xref: /freebsd/lib/libc/stdtime/strptime.c (revision f4b37ed0f8b307b1f3f0f630ca725d68f1dff30d)
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 			len = strlen(tptr->am);
305 			if (strncasecmp_l(buf, tptr->am, len, locale) == 0) {
306 				if (tm->tm_hour > 12)
307 					return (NULL);
308 				if (tm->tm_hour == 12)
309 					tm->tm_hour = 0;
310 				buf += len;
311 				break;
312 			}
313 
314 			len = strlen(tptr->pm);
315 			if (strncasecmp_l(buf, tptr->pm, len, locale) == 0) {
316 				if (tm->tm_hour > 12)
317 					return (NULL);
318 				if (tm->tm_hour != 12)
319 					tm->tm_hour += 12;
320 				buf += len;
321 				break;
322 			}
323 
324 			return (NULL);
325 
326 		case 'A':
327 		case 'a':
328 			for (i = 0; i < asizeof(tptr->weekday); i++) {
329 				len = strlen(tptr->weekday[i]);
330 				if (strncasecmp_l(buf, tptr->weekday[i],
331 						len, locale) == 0)
332 					break;
333 				len = strlen(tptr->wday[i]);
334 				if (strncasecmp_l(buf, tptr->wday[i],
335 						len, locale) == 0)
336 					break;
337 			}
338 			if (i == asizeof(tptr->weekday))
339 				return (NULL);
340 
341 			buf += len;
342 			tm->tm_wday = i;
343 			flags |= FLAG_WDAY;
344 			break;
345 
346 		case 'U':
347 		case 'W':
348 			/*
349 			 * XXX This is bogus, as we can not assume any valid
350 			 * information present in the tm structure at this
351 			 * point to calculate a real value, so just check the
352 			 * range for now.
353 			 */
354 			if (!isdigit_l((unsigned char)*buf, locale))
355 				return (NULL);
356 
357 			len = 2;
358 			for (i = 0; len && *buf != 0 &&
359 			     isdigit_l((unsigned char)*buf, locale); buf++) {
360 				i *= 10;
361 				i += *buf - '0';
362 				len--;
363 			}
364 			if (i > 53)
365 				return (NULL);
366 
367 			if (c == 'U')
368 				day_offset = TM_SUNDAY;
369 			else
370 				day_offset = TM_MONDAY;
371 
372 
373 			week_offset = i;
374 
375 			break;
376 
377 		case 'w':
378 			if (!isdigit_l((unsigned char)*buf, locale))
379 				return (NULL);
380 
381 			i = *buf - '0';
382 			if (i > 6)
383 				return (NULL);
384 
385 			tm->tm_wday = i;
386 			flags |= FLAG_WDAY;
387 
388 			break;
389 
390 		case 'e':
391 			/*
392 			 * With %e format, our strftime(3) adds a blank space
393 			 * before single digits.
394 			 */
395 			if (*buf != 0 &&
396 			    isspace_l((unsigned char)*buf, locale))
397 			       buf++;
398 			/* FALLTHROUGH */
399 		case 'd':
400 			/*
401 			 * The %e specifier was once explicitly documented as
402 			 * not being zero-padded but was later changed to
403 			 * equivalent to %d.  There is no harm in allowing
404 			 * such padding.
405 			 *
406 			 * XXX The %e specifier may gobble one too many
407 			 * digits if used incorrectly.
408 			 */
409 			if (!isdigit_l((unsigned char)*buf, locale))
410 				return (NULL);
411 
412 			len = 2;
413 			for (i = 0; len && *buf != 0 &&
414 			     isdigit_l((unsigned char)*buf, locale); buf++) {
415 				i *= 10;
416 				i += *buf - '0';
417 				len--;
418 			}
419 			if (i > 31)
420 				return (NULL);
421 
422 			tm->tm_mday = i;
423 			flags |= FLAG_MDAY;
424 
425 			break;
426 
427 		case 'B':
428 		case 'b':
429 		case 'h':
430 			for (i = 0; i < asizeof(tptr->month); i++) {
431 				if (Oalternative) {
432 					if (c == 'B') {
433 						len = strlen(tptr->alt_month[i]);
434 						if (strncasecmp_l(buf,
435 								tptr->alt_month[i],
436 								len, locale) == 0)
437 							break;
438 					}
439 				} else {
440 					len = strlen(tptr->month[i]);
441 					if (strncasecmp_l(buf, tptr->month[i],
442 							len, locale) == 0)
443 						break;
444 				}
445 			}
446 			/*
447 			 * Try the abbreviated month name if the full name
448 			 * wasn't found and Oalternative was not requested.
449 			 */
450 			if (i == asizeof(tptr->month) && !Oalternative) {
451 				for (i = 0; i < asizeof(tptr->month); i++) {
452 					len = strlen(tptr->mon[i]);
453 					if (strncasecmp_l(buf, tptr->mon[i],
454 							len, locale) == 0)
455 						break;
456 				}
457 			}
458 			if (i == asizeof(tptr->month))
459 				return (NULL);
460 
461 			tm->tm_mon = i;
462 			buf += len;
463 			flags |= FLAG_MONTH;
464 
465 			break;
466 
467 		case 'm':
468 			if (!isdigit_l((unsigned char)*buf, locale))
469 				return (NULL);
470 
471 			len = 2;
472 			for (i = 0; len && *buf != 0 &&
473 			     isdigit_l((unsigned char)*buf, locale); buf++) {
474 				i *= 10;
475 				i += *buf - '0';
476 				len--;
477 			}
478 			if (i < 1 || i > 12)
479 				return (NULL);
480 
481 			tm->tm_mon = i - 1;
482 			flags |= FLAG_MONTH;
483 
484 			break;
485 
486 		case 's':
487 			{
488 			char *cp;
489 			int sverrno;
490 			long n;
491 			time_t t;
492 
493 			sverrno = errno;
494 			errno = 0;
495 			n = strtol_l(buf, &cp, 10, locale);
496 			if (errno == ERANGE || (long)(t = n) != n) {
497 				errno = sverrno;
498 				return (NULL);
499 			}
500 			errno = sverrno;
501 			buf = cp;
502 			if (gmtime_r(&t, tm) == NULL)
503 				return (NULL);
504 			*GMTp = 1;
505 			flags |= FLAG_YDAY | FLAG_WDAY | FLAG_MONTH |
506 			    FLAG_MDAY | FLAG_YEAR;
507 			}
508 			break;
509 
510 		case 'Y':
511 		case 'y':
512 			if (*buf == 0 ||
513 			    isspace_l((unsigned char)*buf, locale))
514 				break;
515 
516 			if (!isdigit_l((unsigned char)*buf, locale))
517 				return (NULL);
518 
519 			len = (c == 'Y') ? 4 : 2;
520 			for (i = 0; len && *buf != 0 &&
521 			     isdigit_l((unsigned char)*buf, locale); buf++) {
522 				i *= 10;
523 				i += *buf - '0';
524 				len--;
525 			}
526 			if (c == 'Y')
527 				i -= TM_YEAR_BASE;
528 			if (c == 'y' && i < 69)
529 				i += 100;
530 			if (i < 0)
531 				return (NULL);
532 
533 			tm->tm_year = i;
534 			flags |= FLAG_YEAR;
535 
536 			break;
537 
538 		case 'Z':
539 			{
540 			const char *cp;
541 			char *zonestr;
542 
543 			for (cp = buf; *cp &&
544 			     isupper_l((unsigned char)*cp, locale); ++cp) {
545 				/*empty*/}
546 			if (cp - buf) {
547 				zonestr = alloca(cp - buf + 1);
548 				strncpy(zonestr, buf, cp - buf);
549 				zonestr[cp - buf] = '\0';
550 				tzset();
551 				if (0 == strcmp(zonestr, "GMT") ||
552 				    0 == strcmp(zonestr, "UTC")) {
553 				    *GMTp = 1;
554 				} else if (0 == strcmp(zonestr, tzname[0])) {
555 				    tm->tm_isdst = 0;
556 				} else if (0 == strcmp(zonestr, tzname[1])) {
557 				    tm->tm_isdst = 1;
558 				} else {
559 				    return (NULL);
560 				}
561 				buf += cp - buf;
562 			}
563 			}
564 			break;
565 
566 		case 'z':
567 			{
568 			int sign = 1;
569 
570 			if (*buf != '+') {
571 				if (*buf == '-')
572 					sign = -1;
573 				else
574 					return (NULL);
575 			}
576 
577 			buf++;
578 			i = 0;
579 			for (len = 4; len > 0; len--) {
580 				if (isdigit_l((unsigned char)*buf, locale)) {
581 					i *= 10;
582 					i += *buf - '0';
583 					buf++;
584 				} else
585 					return (NULL);
586 			}
587 
588 			tm->tm_hour -= sign * (i / 100);
589 			tm->tm_min  -= sign * (i % 100);
590 			*GMTp = 1;
591 			}
592 			break;
593 
594 		case 'n':
595 		case 't':
596 			while (isspace_l((unsigned char)*buf, locale))
597 				buf++;
598 			break;
599 
600 		default:
601 			return (NULL);
602 		}
603 	}
604 
605 	if (!(flags & FLAG_YDAY) && (flags & FLAG_YEAR)) {
606 		if ((flags & (FLAG_MONTH | FLAG_MDAY)) ==
607 		    (FLAG_MONTH | FLAG_MDAY)) {
608 			tm->tm_yday = start_of_month[isleap(tm->tm_year +
609 			    TM_YEAR_BASE)][tm->tm_mon] + (tm->tm_mday - 1);
610 			flags |= FLAG_YDAY;
611 		} else if (day_offset != -1) {
612 			/* Set the date to the first Sunday (or Monday)
613 			 * of the specified week of the year.
614 			 */
615 			if (!(flags & FLAG_WDAY)) {
616 				tm->tm_wday = day_offset;
617 				flags |= FLAG_WDAY;
618 			}
619 			tm->tm_yday = (7 -
620 			    first_wday_of(tm->tm_year + TM_YEAR_BASE) +
621 			    day_offset) % 7 + (week_offset - 1) * 7 +
622 			    tm->tm_wday - day_offset;
623 			flags |= FLAG_YDAY;
624 		}
625 	}
626 
627 	if ((flags & (FLAG_YEAR | FLAG_YDAY)) == (FLAG_YEAR | FLAG_YDAY)) {
628 		if (!(flags & FLAG_MONTH)) {
629 			i = 0;
630 			while (tm->tm_yday >=
631 			    start_of_month[isleap(tm->tm_year +
632 			    TM_YEAR_BASE)][i])
633 				i++;
634 			if (i > 12) {
635 				i = 1;
636 				tm->tm_yday -=
637 				    start_of_month[isleap(tm->tm_year +
638 				    TM_YEAR_BASE)][12];
639 				tm->tm_year++;
640 			}
641 			tm->tm_mon = i - 1;
642 			flags |= FLAG_MONTH;
643 		}
644 		if (!(flags & FLAG_MDAY)) {
645 			tm->tm_mday = tm->tm_yday -
646 			    start_of_month[isleap(tm->tm_year + TM_YEAR_BASE)]
647 			    [tm->tm_mon] + 1;
648 			flags |= FLAG_MDAY;
649 		}
650 		if (!(flags & FLAG_WDAY)) {
651 			i = 0;
652 			wday_offset = first_wday_of(tm->tm_year);
653 			while (i++ <= tm->tm_yday) {
654 				if (wday_offset++ >= 6)
655 					wday_offset = 0;
656 			}
657 			tm->tm_wday = wday_offset;
658 			flags |= FLAG_WDAY;
659 		}
660 	}
661 
662 	return ((char *)buf);
663 }
664 
665 char *
666 strptime_l(const char * __restrict buf, const char * __restrict fmt,
667     struct tm * __restrict tm, locale_t loc)
668 {
669 	char *ret;
670 	int gmt;
671 	FIX_LOCALE(loc);
672 
673 	gmt = 0;
674 	ret = _strptime(buf, fmt, tm, &gmt, loc);
675 	if (ret && gmt) {
676 		time_t t = timegm(tm);
677 
678 		localtime_r(&t, tm);
679 	}
680 
681 	return (ret);
682 }
683 
684 char *
685 strptime(const char * __restrict buf, const char * __restrict fmt,
686     struct tm * __restrict tm)
687 {
688 	return strptime_l(buf, fmt, tm, __get_locale());
689 }
690