1 /***************************************************************************
3 * Project ___| | | | _ \| |
5 * | (__| |_| | _ <| |___
6 * \___|\___/|_| \_\_____|
8 * Copyright (C) 1998 - 2010, Daniel Stenberg, <daniel@haxx.se>, et al.
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.
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.
18 * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
19 * KIND, either express or implied.
21 ***************************************************************************/
23 A brief summary of the date string formats this parser groks:
27 Sun, 06 Nov 1994 08:49:37 GMT ; RFC 822, updated by RFC 1123
28 Sunday, 06-Nov-94 08:49:37 GMT ; RFC 850, obsoleted by RFC 1036
29 Sun Nov 6 08:49:37 1994 ; ANSI C's asctime() format
31 we support dates without week day name:
33 06 Nov 1994 08:49:37 GMT
34 06-Nov-94 08:49:37 GMT
37 without the time zone:
44 1994 Nov 6 08:49:37 (GNU date fails)
45 GMT 08:49:37 06-Nov-94 Sunday
46 94 6 Nov 08:49:37 (GNU date fails)
59 commonly used time zone names:
61 Sun, 06 Nov 1994 08:49:37 CET
62 06 Nov 1994 08:49:37 EST
64 time zones specified using RFC822 style:
66 Sun, 12 Sep 2004 15:05:58 -0700
67 Sat, 11 Sep 2004 21:32:11 +0200
69 compact numerical date strings:
71 20040912 15:05:58 -0700
81 #include <stdlib.h> /* for strtol() */
84 #include <curl/curl.h>
87 #include "parsedate.h"
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" };
100 int offset; /* +/- in minutes */
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 /* Next up: Military timezone names. RFC822 allowed these, but (as noted in
151 RFC 1123) had their signs wrong. Here we use the correct signs to match
152 actual military usage.
154 {"A", +1 * 60}, /* Alpha */
155 {"B", +2 * 60}, /* Bravo */
156 {"C", +3 * 60}, /* Charlie */
157 {"D", +4 * 60}, /* Delta */
158 {"E", +5 * 60}, /* Echo */
159 {"F", +6 * 60}, /* Foxtrot */
160 {"G", +7 * 60}, /* Golf */
161 {"H", +8 * 60}, /* Hotel */
162 {"I", +9 * 60}, /* India */
163 /* "J", Juliet is not used as a timezone, to indicate the observer's local time */
164 {"K", +10 * 60}, /* Kilo */
165 {"L", +11 * 60}, /* Lima */
166 {"M", +12 * 60}, /* Mike */
167 {"N", -1 * 60}, /* November */
168 {"O", -2 * 60}, /* Oscar */
169 {"P", -3 * 60}, /* Papa */
170 {"Q", -4 * 60}, /* Quebec */
171 {"R", -5 * 60}, /* Romeo */
172 {"S", -6 * 60}, /* Sierra */
173 {"T", -7 * 60}, /* Tango */
174 {"U", -8 * 60}, /* Uniform */
175 {"V", -9 * 60}, /* Victor */
176 {"W", -10 * 60}, /* Whiskey */
177 {"X", -11 * 60}, /* X-ray */
178 {"Y", -12 * 60}, /* Yankee */
179 {"Z", 0}, /* Zulu, zero meridian, a.k.a. UTC */
187 static int checkday(const char *check, size_t len)
190 const char * const *what;
195 what = &Curl_wkday[0];
197 if(Curl_raw_equal(check, what[0])) {
206 static int checkmonth(const char *check)
209 const char * const *what;
212 what = &Curl_month[0];
213 for(i=0; i<12; i++) {
214 if(Curl_raw_equal(check, what[0])) {
220 return found?i:-1; /* return the offset or -1, no real offset is -1 */
223 /* return the time zone offset between GMT and the input one, in number
224 of seconds or -1 if the timezone wasn't found/legal */
226 static int checktz(const char *check)
229 const struct tzinfo *what;
233 for(i=0; i< sizeof(tz)/sizeof(tz[0]); i++) {
234 if(Curl_raw_equal(check, what->name)) {
240 return found?what->offset*60:-1;
243 static void skip(const char **date)
245 /* skip everything that aren't letters or digits */
246 while(**date && !ISALNUM(**date))
256 /* this is a clone of 'struct tm' but with all fields we don't need or use
267 /* struct tm to time since epoch in GMT time zone.
268 * This is similar to the standard mktime function but for GMT only, and
269 * doesn't suffer from the various bugs and portability problems that
270 * some systems' implementations have.
272 static time_t my_timegm(struct my_tm *tm)
274 static const int month_days_cumulative [12] =
275 { 0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334 };
276 int month, year, leap_days;
279 /* we don't support years before 1970 as they will cause this function
280 to return a negative value */
283 year = tm->tm_year + 1900;
286 year += (11 - month) / 12;
287 month = 11 - (11 - month) % 12;
289 else if (month >= 12) {
294 leap_days = year - (tm->tm_mon <= 1);
295 leap_days = ((leap_days / 4) - (leap_days / 100) + (leap_days / 400)
296 - (1969 / 4) + (1969 / 100) - (1969 / 400));
298 return ((((time_t) (year - 1970) * 365
299 + leap_days + month_days_cumulative [month] + tm->tm_mday - 1) * 24
300 + tm->tm_hour) * 60 + tm->tm_min) * 60 + tm->tm_sec;
308 * PARSEDATE_OK - a fine conversion
309 * PARSEDATE_FAIL - failed to convert
310 * PARSEDATE_LATER - time overflow at the far end of time_t
311 * PARSEDATE_SOONER - time underflow at the low end of time_t
314 int Curl_parsedate(const char *date, time_t *output)
317 int wdaynum=-1; /* day of the week number, 0-6 (mon-sun) */
318 int monnum=-1; /* month of the year number, 0-11 */
319 int mdaynum=-1; /* day of month, 1 - 31 */
326 enum assume dignext = DATE_MDAY;
327 const char *indate = date; /* save the original pointer */
328 int part = 0; /* max 6 parts */
330 while(*date && (part < 6)) {
336 /* a name coming up */
339 sscanf(date, "%31[ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz]",
344 wdaynum = checkday(buf, len);
348 if(!found && (monnum == -1)) {
349 monnum = checkmonth(buf);
354 if(!found && (tzoff == -1)) {
355 /* this just must be a time zone string */
356 tzoff = checktz(buf);
362 return PARSEDATE_FAIL; /* bad string */
366 else if(ISDIGIT(*date)) {
371 (3 == sscanf(date, "%02d:%02d:%02d", &hournum, &minnum, &secnum))) {
375 else if((secnum == -1) &&
376 (2 == sscanf(date, "%02d:%02d", &hournum, &minnum))) {
377 /* time stamp without seconds */
382 val = curlx_sltosi(strtol(date, &end, 10));
385 ((end - date) == 4) &&
388 ((date[-1] == '+' || date[-1] == '-'))) {
389 /* four digits and a value less than or equal to 1400 (to take into
390 account all sorts of funny time zone diffs) and it is preceeded
391 with a plus or minus. This is a time zone indication. 1400 is
392 picked since +1300 is frequently used and +1400 is mentioned as
393 an edge number in the document "ISO C 200X Proposal: Timezone
394 Functions" at http://david.tribble.com/text/c0xtimezone.html If
395 anyone has a more authoritative source for the exact maximum time
396 zone offsets, please speak up! */
398 tzoff = (val/100 * 60 + val%100)*60;
400 /* the + and - prefix indicates the local time compared to GMT,
401 this we need ther reversed math to get what we want */
402 tzoff = date[-1]=='+'?-tzoff:tzoff;
405 if(((end - date) == 8) &&
409 /* 8 digits, no year, month or day yet. This is YYYYMMDD */
412 monnum = (val%10000)/100-1; /* month is 0 - 11 */
416 if(!found && (dignext == DATE_MDAY) && (mdaynum == -1)) {
417 if((val > 0) && (val<32)) {
424 if(!found && (dignext == DATE_YEAR) && (yearnum == -1)) {
438 return PARSEDATE_FAIL;
448 secnum = minnum = hournum = 0; /* no time, make it zero */
450 if((-1 == mdaynum) ||
453 /* lacks vital info, fail */
454 return PARSEDATE_FAIL;
456 #if SIZEOF_TIME_T < 5
457 /* 32 bit time_t can only hold dates to the beginning of 2038 */
459 *output = 0x7fffffff;
460 return PARSEDATE_LATER;
466 return PARSEDATE_SOONER;
471 tm.tm_hour = hournum;
472 tm.tm_mday = mdaynum;
474 tm.tm_year = yearnum - 1900;
476 /* my_timegm() returns a time_t. time_t is often 32 bits, even on many
477 architectures that feature 64 bit 'long'.
479 Some systems have 64 bit time_t and deal with years beyond 2038. However,
480 even on some of the systems with 64 bit time_t mktime() returns -1 for
481 dates beyond 03:14:07 UTC, January 19, 2038. (Such as AIX 5100-06)
485 /* time zone adjust (cast t to int to compare to negative one) */
488 /* Add the time zone diff between local time zone and GMT. */
489 long delta = (long)(tzoff!=-1?tzoff:0);
491 if((delta>0) && (t + delta < t))
492 return -1; /* time_t overflow */
502 time_t curl_getdate(const char *p, const time_t *now)
505 int rc = Curl_parsedate(p, &parsed);
506 (void)now; /* legacy argument from the past that we ignore */
510 case PARSEDATE_LATER:
511 case PARSEDATE_SOONER:
514 /* everything else is fail */