Mercurial > hgrepos > Python2 > PyMuPDF

comparison mupdf-source/thirdparty/curl/lib/parsedate.c @ 2:b50eed0cc0ef upstream

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
ADD: MuPDF v1.26.7: the MuPDF source as downloaded by a default build of PyMuPDF 1.26.4. The directory name has changed: no version number in the expanded directory now.
author Franz Glasner <fzglas.hg@dom66.de>
date Mon, 15 Sep 2025 11:43:07 +0200
parents
children
comparison
equal deleted inserted replaced
1:1d09e1dec1d9 2:b50eed0cc0ef
1 /***************************************************************************
2 * _ _ ____ _
3 * Project ___| | | | _ \| |
4 * / __| | | | |_) | |
5 * | (__| |_| | _ <| |___
6 * \___|\___/|_| \_\_____|
7 *
8 * Copyright (C) 1998 - 2019, 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 https://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 ***************************************************************************/
22 /*
23 A brief summary of the date string formats this parser groks:
24
25 RFC 2616 3.3.1
26
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
30
31 we support dates without week day name:
32
33 06 Nov 1994 08:49:37 GMT
34 06-Nov-94 08:49:37 GMT
35 Nov 6 08:49:37 1994
36
37 without the time zone:
38
39 06 Nov 1994 08:49:37
40 06-Nov-94 08:49:37
41
42 weird order:
43
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)
47
48 time left out:
49
50 1994 Nov 6
51 06-Nov-94
52 Sun Nov 6 94
53
54 unusual separators:
55
56 1994.Nov.6
57 Sun/Nov/6/94/GMT
58
59 commonly used time zone names:
60
61 Sun, 06 Nov 1994 08:49:37 CET
62 06 Nov 1994 08:49:37 EST
63
64 time zones specified using RFC822 style:
65
66 Sun, 12 Sep 2004 15:05:58 -0700
67 Sat, 11 Sep 2004 21:32:11 +0200
68
69 compact numerical date strings:
70
71 20040912 15:05:58 -0700
72 20040911 +0200
73
74 */
75
76 #include "curl_setup.h"
77
78 #include <limits.h>
79
80 #include <curl/curl.h>
81 #include "strcase.h"
82 #include "warnless.h"
83 #include "parsedate.h"
84
85 /*
86 * parsedate()
87 *
88 * Returns:
89 *
90 * PARSEDATE_OK - a fine conversion
91 * PARSEDATE_FAIL - failed to convert
92 * PARSEDATE_LATER - time overflow at the far end of time_t
93 * PARSEDATE_SOONER - time underflow at the low end of time_t
94 */
95
96 static int parsedate(const char *date, time_t *output);
97
98 #define PARSEDATE_OK 0
99 #define PARSEDATE_FAIL -1
100 #define PARSEDATE_LATER 1
101 #define PARSEDATE_SOONER 2
102
103 #ifndef CURL_DISABLE_PARSEDATE
104
105 const char * const Curl_wkday[] =
106 {"Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"};
107 static const char * const weekday[] =
108 { "Monday", "Tuesday", "Wednesday", "Thursday",
109 "Friday", "Saturday", "Sunday" };
110 const char * const Curl_month[]=
111 { "Jan", "Feb", "Mar", "Apr", "May", "Jun",
112 "Jul", "Aug", "Sep", "Oct", "Nov", "Dec" };
113
114 struct tzinfo {
115 char name[5];
116 int offset; /* +/- in minutes */
117 };
118
119 /* Here's a bunch of frequently used time zone names. These were supported
120 by the old getdate parser. */
121 #define tDAYZONE -60 /* offset for daylight savings time */
122 static const struct tzinfo tz[]= {
123 {"GMT", 0}, /* Greenwich Mean */
124 {"UT", 0}, /* Universal Time */
125 {"UTC", 0}, /* Universal (Coordinated) */
126 {"WET", 0}, /* Western European */
127 {"BST", 0 tDAYZONE}, /* British Summer */
128 {"WAT", 60}, /* West Africa */
129 {"AST", 240}, /* Atlantic Standard */
130 {"ADT", 240 tDAYZONE}, /* Atlantic Daylight */
131 {"EST", 300}, /* Eastern Standard */
132 {"EDT", 300 tDAYZONE}, /* Eastern Daylight */
133 {"CST", 360}, /* Central Standard */
134 {"CDT", 360 tDAYZONE}, /* Central Daylight */
135 {"MST", 420}, /* Mountain Standard */
136 {"MDT", 420 tDAYZONE}, /* Mountain Daylight */
137 {"PST", 480}, /* Pacific Standard */
138 {"PDT", 480 tDAYZONE}, /* Pacific Daylight */
139 {"YST", 540}, /* Yukon Standard */
140 {"YDT", 540 tDAYZONE}, /* Yukon Daylight */
141 {"HST", 600}, /* Hawaii Standard */
142 {"HDT", 600 tDAYZONE}, /* Hawaii Daylight */
143 {"CAT", 600}, /* Central Alaska */
144 {"AHST", 600}, /* Alaska-Hawaii Standard */
145 {"NT", 660}, /* Nome */
146 {"IDLW", 720}, /* International Date Line West */
147 {"CET", -60}, /* Central European */
148 {"MET", -60}, /* Middle European */
149 {"MEWT", -60}, /* Middle European Winter */
150 {"MEST", -60 tDAYZONE}, /* Middle European Summer */
151 {"CEST", -60 tDAYZONE}, /* Central European Summer */
152 {"MESZ", -60 tDAYZONE}, /* Middle European Summer */
153 {"FWT", -60}, /* French Winter */
154 {"FST", -60 tDAYZONE}, /* French Summer */
155 {"EET", -120}, /* Eastern Europe, USSR Zone 1 */
156 {"WAST", -420}, /* West Australian Standard */
157 {"WADT", -420 tDAYZONE}, /* West Australian Daylight */
158 {"CCT", -480}, /* China Coast, USSR Zone 7 */
159 {"JST", -540}, /* Japan Standard, USSR Zone 8 */
160 {"EAST", -600}, /* Eastern Australian Standard */
161 {"EADT", -600 tDAYZONE}, /* Eastern Australian Daylight */
162 {"GST", -600}, /* Guam Standard, USSR Zone 9 */
163 {"NZT", -720}, /* New Zealand */
164 {"NZST", -720}, /* New Zealand Standard */
165 {"NZDT", -720 tDAYZONE}, /* New Zealand Daylight */
166 {"IDLE", -720}, /* International Date Line East */
167 /* Next up: Military timezone names. RFC822 allowed these, but (as noted in
168 RFC 1123) had their signs wrong. Here we use the correct signs to match
169 actual military usage.
170 */
171 {"A", 1 * 60}, /* Alpha */
172 {"B", 2 * 60}, /* Bravo */
173 {"C", 3 * 60}, /* Charlie */
174 {"D", 4 * 60}, /* Delta */
175 {"E", 5 * 60}, /* Echo */
176 {"F", 6 * 60}, /* Foxtrot */
177 {"G", 7 * 60}, /* Golf */
178 {"H", 8 * 60}, /* Hotel */
179 {"I", 9 * 60}, /* India */
180 /* "J", Juliet is not used as a timezone, to indicate the observer's local
181 time */
182 {"K", 10 * 60}, /* Kilo */
183 {"L", 11 * 60}, /* Lima */
184 {"M", 12 * 60}, /* Mike */
185 {"N", -1 * 60}, /* November */
186 {"O", -2 * 60}, /* Oscar */
187 {"P", -3 * 60}, /* Papa */
188 {"Q", -4 * 60}, /* Quebec */
189 {"R", -5 * 60}, /* Romeo */
190 {"S", -6 * 60}, /* Sierra */
191 {"T", -7 * 60}, /* Tango */
192 {"U", -8 * 60}, /* Uniform */
193 {"V", -9 * 60}, /* Victor */
194 {"W", -10 * 60}, /* Whiskey */
195 {"X", -11 * 60}, /* X-ray */
196 {"Y", -12 * 60}, /* Yankee */
197 {"Z", 0}, /* Zulu, zero meridian, a.k.a. UTC */
198 };
199
200 /* returns:
201 -1 no day
202 0 monday - 6 sunday
203 */
204
205 static int checkday(const char *check, size_t len)
206 {
207 int i;
208 const char * const *what;
209 bool found = FALSE;
210 if(len > 3)
211 what = &weekday[0];
212 else
213 what = &Curl_wkday[0];
214 for(i = 0; i<7; i++) {
215 if(strcasecompare(check, what[0])) {
216 found = TRUE;
217 break;
218 }
219 what++;
220 }
221 return found?i:-1;
222 }
223
224 static int checkmonth(const char *check)
225 {
226 int i;
227 const char * const *what;
228 bool found = FALSE;
229
230 what = &Curl_month[0];
231 for(i = 0; i<12; i++) {
232 if(strcasecompare(check, what[0])) {
233 found = TRUE;
234 break;
235 }
236 what++;
237 }
238 return found?i:-1; /* return the offset or -1, no real offset is -1 */
239 }
240
241 /* return the time zone offset between GMT and the input one, in number
242 of seconds or -1 if the timezone wasn't found/legal */
243
244 static int checktz(const char *check)
245 {
246 unsigned int i;
247 const struct tzinfo *what;
248 bool found = FALSE;
249
250 what = tz;
251 for(i = 0; i< sizeof(tz)/sizeof(tz[0]); i++) {
252 if(strcasecompare(check, what->name)) {
253 found = TRUE;
254 break;
255 }
256 what++;
257 }
258 return found?what->offset*60:-1;
259 }
260
261 static void skip(const char **date)
262 {
263 /* skip everything that aren't letters or digits */
264 while(**date && !ISALNUM(**date))
265 (*date)++;
266 }
267
268 enum assume {
269 DATE_MDAY,
270 DATE_YEAR,
271 DATE_TIME
272 };
273
274 /* this is a clone of 'struct tm' but with all fields we don't need or use
275 cut out */
276 struct my_tm {
277 int tm_sec;
278 int tm_min;
279 int tm_hour;
280 int tm_mday;
281 int tm_mon;
282 int tm_year; /* full year */
283 };
284
285 /* struct tm to time since epoch in GMT time zone.
286 * This is similar to the standard mktime function but for GMT only, and
287 * doesn't suffer from the various bugs and portability problems that
288 * some systems' implementations have.
289 *
290 * Returns 0 on success, otherwise non-zero.
291 */
292 static void my_timegm(struct my_tm *tm, time_t *t)
293 {
294 static const int month_days_cumulative [12] =
295 { 0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334 };
296 int month, year, leap_days;
297
298 year = tm->tm_year;
299 month = tm->tm_mon;
300 if(month < 0) {
301 year += (11 - month) / 12;
302 month = 11 - (11 - month) % 12;
303 }
304 else if(month >= 12) {
305 year -= month / 12;
306 month = month % 12;
307 }
308
309 leap_days = year - (tm->tm_mon <= 1);
310 leap_days = ((leap_days / 4) - (leap_days / 100) + (leap_days / 400)
311 - (1969 / 4) + (1969 / 100) - (1969 / 400));
312
313 *t = ((((time_t) (year - 1970) * 365
314 + leap_days + month_days_cumulative[month] + tm->tm_mday - 1) * 24
315 + tm->tm_hour) * 60 + tm->tm_min) * 60 + tm->tm_sec;
316 }
317
318 /*
319 * parsedate()
320 *
321 * Returns:
322 *
323 * PARSEDATE_OK - a fine conversion
324 * PARSEDATE_FAIL - failed to convert
325 * PARSEDATE_LATER - time overflow at the far end of time_t
326 * PARSEDATE_SOONER - time underflow at the low end of time_t
327 */
328
329 static int parsedate(const char *date, time_t *output)
330 {
331 time_t t = 0;
332 int wdaynum = -1; /* day of the week number, 0-6 (mon-sun) */
333 int monnum = -1; /* month of the year number, 0-11 */
334 int mdaynum = -1; /* day of month, 1 - 31 */
335 int hournum = -1;
336 int minnum = -1;
337 int secnum = -1;
338 int yearnum = -1;
339 int tzoff = -1;
340 struct my_tm tm;
341 enum assume dignext = DATE_MDAY;
342 const char *indate = date; /* save the original pointer */
343 int part = 0; /* max 6 parts */
344
345 while(*date && (part < 6)) {
346 bool found = FALSE;
347
348 skip(&date);
349
350 if(ISALPHA(*date)) {
351 /* a name coming up */
352 char buf[32]="";
353 size_t len;
354 if(sscanf(date, "%31[ABCDEFGHIJKLMNOPQRSTUVWXYZ"
355 "abcdefghijklmnopqrstuvwxyz]", buf))
356 len = strlen(buf);
357 else
358 len = 0;
359
360 if(wdaynum == -1) {
361 wdaynum = checkday(buf, len);
362 if(wdaynum != -1)
363 found = TRUE;
364 }
365 if(!found && (monnum == -1)) {
366 monnum = checkmonth(buf);
367 if(monnum != -1)
368 found = TRUE;
369 }
370
371 if(!found && (tzoff == -1)) {
372 /* this just must be a time zone string */
373 tzoff = checktz(buf);
374 if(tzoff != -1)
375 found = TRUE;
376 }
377
378 if(!found)
379 return PARSEDATE_FAIL; /* bad string */
380
381 date += len;
382 }
383 else if(ISDIGIT(*date)) {
384 /* a digit */
385 int val;
386 char *end;
387 int len = 0;
388 if((secnum == -1) &&
389 (3 == sscanf(date, "%02d:%02d:%02d%n",
390 &hournum, &minnum, &secnum, &len))) {
391 /* time stamp! */
392 date += len;
393 }
394 else if((secnum == -1) &&
395 (2 == sscanf(date, "%02d:%02d%n", &hournum, &minnum, &len))) {
396 /* time stamp without seconds */
397 date += len;
398 secnum = 0;
399 }
400 else {
401 long lval;
402 int error;
403 int old_errno;
404
405 old_errno = errno;
406 errno = 0;
407 lval = strtol(date, &end, 10);
408 error = errno;
409 if(errno != old_errno)
410 errno = old_errno;
411
412 if(error)
413 return PARSEDATE_FAIL;
414
415 #if LONG_MAX != INT_MAX
416 if((lval > (long)INT_MAX) || (lval < (long)INT_MIN))
417 return PARSEDATE_FAIL;
418 #endif
419
420 val = curlx_sltosi(lval);
421
422 if((tzoff == -1) &&
423 ((end - date) == 4) &&
424 (val <= 1400) &&
425 (indate< date) &&
426 ((date[-1] == '+' || date[-1] == '-'))) {
427 /* four digits and a value less than or equal to 1400 (to take into
428 account all sorts of funny time zone diffs) and it is preceded
429 with a plus or minus. This is a time zone indication. 1400 is
430 picked since +1300 is frequently used and +1400 is mentioned as
431 an edge number in the document "ISO C 200X Proposal: Timezone
432 Functions" at http://david.tribble.com/text/c0xtimezone.html If
433 anyone has a more authoritative source for the exact maximum time
434 zone offsets, please speak up! */
435 found = TRUE;
436 tzoff = (val/100 * 60 + val%100)*60;
437
438 /* the + and - prefix indicates the local time compared to GMT,
439 this we need their reversed math to get what we want */
440 tzoff = date[-1]=='+'?-tzoff:tzoff;
441 }
442
443 if(((end - date) == 8) &&
444 (yearnum == -1) &&
445 (monnum == -1) &&
446 (mdaynum == -1)) {
447 /* 8 digits, no year, month or day yet. This is YYYYMMDD */
448 found = TRUE;
449 yearnum = val/10000;
450 monnum = (val%10000)/100-1; /* month is 0 - 11 */
451 mdaynum = val%100;
452 }
453
454 if(!found && (dignext == DATE_MDAY) && (mdaynum == -1)) {
455 if((val > 0) && (val<32)) {
456 mdaynum = val;
457 found = TRUE;
458 }
459 dignext = DATE_YEAR;
460 }
461
462 if(!found && (dignext == DATE_YEAR) && (yearnum == -1)) {
463 yearnum = val;
464 found = TRUE;
465 if(yearnum < 100) {
466 if(yearnum > 70)
467 yearnum += 1900;
468 else
469 yearnum += 2000;
470 }
471 if(mdaynum == -1)
472 dignext = DATE_MDAY;
473 }
474
475 if(!found)
476 return PARSEDATE_FAIL;
477
478 date = end;
479 }
480 }
481
482 part++;
483 }
484
485 if(-1 == secnum)
486 secnum = minnum = hournum = 0; /* no time, make it zero */
487
488 if((-1 == mdaynum) ||
489 (-1 == monnum) ||
490 (-1 == yearnum))
491 /* lacks vital info, fail */
492 return PARSEDATE_FAIL;
493
494 #ifdef HAVE_TIME_T_UNSIGNED
495 if(yearnum < 1970) {
496 /* only positive numbers cannot return earlier */
497 *output = TIME_T_MIN;
498 return PARSEDATE_SOONER;
499 }
500 #endif
501
502 #if (SIZEOF_TIME_T < 5)
503
504 #ifdef HAVE_TIME_T_UNSIGNED
505 /* an unsigned 32 bit time_t can only hold dates to 2106 */
506 if(yearnum > 2105) {
507 *output = TIME_T_MAX;
508 return PARSEDATE_LATER;
509 }
510 #else
511 /* a signed 32 bit time_t can only hold dates to the beginning of 2038 */
512 if(yearnum > 2037) {
513 *output = TIME_T_MAX;
514 return PARSEDATE_LATER;
515 }
516 if(yearnum < 1903) {
517 *output = TIME_T_MIN;
518 return PARSEDATE_SOONER;
519 }
520 #endif
521
522 #else
523 /* The Gregorian calendar was introduced 1582 */
524 if(yearnum < 1583)
525 return PARSEDATE_FAIL;
526 #endif
527
528 if((mdaynum > 31) || (monnum > 11) ||
529 (hournum > 23) || (minnum > 59) || (secnum > 60))
530 return PARSEDATE_FAIL; /* clearly an illegal date */
531
532 tm.tm_sec = secnum;
533 tm.tm_min = minnum;
534 tm.tm_hour = hournum;
535 tm.tm_mday = mdaynum;
536 tm.tm_mon = monnum;
537 tm.tm_year = yearnum;
538
539 /* my_timegm() returns a time_t. time_t is often 32 bits, sometimes even on
540 architectures that feature 64 bit 'long' but ultimately time_t is the
541 correct data type to use.
542 */
543 my_timegm(&tm, &t);
544
545 /* Add the time zone diff between local time zone and GMT. */
546 if(tzoff == -1)
547 tzoff = 0;
548
549 if((tzoff > 0) && (t > TIME_T_MAX - tzoff)) {
550 *output = TIME_T_MAX;
551 return PARSEDATE_LATER; /* time_t overflow */
552 }
553
554 t += tzoff;
555
556 *output = t;
557
558 return PARSEDATE_OK;
559 }
560 #else
561 /* disabled */
562 static int parsedate(const char *date, time_t *output)
563 {
564 (void)date;
565 *output = 0;
566 return PARSEDATE_OK; /* a lie */
567 }
568 #endif
569
570 time_t curl_getdate(const char *p, const time_t *now)
571 {
572 time_t parsed = -1;
573 int rc = parsedate(p, &parsed);
574 (void)now; /* legacy argument from the past that we ignore */
575
576 if(rc == PARSEDATE_OK) {
577 if(parsed == -1)
578 /* avoid returning -1 for a working scenario */
579 parsed++;
580 return parsed;
581 }
582 /* everything else is fail */
583 return -1;
584 }
585
586 /*
587 * Curl_gmtime() is a gmtime() replacement for portability. Do not use the
588 * gmtime_r() or gmtime() functions anywhere else but here.
589 *
590 */
591
592 CURLcode Curl_gmtime(time_t intime, struct tm *store)
593 {
594 const struct tm *tm;
595 #ifdef HAVE_GMTIME_R
596 /* thread-safe version */
597 tm = (struct tm *)gmtime_r(&intime, store);
598 #else
599 tm = gmtime(&intime);
600 if(tm)
601 *store = *tm; /* copy the pointed struct to the local copy */
602 #endif
603
604 if(!tm)
605 return CURLE_BAD_FUNCTION_ARGUMENT;
606 return CURLE_OK;
607 }