xref: /netbsd-src/lib/libc/stdlib/strfmon.c (revision 7d62b00eb9ad855ffcd7da46b41e23feb5476fac)
1 /*	$NetBSD: strfmon.c,v 1.18 2022/08/18 11:05:02 christos Exp $	*/
2 
3 /*-
4  * Copyright (c) 2001 Alexey Zelkin <phantom@FreeBSD.org>
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  *
16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26  * SUCH DAMAGE.
27  *
28  */
29 
30 #include <sys/cdefs.h>
31 #if defined(LIBC_SCCS) && !defined(lint)
32 #if 0
33 __FBSDID("$FreeBSD: src/lib/libc/stdlib/strfmon.c,v 1.14 2003/03/20 08:18:55 ache Exp $");
34 #else
35 __RCSID("$NetBSD: strfmon.c,v 1.18 2022/08/18 11:05:02 christos Exp $");
36 #endif
37 #endif /* LIBC_SCCS and not lint */
38 
39 #include "namespace.h"
40 
41 #include <sys/types.h>
42 #include <assert.h>
43 #include <ctype.h>
44 #include <errno.h>
45 #include <limits.h>
46 #include <locale.h>
47 #include <monetary.h>
48 #include <stdarg.h>
49 #include <stddef.h>
50 #include <stdio.h>
51 #include <stdlib.h>
52 #include <string.h>
53 
54 #include "setlocale_local.h"
55 
56 /* internal flags */
57 #define	NEED_GROUPING		0x01	/* print digits grouped (default) */
58 #define	SIGN_POSN_USED		0x02	/* '+' or '(' usage flag */
59 #define	LOCALE_POSN		0x04	/* use locale defined +/- (default) */
60 #define	PARENTH_POSN		0x08	/* enclose negative amount in () */
61 #define	SUPPRESS_CURR_SYMBOL	0x10	/* suppress the currency from output */
62 #define	LEFT_JUSTIFY		0x20	/* left justify */
63 #define	USE_INTL_CURRENCY	0x40	/* use international currency symbol */
64 #define IS_NEGATIVE		0x80	/* is argument value negative ? */
65 
66 #ifndef NBCHAR_MAX
67 #define NBCHAR_MAX ((unsigned char)CHAR_MAX)
68 #endif
69 
70 /* internal macros */
71 #define PRINT(CH) do {						\
72 	if (dst >= s + maxsize) 				\
73 		goto e2big_error;				\
74 	*dst++ = CH;						\
75 } while (0)
76 
77 #define PRINTS(STR) do {					\
78 	const char *tmps = STR;					\
79 	while (*tmps != '\0')					\
80 		PRINT(*tmps++);					\
81 } while (0)
82 
83 #define GET_NUMBER(VAR)	do {					\
84 	VAR = 0;						\
85 	while (isdigit((unsigned char)*fmt)) {			\
86 		VAR *= 10;					\
87 		VAR += *fmt - '0';				\
88 		if (VAR > 0x00ffffff)				\
89 			goto e2big_error;			\
90 		fmt++;						\
91 	}							\
92 } while (0)
93 
94 #define GRPCPY(howmany) do {					\
95 	int i = howmany;					\
96 	while (i-- > 0) {					\
97 		avalue_size--;					\
98 		*--bufend = *(avalue+avalue_size+padded);	\
99 	}							\
100 } while (0)
101 
102 #define GRPSEP do {						\
103 	bufend -= thousands_sep_size;				\
104 	memcpy(bufend, thousands_sep, thousands_sep_size);	\
105 	groups++;						\
106 } while (0)
107 
108 static void __setup_vars(struct lconv *, int, char *, char *, char *, const char **);
109 static int __calc_left_pad(struct lconv *, int, char *);
110 static char *__format_grouped_double(struct lconv *, double, int *, int, int, int);
111 
112 static ssize_t
113 vstrfmon_l(char * __restrict s, size_t maxsize, locale_t loc,
114     const char * __restrict format, va_list ap)
115 {
116 	char 		*dst;		/* output destination pointer */
117 	const char 	*fmt;		/* current format poistion pointer */
118 	struct lconv 	*lc;		/* pointer to lconv structure */
119 	char		*asciivalue;	/* formatted double pointer */
120 
121 	int		flags;		/* formatting options */
122 	int		pad_char;	/* padding character */
123 	int		pad_size;	/* pad size */
124 	int		width;		/* field width */
125 	int		left_prec;	/* left precision */
126 	int		right_prec;	/* right precision */
127 	double		value;		/* just value */
128 	char		space_char = ' '; /* space after currency */
129 
130 	char		cs_precedes,	/* values gathered from struct lconv */
131 			sep_by_space,
132 			sign_posn,
133 			*currency_symbol;
134 	const char	*signstr;
135 
136 	char		*tmpptr;	/* temporary vars */
137 	int		sverrno;
138 
139 	lc = localeconv_l(loc);
140 	dst = s;
141 	fmt = format;
142 	asciivalue = NULL;
143 	currency_symbol = NULL;
144 	pad_size = 0;
145 
146 	while (*fmt) {
147 		/* pass nonformating characters AS IS */
148 		if (*fmt != '%')
149 			goto literal;
150 
151 		/* '%' found ! */
152 
153 		/* "%%" mean just '%' */
154 		if (*(fmt+1) == '%') {
155 			fmt++;
156 	literal:
157 			PRINT(*fmt++);
158 			continue;
159 		}
160 
161 		/* set up initial values */
162 		flags = (NEED_GROUPING|LOCALE_POSN);
163 		pad_char = ' ';		/* padding character is "space" */
164 		left_prec = -1;		/* no left precision specified */
165 		right_prec = -1;	/* no right precision specified */
166 		width = -1;		/* no width specified */
167 		value = 0;		/* we have no value to print now */
168 
169 		/* Flags */
170 		while (/* CONSTCOND */ 1) {
171 			switch (*++fmt) {
172 				case '=':	/* fill character */
173 					pad_char = *++fmt;
174 					if (pad_char == '\0')
175 						goto format_error;
176 					continue;
177 				case '^':	/* not group currency  */
178 					flags &= ~(NEED_GROUPING);
179 					continue;
180 				case '+':	/* use locale defined signs */
181 					if (flags & SIGN_POSN_USED)
182 						goto format_error;
183 					flags |= (SIGN_POSN_USED|LOCALE_POSN);
184 					continue;
185 				case '(':	/* enclose negatives with () */
186 					if (flags & SIGN_POSN_USED)
187 						goto format_error;
188 					flags |= (SIGN_POSN_USED|PARENTH_POSN);
189 					continue;
190 				case '!':	/* suppress currency symbol */
191 					flags |= SUPPRESS_CURR_SYMBOL;
192 					continue;
193 				case '-':	/* alignment (left)  */
194 					flags |= LEFT_JUSTIFY;
195 					continue;
196 				default:
197 					break;
198 			}
199 			break;
200 		}
201 
202 		/* field Width */
203 		if (isdigit((unsigned char)*fmt)) {
204 			ptrdiff_t d = dst - s;
205 			GET_NUMBER(width);
206 			/* Do we have enough space to put number with
207 			 * required width ?
208 			 */
209 
210 			if ((size_t)(d + width) >= maxsize)
211 				goto e2big_error;
212 		}
213 
214 		/* Left precision */
215 		if (*fmt == '#') {
216 			if (!isdigit((unsigned char)*++fmt))
217 				goto format_error;
218 			GET_NUMBER(left_prec);
219 			if ((unsigned int)left_prec >= maxsize - (dst - s))
220 				goto e2big_error;
221 		}
222 
223 		/* Right precision */
224 		if (*fmt == '.') {
225 			if (!isdigit((unsigned char)*++fmt))
226 				goto format_error;
227 			GET_NUMBER(right_prec);
228 			if ((unsigned int)right_prec >= maxsize - (dst - s) -
229 			    left_prec)
230 				goto e2big_error;
231 		}
232 
233 		/* Conversion Characters */
234 		switch (*fmt++) {
235 			case 'i':	/* use internaltion currency format */
236 				flags |= USE_INTL_CURRENCY;
237 				break;
238 			case 'n':	/* use national currency format */
239 				flags &= ~(USE_INTL_CURRENCY);
240 				break;
241 			default:	/* required character is missing or
242 					   premature EOS */
243 				goto format_error;
244 		}
245 
246 		if (currency_symbol)
247 			free(currency_symbol);
248 		if (flags & USE_INTL_CURRENCY) {
249 			currency_symbol = strdup(lc->int_curr_symbol);
250 			if (currency_symbol != NULL &&
251 			    strlen(currency_symbol) > 3) {
252 				space_char = currency_symbol[3];
253 				currency_symbol[3] = '\0';
254 			}
255 
256 		} else
257 			currency_symbol = strdup(lc->currency_symbol);
258 
259 		if (currency_symbol == NULL)
260 			goto end_error;			/* ENOMEM. */
261 
262 		/* value itself */
263 		value = va_arg(ap, double);
264 
265 		/* detect sign */
266 		if (value < 0) {
267 			flags |= IS_NEGATIVE;
268 			value = -value;
269 		}
270 
271 		/* fill left_prec with amount of padding chars */
272 		if (left_prec >= 0) {
273 			pad_size = __calc_left_pad(lc, (flags ^ IS_NEGATIVE),
274 							currency_symbol) -
275 				   __calc_left_pad(lc, flags, currency_symbol);
276 			if (pad_size < 0)
277 				pad_size = 0;
278 		}
279 
280 		if (asciivalue != NULL)
281 			free(asciivalue);
282 		asciivalue = __format_grouped_double(lc, value, &flags,
283 				left_prec, right_prec, pad_char);
284 		if (asciivalue == NULL)
285 			goto end_error;		/* errno already set     */
286 						/* to ENOMEM by malloc() */
287 
288 		/* set some variables for later use */
289 		__setup_vars(lc, flags, &cs_precedes, &sep_by_space,
290 				&sign_posn, &signstr);
291 
292 		/*
293 		 * Description of some LC_MONETARY's values:
294 		 *
295 		 * p_cs_precedes & n_cs_precedes
296 		 *
297 		 * = 1 - $currency_symbol precedes the value
298 		 *       for a monetary quantity with a non-negative value
299 		 * = 0 - symbol succeeds the value
300 		 *
301 		 * p_sep_by_space & n_sep_by_space
302                  *
303 		 * = 0 - no space separates $currency_symbol
304 		 *       from the value for a monetary quantity with a
305 		 *	 non-negative value
306 		 * = 1 - space separates the symbol from the value
307 		 * = 2 - space separates the symbol and the sign string,
308 		 *       if adjacent.
309                  *
310 		 * p_sign_posn & n_sign_posn
311                  *
312 		 * = 0 - parentheses enclose the quantity and the
313 		 *	 $currency_symbol
314 		 * = 1 - the sign string precedes the quantity and the
315 		 *       $currency_symbol
316 		 * = 2 - the sign string succeeds the quantity and the
317 		 *       $currency_symbol
318 		 * = 3 - the sign string precedes the $currency_symbol
319 		 * = 4 - the sign string succeeds the $currency_symbol
320                  *
321 		 */
322 
323 		tmpptr = dst;
324 
325 		while (pad_size-- > 0)
326 			PRINT(' ');
327 
328 		if (sign_posn == 0 && (flags & IS_NEGATIVE))
329 			PRINT('(');
330 
331 		if (cs_precedes == 1) {
332 			if (sign_posn == 1 || sign_posn == 3) {
333 				PRINTS(signstr);
334 				if (sep_by_space == 2)		/* XXX: ? */
335 					PRINT(' ');
336 			}
337 
338 			if (!(flags & SUPPRESS_CURR_SYMBOL)) {
339 				PRINTS(currency_symbol);
340 
341 				if (sign_posn == 4) {
342 					if (sep_by_space == 2)
343 						PRINT(space_char);
344 					PRINTS(signstr);
345 					if (sep_by_space == 1)
346 						PRINT(' ');
347 				} else if (sep_by_space == 1)
348 					PRINT(space_char);
349 			}
350 		} else if (sign_posn == 1)
351 			PRINTS(signstr);
352 
353 		PRINTS(asciivalue);
354 
355 		if (cs_precedes == 0) {
356 			if (sign_posn == 3) {
357 				if (sep_by_space == 1)
358 					PRINT(' ');
359 				PRINTS(signstr);
360 			}
361 
362 			if (!(flags & SUPPRESS_CURR_SYMBOL)) {
363 				if ((sign_posn == 3 && sep_by_space == 2)
364 				    || (sep_by_space == 1
365 				    && (sign_posn == 0
366 				    || sign_posn == 1
367 				    || sign_posn == 2
368 				    || sign_posn == 4)))
369 					PRINT(space_char);
370 				PRINTS(currency_symbol); /* XXX: len */
371 				if (sign_posn == 4) {
372 					if (sep_by_space == 2)
373 						PRINT(' ');
374 					PRINTS(signstr);
375 				}
376 			}
377 		}
378 
379 		if (sign_posn == 2) {
380 			if (sep_by_space == 2)
381 				PRINT(' ');
382 			PRINTS(signstr);
383 		}
384 
385 		if (sign_posn == 0 && (flags & IS_NEGATIVE))
386 			PRINT(')');
387 
388 		if (dst - tmpptr < width) {
389 			if (flags & LEFT_JUSTIFY) {
390 				while (dst - tmpptr < width)
391 					PRINT(' ');
392 			} else {
393 				_DIAGASSERT(__type_fit(int, dst - tmpptr));
394 				pad_size = dst - tmpptr;
395 				memmove(tmpptr + width-pad_size, tmpptr,
396 				    (size_t) pad_size);
397 				memset(tmpptr, ' ', (size_t) width-pad_size);
398 				dst += width-pad_size;
399 			}
400 		}
401 	}
402 
403 	PRINT('\0');
404 	free(asciivalue);
405 	free(currency_symbol);
406 	return (dst - s - 1);	/* return size of put data except trailing '\0' */
407 
408 e2big_error:
409 	errno = E2BIG;
410 	goto end_error;
411 
412 format_error:
413 	errno = EINVAL;
414 
415 end_error:
416 	sverrno = errno;
417 	if (asciivalue != NULL)
418 		free(asciivalue);
419 	if (currency_symbol != NULL)
420 		free(currency_symbol);
421 	errno = sverrno;
422 	return (-1);
423 }
424 
425 static void
426 __setup_vars(struct lconv *lc, int flags, char *cs_precedes, char *sep_by_space,
427 		char *sign_posn, const char **signstr) {
428 
429 	if ((flags & IS_NEGATIVE) && (flags & USE_INTL_CURRENCY)) {
430 		*cs_precedes = lc->int_n_cs_precedes;
431 		*sep_by_space = lc->int_n_sep_by_space;
432 		*sign_posn = (flags & PARENTH_POSN) ? 0 : lc->int_n_sign_posn;
433 		*signstr = (lc->negative_sign[0] == '\0') ? "-"
434 		    : lc->negative_sign;
435 	} else if (flags & USE_INTL_CURRENCY) {
436 		*cs_precedes = lc->int_p_cs_precedes;
437 		*sep_by_space = lc->int_p_sep_by_space;
438 		*sign_posn = (flags & PARENTH_POSN) ? 0 : lc->int_p_sign_posn;
439 		*signstr = lc->positive_sign;
440 	} else if (flags & IS_NEGATIVE) {
441 		*cs_precedes = lc->n_cs_precedes;
442 		*sep_by_space = lc->n_sep_by_space;
443 		*sign_posn = (flags & PARENTH_POSN) ? 0 : lc->n_sign_posn;
444 		*signstr = (lc->negative_sign[0] == '\0') ? "-"
445 		    : lc->negative_sign;
446 	} else {
447 		*cs_precedes = lc->p_cs_precedes;
448 		*sep_by_space = lc->p_sep_by_space;
449 		*sign_posn = (flags & PARENTH_POSN) ? 0 : lc->p_sign_posn;
450 		*signstr = lc->positive_sign;
451 	}
452 
453 	/* Set default values for unspecified information. */
454 	if (*cs_precedes != 0)
455 		*cs_precedes = 1;
456 	if ((unsigned char)*sep_by_space == NBCHAR_MAX)
457 		*sep_by_space = 1;
458 	if ((unsigned char)*sign_posn == NBCHAR_MAX)
459 		*sign_posn = 0;
460 }
461 
462 static int
463 __calc_left_pad(struct lconv *lc, int flags, char *cur_symb) {
464 
465 	char cs_precedes, sep_by_space, sign_posn;
466 	const char *signstr;
467 	size_t left_chars = 0;
468 
469 	__setup_vars(lc, flags, &cs_precedes, &sep_by_space, &sign_posn, &signstr);
470 
471 	if (cs_precedes != 0) {
472 		left_chars += strlen(cur_symb);
473 		if (sep_by_space != 0)
474 			left_chars++;
475 	}
476 
477 	switch (sign_posn) {
478 		case 1:
479 			left_chars += strlen(signstr);
480 			break;
481 		case 3:
482 		case 4:
483 			if (cs_precedes != 0)
484 				left_chars += strlen(signstr);
485 	}
486 	_DIAGASSERT(__type_fit(int, left_chars));
487 	return (int)left_chars;
488 }
489 
490 static int
491 get_groups(int size, char *grouping) {
492 
493 	int	chars = 0;
494 
495 	if ((unsigned char)*grouping == NBCHAR_MAX || *grouping <= 0)	/* no grouping ? */
496 		return (0);
497 
498 	while (size > (int)*grouping) {
499 		chars++;
500 		size -= (int)*grouping++;
501 		/* no more grouping ? */
502 		if ((unsigned char)*grouping == NBCHAR_MAX)
503 			break;
504 		/* rest grouping with same value ? */
505 		if (*grouping == 0) {
506 			chars += (size - 1) / *(grouping - 1);
507 			break;
508 		}
509 	}
510 	return (chars);
511 }
512 
513 /* convert double to locale-encoded string */
514 static char *
515 __format_grouped_double(struct lconv *lc, double value, int *flags,
516 			int left_prec, int right_prec, int pad_char) {
517 
518 	char		*rslt;
519 	char		*avalue;
520 	int		avalue_size;
521 
522 	size_t		bufsize;
523 	char		*bufend;
524 
525 	int		padded;
526 
527 	char		*grouping;
528 	const char	*decimal_point;
529 	const char	*thousands_sep;
530 	size_t		decimal_point_size;
531 	size_t		thousands_sep_size;
532 
533 	int groups = 0;
534 
535 	grouping = lc->mon_grouping;
536 	decimal_point = lc->mon_decimal_point;
537 	if (*decimal_point == '\0')
538 		decimal_point = lc->decimal_point;
539 	thousands_sep = lc->mon_thousands_sep;
540 	if (*thousands_sep == '\0')
541 		thousands_sep = lc->thousands_sep;
542 
543 	decimal_point_size = strlen(decimal_point);
544 	thousands_sep_size = strlen(thousands_sep);
545 
546 	/* fill left_prec with default value */
547 	if (left_prec == -1)
548 		left_prec = 0;
549 
550 	/* fill right_prec with default value */
551 	if (right_prec == -1) {
552                 if (*flags & USE_INTL_CURRENCY)
553                         right_prec = lc->int_frac_digits;
554                 else
555                         right_prec = lc->frac_digits;
556 
557 		if (right_prec == CHAR_MAX)	/* POSIX locale ? */
558 			right_prec = 2;
559 	}
560 
561 	if (*flags & NEED_GROUPING)
562 		left_prec += get_groups(left_prec, grouping);
563 
564 	/* convert to string */
565 	avalue_size = asprintf(&avalue, "%*.*f", left_prec + right_prec + 1,
566 	    right_prec, value);
567 	if (avalue_size < 0)
568 		return (NULL);
569 
570 	/* make sure that we've enough space for result string */
571 	bufsize = avalue_size * (1 + thousands_sep_size) + decimal_point_size +
572 	    1;
573 	rslt = calloc(1, bufsize);
574 	if (rslt == NULL) {
575 		free(avalue);
576 		return (NULL);
577 	}
578 	bufend = rslt + bufsize - 1;	/* reserve space for trailing '\0' */
579 
580 	/* skip spaces at beginning */
581 	padded = 0;
582 	while (avalue[padded] == ' ') {
583 		padded++;
584 		avalue_size--;
585 	}
586 
587 	if (right_prec > 0) {
588 		bufend -= right_prec;
589 		memcpy(bufend, avalue + avalue_size+padded-right_prec,
590 		    (size_t) right_prec);
591 		bufend -= decimal_point_size;
592 		memcpy(bufend, decimal_point, decimal_point_size);
593 		avalue_size -= (right_prec + 1);
594 	}
595 
596 	if ((*flags & NEED_GROUPING) &&
597 	    thousands_sep_size > 0 &&	/* XXX: need investigation */
598 	    (unsigned char)*grouping != NBCHAR_MAX &&
599 	    *grouping > 0) {
600 		while (avalue_size > (int)*grouping) {
601 			GRPCPY(*grouping);
602 			GRPSEP;
603 			grouping++;
604 
605 			/* no more grouping ? */
606 			if ((unsigned char)*grouping == NBCHAR_MAX)
607 				break;
608 
609 			/* rest grouping with same value ? */
610 			if (*grouping == 0) {
611 				grouping--;
612 				while (avalue_size > *grouping) {
613 					GRPCPY(*grouping);
614 					GRPSEP;
615 				}
616 			}
617 		}
618 		if (avalue_size != 0)
619 			GRPCPY(avalue_size);
620 		padded -= groups;
621 
622 	} else {
623 		bufend -= avalue_size;
624 		memcpy(bufend, avalue+padded, (size_t) avalue_size);
625 		if (right_prec == 0)
626 			padded -= decimal_point_size;
627 	}
628 
629 	/* do padding with pad_char */
630 	if (padded > 0) {
631 		bufend -= padded;
632 		memset(bufend, pad_char, (size_t) padded);
633 	}
634 
635 	bufsize = rslt + bufsize - bufend;
636 	memmove(rslt, bufend, bufsize);
637 	free(avalue);
638 	return (rslt);
639 }
640 
641 ssize_t
642 strfmon(char * __restrict s, size_t maxsize, const char * __restrict format,
643     ...)
644 {
645 	ssize_t rv;
646 	va_list ap;
647 
648 	va_start(ap, format);
649 	rv = vstrfmon_l(s, maxsize, _current_locale(), format, ap);
650 	va_end(ap);
651 
652 	return rv;
653 }
654 
655 ssize_t
656 strfmon_l(char * __restrict s, size_t maxsize, locale_t loc,
657     const char * __restrict format, ...)
658 {
659 	ssize_t rv;
660 	va_list ap;
661 
662 	va_start(ap, format);
663 	rv = vstrfmon_l(s, maxsize, loc, format, ap);
664 	va_end(ap);
665 
666 	return rv;
667 }
668