[tz] strftime %s
Brooks Harris
brooks at edlmax.com
Mon Jan 15 19:38:10 UTC 2024
You might want to have a look at __mktime_internal (). I discovered some
time ago that mktime() did not return correct time_t values in some
cases. For examples America/New_York, 1945-08-14 19:00:00, where the
only difference is the Abbr (tm_zone), from EWT to EPT
Africa/Johannesburg, 1944-03-19 01:00:00, where the Abbr (tm_zone) is
the same, SAST to SAST __mktime_internal () was missing points near
these transitions (and others) and so returning incorrect results. This
was causing much confusion in my testing. The problem was that
__mktime_internal () was comparing only isdstwhere it also needed to
compare Abbr, as explained more in comments in the attached modified
code. Attached is the __mktime_internal () code with my suggested
modifications extracted from glibc-2.38\time\mktime.c as downloaded from
https://mirrors.ibiblio.org/gnu/libc/glibc-2.38.tar.xz
See the two code blocks commented // Modified by Brooks Harris
Since making these changes I have not seen mktime() make any errors at
many thousands of test points where localtime() was populating struct tm.
I hope this might be helpful.
-Brooks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mm.icann.org/pipermail/tz/attachments/20240115/02e0fdb0/attachment.htm>
-------------- next part --------------
/* Convert *TP to a __time64_t value, inverting
the monotonic and mostly-unit-linear conversion function CONVERT.
Use *OFFSET to keep track of a guess at the offset of the result,
compared to what the result would be for UTC without leap seconds.
If *OFFSET's guess is correct, only one CONVERT call is needed.
If successful, set *TP to the canonicalized struct tm;
otherwise leave *TP alone, return ((time_t) -1) and set errno.
This function is external because it is used also by timegm.c. */
__time64_t
__mktime_internal (struct tm *tp,
struct tm *(*convert) (const __time64_t *, struct tm *),
mktime_offset_t *offset)
{
struct tm tm;
/* The maximum number of probes (calls to CONVERT) should be enough
to handle any combinations of time zone rule changes, solar time,
leap seconds, and oscillations around a spring-forward gap.
POSIX.1 prohibits leap seconds, but some hosts have them anyway. */
int remaining_probes = 6;
/* Time requested. Copy it in case CONVERT modifies *TP; this can
occur if TP is localtime's returned value and CONVERT is localtime. */
int sec = tp->tm_sec;
int min = tp->tm_min;
int hour = tp->tm_hour;
int mday = tp->tm_mday;
int mon = tp->tm_mon;
int year_requested = tp->tm_year;
int isdst = tp->tm_isdst;
/* 1 if the previous probe was DST. */
int dst2 = 0;
/* Ensure that mon is in range, and set year accordingly. */
int mon_remainder = mon % 12;
int negative_mon_remainder = mon_remainder < 0;
int mon_years = mon / 12 - negative_mon_remainder;
long_int lyear_requested = year_requested;
long_int year = lyear_requested + mon_years;
/* The other values need not be in range:
the remaining code handles overflows correctly. */
/* Calculate day of year from year, month, and day of month.
The result need not be in range. */
int mon_yday = ((__mon_yday[leapyear (year)]
[mon_remainder + 12 * negative_mon_remainder])
- 1);
long_int lmday = mday;
long_int yday = mon_yday + lmday;
mktime_offset_t off = *offset;
int negative_offset_guess;
int sec_requested = sec;
if (LEAP_SECONDS_POSSIBLE)
{
/* Handle out-of-range seconds specially,
since ydhms_diff assumes every minute has 60 seconds. */
if (sec < 0)
sec = 0;
if (59 < sec)
sec = 59;
}
/* Invert CONVERT by probing. First assume the same offset as last
time. */
INT_SUBTRACT_WRAPV (0, off, &negative_offset_guess);
long_int t0 = ydhms_diff (year, yday, hour, min, sec,
EPOCH_YEAR - TM_YEAR_BASE, 0, 0, 0,
negative_offset_guess);
long_int t = t0, t1 = t0, t2 = t0;
/* Repeatedly use the error to improve the guess. */
while (true)
{
if (! ranged_convert (convert, &t, &tm))
return -1;
long_int dt = tm_diff (year, yday, hour, min, sec, &tm);
if (dt == 0)
break;
if (t == t1 && t != t2
&& (tm.tm_isdst < 0
|| (isdst < 0
? dst2 <= (tm.tm_isdst != 0)
: (isdst != 0) != (tm.tm_isdst != 0))))
/* We can't possibly find a match, as we are oscillating
between two values. The requested time probably falls
within a spring-forward gap of size DT. Follow the common
practice in this case, which is to return a time that is DT
away from the requested time, preferring a time whose
tm_isdst differs from the requested value. (If no tm_isdst
was requested and only one of the two values has a nonzero
tm_isdst, prefer that value.) In practice, this is more
useful than returning -1. */
goto offset_found;
remaining_probes--;
if (remaining_probes == 0)
{
__set_errno (EOVERFLOW);
return -1;
}
t1 = t2, t2 = t, t += dt, dst2 = tm.tm_isdst != 0;
}
/* We have a match. Check whether tm.tm_isdst has the requested
value, if any. */
// Modified by Brooks Harris
//
// if (isdst_differ (isdst, tm.tm_isdst)) <<<<<<<
//
// Checking only isdst_differ() is insufficient in some cases.
// Example America/New_York, 1945-08-14 19:00:00, where the only
// difference is the Abbr (tm_zone), from EWT to EPT
// Check that Abbr is not same OR tm_isdst and tm_gmtoff differ
// However,
// Example Africa/Johannesburg, 1944-03-19 01:00:00, where the
// Abbr (tm_zone) is the same, SAST to SAST
// Check that Abbr is the same AND tm_isdst and tm_gmtoff differ
//
// see below the second block of modified logic
// also commented // Modified by Brooks Harris
if((strcmp(tp->tm_zone, tm.tm_zone) != 0 // Abbr not same
|| (tp->tm_isdst == tm.tm_isdst // tm_isdst differ
&& tp->tm_gmtoff != tm.tm_gmtoff)) // tm_gmtoff differ
|| (strcmp(tp->tm_zone, tm.tm_zone) == 0 // Abbr same
&& tp->tm_isdst != tm.tm_isdst // tm_isdst differ
&& tp->tm_gmtoff != tm.tm_gmtoff)) // tm_gmtoff differ
{
/* tm.tm_isdst has the wrong value. Look for a neighboring
time with the right value, and use its UTC offset.
Heuristic: probe the adjacent timestamps in both directions,
looking for the desired isdst. If none is found within a
reasonable duration bound, assume a one-hour DST difference.
This should work for all real time zone histories in the tz
database. */
/* +1 if we wanted standard time but got DST, -1 if the reverse. */
int dst_difference = (isdst == 0) - (tm.tm_isdst == 0);
/* Distance between probes when looking for a DST boundary. In
tzdata2003a, the shortest period of DST is 601200 seconds
(e.g., America/Recife starting 2000-10-08 01:00), and the
shortest period of non-DST surrounded by DST is 694800
seconds (Africa/Tunis starting 1943-04-17 01:00). Use the
minimum of these two values, so we don't miss these short
periods when probing. */
int stride = 601200;
/* In TZDB 2021e, the longest period of DST (or of non-DST), in
which the DST (or adjacent DST) difference is not one hour,
is 457243209 seconds: e.g., America/Cambridge_Bay with leap
seconds, starting 1965-10-31 00:00 in a switch from
double-daylight time (-05) to standard time (-07), and
continuing to 1980-04-27 02:00 in a switch from standard time
(-07) to daylight time (-06). */
int duration_max = 457243209;
/* Search in both directions, so the maximum distance is half
the duration; add the stride to avoid off-by-1 problems. */
int delta_bound = duration_max / 2 + stride;
int delta, direction;
for (delta = stride; delta < delta_bound; delta += stride)
for (direction = -1; direction <= 1; direction += 2)
{
long_int ot;
if (! INT_ADD_WRAPV (t, delta * direction, &ot))
{
struct tm otm;
if (! ranged_convert (convert, &ot, &otm))
return -1;
if (! isdst_differ (isdst, otm.tm_isdst))
// Modified by Brooks Harris
// if (! isdst_differ (isdst, otm.tm_isdst))
if(strcmp(tp->tm_zone, otm.tm_zone) == 0 // check abbr match instead of isdst
&& (tp->tm_isdst == otm.tm_isdst
&& tp->tm_gmtoff == otm.tm_gmtoff))
{
/* We found the desired tm_isdst.
Extrapolate back to the desired time. */
long_int gt = ot + tm_diff (year, yday, hour, min, sec,
&otm);
if (mktime_min <= gt && gt <= mktime_max)
{
if (convert_time (convert, gt, &tm))
{
t = gt;
goto offset_found;
}
if (errno != EOVERFLOW)
return -1;
}
}
}
}
/* No unusual DST offset was found nearby. Assume one-hour DST. */
t += 60 * 60 * dst_difference;
if (mktime_min <= t && t <= mktime_max && convert_time (convert, t, &tm))
goto offset_found;
__set_errno (EOVERFLOW);
return -1;
}
offset_found:
/* Set *OFFSET to the low-order bits of T - T0 - NEGATIVE_OFFSET_GUESS.
This is just a heuristic to speed up the next mktime call, and
correctness is unaffected if integer overflow occurs here. */
INT_SUBTRACT_WRAPV (t, t0, offset);
INT_SUBTRACT_WRAPV (*offset, negative_offset_guess, offset);
if (LEAP_SECONDS_POSSIBLE && sec_requested != tm.tm_sec)
{
/* Adjust time to reflect the tm_sec requested, not the normalized value.
Also, repair any damage from a false match due to a leap second. */
long_int sec_adjustment = sec == 0 && tm.tm_sec == 60;
sec_adjustment -= sec;
sec_adjustment += sec_requested;
if (INT_ADD_WRAPV (t, sec_adjustment, &t)
|| ! (mktime_min <= t && t <= mktime_max))
{
__set_errno (EOVERFLOW);
return -1;
}
if (! convert_time (convert, t, &tm))
return -1;
}
*tp = tm;
return t;
}
More information about the tz
mailing list