strptime.c revision 1.46
1 /* $NetBSD: strptime.c,v 1.46 2015/07/20 14:37:11 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 1.46 2015/07/20 14:37:11 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 static const u_char *conv_num(const unsigned char *, int *, uint, uint); 54 static const u_char *find_string(const u_char *, int *, const char * const *, 55 const char * const *, int); 56 57 #define _TIME_LOCALE(loc) \ 58 ((_TimeLocale *)((loc)->part_impl[(size_t)LC_TIME])) 59 60 /* 61 * We do not implement alternate representations. However, we always 62 * check whether a given modifier is allowed for a certain conversion. 63 */ 64 #define ALT_E 0x01 65 #define ALT_O 0x02 66 #define LEGAL_ALT(x) { if (alt_format & ~(x)) return NULL; } 67 68 #define S_YEAR (1 << 0) 69 #define S_MON (1 << 1) 70 #define S_YDAY (1 << 2) 71 #define S_MDAY (1 << 3) 72 #define S_WDAY (1 << 4) 73 74 #define HAVE_MDAY(s) (s & S_MDAY) 75 #define HAVE_MON(s) (s & S_MON) 76 #define HAVE_WDAY(s) (s & S_WDAY) 77 #define HAVE_YDAY(s) (s & S_YDAY) 78 #define HAVE_YEAR(s) (s & S_YEAR) 79 80 static char gmt[] = { "GMT" }; 81 static char utc[] = { "UTC" }; 82 /* RFC-822/RFC-2822 */ 83 static const char * const nast[5] = { 84 "EST", "CST", "MST", "PST", "\0\0\0" 85 }; 86 static const char * const nadt[5] = { 87 "EDT", "CDT", "MDT", "PDT", "\0\0\0" 88 }; 89 90 /* 91 * Table to determine the ordinal date for the start of a month. 92 * Ref: http://en.wikipedia.org/wiki/ISO_week_date 93 */ 94 static const int start_of_month[2][13] = { 95 /* non-leap year */ 96 { 0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334, 365 }, 97 /* leap year */ 98 { 0, 31, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335, 366 } 99 }; 100 101 /* 102 * Calculate the week day of the first day of a year. Valid for 103 * the Gregorian calendar, which began Sept 14, 1752 in the UK 104 * and its colonies. Ref: 105 * http://en.wikipedia.org/wiki/Determination_of_the_day_of_the_week 106 */ 107 108 static int 109 first_wday_of(int yr) 110 { 111 return ((2 * (3 - (yr / 100) % 4)) + (yr % 100) + ((yr % 100) / 4) + 112 (isleap(yr) ? 6 : 0) + 1) % 7; 113 } 114 115 char * 116 strptime(const char *buf, const char *fmt, struct tm *tm) 117 { 118 return strptime_l(buf, fmt, tm, _current_locale()); 119 } 120 121 char * 122 strptime_l(const char *buf, const char *fmt, struct tm *tm, locale_t loc) 123 { 124 unsigned char c; 125 const unsigned char *bp, *ep; 126 int alt_format, i, split_year = 0, neg = 0, state = 0, 127 day_offset = -1, week_offset = 0, offs; 128 const char *new_fmt; 129 130 bp = (const u_char *)buf; 131 132 while (bp != NULL && (c = *fmt++) != '\0') { 133 /* Clear `alternate' modifier prior to new conversion. */ 134 alt_format = 0; 135 i = 0; 136 137 /* Eat up white-space. */ 138 if (isspace(c)) { 139 while (isspace(*bp)) 140 bp++; 141 continue; 142 } 143 144 if (c != '%') 145 goto literal; 146 147 148 again: switch (c = *fmt++) { 149 case '%': /* "%%" is converted to "%". */ 150 literal: 151 if (c != *bp++) 152 return NULL; 153 LEGAL_ALT(0); 154 continue; 155 156 /* 157 * "Alternative" modifiers. Just set the appropriate flag 158 * and start over again. 159 */ 160 case 'E': /* "%E?" alternative conversion modifier. */ 161 LEGAL_ALT(0); 162 alt_format |= ALT_E; 163 goto again; 164 165 case 'O': /* "%O?" alternative conversion modifier. */ 166 LEGAL_ALT(0); 167 alt_format |= ALT_O; 168 goto again; 169 170 /* 171 * "Complex" conversion rules, implemented through recursion. 172 */ 173 case 'c': /* Date and time, using the locale's format. */ 174 new_fmt = _TIME_LOCALE(loc)->d_t_fmt; 175 state |= S_WDAY | S_MON | S_MDAY | S_YEAR; 176 goto recurse; 177 178 case 'D': /* The date as "%m/%d/%y". */ 179 new_fmt = "%m/%d/%y"; 180 LEGAL_ALT(0); 181 state |= S_MON | S_MDAY | S_YEAR; 182 goto recurse; 183 184 case 'F': /* The date as "%Y-%m-%d". */ 185 new_fmt = "%Y-%m-%d"; 186 LEGAL_ALT(0); 187 state |= S_MON | S_MDAY | S_YEAR; 188 goto recurse; 189 190 case 'R': /* The time as "%H:%M". */ 191 new_fmt = "%H:%M"; 192 LEGAL_ALT(0); 193 goto recurse; 194 195 case 'r': /* The time in 12-hour clock representation. */ 196 new_fmt = _TIME_LOCALE(loc)->t_fmt_ampm; 197 LEGAL_ALT(0); 198 goto recurse; 199 200 case 'T': /* The time as "%H:%M:%S". */ 201 new_fmt = "%H:%M:%S"; 202 LEGAL_ALT(0); 203 goto recurse; 204 205 case 'X': /* The time, using the locale's format. */ 206 new_fmt = _TIME_LOCALE(loc)->t_fmt; 207 goto recurse; 208 209 case 'x': /* The date, using the locale's format. */ 210 new_fmt = _TIME_LOCALE(loc)->d_fmt; 211 state |= S_MON | S_MDAY | S_YEAR; 212 recurse: 213 bp = (const u_char *)strptime((const char *)bp, 214 new_fmt, tm); 215 LEGAL_ALT(ALT_E); 216 continue; 217 218 /* 219 * "Elementary" conversion rules. 220 */ 221 case 'A': /* The day of week, using the locale's form. */ 222 case 'a': 223 bp = find_string(bp, &tm->tm_wday, 224 _TIME_LOCALE(loc)->day, _TIME_LOCALE(loc)->abday, 7); 225 LEGAL_ALT(0); 226 state |= S_WDAY; 227 continue; 228 229 case 'B': /* The month, using the locale's form. */ 230 case 'b': 231 case 'h': 232 bp = find_string(bp, &tm->tm_mon, 233 _TIME_LOCALE(loc)->mon, _TIME_LOCALE(loc)->abmon, 234 12); 235 LEGAL_ALT(0); 236 state |= S_MON; 237 continue; 238 239 case 'C': /* The century number. */ 240 i = 20; 241 bp = conv_num(bp, &i, 0, 99); 242 243 i = i * 100 - TM_YEAR_BASE; 244 if (split_year) 245 i += tm->tm_year % 100; 246 split_year = 1; 247 tm->tm_year = i; 248 LEGAL_ALT(ALT_E); 249 state |= S_YEAR; 250 continue; 251 252 case 'd': /* The day of month. */ 253 case 'e': 254 bp = conv_num(bp, &tm->tm_mday, 1, 31); 255 LEGAL_ALT(ALT_O); 256 state |= S_MDAY; 257 continue; 258 259 case 'k': /* The hour (24-hour clock representation). */ 260 LEGAL_ALT(0); 261 /* FALLTHROUGH */ 262 case 'H': 263 bp = conv_num(bp, &tm->tm_hour, 0, 23); 264 LEGAL_ALT(ALT_O); 265 continue; 266 267 case 'l': /* The hour (12-hour clock representation). */ 268 LEGAL_ALT(0); 269 /* FALLTHROUGH */ 270 case 'I': 271 bp = conv_num(bp, &tm->tm_hour, 1, 12); 272 if (tm->tm_hour == 12) 273 tm->tm_hour = 0; 274 LEGAL_ALT(ALT_O); 275 continue; 276 277 case 'j': /* The day of year. */ 278 i = 1; 279 bp = conv_num(bp, &i, 1, 366); 280 tm->tm_yday = i - 1; 281 LEGAL_ALT(0); 282 state |= S_YDAY; 283 continue; 284 285 case 'M': /* The minute. */ 286 bp = conv_num(bp, &tm->tm_min, 0, 59); 287 LEGAL_ALT(ALT_O); 288 continue; 289 290 case 'm': /* The month. */ 291 i = 1; 292 bp = conv_num(bp, &i, 1, 12); 293 tm->tm_mon = i - 1; 294 LEGAL_ALT(ALT_O); 295 state |= S_MON; 296 continue; 297 298 case 'p': /* The locale's equivalent of AM/PM. */ 299 bp = find_string(bp, &i, _TIME_LOCALE(loc)->am_pm, 300 NULL, 2); 301 if (tm->tm_hour > 11) 302 return NULL; 303 tm->tm_hour += i * 12; 304 LEGAL_ALT(0); 305 continue; 306 307 case 'S': /* The seconds. */ 308 bp = conv_num(bp, &tm->tm_sec, 0, 61); 309 LEGAL_ALT(ALT_O); 310 continue; 311 312 #ifndef TIME_MAX 313 #define TIME_MAX INT64_MAX 314 #endif 315 case 's': /* seconds since the epoch */ 316 { 317 time_t sse = 0; 318 uint64_t rulim = TIME_MAX; 319 320 if (*bp < '0' || *bp > '9') { 321 bp = NULL; 322 continue; 323 } 324 325 do { 326 sse *= 10; 327 sse += *bp++ - '0'; 328 rulim /= 10; 329 } while ((sse * 10 <= TIME_MAX) && 330 rulim && *bp >= '0' && *bp <= '9'); 331 332 if (sse < 0 || (uint64_t)sse > TIME_MAX) { 333 bp = NULL; 334 continue; 335 } 336 337 if (localtime_r(&sse, tm) == NULL) 338 bp = NULL; 339 else 340 state |= S_YDAY | S_WDAY | 341 S_MON | S_MDAY | S_YEAR; 342 } 343 continue; 344 345 case 'U': /* The week of year, beginning on sunday. */ 346 case 'W': /* The week of year, beginning on monday. */ 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 bp = conv_num(bp, &i, 0, 53); 354 LEGAL_ALT(ALT_O); 355 if (c == 'U') 356 day_offset = TM_SUNDAY; 357 else 358 day_offset = TM_MONDAY; 359 week_offset = i; 360 continue; 361 362 case 'w': /* The day of week, beginning on sunday. */ 363 bp = conv_num(bp, &tm->tm_wday, 0, 6); 364 LEGAL_ALT(ALT_O); 365 state |= S_WDAY; 366 continue; 367 368 case 'u': /* The day of week, monday = 1. */ 369 bp = conv_num(bp, &i, 1, 7); 370 tm->tm_wday = i % 7; 371 LEGAL_ALT(ALT_O); 372 state |= S_WDAY; 373 continue; 374 375 case 'g': /* The year corresponding to the ISO week 376 * number but without the century. 377 */ 378 bp = conv_num(bp, &i, 0, 99); 379 continue; 380 381 case 'G': /* The year corresponding to the ISO week 382 * number with century. 383 */ 384 do 385 bp++; 386 while (isdigit(*bp)); 387 continue; 388 389 case 'V': /* The ISO 8601:1988 week number as decimal */ 390 bp = conv_num(bp, &i, 0, 53); 391 continue; 392 393 case 'Y': /* The year. */ 394 i = TM_YEAR_BASE; /* just for data sanity... */ 395 bp = conv_num(bp, &i, 0, 9999); 396 tm->tm_year = i - TM_YEAR_BASE; 397 LEGAL_ALT(ALT_E); 398 state |= S_YEAR; 399 continue; 400 401 case 'y': /* The year within 100 years of the epoch. */ 402 /* LEGAL_ALT(ALT_E | ALT_O); */ 403 bp = conv_num(bp, &i, 0, 99); 404 405 if (split_year) 406 /* preserve century */ 407 i += (tm->tm_year / 100) * 100; 408 else { 409 split_year = 1; 410 if (i <= 68) 411 i = i + 2000 - TM_YEAR_BASE; 412 else 413 i = i + 1900 - TM_YEAR_BASE; 414 } 415 tm->tm_year = i; 416 state |= S_YEAR; 417 continue; 418 419 case 'Z': 420 tzset(); 421 if (strncmp((const char *)bp, gmt, 3) == 0 || 422 strncmp((const char *)bp, utc, 3) == 0) { 423 tm->tm_isdst = 0; 424 #ifdef TM_GMTOFF 425 tm->TM_GMTOFF = 0; 426 #endif 427 #ifdef TM_ZONE 428 tm->TM_ZONE = gmt; 429 #endif 430 bp += 3; 431 } else { 432 ep = find_string(bp, &i, 433 (const char * const *)tzname, 434 NULL, 2); 435 if (ep != NULL) { 436 tm->tm_isdst = i; 437 #ifdef TM_GMTOFF 438 tm->TM_GMTOFF = -(timezone); 439 #endif 440 #ifdef TM_ZONE 441 tm->TM_ZONE = tzname[i]; 442 #endif 443 } 444 bp = ep; 445 } 446 continue; 447 448 case 'z': 449 /* 450 * We recognize all ISO 8601 formats: 451 * Z = Zulu time/UTC 452 * [+-]hhmm 453 * [+-]hh:mm 454 * [+-]hh 455 * We recognize all RFC-822/RFC-2822 formats: 456 * UT|GMT 457 * North American : UTC offsets 458 * E[DS]T = Eastern : -4 | -5 459 * C[DS]T = Central : -5 | -6 460 * M[DS]T = Mountain: -6 | -7 461 * P[DS]T = Pacific : -7 | -8 462 * Military 463 * [A-IL-M] = -1 ... -9 (J not used) 464 * [N-Y] = +1 ... +12 465 */ 466 while (isspace(*bp)) 467 bp++; 468 469 switch (*bp++) { 470 case 'G': 471 if (*bp++ != 'M') 472 return NULL; 473 /*FALLTHROUGH*/ 474 case 'U': 475 if (*bp++ != 'T') 476 return NULL; 477 /*FALLTHROUGH*/ 478 case 'Z': 479 tm->tm_isdst = 0; 480 #ifdef TM_GMTOFF 481 tm->TM_GMTOFF = 0; 482 #endif 483 #ifdef TM_ZONE 484 tm->TM_ZONE = utc; 485 #endif 486 continue; 487 case '+': 488 neg = 0; 489 break; 490 case '-': 491 neg = 1; 492 break; 493 default: 494 --bp; 495 ep = find_string(bp, &i, nast, NULL, 4); 496 if (ep != NULL) { 497 #ifdef TM_GMTOFF 498 tm->TM_GMTOFF = -5 - i; 499 #endif 500 #ifdef TM_ZONE 501 tm->TM_ZONE = __UNCONST(nast[i]); 502 #endif 503 bp = ep; 504 continue; 505 } 506 ep = find_string(bp, &i, nadt, NULL, 4); 507 if (ep != NULL) { 508 tm->tm_isdst = 1; 509 #ifdef TM_GMTOFF 510 tm->TM_GMTOFF = -4 - i; 511 #endif 512 #ifdef TM_ZONE 513 tm->TM_ZONE = __UNCONST(nadt[i]); 514 #endif 515 bp = ep; 516 continue; 517 } 518 519 if ((*bp >= 'A' && *bp <= 'I') || 520 (*bp >= 'L' && *bp <= 'Y')) { 521 #ifdef TM_GMTOFF 522 /* Argh! No 'J'! */ 523 if (*bp >= 'A' && *bp <= 'I') 524 tm->TM_GMTOFF = 525 ('A' - 1) - (int)*bp; 526 else if (*bp >= 'L' && *bp <= 'M') 527 tm->TM_GMTOFF = 'A' - (int)*bp; 528 else if (*bp >= 'N' && *bp <= 'Y') 529 tm->TM_GMTOFF = (int)*bp - 'M'; 530 #endif 531 #ifdef TM_ZONE 532 tm->TM_ZONE = NULL; /* XXX */ 533 #endif 534 bp++; 535 continue; 536 } 537 return NULL; 538 } 539 offs = 0; 540 for (i = 0; i < 4; ) { 541 if (isdigit(*bp)) { 542 offs = offs * 10 + (*bp++ - '0'); 543 i++; 544 continue; 545 } 546 if (i == 2 && *bp == ':') { 547 bp++; 548 continue; 549 } 550 break; 551 } 552 switch (i) { 553 case 2: 554 offs *= 100; 555 break; 556 case 4: 557 i = offs % 100; 558 if (i >= 60) 559 return NULL; 560 /* Convert minutes into decimal */ 561 offs = (offs / 100) * 100 + (i * 50) / 30; 562 break; 563 default: 564 return NULL; 565 } 566 if (neg) 567 offs = -offs; 568 tm->tm_isdst = 0; /* XXX */ 569 #ifdef TM_GMTOFF 570 tm->TM_GMTOFF = offs; 571 #endif 572 #ifdef TM_ZONE 573 tm->TM_ZONE = NULL; /* XXX */ 574 #endif 575 continue; 576 577 /* 578 * Miscellaneous conversions. 579 */ 580 case 'n': /* Any kind of white-space. */ 581 case 't': 582 while (isspace(*bp)) 583 bp++; 584 LEGAL_ALT(0); 585 continue; 586 587 588 default: /* Unknown/unsupported conversion. */ 589 return NULL; 590 } 591 } 592 593 if (!HAVE_YDAY(state) && HAVE_YEAR(state)) { 594 if (HAVE_MON(state) && HAVE_MDAY(state)) { 595 /* calculate day of year (ordinal date) */ 596 tm->tm_yday = start_of_month[isleap_sum(tm->tm_year, 597 TM_YEAR_BASE)][tm->tm_mon] + (tm->tm_mday - 1); 598 state |= S_YDAY; 599 } else if (day_offset != -1) { 600 /* 601 * Set the date to the first Sunday (or Monday) 602 * of the specified week of the year. 603 */ 604 if (!HAVE_WDAY(state)) { 605 tm->tm_wday = day_offset; 606 state |= S_WDAY; 607 } 608 tm->tm_yday = (7 - 609 first_wday_of(tm->tm_year + TM_YEAR_BASE) + 610 day_offset) % 7 + (week_offset - 1) * 7 + 611 tm->tm_wday - day_offset; 612 state |= S_YDAY; 613 } 614 } 615 616 if (HAVE_YDAY(state) && HAVE_YEAR(state)) { 617 int isleap; 618 619 if (!HAVE_MON(state)) { 620 /* calculate month of day of year */ 621 i = 0; 622 isleap = isleap_sum(tm->tm_year, TM_YEAR_BASE); 623 while (tm->tm_yday >= start_of_month[isleap][i]) 624 i++; 625 if (i > 12) { 626 i = 1; 627 tm->tm_yday -= start_of_month[isleap][12]; 628 tm->tm_year++; 629 } 630 tm->tm_mon = i - 1; 631 state |= S_MON; 632 } 633 634 if (!HAVE_MDAY(state)) { 635 /* calculate day of month */ 636 isleap = isleap_sum(tm->tm_year, TM_YEAR_BASE); 637 tm->tm_mday = tm->tm_yday - 638 start_of_month[isleap][tm->tm_mon] + 1; 639 state |= S_MDAY; 640 } 641 642 if (!HAVE_WDAY(state)) { 643 /* calculate day of week */ 644 i = 0; 645 week_offset = first_wday_of(tm->tm_year); 646 while (i++ <= tm->tm_yday) { 647 if (week_offset++ >= 6) 648 week_offset = 0; 649 } 650 tm->tm_wday = week_offset; 651 state |= S_WDAY; 652 } 653 } 654 655 return __UNCONST(bp); 656 } 657 658 659 static const u_char * 660 conv_num(const unsigned char *buf, int *dest, uint llim, uint ulim) 661 { 662 uint result = 0; 663 unsigned char ch; 664 665 /* The limit also determines the number of valid digits. */ 666 uint rulim = ulim; 667 668 ch = *buf; 669 if (ch < '0' || ch > '9') 670 return NULL; 671 672 do { 673 result *= 10; 674 result += ch - '0'; 675 rulim /= 10; 676 ch = *++buf; 677 } while ((result * 10 <= ulim) && rulim && ch >= '0' && ch <= '9'); 678 679 if (result < llim || result > ulim) 680 return NULL; 681 682 *dest = result; 683 return buf; 684 } 685 686 static const u_char * 687 find_string(const u_char *bp, int *tgt, const char * const *n1, 688 const char * const *n2, int c) 689 { 690 int i; 691 size_t len; 692 693 /* check full name - then abbreviated ones */ 694 for (; n1 != NULL; n1 = n2, n2 = NULL) { 695 for (i = 0; i < c; i++, n1++) { 696 len = strlen(*n1); 697 if (strncasecmp(*n1, (const char *)bp, len) == 0) { 698 *tgt = i; 699 return bp + len; 700 } 701 } 702 } 703 704 /* Nothing matched */ 705 return NULL; 706 } 707
Indexes created Thu Sep 25 02:09:54 GMT 2025