strptime.c revision 1.43
1 /* $NetBSD: strptime.c,v 1.43 2015/07/13 17:45:16 ginsbach Exp $ */ 2 3 /*- 4 * Copyright (c) 1997, 1998, 2005, 2008 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code was contributed to The NetBSD Foundation by Klaus Klein. 8 * Heavily optimised by David Laight 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 20 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 21 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 22 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 23 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 24 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 25 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 26 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 27 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 28 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 29 * POSSIBILITY OF SUCH DAMAGE. 30 */ 31 32 #include <sys/cdefs.h> 33 #if defined(LIBC_SCCS) && !defined(lint) 34 __RCSID("$NetBSD: strptime.c,v 1.43 2015/07/13 17:45:16 ginsbach Exp $"); 35 #endif 36 37 #include "namespace.h" 38 #include <sys/localedef.h> 39 #include <sys/types.h> 40 #include <ctype.h> 41 #include <locale.h> 42 #include <string.h> 43 #include <time.h> 44 #include <tzfile.h> 45 #include "private.h" 46 #include "setlocale_local.h" 47 48 #ifdef __weak_alias 49 __weak_alias(strptime,_strptime) 50 __weak_alias(strptime_l, _strptime_l) 51 #endif 52 53 #define _TIME_LOCALE(loc) \ 54 ((_TimeLocale *)((loc)->part_impl[(size_t)LC_TIME])) 55 56 /* 57 * We do not implement alternate representations. However, we always 58 * check whether a given modifier is allowed for a certain conversion. 59 */ 60 #define ALT_E 0x01 61 #define ALT_O 0x02 62 #define LEGAL_ALT(x) { if (alt_format & ~(x)) return NULL; } 63 64 #define S_YEAR (1 << 0) 65 #define S_MON (1 << 1) 66 #define S_YDAY (1 << 2) 67 #define S_MDAY (1 << 3) 68 #define S_WDAY (1 << 4) 69 70 #define HAVE_MDAY(s) (s & S_MDAY) 71 #define HAVE_MON(s) (s & S_MON) 72 #define HAVE_WDAY(s) (s & S_WDAY) 73 #define HAVE_YDAY(s) (s & S_YDAY) 74 #define HAVE_YEAR(s) (s & S_YEAR) 75 76 static char gmt[] = { "GMT" }; 77 static char utc[] = { "UTC" }; 78 /* RFC-822/RFC-2822 */ 79 static const char * const nast[5] = { 80 "EST", "CST", "MST", "PST", "\0\0\0" 81 }; 82 static const char * const nadt[5] = { 83 "EDT", "CDT", "MDT", "PDT", "\0\0\0" 84 }; 85 86 static const u_char *conv_num(const unsigned char *, int *, uint, uint); 87 static const u_char *find_string(const u_char *, int *, const char * const *, 88 const char * const *, int); 89 90 static const int start_of_month[2][13] = { 91 { 0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334, 365 }, 92 { 0, 31, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335, 366 } 93 }; 94 95 /* 96 * Calculate the week day of the first day of a year. Valid for 97 * the Gregorian calendar, which began Sept 14, 1752 in the UK 98 * and its colonies. Ref: 99 * http://en.wikipedia.org/wiki/Determination_of_the_day_of_the_week 100 */ 101 102 static int 103 first_wday_of(int yr) 104 { 105 return ((2 * (3 - (yr / 100) % 4)) + (yr % 100) + ((yr % 100) / 4) + 106 (isleap(yr) ? 6 : 0) + 1) % 7; 107 } 108 109 char * 110 strptime(const char *buf, const char *fmt, struct tm *tm) 111 { 112 return strptime_l(buf, fmt, tm, _current_locale()); 113 } 114 115 char * 116 strptime_l(const char *buf, const char *fmt, struct tm *tm, locale_t loc) 117 { 118 unsigned char c; 119 const unsigned char *bp, *ep; 120 int alt_format, i, split_year = 0, neg = 0, state = 0, 121 day_offset = -1, week_offset = 0, offs; 122 const char *new_fmt; 123 124 bp = (const u_char *)buf; 125 126 while (bp != NULL && (c = *fmt++) != '\0') { 127 /* Clear `alternate' modifier prior to new conversion. */ 128 alt_format = 0; 129 i = 0; 130 131 /* Eat up white-space. */ 132 if (isspace(c)) { 133 while (isspace(*bp)) 134 bp++; 135 continue; 136 } 137 138 if (c != '%') 139 goto literal; 140 141 142 again: switch (c = *fmt++) { 143 case '%': /* "%%" is converted to "%". */ 144 literal: 145 if (c != *bp++) 146 return NULL; 147 LEGAL_ALT(0); 148 continue; 149 150 /* 151 * "Alternative" modifiers. Just set the appropriate flag 152 * and start over again. 153 */ 154 case 'E': /* "%E?" alternative conversion modifier. */ 155 LEGAL_ALT(0); 156 alt_format |= ALT_E; 157 goto again; 158 159 case 'O': /* "%O?" alternative conversion modifier. */ 160 LEGAL_ALT(0); 161 alt_format |= ALT_O; 162 goto again; 163 164 /* 165 * "Complex" conversion rules, implemented through recursion. 166 */ 167 case 'c': /* Date and time, using the locale's format. */ 168 new_fmt = _TIME_LOCALE(loc)->d_t_fmt; 169 state |= S_WDAY | S_MON | S_MDAY | S_YEAR; 170 goto recurse; 171 172 case 'D': /* The date as "%m/%d/%y". */ 173 new_fmt = "%m/%d/%y"; 174 LEGAL_ALT(0); 175 state |= S_MON | S_MDAY | S_YEAR; 176 goto recurse; 177 178 case 'F': /* The date as "%Y-%m-%d". */ 179 new_fmt = "%Y-%m-%d"; 180 LEGAL_ALT(0); 181 state |= S_MON | S_MDAY | S_YEAR; 182 goto recurse; 183 184 case 'R': /* The time as "%H:%M". */ 185 new_fmt = "%H:%M"; 186 LEGAL_ALT(0); 187 goto recurse; 188 189 case 'r': /* The time in 12-hour clock representation. */ 190 new_fmt = _TIME_LOCALE(loc)->t_fmt_ampm; 191 LEGAL_ALT(0); 192 goto recurse; 193 194 case 'T': /* The time as "%H:%M:%S". */ 195 new_fmt = "%H:%M:%S"; 196 LEGAL_ALT(0); 197 goto recurse; 198 199 case 'X': /* The time, using the locale's format. */ 200 new_fmt = _TIME_LOCALE(loc)->t_fmt; 201 goto recurse; 202 203 case 'x': /* The date, using the locale's format. */ 204 new_fmt = _TIME_LOCALE(loc)->d_fmt; 205 state |= S_MON | S_MDAY | S_YEAR; 206 recurse: 207 bp = (const u_char *)strptime((const char *)bp, 208 new_fmt, tm); 209 LEGAL_ALT(ALT_E); 210 continue; 211 212 /* 213 * "Elementary" conversion rules. 214 */ 215 case 'A': /* The day of week, using the locale's form. */ 216 case 'a': 217 bp = find_string(bp, &tm->tm_wday, 218 _TIME_LOCALE(loc)->day, _TIME_LOCALE(loc)->abday, 7); 219 LEGAL_ALT(0); 220 state |= S_WDAY; 221 continue; 222 223 case 'B': /* The month, using the locale's form. */ 224 case 'b': 225 case 'h': 226 bp = find_string(bp, &tm->tm_mon, 227 _TIME_LOCALE(loc)->mon, _TIME_LOCALE(loc)->abmon, 228 12); 229 LEGAL_ALT(0); 230 state |= S_MON; 231 continue; 232 233 case 'C': /* The century number. */ 234 i = 20; 235 bp = conv_num(bp, &i, 0, 99); 236 237 i = i * 100 - TM_YEAR_BASE; 238 if (split_year) 239 i += tm->tm_year % 100; 240 split_year = 1; 241 tm->tm_year = i; 242 LEGAL_ALT(ALT_E); 243 state |= S_YEAR; 244 continue; 245 246 case 'd': /* The day of month. */ 247 case 'e': 248 bp = conv_num(bp, &tm->tm_mday, 1, 31); 249 LEGAL_ALT(ALT_O); 250 state |= S_MDAY; 251 continue; 252 253 case 'k': /* The hour (24-hour clock representation). */ 254 LEGAL_ALT(0); 255 /* FALLTHROUGH */ 256 case 'H': 257 bp = conv_num(bp, &tm->tm_hour, 0, 23); 258 LEGAL_ALT(ALT_O); 259 continue; 260 261 case 'l': /* The hour (12-hour clock representation). */ 262 LEGAL_ALT(0); 263 /* FALLTHROUGH */ 264 case 'I': 265 bp = conv_num(bp, &tm->tm_hour, 1, 12); 266 if (tm->tm_hour == 12) 267 tm->tm_hour = 0; 268 LEGAL_ALT(ALT_O); 269 continue; 270 271 case 'j': /* The day of year. */ 272 i = 1; 273 bp = conv_num(bp, &i, 1, 366); 274 tm->tm_yday = i - 1; 275 LEGAL_ALT(0); 276 state |= S_YDAY; 277 continue; 278 279 case 'M': /* The minute. */ 280 bp = conv_num(bp, &tm->tm_min, 0, 59); 281 LEGAL_ALT(ALT_O); 282 continue; 283 284 case 'm': /* The month. */ 285 i = 1; 286 bp = conv_num(bp, &i, 1, 12); 287 tm->tm_mon = i - 1; 288 LEGAL_ALT(ALT_O); 289 state |= S_MON; 290 continue; 291 292 case 'p': /* The locale's equivalent of AM/PM. */ 293 bp = find_string(bp, &i, _TIME_LOCALE(loc)->am_pm, 294 NULL, 2); 295 if (tm->tm_hour > 11) 296 return NULL; 297 tm->tm_hour += i * 12; 298 LEGAL_ALT(0); 299 continue; 300 301 case 'S': /* The seconds. */ 302 bp = conv_num(bp, &tm->tm_sec, 0, 61); 303 LEGAL_ALT(ALT_O); 304 continue; 305 306 #ifndef TIME_MAX 307 #define TIME_MAX INT64_MAX 308 #endif 309 case 's': /* seconds since the epoch */ 310 { 311 time_t sse = 0; 312 uint64_t rulim = TIME_MAX; 313 314 if (*bp < '0' || *bp > '9') { 315 bp = NULL; 316 continue; 317 } 318 319 do { 320 sse *= 10; 321 sse += *bp++ - '0'; 322 rulim /= 10; 323 } while ((sse * 10 <= TIME_MAX) && 324 rulim && *bp >= '0' && *bp <= '9'); 325 326 if (sse < 0 || (uint64_t)sse > TIME_MAX) { 327 bp = NULL; 328 continue; 329 } 330 331 if (localtime_r(&sse, tm) == NULL) 332 bp = NULL; 333 else 334 state |= S_YDAY | S_WDAY | 335 S_MON | S_MDAY | S_YEAR; 336 } 337 continue; 338 339 case 'U': /* The week of year, beginning on sunday. */ 340 case 'W': /* The week of year, beginning on monday. */ 341 /* 342 * XXX This is bogus, as we can not assume any valid 343 * information present in the tm structure at this 344 * point to calculate a real value, so just check the 345 * range for now. 346 */ 347 bp = conv_num(bp, &i, 0, 53); 348 LEGAL_ALT(ALT_O); 349 if (c == 'U') 350 day_offset = TM_SUNDAY; 351 else 352 day_offset = TM_MONDAY; 353 week_offset = i; 354 continue; 355 356 case 'w': /* The day of week, beginning on sunday. */ 357 bp = conv_num(bp, &tm->tm_wday, 0, 6); 358 LEGAL_ALT(ALT_O); 359 state |= S_WDAY; 360 continue; 361 362 case 'u': /* The day of week, monday = 1. */ 363 bp = conv_num(bp, &i, 1, 7); 364 tm->tm_wday = i % 7; 365 LEGAL_ALT(ALT_O); 366 continue; 367 368 case 'g': /* The year corresponding to the ISO week 369 * number but without the century. 370 */ 371 bp = conv_num(bp, &i, 0, 99); 372 continue; 373 374 case 'G': /* The year corresponding to the ISO week 375 * number with century. 376 */ 377 do 378 bp++; 379 while (isdigit(*bp)); 380 continue; 381 382 case 'V': /* The ISO 8601:1988 week number as decimal */ 383 bp = conv_num(bp, &i, 0, 53); 384 continue; 385 386 case 'Y': /* The year. */ 387 i = TM_YEAR_BASE; /* just for data sanity... */ 388 bp = conv_num(bp, &i, 0, 9999); 389 tm->tm_year = i - TM_YEAR_BASE; 390 LEGAL_ALT(ALT_E); 391 state |= S_YEAR; 392 continue; 393 394 case 'y': /* The year within 100 years of the epoch. */ 395 /* LEGAL_ALT(ALT_E | ALT_O); */ 396 bp = conv_num(bp, &i, 0, 99); 397 398 if (split_year) 399 /* preserve century */ 400 i += (tm->tm_year / 100) * 100; 401 else { 402 split_year = 1; 403 if (i <= 68) 404 i = i + 2000 - TM_YEAR_BASE; 405 else 406 i = i + 1900 - TM_YEAR_BASE; 407 } 408 tm->tm_year = i; 409 state |= S_YEAR; 410 continue; 411 412 case 'Z': 413 tzset(); 414 if (strncmp((const char *)bp, gmt, 3) == 0 || 415 strncmp((const char *)bp, utc, 3) == 0) { 416 tm->tm_isdst = 0; 417 #ifdef TM_GMTOFF 418 tm->TM_GMTOFF = 0; 419 #endif 420 #ifdef TM_ZONE 421 tm->TM_ZONE = gmt; 422 #endif 423 bp += 3; 424 } else { 425 ep = find_string(bp, &i, 426 (const char * const *)tzname, 427 NULL, 2); 428 if (ep != NULL) { 429 tm->tm_isdst = i; 430 #ifdef TM_GMTOFF 431 tm->TM_GMTOFF = -(timezone); 432 #endif 433 #ifdef TM_ZONE 434 tm->TM_ZONE = tzname[i]; 435 #endif 436 } 437 bp = ep; 438 } 439 continue; 440 441 case 'z': 442 /* 443 * We recognize all ISO 8601 formats: 444 * Z = Zulu time/UTC 445 * [+-]hhmm 446 * [+-]hh:mm 447 * [+-]hh 448 * We recognize all RFC-822/RFC-2822 formats: 449 * UT|GMT 450 * North American : UTC offsets 451 * E[DS]T = Eastern : -4 | -5 452 * C[DS]T = Central : -5 | -6 453 * M[DS]T = Mountain: -6 | -7 454 * P[DS]T = Pacific : -7 | -8 455 * Military 456 * [A-IL-M] = -1 ... -9 (J not used) 457 * [N-Y] = +1 ... +12 458 */ 459 while (isspace(*bp)) 460 bp++; 461 462 switch (*bp++) { 463 case 'G': 464 if (*bp++ != 'M') 465 return NULL; 466 /*FALLTHROUGH*/ 467 case 'U': 468 if (*bp++ != 'T') 469 return NULL; 470 /*FALLTHROUGH*/ 471 case 'Z': 472 tm->tm_isdst = 0; 473 #ifdef TM_GMTOFF 474 tm->TM_GMTOFF = 0; 475 #endif 476 #ifdef TM_ZONE 477 tm->TM_ZONE = utc; 478 #endif 479 continue; 480 case '+': 481 neg = 0; 482 state |= S_WDAY | S_MON | S_MDAY | S_YEAR; 483 break; 484 case '-': 485 neg = 1; 486 break; 487 default: 488 --bp; 489 ep = find_string(bp, &i, nast, NULL, 4); 490 if (ep != NULL) { 491 #ifdef TM_GMTOFF 492 tm->TM_GMTOFF = -5 - i; 493 #endif 494 #ifdef TM_ZONE 495 tm->TM_ZONE = __UNCONST(nast[i]); 496 #endif 497 bp = ep; 498 continue; 499 } 500 ep = find_string(bp, &i, nadt, NULL, 4); 501 if (ep != NULL) { 502 tm->tm_isdst = 1; 503 #ifdef TM_GMTOFF 504 tm->TM_GMTOFF = -4 - i; 505 #endif 506 #ifdef TM_ZONE 507 tm->TM_ZONE = __UNCONST(nadt[i]); 508 #endif 509 bp = ep; 510 continue; 511 } 512 513 if ((*bp >= 'A' && *bp <= 'I') || 514 (*bp >= 'L' && *bp <= 'Y')) { 515 #ifdef TM_GMTOFF 516 /* Argh! No 'J'! */ 517 if (*bp >= 'A' && *bp <= 'I') 518 tm->TM_GMTOFF = 519 ('A' - 1) - (int)*bp; 520 else if (*bp >= 'L' && *bp <= 'M') 521 tm->TM_GMTOFF = 'A' - (int)*bp; 522 else if (*bp >= 'N' && *bp <= 'Y') 523 tm->TM_GMTOFF = (int)*bp - 'M'; 524 #endif 525 #ifdef TM_ZONE 526 tm->TM_ZONE = NULL; /* XXX */ 527 #endif 528 bp++; 529 continue; 530 } 531 return NULL; 532 } 533 offs = 0; 534 for (i = 0; i < 4; ) { 535 if (isdigit(*bp)) { 536 offs = offs * 10 + (*bp++ - '0'); 537 i++; 538 continue; 539 } 540 if (i == 2 && *bp == ':') { 541 bp++; 542 continue; 543 } 544 break; 545 } 546 switch (i) { 547 case 2: 548 offs *= 100; 549 break; 550 case 4: 551 i = offs % 100; 552 if (i >= 60) 553 return NULL; 554 /* Convert minutes into decimal */ 555 offs = (offs / 100) * 100 + (i * 50) / 30; 556 break; 557 default: 558 return NULL; 559 } 560 if (neg) 561 offs = -offs; 562 tm->tm_isdst = 0; /* XXX */ 563 #ifdef TM_GMTOFF 564 tm->TM_GMTOFF = offs; 565 #endif 566 #ifdef TM_ZONE 567 tm->TM_ZONE = NULL; /* XXX */ 568 #endif 569 continue; 570 571 /* 572 * Miscellaneous conversions. 573 */ 574 case 'n': /* Any kind of white-space. */ 575 case 't': 576 while (isspace(*bp)) 577 bp++; 578 LEGAL_ALT(0); 579 continue; 580 581 582 default: /* Unknown/unsupported conversion. */ 583 return NULL; 584 } 585 } 586 587 if (!HAVE_YDAY(state) && HAVE_YEAR(state)) { 588 if (HAVE_MON(state) && HAVE_MDAY(state)) { 589 tm->tm_yday = start_of_month[isleap_sum(tm->tm_year, 590 TM_YEAR_BASE)][tm->tm_mon] + (tm->tm_mday - 1); 591 state |= S_YDAY; 592 } else if (day_offset != -1) { 593 /* Set the date to the first Sunday (or Monday) 594 * of the specified week of the year. 595 */ 596 if (!HAVE_WDAY(state)) { 597 tm->tm_wday = day_offset; 598 state |= S_WDAY; 599 } 600 tm->tm_yday = (7 - 601 first_wday_of(tm->tm_year + TM_YEAR_BASE) + 602 day_offset) % 7 + (week_offset - 1) * 7 + 603 tm->tm_wday - day_offset; 604 state |= S_YDAY; 605 } 606 } 607 608 if (HAVE_YDAY(state) && HAVE_YEAR(state)) { 609 int isleap; 610 if (!HAVE_MON(state)) { 611 i = 0; 612 isleap = isleap_sum(tm->tm_year, TM_YEAR_BASE); 613 while (tm->tm_yday >= start_of_month[isleap][i]) 614 i++; 615 if (i > 12) { 616 i = 1; 617 tm->tm_yday -= start_of_month[isleap][12]; 618 tm->tm_year++; 619 } 620 tm->tm_mon = i - 1; 621 state |= S_MON; 622 } 623 if (!HAVE_MDAY(state)) { 624 isleap = isleap_sum(tm->tm_year, TM_YEAR_BASE); 625 tm->tm_mday = tm->tm_yday - 626 start_of_month[isleap][tm->tm_mon] + 1; 627 state |= S_MDAY; 628 } 629 if (!HAVE_WDAY(state)) { 630 i = 0; 631 week_offset = first_wday_of(tm->tm_year); 632 while (i++ <= tm->tm_yday) { 633 if (week_offset++ >= 6) 634 week_offset = 0; 635 } 636 tm->tm_wday = week_offset; 637 state |= S_WDAY; 638 } 639 } 640 641 return __UNCONST(bp); 642 } 643 644 645 static const u_char * 646 conv_num(const unsigned char *buf, int *dest, uint llim, uint ulim) 647 { 648 uint result = 0; 649 unsigned char ch; 650 651 /* The limit also determines the number of valid digits. */ 652 uint rulim = ulim; 653 654 ch = *buf; 655 if (ch < '0' || ch > '9') 656 return NULL; 657 658 do { 659 result *= 10; 660 result += ch - '0'; 661 rulim /= 10; 662 ch = *++buf; 663 } while ((result * 10 <= ulim) && rulim && ch >= '0' && ch <= '9'); 664 665 if (result < llim || result > ulim) 666 return NULL; 667 668 *dest = result; 669 return buf; 670 } 671 672 static const u_char * 673 find_string(const u_char *bp, int *tgt, const char * const *n1, 674 const char * const *n2, int c) 675 { 676 int i; 677 size_t len; 678 679 /* check full name - then abbreviated ones */ 680 for (; n1 != NULL; n1 = n2, n2 = NULL) { 681 for (i = 0; i < c; i++, n1++) { 682 len = strlen(*n1); 683 if (strncasecmp(*n1, (const char *)bp, len) == 0) { 684 *tgt = i; 685 return bp + len; 686 } 687 } 688 } 689 690 /* Nothing matched */ 691 return NULL; 692 } 693
Indexes created Fri Sep 26 20:09:58 GMT 2025