 	}
407 	printf("LC_ALL=%s\n", vval);
408 }
409 
410 /*
411  * keyword value lookup helper (via localeconv())
412  */
413 char *
414 kwval_lconv(int id)
415 {
416 	struct lconv *lc;
417 	char *rval;
418 
419 	rval = NULL;
420 	lc = localeconv();
421 	switch (id) {
422 		case KW_GROUPING:
423 			rval = lc->grouping;
424 			break;
425 		case KW_INT_CURR_SYMBOL:
426 			rval = lc->int_curr_symbol;
427 			break;
428 		case KW_CURRENCY_SYMBOL:
429 			rval = lc->currency_symbol;
430 			break;
431 		case KW_MON_DECIMAL_POINT:
432 			rval = lc->mon_decimal_point;
433 			break;
434 		case KW_MON_THOUSANDS_SEP:
435 			rval = lc->mon_thousands_sep;
436 			break;
437 		case KW_MON_GROUPING:
438 			rval = lc->mon_grouping;
439 			break;
440 		case KW_POSITIVE_SIGN:
441 			rval = lc->positive_sign;
442 			break;
443 		case KW_NEGATIVE_SIGN:
444 			rval = lc->negative_sign;
445 			break;
446 		case KW_INT_FRAC_DIGITS:
447 			rval = &(lc->int_frac_digits);
448 			break;
449 		case KW_FRAC_DIGITS:
450 			rval = &(lc->frac_digits);
451 			break;
452 		case KW_P_CS_PRECEDES:
453 			rval = &(lc->p_cs_precedes);
454 			break;
455 		case KW_P_SEP_BY_SPACE:
456 			rval = &(lc->p_sep_by_space);
457 			break;
458 		case KW_N_CS_PRECEDES:
459 			rval = &(lc->n_cs_precedes);
460 			break;
461 		case KW_N_SEP_BY_SPACE:
462 			rval = &(lc->n_sep_by_space);
463 			break;
464 		case KW_P_SIGN_POSN:
465 			rval = &(lc->p_sign_posn);
466 			break;
467 		case KW_N_SIGN_POSN:
468 			rval = &(lc->n_sign_posn);
469 			break;
470 		default:
471 			break;
472 	}
473 	return (rval);
474 }
475 
476 /*
477  * keyword value and properties lookup
478  */
479 int
480 kwval_lookup(char *kwname, char **kwval, int *cat, int *isstr)
481 {
482 	int	rval;
483 	size_t	i;
484 
485 	rval = 0;
486 	for (i = 0; i < NKWINFO; i++) {
487 		if (strcasecmp(kwname, kwinfo[i].name) == 0) {
488 			rval = 1;
489 			*cat = kwinfo[i].catid;
490 			*isstr = kwinfo[i].isstr;
491 			if (kwinfo[i].value_ref < KW_ZERO) {
492 				*kwval = nl_langinfo(kwinfo[i].value_ref);
493 			} else {
494 				*kwval = kwval_lconv(kwinfo[i].value_ref);
495 			}
496 			break;
497 		}
498 	}
499 
500 	return (rval);
501 }
502 
503 /*
504  * Show details about requested keyword according to '-k' and/or '-c'
505  * command line options specified.
506  */
507 void
508 showdetails(char *kw)
509 {
510 	int	isstr, cat, tmpval;
511 	char	*kwval;
512 
513 	if (kwval_lookup(kw, &kwval, &cat, &isstr) == 0) {
514 		/*
515 		 * invalid keyword specified.
516 		 * XXX: any actions?
517 		 */
518 		return;
519 	}
520 
521 	if (prt_categories) {
522 		printf("%s\n", lookup_localecat(cat));
523 	}
524 
525 	if (prt_keywords) {
526 		if (isstr) {
527 			printf("%s=\"%s\"\n", kw, kwval);
528 		} else {
529 			tmpval = (char) *kwval;
530 			printf("%s=%d\n", kw, tmpval);
531 		}
532 	}
533 
534 	if (!prt_categories && !prt_keywords) {
535 		if (isstr) {
536 			printf("%s\n", kwval);
537 		} else {
538 			tmpval = (char) *kwval;
539 			printf("%d\n", tmpval);
540 		}
541 	}
542 }
543 
544 /*
545  * Convert locale category id into string
546  */
547 const char *
548 lookup_localecat(int cat)
549 {
550 	size_t	i;
551 
552 	for (i = 0; i < NLCINFO; i++)
553 		if (lcinfo[i].id == cat) {
554 			return (lcinfo[i].name);
555 		}
556 	return ("UNKNOWN");
557 }
558 
559 /*
560  * Show list of keywords
561  */
562 void
563 showkeywordslist(void)
564 {
565 	size_t	i;
566 
567 #define FMT "%-20s %-12s %-7s %-20s\n"
568 
569 	printf("List of available keywords\n\n");
570 	printf(FMT, "Keyword", "Category", "Type", "Comment");
571 	printf("-------------------- ------------ ------- --------------------\n");
572 	for (i = 0; i < NKWINFO; i++) {
573 		printf(FMT,
574 			kwinfo[i].name,
575 			lookup_localecat(kwinfo[i].catid),
576 			(kwinfo[i].isstr == 0) ? "number" : "string",
577 			kwinfo[i].comment);
578 	}
579 }
580