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
43 static char copyright[] __unused =
44 "@(#) Copyright (c) 1994 Powerdog Industries. All rights reserved.";
45 static char sccsid[] __unused = "@(#)strptime.c 0.1 (Powerdog) 94/03/27";
46 #endif /* !defined NOID */
47 #endif /* not lint */
48 __FBSDID("$FreeBSD$");
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
62 static 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
80 static int
first_wday_of(int year)81 first_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
87 static char *
_strptime(const char * buf,const char * fmt,struct tm * tm,int * GMTp,locale_t locale)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 int century, year;
99 const struct lc_time_T *tptr = __get_current_time_locale(locale);
100 static int start_of_month[2][13] = {
101 {0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334, 365},
102 {0, 31, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335, 366}
103 };
104
105 flags = FLAG_NONE;
106 century = -1;
107 year = -1;
108
109 ptr = fmt;
110 while (*ptr != 0) {
111 c = *ptr++;
112
113 if (c != '%') {
114 if (isspace_l((unsigned char)c, locale))
115 while (*buf != 0 &&
116 isspace_l((unsigned char)*buf, locale))
117 buf++;
118 else if (c != *buf++)
119 return (NULL);
120 continue;
121 }
122
123 Ealternative = 0;
124 Oalternative = 0;
125 label:
126 c = *ptr++;
127 switch (c) {
128 case '%':
129 if (*buf++ != '%')
130 return (NULL);
131 break;
132
133 case '+':
134 buf = _strptime(buf, tptr->date_fmt, tm, GMTp, locale);
135 if (buf == NULL)
136 return (NULL);
137 flags |= FLAG_WDAY | FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
138 break;
139
140 case 'C':
141 if (!isdigit_l((unsigned char)*buf, locale))
142 return (NULL);
143
144 /* XXX This will break for 3-digit centuries. */
145 len = 2;
146 for (i = 0; len && *buf != 0 &&
147 isdigit_l((unsigned char)*buf, locale); buf++) {
148 i *= 10;
149 i += *buf - '0';
150 len--;
151 }
152
153 century = i;
154 flags |= FLAG_YEAR;
155
156 break;
157
158 case 'c':
159 buf = _strptime(buf, tptr->c_fmt, tm, GMTp, locale);
160 if (buf == NULL)
161 return (NULL);
162 flags |= FLAG_WDAY | FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
163 break;
164
165 case 'D':
166 buf = _strptime(buf, "%m/%d/%y", tm, GMTp, locale);
167 if (buf == NULL)
168 return (NULL);
169 flags |= FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
170 break;
171
172 case 'E':
173 if (Ealternative || Oalternative)
174 break;
175 Ealternative++;
176 goto label;
177
178 case 'O':
179 if (Ealternative || Oalternative)
180 break;
181 Oalternative++;
182 goto label;
183
184 case 'F':
185 buf = _strptime(buf, "%Y-%m-%d", tm, GMTp, locale);
186 if (buf == NULL)
187 return (NULL);
188 flags |= FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
189 break;
190
191 case 'R':
192 buf = _strptime(buf, "%H:%M", tm, GMTp, locale);
193 if (buf == NULL)
194 return (NULL);
195 break;
196
197 case 'r':
198 buf = _strptime(buf, tptr->ampm_fmt, tm, GMTp, locale);
199 if (buf == NULL)
200 return (NULL);
201 break;
202
203 case 'T':
204 buf = _strptime(buf, "%H:%M:%S", tm, GMTp, locale);
205 if (buf == NULL)
206 return (NULL);
207 break;
208
209 case 'X':
210 buf = _strptime(buf, tptr->X_fmt, tm, GMTp, locale);
211 if (buf == NULL)
212 return (NULL);
213 break;
214
215 case 'x':
216 buf = _strptime(buf, tptr->x_fmt, tm, GMTp, locale);
217 if (buf == NULL)
218 return (NULL);
219 flags |= FLAG_MONTH | FLAG_MDAY | FLAG_YEAR;
220 break;
221
222 case 'j':
223 if (!isdigit_l((unsigned char)*buf, locale))
224 return (NULL);
225
226 len = 3;
227 for (i = 0; len && *buf != 0 &&
228 isdigit_l((unsigned char)*buf, locale); buf++){
229 i *= 10;
230 i += *buf - '0';
231 len--;
232 }
233 if (i < 1 || i > 366)
234 return (NULL);
235
236 tm->tm_yday = i - 1;
237 flags |= FLAG_YDAY;
238
239 break;
240
241 case 'M':
242 case 'S':
243 if (*buf == 0 ||
244 isspace_l((unsigned char)*buf, locale))
245 break;
246
247 if (!isdigit_l((unsigned char)*buf, locale))
248 return (NULL);
249
250 len = 2;
251 for (i = 0; len && *buf != 0 &&
252 isdigit_l((unsigned char)*buf, locale); buf++){
253 i *= 10;
254 i += *buf - '0';
255 len--;
256 }
257
258 if (c == 'M') {
259 if (i > 59)
260 return (NULL);
261 tm->tm_min = i;
262 } else {
263 if (i > 60)
264 return (NULL);
265 tm->tm_sec = i;
266 }
267
268 break;
269
270 case 'H':
271 case 'I':
272 case 'k':
273 case 'l':
274 /*
275 * %k and %l specifiers are documented as being
276 * blank-padded. However, there is no harm in
277 * allowing zero-padding.
278 *
279 * XXX %k and %l specifiers may gobble one too many
280 * digits if used incorrectly.
281 */
282
283 len = 2;
284 if ((c == 'k' || c == 'l') &&
285 isblank_l((unsigned char)*buf, locale)) {
286 buf++;
287 len = 1;
288 }
289
290 if (!isdigit_l((unsigned char)*buf, locale))
291 return (NULL);
292
293 for (i = 0; len && *buf != 0 &&
294 isdigit_l((unsigned char)*buf, locale); buf++) {
295 i *= 10;
296 i += *buf - '0';
297 len--;
298 }
299 if (c == 'H' || c == 'k') {
300 if (i > 23)
301 return (NULL);
302 } else if (i == 0 || i > 12)
303 return (NULL);
304
305 tm->tm_hour = i;
306
307 break;
308
309 case 'p':
310 /*
311 * XXX This is bogus if parsed before hour-related
312 * specifiers.
313 */
314 if (tm->tm_hour > 12)
315 return (NULL);
316
317 len = strlen(tptr->am);
318 if (strncasecmp_l(buf, tptr->am, len, locale) == 0) {
319 if (tm->tm_hour == 12)
320 tm->tm_hour = 0;
321 buf += len;
322 break;
323 }
324
325 len = strlen(tptr->pm);
326 if (strncasecmp_l(buf, tptr->pm, len, locale) == 0) {
327 if (tm->tm_hour != 12)
328 tm->tm_hour += 12;
329 buf += len;
330 break;
331 }
332
333 return (NULL);
334
335 case 'A':
336 case 'a':
337 for (i = 0; i < asizeof(tptr->weekday); i++) {
338 len = strlen(tptr->weekday[i]);
339 if (strncasecmp_l(buf, tptr->weekday[i],
340 len, locale) == 0)
341 break;
342 len = strlen(tptr->wday[i]);
343 if (strncasecmp_l(buf, tptr->wday[i],
344 len, locale) == 0)
345 break;
346 }
347 if (i == asizeof(tptr->weekday))
348 return (NULL);
349
350 buf += len;
351 tm->tm_wday = i;
352 flags |= FLAG_WDAY;
353 break;
354
355 case 'U':
356 case 'W':
357 /*
358 * XXX This is bogus, as we can not assume any valid
359 * information present in the tm structure at this
360 * point to calculate a real value, so just check the
361 * range for now.
362 */
363 if (!isdigit_l((unsigned char)*buf, locale))
364 return (NULL);
365
366 len = 2;
367 for (i = 0; len && *buf != 0 &&
368 isdigit_l((unsigned char)*buf, locale); buf++) {
369 i *= 10;
370 i += *buf - '0';
371 len--;
372 }
373 if (i > 53)
374 return (NULL);
375
376 if (c == 'U')
377 day_offset = TM_SUNDAY;
378 else
379 day_offset = TM_MONDAY;
380
381
382 week_offset = i;
383
384 break;
385
386 case 'u':
387 case 'w':
388 if (!isdigit_l((unsigned char)*buf, locale))
389 return (NULL);
390
391 i = *buf++ - '0';
392 if (i < 0 || i > 7 || (c == 'u' && i < 1) ||
393 (c == 'w' && i > 6))
394 return (NULL);
395
396 tm->tm_wday = i % 7;
397 flags |= FLAG_WDAY;
398
399 break;
400
401 case 'e':
402 /*
403 * With %e format, our strftime(3) adds a blank space
404 * before single digits.
405 */
406 if (*buf != 0 &&
407 isspace_l((unsigned char)*buf, locale))
408 buf++;
409 /* FALLTHROUGH */
410 case 'd':
411 /*
412 * The %e specifier was once explicitly documented as
413 * not being zero-padded but was later changed to
414 * equivalent to %d. There is no harm in allowing
415 * such padding.
416 *
417 * XXX The %e specifier may gobble one too many
418 * digits if used incorrectly.
419 */
420 if (!isdigit_l((unsigned char)*buf, locale))
421 return (NULL);
422
423 len = 2;
424 for (i = 0; len && *buf != 0 &&
425 isdigit_l((unsigned char)*buf, locale); buf++) {
426 i *= 10;
427 i += *buf - '0';
428 len--;
429 }
430 if (i == 0 || i > 31)
431 return (NULL);
432
433 tm->tm_mday = i;
434 flags |= FLAG_MDAY;
435
436 break;
437
438 case 'B':
439 case 'b':
440 case 'h':
441 for (i = 0; i < asizeof(tptr->month); i++) {
442 if (Oalternative) {
443 if (c == 'B') {
444 len = strlen(tptr->alt_month[i]);
445 if (strncasecmp_l(buf,
446 tptr->alt_month[i],
447 len, locale) == 0)
448 break;
449 }
450 } else {
451 len = strlen(tptr->month[i]);
452 if (strncasecmp_l(buf, tptr->month[i],
453 len, locale) == 0)
454 break;
455 }
456 }
457 /*
458 * Try the abbreviated month name if the full name
459 * wasn't found and Oalternative was not requested.
460 */
461 if (i == asizeof(tptr->month) && !Oalternative) {
462 for (i = 0; i < asizeof(tptr->month); i++) {
463 len = strlen(tptr->mon[i]);
464 if (strncasecmp_l(buf, tptr->mon[i],
465 len, locale) == 0)
466 break;
467 }
468 }
469 if (i == asizeof(tptr->month))
470 return (NULL);
471
472 tm->tm_mon = i;
473 buf += len;
474 flags |= FLAG_MONTH;
475
476 break;
477
478 case 'm':
479 if (!isdigit_l((unsigned char)*buf, locale))
480 return (NULL);
481
482 len = 2;
483 for (i = 0; len && *buf != 0 &&
484 isdigit_l((unsigned char)*buf, locale); buf++) {
485 i *= 10;
486 i += *buf - '0';
487 len--;
488 }
489 if (i < 1 || i > 12)
490 return (NULL);
491
492 tm->tm_mon = i - 1;
493 flags |= FLAG_MONTH;
494
495 break;
496
497 case 's':
498 {
499 char *cp;
500 int sverrno;
501 long n;
502 time_t t;
503
504 sverrno = errno;
505 errno = 0;
506 n = strtol_l(buf, &cp, 10, locale);
507 if (errno == ERANGE || (long)(t = n) != n) {
508 errno = sverrno;
509 return (NULL);
510 }
511 errno = sverrno;
512 buf = cp;
513 if (gmtime_r(&t, tm) == NULL)
514 return (NULL);
515 *GMTp = 1;
516 flags |= FLAG_YDAY | FLAG_WDAY | FLAG_MONTH |
517 FLAG_MDAY | FLAG_YEAR;
518 }
519 break;
520
521 case 'Y':
522 case 'y':
523 if (*buf == 0 ||
524 isspace_l((unsigned char)*buf, locale))
525 break;
526
527 if (!isdigit_l((unsigned char)*buf, locale))
528 return (NULL);
529
530 len = (c == 'Y') ? 4 : 2;
531 for (i = 0; len && *buf != 0 &&
532 isdigit_l((unsigned char)*buf, locale); buf++) {
533 i *= 10;
534 i += *buf - '0';
535 len--;
536 }
537 if (c == 'Y')
538 century = i / 100;
539 year = i % 100;
540
541 flags |= FLAG_YEAR;
542
543 break;
544
545 case 'Z':
546 {
547 const char *cp;
548 char *zonestr;
549
550 for (cp = buf; *cp &&
551 isupper_l((unsigned char)*cp, locale); ++cp) {
552 /*empty*/}
553 if (cp - buf) {
554 zonestr = alloca(cp - buf + 1);
555 strncpy(zonestr, buf, cp - buf);
556 zonestr[cp - buf] = '\0';
557 tzset();
558 if (0 == strcmp(zonestr, "GMT") ||
559 0 == strcmp(zonestr, "UTC")) {
560 *GMTp = 1;
561 } else if (0 == strcmp(zonestr, tzname[0])) {
562 tm->tm_isdst = 0;
563 } else if (0 == strcmp(zonestr, tzname[1])) {
564 tm->tm_isdst = 1;
565 } else {
566 return (NULL);
567 }
568 buf += cp - buf;
569 }
570 }
571 break;
572
573 case 'z':
574 {
575 int sign = 1;
576
577 if (*buf != '+') {
578 if (*buf == '-')
579 sign = -1;
580 else
581 return (NULL);
582 }
583
584 buf++;
585 i = 0;
586 for (len = 4; len > 0; len--) {
587 if (isdigit_l((unsigned char)*buf, locale)) {
588 i *= 10;
589 i += *buf - '0';
590 buf++;
591 } else if (len == 2) {
592 i *= 100;
593 break;
594 } else
595 return (NULL);
596 }
597
598 if (i > 1400 || (sign == -1 && i > 1200) ||
599 (i % 100) >= 60)
600 return (NULL);
601 tm->tm_hour -= sign * (i / 100);
602 tm->tm_min -= sign * (i % 100);
603 *GMTp = 1;
604 }
605 break;
606
607 case 'n':
608 case 't':
609 while (isspace_l((unsigned char)*buf, locale))
610 buf++;
611 break;
612
613 default:
614 return (NULL);
615 }
616 }
617
618 if (century != -1 || year != -1) {
619 if (year == -1)
620 year = 0;
621 if (century == -1) {
622 if (year < 69)
623 year += 100;
624 } else
625 year += century * 100 - TM_YEAR_BASE;
626 tm->tm_year = year;
627 }
628
629 if (!(flags & FLAG_YDAY) && (flags & FLAG_YEAR)) {
630 if ((flags & (FLAG_MONTH | FLAG_MDAY)) ==
631 (FLAG_MONTH | FLAG_MDAY)) {
632 tm->tm_yday = start_of_month[isleap(tm->tm_year +
633 TM_YEAR_BASE)][tm->tm_mon] + (tm->tm_mday - 1);
634 flags |= FLAG_YDAY;
635 } else if (day_offset != -1) {
636 int tmpwday, tmpyday, fwo;
637
638 fwo = first_wday_of(tm->tm_year + TM_YEAR_BASE);
639 /* No incomplete week (week 0). */
640 if (week_offset == 0 && fwo == day_offset)
641 return (NULL);
642
643 /* Set the date to the first Sunday (or Monday)
644 * of the specified week of the year.
645 */
646 tmpwday = (flags & FLAG_WDAY) ? tm->tm_wday :
647 day_offset;
648 tmpyday = (7 - fwo + day_offset) % 7 +
649 (week_offset - 1) * 7 +
650 (tmpwday - day_offset + 7) % 7;
651 /* Impossible yday for incomplete week (week 0). */
652 if (tmpyday < 0) {
653 if (flags & FLAG_WDAY)
654 return (NULL);
655 tmpyday = 0;
656 }
657 tm->tm_yday = tmpyday;
658 flags |= FLAG_YDAY;
659 }
660 }
661
662 if ((flags & (FLAG_YEAR | FLAG_YDAY)) == (FLAG_YEAR | FLAG_YDAY)) {
663 if (!(flags & FLAG_MONTH)) {
664 i = 0;
665 while (tm->tm_yday >=
666 start_of_month[isleap(tm->tm_year +
667 TM_YEAR_BASE)][i])
668 i++;
669 if (i > 12) {
670 i = 1;
671 tm->tm_yday -=
672 start_of_month[isleap(tm->tm_year +
673 TM_YEAR_BASE)][12];
674 tm->tm_year++;
675 }
676 tm->tm_mon = i - 1;
677 flags |= FLAG_MONTH;
678 }
679 if (!(flags & FLAG_MDAY)) {
680 tm->tm_mday = tm->tm_yday -
681 start_of_month[isleap(tm->tm_year + TM_YEAR_BASE)]
682 [tm->tm_mon] + 1;
683 flags |= FLAG_MDAY;
684 }
685 if (!(flags & FLAG_WDAY)) {
686 i = 0;
687 wday_offset = first_wday_of(tm->tm_year);
688 while (i++ <= tm->tm_yday) {
689 if (wday_offset++ >= 6)
690 wday_offset = 0;
691 }
692 tm->tm_wday = wday_offset;
693 flags |= FLAG_WDAY;
694 }
695 }
696
697 return ((char *)buf);
698 }
699
700 char *
strptime_l(const char * __restrict buf,const char * __restrict fmt,struct tm * __restrict tm,locale_t loc)701 strptime_l(const char * __restrict buf, const char * __restrict fmt,
702 struct tm * __restrict tm, locale_t loc)
703 {
704 char *ret;
705 int gmt;
706 FIX_LOCALE(loc);
707
708 gmt = 0;
709 ret = _strptime(buf, fmt, tm, &gmt, loc);
710 if (ret && gmt) {
711 time_t t = timegm(tm);
712
713 localtime_r(&t, tm);
714 }
715
716 return (ret);
717 }
718
719 char *
strptime(const char * __restrict buf,const char * __restrict fmt,struct tm * __restrict tm)720 strptime(const char * __restrict buf, const char * __restrict fmt,
721 struct tm * __restrict tm)
722 {
723 return strptime_l(buf, fmt, tm, __get_locale());
724 }
725