5158 sed dumps core in new multibyte code
[illumos-gate.git] / usr / src / lib / libc / port / locale / strfmon.c
blobcc90ddf9f12fe98686f0e5747401236b75260af3
1 /*
2 * Copyright 2013 Garrett D'Amore <garrett@damore.org>
3 * Copyright 2010 Nexenta Systems, Inc. All rights reserved.
4 * Copyright (c) 2001 Alexey Zelkin <phantom@FreeBSD.org>
5 * All rights reserved.
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.
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.
30 #ifndef _LCONV_C99
31 #define _LCONV_C99
32 #endif
34 #include "lint.h"
35 #include <sys/types.h>
36 #include <ctype.h>
37 #include <errno.h>
38 #include <limits.h>
39 #include <locale.h>
40 #include <monetary.h>
41 #include <stdarg.h>
42 #include <stdio.h>
43 #include <stdlib.h>
44 #include <string.h>
45 #include "localeimpl.h"
46 #include "lmonetary.h"
47 #include "lnumeric.h"
49 /* internal flags */
50 #define NEED_GROUPING 0x01 /* print digits grouped (default) */
51 #define SIGN_POSN_USED 0x02 /* '+' or '(' usage flag */
52 #define LOCALE_POSN 0x04 /* use locale defined +/- (default) */
53 #define PARENTH_POSN 0x08 /* enclose negative amount in () */
54 #define SUPRESS_CURR_SYMBOL 0x10 /* supress the currency from output */
55 #define LEFT_JUSTIFY 0x20 /* left justify */
56 #define USE_INTL_CURRENCY 0x40 /* use international currency symbol */
57 #define IS_NEGATIVE 0x80 /* is argument value negative ? */
59 /* internal macros */
60 #define PRINT(CH) { \
61 if (dst >= s + maxsize) \
62 goto e2big_error; \
63 *dst++ = CH; \
66 #define PRINTS(STR) { \
67 const char *tmps = STR; \
68 while (*tmps != '\0') \
69 PRINT(*tmps++); \
72 #define GET_NUMBER(VAR) { \
73 VAR = 0; \
74 while (isdigit((unsigned char)*fmt)) { \
75 if (VAR > INT_MAX / 10) \
76 goto e2big_error; \
77 VAR *= 10; \
78 VAR += *fmt - '0'; \
79 if (VAR < 0) \
80 goto e2big_error; \
81 fmt++; \
82 } \
85 #define GRPCPY(howmany) { \
86 int i = howmany; \
87 while (i-- > 0) { \
88 avalue_size--; \
89 *--bufend = *(avalue+avalue_size+padded); \
90 } \
93 #define GRPSEP { \
94 bufend -= thousands_len; \
95 (void) memcpy(bufend, thousands_sep, thousands_len); \
96 groups++; \
99 static void setup_vars(const struct lc_monetary *, int, char *, char *, char *,
100 const char **);
101 static int calc_left_pad(const struct lc_monetary *, int, const char *);
102 static char *format_grouped_double(const struct lc_monetary *,
103 const struct lc_numeric *, double, int *, int, int, int);
105 ssize_t
106 strfmon_impl(char *_RESTRICT_KYWD s, size_t maxsize, locale_t loc,
107 const char *_RESTRICT_KYWD format, va_list ap)
109 char *dst; /* output destination pointer */
110 const char *fmt; /* current format poistion pointer */
111 char *asciivalue; /* formatted double pointer */
113 int flags; /* formatting options */
114 int pad_char; /* padding character */
115 int pad_size; /* pad size */
116 int width; /* field width */
117 int left_prec; /* left precision */
118 int right_prec; /* right precision */
119 double value; /* just value */
120 char space_char = ' '; /* space after currency */
122 char cs_precedes; /* values from struct lc_monetary */
123 char sep_by_space;
124 char sign_posn;
125 const char *signstr;
126 const char *currency_symbol;
128 char *tmpptr; /* temporary vars */
129 int sverrno;
130 const struct lc_monetary *lmon; /* monetary structure */
131 const struct lc_numeric *lnum; /* numeric structure */
133 lmon = loc->monetary;
134 lnum = loc->numeric;
136 dst = s;
137 fmt = format;
138 asciivalue = NULL;
139 currency_symbol = NULL;
140 pad_size = 0;
142 while (*fmt) {
143 /* pass nonformating characters AS IS */
144 if (*fmt != '%')
145 goto literal;
147 /* '%' found ! */
149 /* "%%" mean just '%' */
150 if (*(fmt+1) == '%') {
151 fmt++;
152 literal:
153 PRINT(*fmt++);
154 continue;
157 /* set up initial values */
158 flags = (NEED_GROUPING|LOCALE_POSN);
159 pad_char = ' '; /* padding character is "space" */
160 left_prec = -1; /* no left precision specified */
161 right_prec = -1; /* no right precision specified */
162 width = -1; /* no width specified */
163 value = 0; /* we have no value to print now */
165 /* Flags */
166 for (;;) {
167 switch (*++fmt) {
168 case '=': /* fill character */
169 pad_char = *++fmt;
170 if (pad_char == '\0')
171 goto format_error;
172 continue;
173 case '^': /* not group currency */
174 flags &= ~(NEED_GROUPING);
175 continue;
176 case '+': /* use locale defined signs */
177 if (flags & SIGN_POSN_USED)
178 goto format_error;
179 flags |= (SIGN_POSN_USED|LOCALE_POSN);
180 continue;
181 case '(': /* enclose negatives with () */
182 if (flags & SIGN_POSN_USED)
183 goto format_error;
184 flags |= (SIGN_POSN_USED|PARENTH_POSN);
185 continue;
186 case '!': /* suppress currency symbol */
187 flags |= SUPRESS_CURR_SYMBOL;
188 continue;
189 case '-': /* alignment (left) */
190 flags |= LEFT_JUSTIFY;
191 continue;
192 default:
193 break;
195 break;
198 /* field Width */
199 if (isdigit((unsigned char)*fmt)) {
200 GET_NUMBER(width);
202 * Do we have enough space to put number with
203 * required width ?
205 if ((unsigned int)width >= maxsize - (dst - s))
206 goto e2big_error;
209 /* Left precision */
210 if (*fmt == '#') {
211 if (!isdigit((unsigned char)*++fmt))
212 goto format_error;
213 GET_NUMBER(left_prec);
214 if ((unsigned int)left_prec >= maxsize - (dst - s))
215 goto e2big_error;
218 /* Right precision */
219 if (*fmt == '.') {
220 if (!isdigit((unsigned char)*++fmt))
221 goto format_error;
222 GET_NUMBER(right_prec);
223 if ((unsigned int)right_prec >= maxsize - (dst - s) -
224 left_prec)
225 goto e2big_error;
228 /* Conversion Characters */
229 switch (*fmt++) {
230 case 'i': /* use internaltion currency format */
231 flags |= USE_INTL_CURRENCY;
232 break;
233 case 'n': /* use national currency format */
234 flags &= ~(USE_INTL_CURRENCY);
235 break;
236 default:
237 /* required char missing or premature EOS */
238 goto format_error;
241 if (flags & USE_INTL_CURRENCY) {
242 currency_symbol = lmon->int_curr_symbol;
243 /* by definition three letters followed by a space */
244 if (currency_symbol != NULL)
245 space_char = currency_symbol[3];
246 } else
247 currency_symbol = lmon->currency_symbol;
249 /* value itself */
250 value = va_arg(ap, double);
252 /* detect sign */
253 if (value < 0) {
254 flags |= IS_NEGATIVE;
255 value = -value;
258 /* fill left_prec with amount of padding chars */
259 if (left_prec >= 0) {
260 pad_size = calc_left_pad(lmon, (flags ^ IS_NEGATIVE),
261 currency_symbol) -
262 calc_left_pad(lmon, flags, currency_symbol);
263 if (pad_size < 0)
264 pad_size = 0;
267 if (asciivalue != NULL)
268 free(asciivalue);
269 asciivalue = format_grouped_double(lmon, lnum, value, &flags,
270 left_prec, right_prec, pad_char);
271 if (asciivalue == NULL)
272 goto end_error; /* errno already set */
273 /* to ENOMEM by malloc() */
275 /* set some variables for later use */
276 setup_vars(lmon, flags, &cs_precedes, &sep_by_space,
277 &sign_posn, &signstr);
280 * Description of some LC_MONETARY's values:
282 * p_cs_precedes & n_cs_precedes
284 * = 1 - $currency_symbol precedes the value
285 * for a monetary quantity with a non-negative value
286 * = 0 - symbol succeeds the value
288 * p_sep_by_space & n_sep_by_space
290 * = 0 - no space separates $currency_symbol
291 * from the value for a monetary quantity with a
292 * non-negative value
293 * = 1 - space separates the symbol from the value
294 * = 2 - space separates the symbol and the sign string,
295 * if adjacent.
297 * p_sign_posn & n_sign_posn
299 * = 0 - parentheses enclose the quantity and the
300 * $currency_symbol
301 * = 1 - the sign string precedes the quantity and the
302 * $currency_symbol
303 * = 2 - the sign string succeeds the quantity and the
304 * $currency_symbol
305 * = 3 - the sign string precedes the $currency_symbol
306 * = 4 - the sign string succeeds the $currency_symbol
310 tmpptr = dst;
312 while (pad_size-- > 0)
313 PRINT(' ');
315 if (sign_posn == 0 && (flags & IS_NEGATIVE))
316 PRINT('(');
318 if (cs_precedes == 1) {
319 if (sign_posn == 1 || sign_posn == 3) {
320 PRINTS(signstr);
321 if (sep_by_space == 2)
322 PRINT(' ');
325 if (!(flags & SUPRESS_CURR_SYMBOL)) {
326 PRINTS(currency_symbol);
328 if (sign_posn == 4) {
329 if (sep_by_space == 2)
330 PRINT(space_char);
331 PRINTS(signstr);
332 if (sep_by_space == 1)
333 PRINT(' ');
334 } else if (sep_by_space == 1)
335 PRINT(space_char);
337 } else if (sign_posn == 1)
338 PRINTS(signstr);
340 PRINTS(asciivalue);
342 if (cs_precedes == 0) {
343 if (sign_posn == 3) {
344 if (sep_by_space == 1)
345 PRINT(' ');
346 PRINTS(signstr);
349 if (!(flags & SUPRESS_CURR_SYMBOL)) {
350 if ((sign_posn == 3 && sep_by_space == 2) ||
351 (sep_by_space == 1 && (sign_posn == 0 ||
352 sign_posn == 1 || sign_posn == 2 ||
353 sign_posn == 4)))
354 PRINT(space_char);
355 PRINTS(currency_symbol); /* XXX: len */
356 if (sign_posn == 4) {
357 if (sep_by_space == 2)
358 PRINT(' ');
359 PRINTS(signstr);
364 if (sign_posn == 2) {
365 if (sep_by_space == 2)
366 PRINT(' ');
367 PRINTS(signstr);
370 if (sign_posn == 0 && (flags & IS_NEGATIVE))
371 PRINT(')');
373 if (dst - tmpptr < width) {
374 if (flags & LEFT_JUSTIFY) {
375 while (dst - tmpptr < width)
376 PRINT(' ');
377 } else {
378 pad_size = dst-tmpptr;
379 (void) memmove(tmpptr + width-pad_size, tmpptr,
380 pad_size);
381 (void) memset(tmpptr, ' ', width-pad_size);
382 dst += width-pad_size;
387 PRINT('\0');
388 free(asciivalue);
389 return (dst - s - 1); /* size of put data except trailing '\0' */
391 e2big_error:
392 errno = E2BIG;
393 goto end_error;
395 format_error:
396 errno = EINVAL;
398 end_error:
399 sverrno = errno;
400 if (asciivalue != NULL)
401 free(asciivalue);
402 errno = sverrno;
403 return (-1);
406 ssize_t
407 strfmon(char *_RESTRICT_KYWD s, size_t maxsize,
408 const char *_RESTRICT_KYWD format, ...)
410 va_list ap;
411 ssize_t ret;
413 va_start(ap, format);
414 ret = strfmon_impl(s, maxsize, uselocale(NULL), format, ap);
415 va_end(ap);
416 return (ret);
419 ssize_t
420 strfmon_l(char *_RESTRICT_KYWD s, size_t maxsize, locale_t loc,
421 const char *_RESTRICT_KYWD format, ...)
423 ssize_t ret;
424 va_list ap;
425 va_start(ap, format);
426 ret = strfmon_impl(s, maxsize, loc, format, ap);
427 va_end(ap);
428 return (ret);
431 static void
432 setup_vars(const struct lc_monetary *lmon, int flags, char *cs_precedes,
433 char *sep_by_space, char *sign_posn, const char **signstr)
435 if ((flags & IS_NEGATIVE) && (flags & USE_INTL_CURRENCY)) {
436 *cs_precedes = lmon->int_n_cs_precedes[0];
437 *sep_by_space = lmon->int_n_sep_by_space[0];
438 *sign_posn = (flags & PARENTH_POSN) ? 0 :
439 lmon->int_n_sign_posn[0];
440 *signstr = (lmon->negative_sign[0] == '\0') ? "-" :
441 lmon->negative_sign;
442 } else if (flags & USE_INTL_CURRENCY) {
443 *cs_precedes = lmon->int_p_cs_precedes[0];
444 *sep_by_space = lmon->int_p_sep_by_space[0];
445 *sign_posn = (flags & PARENTH_POSN) ? 0 :
446 lmon->int_p_sign_posn[0];
447 *signstr = lmon->positive_sign;
448 } else if (flags & IS_NEGATIVE) {
449 *cs_precedes = lmon->n_cs_precedes[0];
450 *sep_by_space = lmon->n_sep_by_space[0];
451 *sign_posn = (flags & PARENTH_POSN) ? 0 : lmon->n_sign_posn[0];
452 *signstr = (lmon->negative_sign[0] == '\0') ? "-" :
453 lmon->negative_sign;
454 } else {
455 *cs_precedes = lmon->p_cs_precedes[0];
456 *sep_by_space = lmon->p_sep_by_space[0];
457 *sign_posn = (flags & PARENTH_POSN) ? 0 : lmon->p_sign_posn[0];
458 *signstr = lmon->positive_sign;
461 /* Set default values for unspecified information. */
462 if (*cs_precedes != 0)
463 *cs_precedes = 1;
464 if (*sep_by_space == CHAR_MAX)
465 *sep_by_space = 0;
466 if (*sign_posn == CHAR_MAX)
467 *sign_posn = 0;
470 static int
471 calc_left_pad(const struct lc_monetary *lmon, int flags, const char *cur_symb)
473 char cs_precedes, sep_by_space, sign_posn;
474 const char *signstr;
475 int left_chars = 0;
477 setup_vars(lmon, flags, &cs_precedes, &sep_by_space, &sign_posn,
478 &signstr);
480 if (cs_precedes != 0) {
481 left_chars += strlen(cur_symb);
482 if (sep_by_space != 0)
483 left_chars++;
486 switch (sign_posn) {
487 case 1:
488 left_chars += strlen(signstr);
489 break;
490 case 3:
491 case 4:
492 if (cs_precedes != 0)
493 left_chars += strlen(signstr);
495 return (left_chars);
498 static int
499 get_groups(int size, const char *grouping)
502 int chars = 0;
504 if (*grouping == CHAR_MAX || *grouping <= 0) /* no grouping ? */
505 return (0);
507 while (size > (int)*grouping) {
508 chars++;
509 size -= (int)*grouping++;
510 /* no more grouping ? */
511 if (*grouping == CHAR_MAX)
512 break;
513 /* rest grouping with same value ? */
514 if (*grouping == 0) {
515 chars += (size - 1) / *(grouping - 1);
516 break;
519 return (chars);
522 /* convert double to ASCII */
523 static char *
524 format_grouped_double(const struct lc_monetary *lmon,
525 const struct lc_numeric *lnum,
526 double value, int *flags, int left_prec, int right_prec, int pad_char)
529 char *rslt;
530 char *avalue;
531 int avalue_size;
532 char fmt[32];
534 size_t bufsize;
535 char *bufend;
537 int padded;
539 const char *grouping;
540 const char *decimal_point;
541 const char *thousands_sep;
542 int decimal_len;
543 int thousands_len;
545 int groups = 0;
547 grouping = lmon->mon_grouping;
548 decimal_point = lmon->mon_decimal_point;
549 if (*decimal_point == '\0')
550 decimal_point = lnum->decimal_point;
551 thousands_sep = lmon->mon_thousands_sep;
552 if (*thousands_sep == '\0')
553 thousands_sep = lnum->thousands_sep;
555 decimal_len = strlen(decimal_point); /* usually 1 */
556 thousands_len = strlen(thousands_sep); /* 0 or 1 usually */
558 /* fill left_prec with default value */
559 if (left_prec == -1)
560 left_prec = 0;
562 /* fill right_prec with default value */
563 if (right_prec == -1) {
564 if (*flags & USE_INTL_CURRENCY)
565 right_prec = lmon->int_frac_digits[0];
566 else
567 right_prec = lmon->frac_digits[0];
569 if (right_prec == CHAR_MAX) /* POSIX locale ? */
570 right_prec = 2;
573 if (*flags & NEED_GROUPING)
574 left_prec += get_groups(left_prec, grouping);
576 /* convert to string */
577 (void) snprintf(fmt, sizeof (fmt), "%%%d.%df",
578 left_prec + right_prec + 1, right_prec);
579 avalue_size = asprintf(&avalue, fmt, value);
580 if (avalue_size < 0)
581 return (NULL);
584 * Make sure that we've enough space for result string.
585 * This assumes that digits take up at least much space as
586 * grouping and radix characters. The worst case currently known
587 * is for Arabic, where two-byte UTF-8 sequences are used for both
588 * decimal and thousands seperators, and groups can be a small as two
589 * decimal digits. This will do no worse than doubling the storage
590 * requirement.
592 bufsize = strlen(avalue)*2+1;
593 rslt = calloc(1, bufsize);
594 if (rslt == NULL) {
595 free(avalue);
596 return (NULL);
598 bufend = rslt + bufsize - 1; /* reserve space for trailing '\0' */
600 /* skip spaces at beginning */
601 padded = 0;
602 while (avalue[padded] == ' ') {
603 padded++;
604 avalue_size--;
607 if (right_prec > 0) {
608 bufend -= right_prec;
609 (void) memcpy(bufend, avalue + avalue_size+padded-right_prec,
610 right_prec);
611 bufend -= decimal_len;
612 (void) memcpy(bufend, decimal_point, decimal_len);
613 avalue_size -= (right_prec + decimal_len);
616 if ((*flags & NEED_GROUPING) &&
617 thousands_len != 0 &&
618 *grouping != CHAR_MAX &&
619 *grouping > 0) {
620 while (avalue_size > (int)*grouping) {
621 GRPCPY(*grouping);
622 GRPSEP;
623 grouping++;
625 /* no more grouping ? */
626 if (*grouping == CHAR_MAX)
627 break;
629 /* rest grouping with same value ? */
630 if (*grouping == 0) {
631 grouping--;
632 while (avalue_size > *grouping) {
633 GRPCPY(*grouping);
634 GRPSEP;
638 if (avalue_size != 0)
639 GRPCPY(avalue_size);
640 padded -= groups;
642 } else {
643 bufend -= avalue_size;
644 (void) memcpy(bufend, avalue+padded, avalue_size);
645 if (right_prec == 0)
646 padded--; /* decrease assumed $decimal_point */
649 /* do padding with pad_char */
650 if (padded > 0) {
651 bufend -= padded;
652 (void) memset(bufend, pad_char, padded);
655 bufsize = bufsize - (bufend - rslt) + 1;
656 (void) memmove(rslt, bufend, bufsize);
657 free(avalue);
658 return (rslt);