1 /***************************************************************************
3 * Project ___| | | | _ \| |
5 * | (__| |_| | _ <| |___
6 * \___|\___/|_| \_\_____|
8 * Copyright (C) 1998 - 2008, 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.
22 ***************************************************************************/
24 A brief summary of the date string formats this parser groks:
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
32 we support dates without week day name:
34 06 Nov 1994 08:49:37 GMT
35 06-Nov-94 08:49:37 GMT
38 without the time zone:
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)
60 commonly used time zone names:
62 Sun, 06 Nov 1994 08:49:37 CET
63 06 Nov 1994 08:49:37 EST
65 time zones specified using RFC822 style:
67 Sun, 12 Sep 2004 15:05:58 -0700
68 Sat, 11 Sep 2004 21:32:11 +0200
70 compact numerical date strings:
72 20040912 15:05:58 -0700
82 #include <stdlib.h> /* for strtol() */
85 #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 */
157 static int checkday(const char *check, size_t len)
160 const char * const *what;
165 what = &Curl_wkday[0];
167 if(Curl_raw_equal(check, what[0])) {
176 static int checkmonth(const char *check)
179 const char * const *what;
182 what = &Curl_month[0];
183 for(i=0; i<12; i++) {
184 if(Curl_raw_equal(check, what[0])) {
190 return found?i:-1; /* return the offset or -1, no real offset is -1 */
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 */
196 static int checktz(const char *check)
199 const struct tzinfo *what;
203 for(i=0; i< sizeof(tz)/sizeof(tz[0]); i++) {
204 if(Curl_raw_equal(check, what->name)) {
210 return found?what->offset*60:-1;
213 static void skip(const char **date)
215 /* skip everything that aren't letters or digits */
216 while(**date && !ISALNUM(**date))
226 /* this is a clone of 'struct tm' but with all fields we don't need or use
237 /* struct tm to time since epoch in GMT time zone.
238 * This is similar to the standard mktime function but for GMT only, and
239 * doesn't suffer from the various bugs and portability problems that
240 * some systems' implementations have.
242 static time_t my_timegm(struct my_tm *tm)
244 static const int month_days_cumulative [12] =
245 { 0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334 };
246 int month, year, leap_days;
249 /* we don't support years before 1970 as they will cause this function
250 to return a negative value */
253 year = tm->tm_year + 1900;
256 year += (11 - month) / 12;
257 month = 11 - (11 - month) % 12;
259 else if (month >= 12) {
264 leap_days = year - (tm->tm_mon <= 1);
265 leap_days = ((leap_days / 4) - (leap_days / 100) + (leap_days / 400)
266 - (1969 / 4) + (1969 / 100) - (1969 / 400));
268 return ((((time_t) (year - 1970) * 365
269 + leap_days + month_days_cumulative [month] + tm->tm_mday - 1) * 24
270 + tm->tm_hour) * 60 + tm->tm_min) * 60 + tm->tm_sec;
273 static time_t parsedate(const char *date)
276 int wdaynum=-1; /* day of the week number, 0-6 (mon-sun) */
277 int monnum=-1; /* month of the year number, 0-11 */
278 int mdaynum=-1; /* day of month, 1 - 31 */
285 enum assume dignext = DATE_MDAY;
286 const char *indate = date; /* save the original pointer */
287 int part = 0; /* max 6 parts */
289 while(*date && (part < 6)) {
295 /* a name coming up */
298 sscanf(date, "%31[ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz]",
303 wdaynum = checkday(buf, len);
307 if(!found && (monnum == -1)) {
308 monnum = checkmonth(buf);
313 if(!found && (tzoff == -1)) {
314 /* this just must be a time zone string */
315 tzoff = checktz(buf);
321 return -1; /* bad string */
325 else if(ISDIGIT(*date)) {
330 (3 == sscanf(date, "%02d:%02d:%02d", &hournum, &minnum, &secnum))) {
336 val = (int)strtol(date, &end, 10);
339 ((end - date) == 4) &&
342 ((date[-1] == '+' || date[-1] == '-'))) {
343 /* four digits and a value less than or equal to 1400 (to take into
344 account all sorts of funny time zone diffs) and it is preceeded
345 with a plus or minus. This is a time zone indication. 1400 is
346 picked since +1300 is frequently used and +1400 is mentioned as
347 an edge number in the document "ISO C 200X Proposal: Timezone
348 Functions" at http://david.tribble.com/text/c0xtimezone.html If
349 anyone has a more authoritative source for the exact maximum time
350 zone offsets, please speak up! */
352 tzoff = (val/100 * 60 + val%100)*60;
354 /* the + and - prefix indicates the local time compared to GMT,
355 this we need ther reversed math to get what we want */
356 tzoff = date[-1]=='+'?-tzoff:tzoff;
359 if(((end - date) == 8) &&
363 /* 8 digits, no year, month or day yet. This is YYYYMMDD */
366 monnum = (val%10000)/100-1; /* month is 0 - 11 */
370 if(!found && (dignext == DATE_MDAY) && (mdaynum == -1)) {
371 if((val > 0) && (val<32)) {
378 if(!found && (dignext == DATE_YEAR) && (yearnum == -1)) {
402 secnum = minnum = hournum = 0; /* no time, make it zero */
404 if((-1 == mdaynum) ||
407 /* lacks vital info, fail */
410 #if SIZEOF_TIME_T < 5
411 /* 32 bit time_t can only hold dates to the beginning of 2038 */
418 tm.tm_hour = hournum;
419 tm.tm_mday = mdaynum;
421 tm.tm_year = yearnum - 1900;
423 /* my_timegm() returns a time_t. time_t is often 32 bits, even on many
424 architectures that feature 64 bit 'long'.
426 Some systems have 64 bit time_t and deal with years beyond 2038. However,
427 even on some of the systems with 64 bit time_t mktime() returns -1 for
428 dates beyond 03:14:07 UTC, January 19, 2038. (Such as AIX 5100-06)
432 /* time zone adjust (cast t to int to compare to negative one) */
435 /* Add the time zone diff between local time zone and GMT. */
436 long delta = (long)(tzoff!=-1?tzoff:0);
438 if((delta>0) && (t + delta < t))
439 return -1; /* time_t overflow */
447 time_t curl_getdate(const char *p, const time_t *now)