xref: /freebsd/lib/libc/stdtime/strptime.c (revision 3a970562d7f8b2506f7bcbe378e3c02a408425d7)
1 /*-
2  * Copyright (c) 2014 Gary Mills
3  * Copyright 2011, Nexenta Systems, Inc.  All rights reserved.
4  * Copyright (c) 1994 Powerdog Industries.  All rights reserved.
5  *
6  * Copyright (c) 2011 The FreeBSD Foundation
7  * All rights reserved.
8  * Portions of this software were developed by David Chisnall
9  * under sponsorship from the FreeBSD Foundation.
10  *
11  * Redistribution and use in source and binary forms, with or without
12  * modification, are permitted provided that the following conditions
13  * are met:
14  * 1. Redistributions of source code must retain the above copyright
15  *    notice, this list of conditions and the following disclaimer.
16  * 2. Redistributions in binary form must reproduce the above copyright
17  *    notice, this list of conditions and the following disclaimer
18  *    in the documentation and/or other materials provided with the
19  *    distribution.
20  *
21  * THIS SOFTWARE IS PROVIDED BY POWERDOG INDUSTRIES ``AS IS'' AND ANY
22  * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
24  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE POWERDOG INDUSTRIES BE
25  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
26  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
27  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
28  * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
29  * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
30  * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
31  * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
32  *
33  * The views and conclusions contained in the software and documentation
34  * are those of the authors and should not be interpreted as representing
35  * official policies, either expressed or implied, of Powerdog Industries.
36  */
37 
38 #include <sys/cdefs.h>
39 #ifndef lint
40 #ifndef NOID
41 static char copyright[] __unused =
42 "@(#) Copyright (c) 1994 Powerdog Industries.  All rights reserved.";
43 static char sccsid[] __unused = "@(#)strptime.c	0.1 (Powerdog) 94/03/27";
44 #endif /* !defined NOID */
45 #endif /* not lint */
46 __FBSDID("$FreeBSD$");
47 
48 #include "namespace.h"
49 #include <time.h>
50 #include <ctype.h>
51 #include <errno.h>
52 #include <stdlib.h>
53 #include <string.h>
54 #include <pthread.h>
55 #include "un-namespace.h"
56 #include "libc_private.h"
57 #include "timelocal.h"
58 #include "tzfile.h"
59 
60 static char * _strptime(const char *, const char *, struct tm *, int *, locale_t);
61 
62 #define	asizeof(a)	(sizeof(a) / sizeof((a)[0]))
63 
64 #define	FLAG_NONE	(1 << 0)
65 #define	FLAG_YEAR	(1 << 1)
66 #define	FLAG_MONTH	(1 << 2)
67 #define	FLAG_YDAY	(1 << 3)
68 #define	FLAG_MDAY	(1 << 4)
69 #define	FLAG_WDAY	(1 << 5)
70 
71 /*
72  * Calculate the week day of the first day of a year. Valid for
73  * the Gregorian calendar, which began Sept 14, 1752 in the UK
74  * and its colonies. Ref:
75  * http://en.wikipedia.org/wiki/Determination_of_the_day_of_the_week
76  */
77 
78 static int
79 first_wday_of(int year)
80 {
81 	return (((2 * (3 - (year / 100) % 4)) + (year % 100) +
82 		((year % 100) / 4) + (isleap(year) ? 6 : 0) + 1) % 7);
83 }
84 
85 static char *
86 _strptime(const char *buf, const char *fmt, struct tm *tm, int *GMTp,
87 		locale_t locale)
88 {
89 	char	c;
90 	const char *ptr;
91 	int	day_offset = -1, wday_offset;
92 	int week_offset;
93 	int	i, len;
94 	int flags;
95 	int Ealternative, Oalternative;
96 	const struct lc_time_T *tptr = __get_current_time_locale(locale);
97 	static int start_of_month[2][13] = {
98 		{0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334, 365},
99 		{0, 31, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335, 366}
100 	};
101 
102 	flags = FLAG_NONE;
103 
104 	ptr = fmt;
105 	while (*ptr != 0) {
106 		c = *ptr++;
107 
108 		if (c != '%') {
109 			if (isspace_l((unsigned char)c, locale))
110 				while (*buf != 0 &&
111 				       isspace_l((unsigned char)*buf, locale))
112 					buf++;
113 			else if (c != *buf++)
114 				return (NULL);
115 			continue;
116 		}
117 
118 		Ealternative = 0;
119 		Oalternative = 0;
120 label:
121 		c = *ptr++;
122 		switch (c) {
123 		case '%':
124 			if (*buf++ != '%')
125 				return (NULL);
126 			break;
127 
128 		case '+':
129 			buf = _strptime(buf, tptr->date_fmt, tm, GMTp, locale);
130 			if (buf == NULL)
131 				return (NULL);
132 			flags |= FLAG_WDAY | FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
133 			break;
134 
135 		case 'C':
136 			if (!isdigit_l((unsigned char)*buf, locale))
137 				return (NULL);
138 
139 			/* XXX This will break for 3-digit centuries. */
140 			len = 2;
141 			for (i = 0; len && *buf != 0 &&
142 			     isdigit_l((unsigned char)*buf, locale); buf++) {
143 				i *= 10;
144 				i += *buf - '0';
145 				len--;
146 			}
147 			if (i < 19)
148 				return (NULL);
149 
150 			tm->tm_year = i * 100 - TM_YEAR_BASE;
151 			flags |= FLAG_YEAR;
152 
153 			break;
154 
155 		case 'c':
156 			buf = _strptime(buf, tptr->c_fmt, tm, GMTp, locale);
157 			if (buf == NULL)
158 				return (NULL);
159 			flags |= FLAG_WDAY | FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
160 			break;
161 
162 		case 'D':
163 			buf = _strptime(buf, "%m/%d/%y", tm, GMTp, locale);
164 			if (buf == NULL)
165 				return (NULL);
166 			flags |= FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
167 			break;
168 
169 		case 'E':
170 			if (Ealternative || Oalternative)
171 				break;
172 			Ealternative++;
173 			goto label;
174 
175 		case 'O':
176 			if (Ealternative || Oalternative)
177 				break;
178 			Oalternative++;
179 			goto label;
180 
181 		case 'F':
182 			buf = _strptime(buf, "%Y-%m-%d", tm, GMTp, locale);
183 			if (buf == NULL)
184 				return (NULL);
185 			flags |= FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
186 			break;
187 
188 		case 'R':
189 			buf = _strptime(buf, "%H:%M", tm, GMTp, locale);
190 			if (buf == NULL)
191 				return (NULL);
192 			break;
193 
194 		case 'r':
195 			buf = _strptime(buf, tptr->ampm_fmt, tm, GMTp, locale);
196 			if (buf == NULL)
197 				return (NULL);
198 			break;
199 
200 		case 'T':
201 			buf = _strptime(buf, "%H:%M:%S", tm, GMTp, locale);
202 			if (buf == NULL)
203 				return (NULL);
204 			break;
205 
206 		case 'X':
207 			buf = _strptime(buf, tptr->X_fmt, tm, GMTp, locale);
208 			if (buf == NULL)
209 				return (NULL);
210 			break;
211 
212 		case 'x':
213 			buf = _strptime(buf, tptr->x_fmt, tm, GMTp, locale);
214 			if (buf == NULL)
215 				return (NULL);
216 			flags |= FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
217 			break;
218 
219 		case 'j':
220 			if (!isdigit_l((unsigned char)*buf, locale))
221 				return (NULL);
222 
223 			len = 3;
224 			for (i = 0; len && *buf != 0 &&
225 			     isdigit_l((unsigned char)*buf, locale); buf++){
226 				i *= 10;
227 				i += *buf - '0';
228 				len--;
229 			}
230 			if (i < 1 || i > 366)
231 				return (NULL);
232 
233 			tm->tm_yday = i - 1;
234 			flags |= FLAG_YDAY;
235 
236 			break;
237 
238 		case 'M':
239 		case 'S':
240 			if (*buf == 0 ||
241 				isspace_l((unsigned char)*buf, locale))
242 				break;
243 
244 			if (!isdigit_l((unsigned char)*buf, locale))
245 				return (NULL);
246 
247 			len = 2;
248 			for (i = 0; len && *buf != 0 &&
249 				isdigit_l((unsigned char)*buf, locale); buf++){
250 				i *= 10;
251 				i += *buf - '0';
252 				len--;
253 			}
254 
255 			if (c == 'M') {
256 				if (i > 59)
257 					return (NULL);
258 				tm->tm_min = i;
259 			} else {
260 				if (i > 60)
261 					return (NULL);
262 				tm->tm_sec = i;
263 			}
264 
265 			break;
266 
267 		case 'H':
268 		case 'I':
269 		case 'k':
270 		case 'l':
271 			/*
272 			 * Of these, %l is the only specifier explicitly
273 			 * documented as not being zero-padded.  However,
274 			 * there is no harm in allowing zero-padding.
275 			 *
276 			 * XXX The %l specifier may gobble one too many
277 			 * digits if used incorrectly.
278 			 */
279 			if (!isdigit_l((unsigned char)*buf, locale))
280 				return (NULL);
281 
282 			len = 2;
283 			for (i = 0; len && *buf != 0 &&
284 			     isdigit_l((unsigned char)*buf, locale); buf++) {
285 				i *= 10;
286 				i += *buf - '0';
287 				len--;
288 			}
289 			if (c == 'H' || c == 'k') {
290 				if (i > 23)
291 					return (NULL);
292 			} else if (i > 12)
293 				return (NULL);
294 
295 			tm->tm_hour = i;
296 
297 			break;
298 
299 		case 'p':
300 			/*
301 			 * XXX This is bogus if parsed before hour-related
302 			 * specifiers.
303 			 */
304 			if (tm->tm_hour > 12)
305 				return (NULL);
306 
307 			len = strlen(tptr->am);
308 			if (strncasecmp_l(buf, tptr->am, len, locale) == 0) {
309 				if (tm->tm_hour == 12)
310 					tm->tm_hour = 0;
311 				buf += len;
312 				break;
313 			}
314 
315 			len = strlen(tptr->pm);
316 			if (strncasecmp_l(buf, tptr->pm, len, locale) == 0) {
317 				if (tm->tm_hour != 12)
318 					tm->tm_hour += 12;
319 				buf += len;
320 				break;
321 			}
322 
323 			return (NULL);
324 
325 		case 'A':
326 		case 'a':
327 			for (i = 0; i < asizeof(tptr->weekday); i++) {
328 				len = strlen(tptr->weekday[i]);
329 				if (strncasecmp_l(buf, tptr->weekday[i],
330 						len, locale) == 0)
331 					break;
332 				len = strlen(tptr->wday[i]);
333 				if (strncasecmp_l(buf, tptr->wday[i],
334 						len, locale) == 0)
335 					break;
336 			}
337 			if (i == asizeof(tptr->weekday))
338 				return (NULL);
339 
340 			buf += len;
341 			tm->tm_wday = i;
342 			flags |= FLAG_WDAY;
343 			break;
344 
345 		case 'U':
346 		case 'W':
347 			/*
348 			 * XXX This is bogus, as we can not assume any valid
349 			 * information present in the tm structure at this
350 			 * point to calculate a real value, so just check the
351 			 * range for now.
352 			 */
353 			if (!isdigit_l((unsigned char)*buf, locale))
354 				return (NULL);
355 
356 			len = 2;
357 			for (i = 0; len && *buf != 0 &&
358 			     isdigit_l((unsigned char)*buf, locale); buf++) {
359 				i *= 10;
360 				i += *buf - '0';
361 				len--;
362 			}
363 			if (i > 53)
364 				return (NULL);
365 
366 			if (c == 'U')
367 				day_offset = TM_SUNDAY;
368 			else
369 				day_offset = TM_MONDAY;
370 
371 
372 			week_offset = i;
373 
374 			break;
375 
376 		case 'u':
377 		case 'w':
378 			if (!isdigit_l((unsigned char)*buf, locale))
379 				return (NULL);
380 
381 			i = *buf++ - '0';
382 			if (i < 0 || i > 7 || (c == 'u' && i < 1) ||
383 			    (c == 'w' && i > 6))
384 				return (NULL);
385 
386 			tm->tm_wday = i % 7;
387 			flags |= FLAG_WDAY;
388 
389 			break;
390 
391 		case 'e':
392 			/*
393 			 * With %e format, our strftime(3) adds a blank space
394 			 * before single digits.
395 			 */
396 			if (*buf != 0 &&
397 			    isspace_l((unsigned char)*buf, locale))
398 			       buf++;
399 			/* FALLTHROUGH */
400 		case 'd':
401 			/*
402 			 * The %e specifier was once explicitly documented as
403 			 * not being zero-padded but was later changed to
404 			 * equivalent to %d.  There is no harm in allowing
405 			 * such padding.
406 			 *
407 			 * XXX The %e specifier may gobble one too many
408 			 * digits if used incorrectly.
409 			 */
410 			if (!isdigit_l((unsigned char)*buf, locale))
411 				return (NULL);
412 
413 			len = 2;
414 			for (i = 0; len && *buf != 0 &&
415 			     isdigit_l((unsigned char)*buf, locale); buf++) {
416 				i *= 10;
417 				i += *buf - '0';
418 				len--;
419 			}
420 			if (i > 31)
421 				return (NULL);
422 
423 			tm->tm_mday = i;
424 			flags |= FLAG_MDAY;
425 
426 			break;
427 
428 		case 'B':
429 		case 'b':
430 		case 'h':
431 			for (i = 0; i < asizeof(tptr->month); i++) {
432 				if (Oalternative) {
433 					if (c == 'B') {
434 						len = strlen(tptr->alt_month[i]);
435 						if (strncasecmp_l(buf,
436 								tptr->alt_month[i],
437 								len, locale) == 0)
438 							break;
439 					}
440 				} else {
441 					len = strlen(tptr->month[i]);
442 					if (strncasecmp_l(buf, tptr->month[i],
443 							len, locale) == 0)
444 						break;
445 				}
446 			}
447 			/*
448 			 * Try the abbreviated month name if the full name
449 			 * wasn't found and Oalternative was not requested.
450 			 */
451 			if (i == asizeof(tptr->month) && !Oalternative) {
452 				for (i = 0; i < asizeof(tptr->month); i++) {
453 					len = strlen(tptr->mon[i]);
454 					if (strncasecmp_l(buf, tptr->mon[i],
455 							len, locale) == 0)
456 						break;
457 				}
458 			}
459 			if (i == asizeof(tptr->month))
460 				return (NULL);
461 
462 			tm->tm_mon = i;
463 			buf += len;
464 			flags |= FLAG_MONTH;
465 
466 			break;
467 
468 		case 'm':
469 			if (!isdigit_l((unsigned char)*buf, locale))
470 				return (NULL);
471 
472 			len = 2;
473 			for (i = 0; len && *buf != 0 &&
474 			     isdigit_l((unsigned char)*buf, locale); buf++) {
475 				i *= 10;
476 				i += *buf - '0';
477 				len--;
478 			}
479 			if (i < 1 || i > 12)
480 				return (NULL);
481 
482 			tm->tm_mon = i - 1;
483 			flags |= FLAG_MONTH;
484 
485 			break;
486 
487 		case 's':
488 			{
489 			char *cp;
490 			int sverrno;
491 			long n;
492 			time_t t;
493 
494 			sverrno = errno;
495 			errno = 0;
496 			n = strtol_l(buf, &cp, 10, locale);
497 			if (errno == ERANGE || (long)(t = n) != n) {
498 				errno = sverrno;
499 				return (NULL);
500 			}
501 			errno = sverrno;
502 			buf = cp;
503 			if (gmtime_r(&t, tm) == NULL)
504 				return (NULL);
505 			*GMTp = 1;
506 			flags |= FLAG_YDAY | FLAG_WDAY | FLAG_MONTH |
507 			    FLAG_MDAY | FLAG_YEAR;
508 			}
509 			break;
510 
511 		case 'Y':
512 		case 'y':
513 			if (*buf == 0 ||
514 			    isspace_l((unsigned char)*buf, locale))
515 				break;
516 
517 			if (!isdigit_l((unsigned char)*buf, locale))
518 				return (NULL);
519 
520 			len = (c == 'Y') ? 4 : 2;
521 			for (i = 0; len && *buf != 0 &&
522 			     isdigit_l((unsigned char)*buf, locale); buf++) {
523 				i *= 10;
524 				i += *buf - '0';
525 				len--;
526 			}
527 			if (c == 'Y')
528 				i -= TM_YEAR_BASE;
529 			if (c == 'y' && i < 69)
530 				i += 100;
531 			if (i < 0)
532 				return (NULL);
533 
534 			tm->tm_year = i;
535 			flags |= FLAG_YEAR;
536 
537 			break;
538 
539 		case 'Z':
540 			{
541 			const char *cp;
542 			char *zonestr;
543 
544 			for (cp = buf; *cp &&
545 			     isupper_l((unsigned char)*cp, locale); ++cp) {
546 				/*empty*/}
547 			if (cp - buf) {
548 				zonestr = alloca(cp - buf + 1);
549 				strncpy(zonestr, buf, cp - buf);
550 				zonestr[cp - buf] = '\0';
551 				tzset();
552 				if (0 == strcmp(zonestr, "GMT") ||
553 				    0 == strcmp(zonestr, "UTC")) {
554 				    *GMTp = 1;
555 				} else if (0 == strcmp(zonestr, tzname[0])) {
556 				    tm->tm_isdst = 0;
557 				} else if (0 == strcmp(zonestr, tzname[1])) {
558 				    tm->tm_isdst = 1;
559 				} else {
560 				    return (NULL);
561 				}
562 				buf += cp - buf;
563 			}
564 			}
565 			break;
566 
567 		case 'z':
568 			{
569 			int sign = 1;
570 
571 			if (*buf != '+') {
572 				if (*buf == '-')
573 					sign = -1;
574 				else
575 					return (NULL);
576 			}
577 
578 			buf++;
579 			i = 0;
580 			for (len = 4; len > 0; len--) {
581 				if (isdigit_l((unsigned char)*buf, locale)) {
582 					i *= 10;
583 					i += *buf - '0';
584 					buf++;
585 				} else if (len == 2) {
586 					i *= 100;
587 					break;
588 				} else
589 					return (NULL);
590 			}
591 
592 			if (i > 1400 || (sign == -1 && i > 1200) ||
593 			    (i % 100) >= 60)
594 				return (NULL);
595 			tm->tm_hour -= sign * (i / 100);
596 			tm->tm_min  -= sign * (i % 100);
597 			*GMTp = 1;
598 			}
599 			break;
600 
601 		case 'n':
602 		case 't':
603 			while (isspace_l((unsigned char)*buf, locale))
604 				buf++;
605 			break;
606 
607 		default:
608 			return (NULL);
609 		}
610 	}
611 
612 	if (!(flags & FLAG_YDAY) && (flags & FLAG_YEAR)) {
613 		if ((flags & (FLAG_MONTH | FLAG_MDAY)) ==
614 		    (FLAG_MONTH | FLAG_MDAY)) {
615 			tm->tm_yday = start_of_month[isleap(tm->tm_year +
616 			    TM_YEAR_BASE)][tm->tm_mon] + (tm->tm_mday - 1);
617 			flags |= FLAG_YDAY;
618 		} else if (day_offset != -1) {
619 			int tmpwday, tmpyday, fwo;
620 
621 			fwo = first_wday_of(tm->tm_year + TM_YEAR_BASE);
622 			/* No incomplete week (week 0). */
623 			if (week_offset == 0 && fwo == day_offset)
624 				return (NULL);
625 
626 			/* Set the date to the first Sunday (or Monday)
627 			 * of the specified week of the year.
628 			 */
629 			tmpwday = (flags & FLAG_WDAY) ? tm->tm_wday :
630 			    day_offset;
631 			tmpyday = (7 - fwo + day_offset) % 7 +
632 			    (week_offset - 1) * 7 +
633 			    (tmpwday - day_offset + 7) % 7;
634 			/* Impossible yday for incomplete week (week 0). */
635 			if (tmpyday < 0) {
636 				if (flags & FLAG_WDAY)
637 					return (NULL);
638 				tmpyday = 0;
639 			}
640 			tm->tm_yday = tmpyday;
641 			flags |= FLAG_YDAY;
642 		}
643 	}
644 
645 	if ((flags & (FLAG_YEAR | FLAG_YDAY)) == (FLAG_YEAR | FLAG_YDAY)) {
646 		if (!(flags & FLAG_MONTH)) {
647 			i = 0;
648 			while (tm->tm_yday >=
649 			    start_of_month[isleap(tm->tm_year +
650 			    TM_YEAR_BASE)][i])
651 				i++;
652 			if (i > 12) {
653 				i = 1;
654 				tm->tm_yday -=
655 				    start_of_month[isleap(tm->tm_year +
656 				    TM_YEAR_BASE)][12];
657 				tm->tm_year++;
658 			}
659 			tm->tm_mon = i - 1;
660 			flags |= FLAG_MONTH;
661 		}
662 		if (!(flags & FLAG_MDAY)) {
663 			tm->tm_mday = tm->tm_yday -
664 			    start_of_month[isleap(tm->tm_year + TM_YEAR_BASE)]
665 			    [tm->tm_mon] + 1;
666 			flags |= FLAG_MDAY;
667 		}
668 		if (!(flags & FLAG_WDAY)) {
669 			i = 0;
670 			wday_offset = first_wday_of(tm->tm_year);
671 			while (i++ <= tm->tm_yday) {
672 				if (wday_offset++ >= 6)
673 					wday_offset = 0;
674 			}
675 			tm->tm_wday = wday_offset;
676 			flags |= FLAG_WDAY;
677 		}
678 	}
679 
680 	return ((char *)buf);
681 }
682 
683 char *
684 strptime_l(const char * __restrict buf, const char * __restrict fmt,
685     struct tm * __restrict tm, locale_t loc)
686 {
687 	char *ret;
688 	int gmt;
689 	FIX_LOCALE(loc);
690 
691 	gmt = 0;
692 	ret = _strptime(buf, fmt, tm, &gmt, loc);
693 	if (ret && gmt) {
694 		time_t t = timegm(tm);
695 
696 		localtime_r(&t, tm);
697 	}
698 
699 	return (ret);
700 }
701 
702 char *
703 strptime(const char * __restrict buf, const char * __restrict fmt,
704     struct tm * __restrict tm)
705 {
706 	return strptime_l(buf, fmt, tm, __get_locale());
707 }
708