Utilities/cmcurl/parsedate.c

   1 /***************************************************************************
   2  *                                  _   _ ____  _
   3  *  Project                     ___| | | |  _ \| |
   4  *                             / __| | | | |_) | |
   5  *                            | (__| |_| |  _ <| |___
   6  *                             \___|\___/|_| \_\_____|
   7  *
   8  * Copyright (C) 1998 - 2006, Daniel Stenberg, <daniel@haxx.se>, et al.
   9  *
  10  * This software is licensed as described in the file COPYING, which
  11  * you should have received as part of this distribution. The terms
  12  * are also available at http://curl.haxx.se/docs/copyright.html.
  13  *
  14  * You may opt to use, copy, modify, merge, publish, distribute and/or sell
  15  * copies of the Software, and permit persons to whom the Software is
  16  * furnished to do so, under the terms of the COPYING file.
  17  *
  18  * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
  19  * KIND, either express or implied.
  20  *
  21  * $Id: parsedate.c,v 1.2 2007-03-15 19:22:13 andy Exp $
  22  ***************************************************************************/
  23 /*
  24   A brief summary of the date string formats this parser groks:
  25
  26   RFC 2616 3.3.1
  27
  28   Sun, 06 Nov 1994 08:49:37 GMT  ; RFC 822, updated by RFC 1123
  29   Sunday, 06-Nov-94 08:49:37 GMT ; RFC 850, obsoleted by RFC 1036
  30   Sun Nov  6 08:49:37 1994       ; ANSI C's asctime() format
  31
  32   we support dates without week day name:
  33
  34   06 Nov 1994 08:49:37 GMT
  35   06-Nov-94 08:49:37 GMT
  36   Nov  6 08:49:37 1994
  37
  38   without the time zone:
  39
  40   06 Nov 1994 08:49:37
  41   06-Nov-94 08:49:37
  42
  43   weird order:
  44
  45   1994 Nov 6 08:49:37  (GNU date fails)
  46   GMT 08:49:37 06-Nov-94 Sunday
  47   94 6 Nov 08:49:37    (GNU date fails)
  48
  49   time left out:
  50
  51   1994 Nov 6
  52   06-Nov-94
  53   Sun Nov 6 94
  54
  55   unusual separators:
  56
  57   1994.Nov.6
  58   Sun/Nov/6/94/GMT
  59
  60   commonly used time zone names:
  61
  62   Sun, 06 Nov 1994 08:49:37 CET
  63   06 Nov 1994 08:49:37 EST
  64
  65   time zones specified using RFC822 style:
  66
  67   Sun, 12 Sep 2004 15:05:58 -0700
  68   Sat, 11 Sep 2004 21:32:11 +0200
  69
  70   compact numerical date strings:
  71
  72   20040912 15:05:58 -0700
  73   20040911 +0200
  74
  75 */
  76 #include "setup.h"
  77 #include <stdio.h>
  78 #include <ctype.h>
  79 #include <string.h>
  80
  81 #ifdef HAVE_STDLIB_H
  82 #include <stdlib.h> /* for strtol() */
  83 #endif
  84
  85 #include <curl/curl.h>
  86
  87 static time_t Curl_parsedate(const char *date);
  88
  89 const char * const Curl_wkday[] =
  90 {"Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"};
  91 static const char * const weekday[] =
  92 { "Monday", "Tuesday", "Wednesday", "Thursday",
  93   "Friday", "Saturday", "Sunday" };
  94 const char * const Curl_month[]=
  95 { "Jan", "Feb", "Mar", "Apr", "May", "Jun",
  96   "Jul", "Aug", "Sep", "Oct", "Nov", "Dec" };
  97
  98 struct tzinfo {
  99   const char *name;
 100   int offset; /* +/- in minutes */
 101 };
 102
 103 /* Here's a bunch of frequently used time zone names. These were supported
 104    by the old getdate parser. */
 105 #define tDAYZONE -60       /* offset for daylight savings time */
 106 static const struct tzinfo tz[]= {
 107   {"GMT", 0},              /* Greenwich Mean */
 108   {"UTC", 0},              /* Universal (Coordinated) */
 109   {"WET", 0},              /* Western European */
 110   {"BST", 0 tDAYZONE},     /* British Summer */
 111   {"WAT", 60},             /* West Africa */
 112   {"AST", 240},            /* Atlantic Standard */
 113   {"ADT", 240 tDAYZONE},   /* Atlantic Daylight */
 114   {"EST", 300},            /* Eastern Standard */
 115   {"EDT", 300 tDAYZONE},   /* Eastern Daylight */
 116   {"CST", 360},            /* Central Standard */
 117   {"CDT", 360 tDAYZONE},   /* Central Daylight */
 118   {"MST", 420},            /* Mountain Standard */
 119   {"MDT", 420 tDAYZONE},   /* Mountain Daylight */
 120   {"PST", 480},            /* Pacific Standard */
 121   {"PDT", 480 tDAYZONE},   /* Pacific Daylight */
 122   {"YST", 540},            /* Yukon Standard */
 123   {"YDT", 540 tDAYZONE},   /* Yukon Daylight */
 124   {"HST", 600},            /* Hawaii Standard */
 125   {"HDT", 600 tDAYZONE},   /* Hawaii Daylight */
 126   {"CAT", 600},            /* Central Alaska */
 127   {"AHST", 600},           /* Alaska-Hawaii Standard */
 128   {"NT",  660},            /* Nome */
 129   {"IDLW", 720},           /* International Date Line West */
 130   {"CET", -60},            /* Central European */
 131   {"MET", -60},            /* Middle European */
 132   {"MEWT", -60},           /* Middle European Winter */
 133   {"MEST", -60 tDAYZONE},  /* Middle European Summer */
 134   {"CEST", -60 tDAYZONE},  /* Central European Summer */
 135   {"MESZ", -60 tDAYZONE},  /* Middle European Summer */
 136   {"FWT", -60},            /* French Winter */
 137   {"FST", -60 tDAYZONE},   /* French Summer */
 138   {"EET", -120},           /* Eastern Europe, USSR Zone 1 */
 139   {"WAST", -420},          /* West Australian Standard */
 140   {"WADT", -420 tDAYZONE}, /* West Australian Daylight */
 141   {"CCT", -480},           /* China Coast, USSR Zone 7 */
 142   {"JST", -540},           /* Japan Standard, USSR Zone 8 */
 143   {"EAST", -600},          /* Eastern Australian Standard */
 144   {"EADT", -600 tDAYZONE}, /* Eastern Australian Daylight */
 145   {"GST", -600},           /* Guam Standard, USSR Zone 9 */
 146   {"NZT", -720},           /* New Zealand */
 147   {"NZST", -720},          /* New Zealand Standard */
 148   {"NZDT", -720 tDAYZONE}, /* New Zealand Daylight */
 149   {"IDLE", -720},          /* International Date Line East */
 150 };
 151
 152 /* returns:
 153    -1 no day
 154    0 monday - 6 sunday
 155 */
 156
 157 static int checkday(char *check, size_t len)
 158 {
 159   int i;
 160   const char * const *what;
 161   bool found= FALSE;
 162   if(len > 3)
 163     what = &weekday[0];
 164   else
 165     what = &Curl_wkday[0];
 166   for(i=0; i<7; i++) {
 167     if(curl_strequal(check, what[0])) {
 168       found=TRUE;
 169       break;
 170     }
 171     what++;
 172   }
 173   return found?i:-1;
 174 }
 175
 176 static int checkmonth(char *check)
 177 {
 178   int i;
 179   const char * const *what;
 180   bool found= FALSE;
 181
 182   what = &Curl_month[0];
 183   for(i=0; i<12; i++) {
 184     if(curl_strequal(check, what[0])) {
 185       found=TRUE;
 186       break;
 187     }
 188     what++;
 189   }
 190   return found?i:-1; /* return the offset or -1, no real offset is -1 */
 191 }
 192
 193 /* return the time zone offset between GMT and the input one, in number
 194    of seconds or -1 if the timezone wasn't found/legal */
 195
 196 static int checktz(char *check)
 197 {
 198   unsigned int i;
 199   const struct tzinfo *what;
 200   bool found= FALSE;
 201
 202   what = tz;
 203   for(i=0; i< sizeof(tz)/sizeof(tz[0]); i++) {
 204     if(curl_strequal(check, what->name)) {
 205       found=TRUE;
 206       break;
 207     }
 208     what++;
 209   }
 210   return found?what->offset*60:-1;
 211 }
 212
 213 static void skip(const char **date)
 214 {
 215   /* skip everything that aren't letters or digits */
 216   while(**date && !ISALNUM(**date))
 217     (*date)++;
 218 }
 219
 220 enum assume {
 221   DATE_MDAY,
 222   DATE_YEAR,
 223   DATE_TIME
 224 };
 225
 226 static time_t Curl_parsedate(const char *date)
 227 {
 228   time_t t = 0;
 229   int wdaynum=-1;  /* day of the week number, 0-6 (mon-sun) */
 230   int monnum=-1;   /* month of the year number, 0-11 */
 231   int mdaynum=-1; /* day of month, 1 - 31 */
 232   int hournum=-1;
 233   int minnum=-1;
 234   int secnum=-1;
 235   int yearnum=-1;
 236   int tzoff=-1;
 237   struct tm tm;
 238   enum assume dignext = DATE_MDAY;
 239   const char *indate = date; /* save the original pointer */
 240   int part = 0; /* max 6 parts */
 241
 242   while(*date && (part < 6)) {
 243     bool found=FALSE;
 244
 245     skip(&date);
 246
 247     if(ISALPHA(*date)) {
 248       /* a name coming up */
 249       char buf[32]="";
 250       size_t len;
 251       sscanf(date, "%31[A-Za-z]", buf);
 252       len = strlen(buf);
 253
 254       if(wdaynum == -1) {
 255         wdaynum = checkday(buf, len);
 256         if(wdaynum != -1)
 257           found = TRUE;
 258       }
 259       if(!found && (monnum == -1)) {
 260         monnum = checkmonth(buf);
 261         if(monnum != -1)
 262           found = TRUE;
 263       }
 264
 265       if(!found && (tzoff == -1)) {
 266         /* this just must be a time zone string */
 267         tzoff = checktz(buf);
 268         if(tzoff != -1)
 269           found = TRUE;
 270       }
 271
 272       if(!found)
 273         return -1; /* bad string */
 274
 275       date += len;
 276     }
 277     else if(ISDIGIT(*date)) {
 278       /* a digit */
 279       int val;
 280       char *end;
 281       if((secnum == -1) &&
 282          (3 == sscanf(date, "%02d:%02d:%02d", &hournum, &minnum, &secnum))) {
 283         /* time stamp! */
 284         date += 8;
 285         found = TRUE;
 286       }
 287       else {
 288         val = (int)strtol(date, &end, 10);
 289
 290         if((tzoff == -1) &&
 291            ((end - date) == 4) &&
 292            (val < 1300) &&
 293            (indate< date) &&
 294            ((date[-1] == '+' || date[-1] == '-'))) {
 295           /* four digits and a value less than 1300 and it is preceeded with
 296              a plus or minus. This is a time zone indication. */
 297           found = TRUE;
 298           tzoff = (val/100 * 60 + val%100)*60;
 299
 300           /* the + and - prefix indicates the local time compared to GMT,
 301              this we need ther reversed math to get what we want */
 302           tzoff = date[-1]=='+'?-tzoff:tzoff;
 303         }
 304
 305         if(((end - date) == 8) &&
 306            (yearnum == -1) &&
 307            (monnum == -1) &&
 308            (mdaynum == -1)) {
 309           /* 8 digits, no year, month or day yet. This is YYYYMMDD */
 310           found = TRUE;
 311           yearnum = val/10000;
 312           monnum = (val%10000)/100-1; /* month is 0 - 11 */
 313           mdaynum = val%100;
 314         }
 315
 316         if(!found && (dignext == DATE_MDAY) && (mdaynum == -1)) {
 317           if((val > 0) && (val<32)) {
 318             mdaynum = val;
 319             found = TRUE;
 320           }
 321           dignext = DATE_YEAR;
 322         }
 323
 324         if(!found && (dignext == DATE_YEAR) && (yearnum == -1)) {
 325           yearnum = val;
 326           found = TRUE;
 327           if(yearnum < 1900) {
 328             if (yearnum > 70)
 329               yearnum += 1900;
 330             else
 331               yearnum += 2000;
 332           }
 333           if(mdaynum == -1)
 334             dignext = DATE_MDAY;
 335         }
 336
 337         if(!found)
 338           return -1;
 339
 340         date = end;
 341       }
 342     }
 343
 344     part++;
 345   }
 346
 347   if(-1 == secnum)
 348     secnum = minnum = hournum = 0; /* no time, make it zero */
 349
 350   if((-1 == mdaynum) ||
 351      (-1 == monnum) ||
 352      (-1 == yearnum))
 353     /* lacks vital info, fail */
 354     return -1;
 355
 356 #if SIZEOF_TIME_T < 5
 357   /* 32 bit time_t can only hold dates to the beginning of 2038 */
 358   if(yearnum > 2037)
 359     return 0x7fffffff;
 360 #endif
 361
 362   tm.tm_sec = secnum;
 363   tm.tm_min = minnum;
 364   tm.tm_hour = hournum;
 365   tm.tm_mday = mdaynum;
 366   tm.tm_mon = monnum;
 367   tm.tm_year = yearnum - 1900;
 368   tm.tm_wday = 0;
 369   tm.tm_yday = 0;
 370   tm.tm_isdst = 0;
 371
 372   /* mktime() returns a time_t. time_t is often 32 bits, even on many
 373      architectures that feature 64 bit 'long'.
 374
 375      Some systems have 64 bit time_t and deal with years beyond 2038. However,
 376      even some of the systems with 64 bit time_t returns -1 for dates beyond
 377      03:14:07 UTC, January 19, 2038. (Such as AIX 5100-06)
 378   */
 379   t = mktime(&tm);
 380
 381   /* time zone adjust (cast t to int to compare to negative one) */
 382   if(-1 != (int)t) {
 383     struct tm *gmt;
 384     long delta;
 385     time_t t2;
 386
 387 #ifdef HAVE_GMTIME_R
 388     /* thread-safe version */
 389     struct tm keeptime2;
 390     gmt = (struct tm *)gmtime_r(&t, &keeptime2);
 391     if(!gmt)
 392       return -1; /* illegal date/time */
 393     t2 = mktime(gmt);
 394 #else
 395     /* It seems that at least the MSVC version of mktime() doesn't work
 396        properly if it gets the 'gmt' pointer passed in (which is a pointer
 397        returned from gmtime() pointing to static memory), so instead we copy
 398        the tm struct to a local struct and pass a pointer to that struct as
 399        input to mktime(). */
 400     struct tm gmt2;
 401     gmt = gmtime(&t); /* use gmtime_r() if available */
 402     if(!gmt)
 403       return -1; /* illegal date/time */
 404     gmt2 = *gmt;
 405     t2 = mktime(&gmt2);
 406 #endif
 407
 408     /* Add the time zone diff (between the given timezone and GMT) and the
 409        diff between the local time zone and GMT. */
 410     delta = (long)((tzoff!=-1?tzoff:0) + (t - t2));
 411
 412     if((delta>0) && (t + delta < t))
 413       return -1; /* time_t overflow */
 414
 415     t += delta;
 416   }
 417
 418   return t;
 419 }
 420
 421 time_t curl_getdate(const char *p, const time_t *now)
 422 {
 423   (void)now;
 424   return Curl_parsedate(p);
 425 }