strptime.c revision 330897 
1/*-
2 * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
3 *
4 * Copyright (c) 2014 Gary Mills
5 * Copyright 2011, Nexenta Systems, Inc.  All rights reserved.
6 * Copyright (c) 1994 Powerdog Industries.  All rights reserved.
7 *
8 * Copyright (c) 2011 The FreeBSD Foundation
9 * All rights reserved.
10 * Portions of this software were developed by David Chisnall
11 * under sponsorship from the FreeBSD Foundation.
12 *
13 * Redistribution and use in source and binary forms, with or without
14 * modification, are permitted provided that the following conditions
15 * are met:
16 * 1. Redistributions of source code must retain the above copyright
17 *    notice, this list of conditions and the following disclaimer.
18 * 2. Redistributions in binary form must reproduce the above copyright
19 *    notice, this list of conditions and the following disclaimer
20 *    in the documentation and/or other materials provided with the
21 *    distribution.
22 *
23 * THIS SOFTWARE IS PROVIDED BY POWERDOG INDUSTRIES ``AS IS'' AND ANY
24 * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
26 * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE POWERDOG INDUSTRIES BE
27 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
28 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
29 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
30 * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
31 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
32 * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
33 * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
34 *
35 * The views and conclusions contained in the software and documentation
36 * are those of the authors and should not be interpreted as representing
37 * official policies, either expressed or implied, of Powerdog Industries.
38 */
39
40#include <sys/cdefs.h>
41#ifndef lint
42#ifndef NOID
43static char copyright[] __unused =
44"@(#) Copyright (c) 1994 Powerdog Industries.  All rights reserved.";
45static char sccsid[] __unused = "@(#)strptime.c	0.1 (Powerdog) 94/03/27";
46#endif /* !defined NOID */
47#endif /* not lint */
48__FBSDID("$FreeBSD: stable/11/lib/libc/stdtime/strptime.c 330897 2018-03-14 03:19:51Z eadler $");
49
50#include "namespace.h"
51#include <time.h>
52#include <ctype.h>
53#include <errno.h>
54#include <stdlib.h>
55#include <string.h>
56#include <pthread.h>
57#include "un-namespace.h"
58#include "libc_private.h"
59#include "timelocal.h"
60#include "tzfile.h"
61
62static char * _strptime(const char *, const char *, struct tm *, int *, locale_t);
63
64#define	asizeof(a)	(sizeof(a) / sizeof((a)[0]))
65
66#define	FLAG_NONE	(1 << 0)
67#define	FLAG_YEAR	(1 << 1)
68#define	FLAG_MONTH	(1 << 2)
69#define	FLAG_YDAY	(1 << 3)
70#define	FLAG_MDAY	(1 << 4)
71#define	FLAG_WDAY	(1 << 5)
72
73/*
74 * Calculate the week day of the first day of a year. Valid for
75 * the Gregorian calendar, which began Sept 14, 1752 in the UK
76 * and its colonies. Ref:
77 * http://en.wikipedia.org/wiki/Determination_of_the_day_of_the_week
78 */
79
80static int
81first_wday_of(int year)
82{
83	return (((2 * (3 - (year / 100) % 4)) + (year % 100) +
84		((year % 100) / 4) + (isleap(year) ? 6 : 0) + 1) % 7);
85}
86
87static char *
88_strptime(const char *buf, const char *fmt, struct tm *tm, int *GMTp,
89		locale_t locale)
90{
91	char	c;
92	const char *ptr;
93	int	day_offset = -1, wday_offset;
94	int week_offset;
95	int	i, len;
96	int flags;
97	int Ealternative, Oalternative;
98	const struct lc_time_T *tptr = __get_current_time_locale(locale);
99	static int start_of_month[2][13] = {
100		{0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334, 365},
101		{0, 31, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335, 366}
102	};
103
104	flags = FLAG_NONE;
105
106	ptr = fmt;
107	while (*ptr != 0) {
108		c = *ptr++;
109
110		if (c != '%') {
111			if (isspace_l((unsigned char)c, locale))
112				while (*buf != 0 &&
113				       isspace_l((unsigned char)*buf, locale))
114					buf++;
115			else if (c != *buf++)
116				return (NULL);
117			continue;
118		}
119
120		Ealternative = 0;
121		Oalternative = 0;
122label:
123		c = *ptr++;
124		switch (c) {
125		case '%':
126			if (*buf++ != '%')
127				return (NULL);
128			break;
129
130		case '+':
131			buf = _strptime(buf, tptr->date_fmt, tm, GMTp, locale);
132			if (buf == NULL)
133				return (NULL);
134			flags |= FLAG_WDAY | FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
135			break;
136
137		case 'C':
138			if (!isdigit_l((unsigned char)*buf, locale))
139				return (NULL);
140
141			/* XXX This will break for 3-digit centuries. */
142			len = 2;
143			for (i = 0; len && *buf != 0 &&
144			     isdigit_l((unsigned char)*buf, locale); buf++) {
145				i *= 10;
146				i += *buf - '0';
147				len--;
148			}
149			if (i < 19)
150				return (NULL);
151
152			tm->tm_year = i * 100 - TM_YEAR_BASE;
153			flags |= FLAG_YEAR;
154
155			break;
156
157		case 'c':
158			buf = _strptime(buf, tptr->c_fmt, tm, GMTp, locale);
159			if (buf == NULL)
160				return (NULL);
161			flags |= FLAG_WDAY | FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
162			break;
163
164		case 'D':
165			buf = _strptime(buf, "%m/%d/%y", tm, GMTp, locale);
166			if (buf == NULL)
167				return (NULL);
168			flags |= FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
169			break;
170
171		case 'E':
172			if (Ealternative || Oalternative)
173				break;
174			Ealternative++;
175			goto label;
176
177		case 'O':
178			if (Ealternative || Oalternative)
179				break;
180			Oalternative++;
181			goto label;
182
183		case 'F':
184			buf = _strptime(buf, "%Y-%m-%d", tm, GMTp, locale);
185			if (buf == NULL)
186				return (NULL);
187			flags |= FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
188			break;
189
190		case 'R':
191			buf = _strptime(buf, "%H:%M", tm, GMTp, locale);
192			if (buf == NULL)
193				return (NULL);
194			break;
195
196		case 'r':
197			buf = _strptime(buf, tptr->ampm_fmt, tm, GMTp, locale);
198			if (buf == NULL)
199				return (NULL);
200			break;
201
202		case 'T':
203			buf = _strptime(buf, "%H:%M:%S", tm, GMTp, locale);
204			if (buf == NULL)
205				return (NULL);
206			break;
207
208		case 'X':
209			buf = _strptime(buf, tptr->X_fmt, tm, GMTp, locale);
210			if (buf == NULL)
211				return (NULL);
212			break;
213
214		case 'x':
215			buf = _strptime(buf, tptr->x_fmt, tm, GMTp, locale);
216			if (buf == NULL)
217				return (NULL);
218			flags |= FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
219			break;
220
221		case 'j':
222			if (!isdigit_l((unsigned char)*buf, locale))
223				return (NULL);
224
225			len = 3;
226			for (i = 0; len && *buf != 0 &&
227			     isdigit_l((unsigned char)*buf, locale); buf++){
228				i *= 10;
229				i += *buf - '0';
230				len--;
231			}
232			if (i < 1 || i > 366)
233				return (NULL);
234
235			tm->tm_yday = i - 1;
236			flags |= FLAG_YDAY;
237
238			break;
239
240		case 'M':
241		case 'S':
242			if (*buf == 0 ||
243				isspace_l((unsigned char)*buf, locale))
244				break;
245
246			if (!isdigit_l((unsigned char)*buf, locale))
247				return (NULL);
248
249			len = 2;
250			for (i = 0; len && *buf != 0 &&
251				isdigit_l((unsigned char)*buf, locale); buf++){
252				i *= 10;
253				i += *buf - '0';
254				len--;
255			}
256
257			if (c == 'M') {
258				if (i > 59)
259					return (NULL);
260				tm->tm_min = i;
261			} else {
262				if (i > 60)
263					return (NULL);
264				tm->tm_sec = i;
265			}
266
267			break;
268
269		case 'H':
270		case 'I':
271		case 'k':
272		case 'l':
273			/*
274			 * Of these, %l is the only specifier explicitly
275			 * documented as not being zero-padded.  However,
276			 * there is no harm in allowing zero-padding.
277			 *
278			 * XXX The %l specifier may gobble one too many
279			 * digits if used incorrectly.
280			 */
281			if (!isdigit_l((unsigned char)*buf, locale))
282				return (NULL);
283
284			len = 2;
285			for (i = 0; len && *buf != 0 &&
286			     isdigit_l((unsigned char)*buf, locale); buf++) {
287				i *= 10;
288				i += *buf - '0';
289				len--;
290			}
291			if (c == 'H' || c == 'k') {
292				if (i > 23)
293					return (NULL);
294			} else if (i > 12)
295				return (NULL);
296
297			tm->tm_hour = i;
298
299			break;
300
301		case 'p':
302			/*
303			 * XXX This is bogus if parsed before hour-related
304			 * specifiers.
305			 */
306			if (tm->tm_hour > 12)
307				return (NULL);
308
309			len = strlen(tptr->am);
310			if (strncasecmp_l(buf, tptr->am, len, locale) == 0) {
311				if (tm->tm_hour == 12)
312					tm->tm_hour = 0;
313				buf += len;
314				break;
315			}
316
317			len = strlen(tptr->pm);
318			if (strncasecmp_l(buf, tptr->pm, len, locale) == 0) {
319				if (tm->tm_hour != 12)
320					tm->tm_hour += 12;
321				buf += len;
322				break;
323			}
324
325			return (NULL);
326
327		case 'A':
328		case 'a':
329			for (i = 0; i < asizeof(tptr->weekday); i++) {
330				len = strlen(tptr->weekday[i]);
331				if (strncasecmp_l(buf, tptr->weekday[i],
332						len, locale) == 0)
333					break;
334				len = strlen(tptr->wday[i]);
335				if (strncasecmp_l(buf, tptr->wday[i],
336						len, locale) == 0)
337					break;
338			}
339			if (i == asizeof(tptr->weekday))
340				return (NULL);
341
342			buf += len;
343			tm->tm_wday = i;
344			flags |= FLAG_WDAY;
345			break;
346
347		case 'U':
348		case 'W':
349			/*
350			 * XXX This is bogus, as we can not assume any valid
351			 * information present in the tm structure at this
352			 * point to calculate a real value, so just check the
353			 * range for now.
354			 */
355			if (!isdigit_l((unsigned char)*buf, locale))
356				return (NULL);
357
358			len = 2;
359			for (i = 0; len && *buf != 0 &&
360			     isdigit_l((unsigned char)*buf, locale); buf++) {
361				i *= 10;
362				i += *buf - '0';
363				len--;
364			}
365			if (i > 53)
366				return (NULL);
367
368			if (c == 'U')
369				day_offset = TM_SUNDAY;
370			else
371				day_offset = TM_MONDAY;
372
373
374			week_offset = i;
375
376			break;
377
378		case 'u':
379		case 'w':
380			if (!isdigit_l((unsigned char)*buf, locale))
381				return (NULL);
382
383			i = *buf++ - '0';
384			if (i < 0 || i > 7 || (c == 'u' && i < 1) ||
385			    (c == 'w' && i > 6))
386				return (NULL);
387
388			tm->tm_wday = i % 7;
389			flags |= FLAG_WDAY;
390
391			break;
392
393		case 'e':
394			/*
395			 * With %e format, our strftime(3) adds a blank space
396			 * before single digits.
397			 */
398			if (*buf != 0 &&
399			    isspace_l((unsigned char)*buf, locale))
400			       buf++;
401			/* FALLTHROUGH */
402		case 'd':
403			/*
404			 * The %e specifier was once explicitly documented as
405			 * not being zero-padded but was later changed to
406			 * equivalent to %d.  There is no harm in allowing
407			 * such padding.
408			 *
409			 * XXX The %e specifier may gobble one too many
410			 * digits if used incorrectly.
411			 */
412			if (!isdigit_l((unsigned char)*buf, locale))
413				return (NULL);
414
415			len = 2;
416			for (i = 0; len && *buf != 0 &&
417			     isdigit_l((unsigned char)*buf, locale); buf++) {
418				i *= 10;
419				i += *buf - '0';
420				len--;
421			}
422			if (i > 31)
423				return (NULL);
424
425			tm->tm_mday = i;
426			flags |= FLAG_MDAY;
427
428			break;
429
430		case 'B':
431		case 'b':
432		case 'h':
433			for (i = 0; i < asizeof(tptr->month); i++) {
434				if (Oalternative) {
435					if (c == 'B') {
436						len = strlen(tptr->alt_month[i]);
437						if (strncasecmp_l(buf,
438								tptr->alt_month[i],
439								len, locale) == 0)
440							break;
441					}
442				} else {
443					len = strlen(tptr->month[i]);
444					if (strncasecmp_l(buf, tptr->month[i],
445							len, locale) == 0)
446						break;
447				}
448			}
449			/*
450			 * Try the abbreviated month name if the full name
451			 * wasn't found and Oalternative was not requested.
452			 */
453			if (i == asizeof(tptr->month) && !Oalternative) {
454				for (i = 0; i < asizeof(tptr->month); i++) {
455					len = strlen(tptr->mon[i]);
456					if (strncasecmp_l(buf, tptr->mon[i],
457							len, locale) == 0)
458						break;
459				}
460			}
461			if (i == asizeof(tptr->month))
462				return (NULL);
463
464			tm->tm_mon = i;
465			buf += len;
466			flags |= FLAG_MONTH;
467
468			break;
469
470		case 'm':
471			if (!isdigit_l((unsigned char)*buf, locale))
472				return (NULL);
473
474			len = 2;
475			for (i = 0; len && *buf != 0 &&
476			     isdigit_l((unsigned char)*buf, locale); buf++) {
477				i *= 10;
478				i += *buf - '0';
479				len--;
480			}
481			if (i < 1 || i > 12)
482				return (NULL);
483
484			tm->tm_mon = i - 1;
485			flags |= FLAG_MONTH;
486
487			break;
488
489		case 's':
490			{
491			char *cp;
492			int sverrno;
493			long n;
494			time_t t;
495
496			sverrno = errno;
497			errno = 0;
498			n = strtol_l(buf, &cp, 10, locale);
499			if (errno == ERANGE || (long)(t = n) != n) {
500				errno = sverrno;
501				return (NULL);
502			}
503			errno = sverrno;
504			buf = cp;
505			if (gmtime_r(&t, tm) == NULL)
506				return (NULL);
507			*GMTp = 1;
508			flags |= FLAG_YDAY | FLAG_WDAY | FLAG_MONTH |
509			    FLAG_MDAY | FLAG_YEAR;
510			}
511			break;
512
513		case 'Y':
514		case 'y':
515			if (*buf == 0 ||
516			    isspace_l((unsigned char)*buf, locale))
517				break;
518
519			if (!isdigit_l((unsigned char)*buf, locale))
520				return (NULL);
521
522			len = (c == 'Y') ? 4 : 2;
523			for (i = 0; len && *buf != 0 &&
524			     isdigit_l((unsigned char)*buf, locale); buf++) {
525				i *= 10;
526				i += *buf - '0';
527				len--;
528			}
529			if (c == 'Y')
530				i -= TM_YEAR_BASE;
531			if (c == 'y' && i < 69)
532				i += 100;
533			if (i < 0)
534				return (NULL);
535
536			tm->tm_year = i;
537			flags |= FLAG_YEAR;
538
539			break;
540
541		case 'Z':
542			{
543			const char *cp;
544			char *zonestr;
545
546			for (cp = buf; *cp &&
547			     isupper_l((unsigned char)*cp, locale); ++cp) {
548				/*empty*/}
549			if (cp - buf) {
550				zonestr = alloca(cp - buf + 1);
551				strncpy(zonestr, buf, cp - buf);
552				zonestr[cp - buf] = '\0';
553				tzset();
554				if (0 == strcmp(zonestr, "GMT") ||
555				    0 == strcmp(zonestr, "UTC")) {
556				    *GMTp = 1;
557				} else if (0 == strcmp(zonestr, tzname[0])) {
558				    tm->tm_isdst = 0;
559				} else if (0 == strcmp(zonestr, tzname[1])) {
560				    tm->tm_isdst = 1;
561				} else {
562				    return (NULL);
563				}
564				buf += cp - buf;
565			}
566			}
567			break;
568
569		case 'z':
570			{
571			int sign = 1;
572
573			if (*buf != '+') {
574				if (*buf == '-')
575					sign = -1;
576				else
577					return (NULL);
578			}
579
580			buf++;
581			i = 0;
582			for (len = 4; len > 0; len--) {
583				if (isdigit_l((unsigned char)*buf, locale)) {
584					i *= 10;
585					i += *buf - '0';
586					buf++;
587				} else if (len == 2) {
588					i *= 100;
589					break;
590				} else
591					return (NULL);
592			}
593
594			if (i > 1400 || (sign == -1 && i > 1200) ||
595			    (i % 100) >= 60)
596				return (NULL);
597			tm->tm_hour -= sign * (i / 100);
598			tm->tm_min  -= sign * (i % 100);
599			*GMTp = 1;
600			}
601			break;
602
603		case 'n':
604		case 't':
605			while (isspace_l((unsigned char)*buf, locale))
606				buf++;
607			break;
608
609		default:
610			return (NULL);
611		}
612	}
613
614	if (!(flags & FLAG_YDAY) && (flags & FLAG_YEAR)) {
615		if ((flags & (FLAG_MONTH | FLAG_MDAY)) ==
616		    (FLAG_MONTH | FLAG_MDAY)) {
617			tm->tm_yday = start_of_month[isleap(tm->tm_year +
618			    TM_YEAR_BASE)][tm->tm_mon] + (tm->tm_mday - 1);
619			flags |= FLAG_YDAY;
620		} else if (day_offset != -1) {
621			int tmpwday, tmpyday, fwo;
622
623			fwo = first_wday_of(tm->tm_year + TM_YEAR_BASE);
624			/* No incomplete week (week 0). */
625			if (week_offset == 0 && fwo == day_offset)
626				return (NULL);
627
628			/* Set the date to the first Sunday (or Monday)
629			 * of the specified week of the year.
630			 */
631			tmpwday = (flags & FLAG_WDAY) ? tm->tm_wday :
632			    day_offset;
633			tmpyday = (7 - fwo + day_offset) % 7 +
634			    (week_offset - 1) * 7 +
635			    (tmpwday - day_offset + 7) % 7;
636			/* Impossible yday for incomplete week (week 0). */
637			if (tmpyday < 0) {
638				if (flags & FLAG_WDAY)
639					return (NULL);
640				tmpyday = 0;
641			}
642			tm->tm_yday = tmpyday;
643			flags |= FLAG_YDAY;
644		}
645	}
646
647	if ((flags & (FLAG_YEAR | FLAG_YDAY)) == (FLAG_YEAR | FLAG_YDAY)) {
648		if (!(flags & FLAG_MONTH)) {
649			i = 0;
650			while (tm->tm_yday >=
651			    start_of_month[isleap(tm->tm_year +
652			    TM_YEAR_BASE)][i])
653				i++;
654			if (i > 12) {
655				i = 1;
656				tm->tm_yday -=
657				    start_of_month[isleap(tm->tm_year +
658				    TM_YEAR_BASE)][12];
659				tm->tm_year++;
660			}
661			tm->tm_mon = i - 1;
662			flags |= FLAG_MONTH;
663		}
664		if (!(flags & FLAG_MDAY)) {
665			tm->tm_mday = tm->tm_yday -
666			    start_of_month[isleap(tm->tm_year + TM_YEAR_BASE)]
667			    [tm->tm_mon] + 1;
668			flags |= FLAG_MDAY;
669		}
670		if (!(flags & FLAG_WDAY)) {
671			i = 0;
672			wday_offset = first_wday_of(tm->tm_year);
673			while (i++ <= tm->tm_yday) {
674				if (wday_offset++ >= 6)
675					wday_offset = 0;
676			}
677			tm->tm_wday = wday_offset;
678			flags |= FLAG_WDAY;
679		}
680	}
681
682	return ((char *)buf);
683}
684
685char *
686strptime_l(const char * __restrict buf, const char * __restrict fmt,
687    struct tm * __restrict tm, locale_t loc)
688{
689	char *ret;
690	int gmt;
691	FIX_LOCALE(loc);
692
693	gmt = 0;
694	ret = _strptime(buf, fmt, tm, &gmt, loc);
695	if (ret && gmt) {
696		time_t t = timegm(tm);
697
698		localtime_r(&t, tm);
699	}
700
701	return (ret);
702}
703
704char *
705strptime(const char * __restrict buf, const char * __restrict fmt,
706    struct tm * __restrict tm)
707{
708	return strptime_l(buf, fmt, tm, __get_locale());
709}
710