Read dates after 2038 correctly on 32-bit builds

src/htsbauth.c

@@ -350,31 +350,14 @@ hts_boolean cookie_max_age_time(const char *value, int64_t now, int64_t *when) {
   return HTS_TRUE;
 }
 
-/* Return the Unix time of the UTC date in tm, in 64 bits because a 32-bit
-   time_t ends in 2038. The calendar year must be 1 or later, and tm_mon 0 to
-   11, as convert_time_rfc822_pivot yields them. */
-static int64_t cookie_utc_seconds(const struct tm *tm) {
-  // days from the civil date, counting years from March (Howard Hinnant)
-  const int64_t y = (int64_t) tm->tm_year + 1900 - (tm->tm_mon < 2 ? 1 : 0);
-  const int64_t era = y / 400;
-  const int64_t yoe = y - era * 400;
-  const int64_t doy = (153 * (int64_t) ((tm->tm_mon + 10) % 12) + 2) / 5 +
-                      (int64_t) tm->tm_mday - 1;
-  const int64_t days =
-      era * 146097 + yoe * 365 + yoe / 4 - yoe / 100 + doy - 719468;
-
-  return days * 86400 + (int64_t) tm->tm_hour * 3600 +
-         (int64_t) tm->tm_min * 60 + tm->tm_sec;
-}
-
 /* See htsbauth.h. */
 hts_boolean cookie_expires_time(const char *value, int64_t now, int64_t *when) {
   struct tm tm;
 
   // RFC 6265 5.1.1 reads a two-digit year below 70 as 20xx
   if (convert_time_rfc822_pivot(&tm, value, 69) == NULL)
     return HTS_FALSE;
-  *when = cookie_bound_expiry(cookie_utc_seconds(&tm), now);
+  *when = cookie_bound_expiry(hts_utc_seconds(&tm), now);
   return HTS_TRUE;
 }
 

src/htscoremain.c

@@ -371,7 +371,7 @@ static hts_boolean write_progress_lock(httrackp *opt, const char *path,
   fprintf(fp, "UID=%d\n", (int) getuid());
   fprintf(fp, "GID=%d\n", (int) getuid());
 #endif
-  fprintf(fp, "START=%d\n", (int) time(NULL));
+  fprintf(fp, "START=" LLintP "\n", (LLint) time(NULL));
 #endif
   fclose(fp);
   /* Whole-second filesystems tie two files written in the same second, and the

src/htsdate.c

@@ -38,6 +38,7 @@ Please visit our Website: http://www.httrack.com
 
 #include <ctype.h>
 #include <limits.h>
+#include <stdint.h>
 #include <string.h>
 
 /* hts_lowcase() lives in the library proxytrack does not link. */
@@ -150,3 +151,56 @@ struct tm *convert_time_rfc822_pivot(struct tm *result, const char *s,
   }
   return NULL;
 }
+
+/* See htsdate.h. */
+int64_t hts_utc_seconds(const struct tm *tm) {
+  // days from the civil date, counting years from March (Howard Hinnant)
+  int64_t y = (int64_t) tm->tm_year + 1900 + tm->tm_mon / 12;
+  int64_t m = tm->tm_mon % 12;
+  int64_t era, yoe, doy, days;
+
+  if (m < 0) {
+    m += 12;
+    y--;
+  }
+  if (m < 2)
+    y--;
+  era = (y >= 0 ? y : y - 399) / 400;
+  yoe = y - era * 400;
+  doy = (153 * ((m + 10) % 12) + 2) / 5 + (int64_t) tm->tm_mday - 1;
+  days = era * 146097 + yoe * 365 + yoe / 4 - yoe / 100 + doy - 719468;
+  return days * 86400 + (int64_t) tm->tm_hour * 3600 +
+         (int64_t) tm->tm_min * 60 + tm->tm_sec;
+}
+
+/* See htsdate.h. */
+time_t hts_time_clip(int64_t t) {
+  if (sizeof(time_t) < sizeof(int64_t)) {
+    if (t > INT32_MAX)
+      return (time_t) INT32_MAX;
+    if (t < INT32_MIN)
+      return (time_t) INT32_MIN;
+  }
+  return (time_t) t;
+}
+
+/* See htsdate.h. */
+hts_boolean hts_time_exact(int64_t t, time_t *out) {
+  const time_t clipped = hts_time_clip(t);
+
+  if ((int64_t) clipped != t)
+    return HTS_FALSE;
+  *out = clipped;
+  return HTS_TRUE;
+}
+
+/* See htsdate.h. */
+hts_boolean hts_tm_filetime(const struct tm *tm, time_t *out) {
+  const int64_t t = hts_utc_seconds(tm);
+
+  if (t == -1 || t == 0)
+    return HTS_FALSE;
+  // A date past what time_t holds stamps its last second, never 1901.
+  *out = hts_time_clip(t);
+  return HTS_TRUE;
+}

src/htsdate.h

@@ -33,17 +33,35 @@ Please visit our Website: http://www.httrack.com
 #ifndef HTSDATE_DEFH
 #define HTSDATE_DEFH
 
+#include "htsglobal.h"
+
+#include <stdint.h>
 #include <time.h>
 
 /* Parse an RFC822 date such as "Sun, 06 Nov 1994 08:49:37 GMT" into "result",
    which is returned, or NULL if "s" does not parse. The fields are GMT, so a
-   caller wanting a time_t uses timegm(), never mktime() (#1491). Its own file
-   so proxytrack, which does not link the library, can share it. */
+   caller wanting seconds uses hts_utc_seconds(), never mktime() (#1491). This
+   file is separate so proxytrack, which does not link the library, can share
+   it. */
 struct tm *convert_time_rfc822(struct tm *result, const char *s);
 
 /* As convert_time_rfc822, but a two-digit year up to PIVOT means 20xx and one
    above it 19xx. HTTP dates use 50, and cookies use 69 (RFC 6265 5.1.1). */
 struct tm *convert_time_rfc822_pivot(struct tm *result, const char *s,
                                      int pivot);
 
+/* Return the Unix time of the UTC date in tm, in 64 bits because a 32-bit
+   time_t ends in 2038. Out-of-range fields carry over as with timegm(). */
+int64_t hts_utc_seconds(const struct tm *tm);
+
+/* Return t as a time_t, clipped to the range this platform's time_t holds. */
+time_t hts_time_clip(int64_t t);
+
+/* Does t fit this platform's time_t exactly? If so, store it in *out. */
+hts_boolean hts_time_exact(int64_t t, time_t *out);
+
+/* Store the file time for the UTC date in tm in *out, clipped to time_t. False
+   when the date is the epoch or one second before it, which mean no date. */
+hts_boolean hts_tm_filetime(const struct tm *tm, time_t *out);
+
 #endif

src/htsftp.c

@@ -1025,7 +1025,10 @@ int run_launch_ftp(FTPDownloadStruct * pStruct) {
                     strlncatbuff(back->r.lastmodified, date,
                                  sizeof(back->r.lastmodified),
                                  sizeof(back->r.lastmodified) - 1);
-                    remote_mtime = timegm(&remote_tm);
+                    // a clipped date cannot prove the copy matches
+                    if (!hts_time_exact(hts_utc_seconds(&remote_tm),
+                                        &remote_mtime))
+                      remote_mtime = (time_t) -1;
                   }
                 }
 

src/htslib.c

@@ -3731,22 +3731,12 @@ void time_local_rfc822(char *s) {
     s[0] = '\0';
 }
 
-/* convertir une chaine en temps */
-static time_t getGMT(struct tm *tm) {
-  time_t t = timegm(tm);
-
-  if (t != (time_t) - 1 && t != (time_t) 0) {
-    return (time_t) t;
-  }
-  return (time_t) -1;
-}
-
 /* sets file time. -1 if error */
 /* Note: utf-8 */
 int set_filetime(const char *file, struct tm *tm_time) {
-  time_t t = getGMT(tm_time);
+  time_t t;
 
-  if (t != (time_t) - 1) {
+  if (hts_tm_filetime(tm_time, &t)) {
     STRUCT_UTIMBUF tim;
 
     memset(&tim, 0, sizeof(tim));
@@ -3779,26 +3769,24 @@ int hts_parse_retry_after(const char *value, time_t now) {
   }
 
   {
-    const time_t when = hts_parse_date(p, 50);
+    const int64_t when = hts_parse_date(p, 50);
 
-    if (when == (time_t) -1)
+    if (when == -1)
       return -1;
-    if (when <= now)
+    if (when <= (int64_t) now)
       return 0;
-    /* widen before subtracting: two time_t values can differ by more than an
-       int holds */
-    if ((LLint) when - (LLint) now > (LLint) INT_MAX)
+    if (when - (int64_t) now > INT_MAX)
       return INT_MAX;
-    return (int) (when - now);
+    return (int) (when - (int64_t) now);
   }
 }
 
 /* See htslib.h. */
-time_t hts_parse_date(const char *s, int pivot) {
+int64_t hts_parse_date(const char *s, int pivot) {
   struct tm buffer;
   struct tm *const tm_s = convert_time_rfc822_pivot(&buffer, s, pivot);
 
-  return tm_s != NULL ? timegm(tm_s) : (time_t) -1;
+  return tm_s != NULL ? hts_utc_seconds(tm_s) : -1;
 }
 
 /* sets file time from RFC822 date+time, -1 if error*/

src/htslib.h

@@ -246,9 +246,9 @@ void hts_init_htsblk(htsblk * r);
    no usable delay. `now` dates the HTTP-date form, and the caller caps. */
 int hts_parse_retry_after(const char *value, time_t now);
 
-/* Read the HTTP date S as a time_t, with a two-digit year up to PIVOT in 20xx
-   (see convert_time_rfc822_pivot). (time_t) -1 when S is not a date. */
-time_t hts_parse_date(const char *s, int pivot);
+/* Read the HTTP date S as Unix seconds, with a two-digit year up to PIVOT in
+   20xx (see convert_time_rfc822_pivot). -1 when S is not a date. */
+int64_t hts_parse_date(const char *s, int pivot);
 
 // attach specific project log to hachtable logger
 void hts_set_hash_handler(coucal hashtable, httrackp *opt);
@@ -843,7 +843,6 @@ HTS_STATIC int strcmpnocase(const char *a, const char *b) {
 #ifdef _MSC_VER
 #define fseeko _fseeki64
 #define ftello _ftelli64
-#define timegm _mkgmtime
 #endif
 
 #define strfield2(f,s) ( (strlen(f)!=strlen(s)) ? 0 : (strfield(f,s)) )