> On Oct 15, 2026, at 2:57 PM, Bradley White <[email protected]> wrote:
> 
> On Thu, Oct 15, 2026 at 5:04 PM Guy Harris via tz <[email protected] 
> <mailto:[email protected]>> wrote:
>> The one thing for which the leap second data could be useful is converting 
>> "seconds that have elapsed since the POSIX epoch, yeah, even counting leap 
>> seconds" time stamps for values in the past to UTC.
> 
> See time2posix(), which converts a "right" time_t (seconds that have elapsed 
> since the POSIX epoch, yeah, even counting leap seconds) to a "posix" time_t, 
> which can then easily be converted to UTC.

...as long as the "right" time_t value didn't correspond to a positive leap 
second, as converting a "posix" time_t using gmtime() will not produce a tm_sec 
value of 60, as would be correct for a UTC representation of that time stamp. 
(Similar problems would occur for negative leap seconds.)

Reply via email to