[buug] Leap second cometh: 2016-12-31T23:59:60+0000 2016-12-31T15:59:60-0800

[Michael Paoli]

Leap second cometh.
Ooooh, and smears and not.
Now we can have many millions or more clocks disagree about what time it is
for up to 24 hours.
2016-12-31T23:59:57+0000 2016-12-31T15:59:57-0800
2016-12-31T23:59:58+0000 2016-12-31T15:59:58-0800
2016-12-31T23:59:59+0000 2016-12-31T15:59:59-0800
2016-12-31T23:59:60+0000 2016-12-31T15:59:60-0800
2017-01-01T00:00:00+0000 2016-12-31T16:00:00-0800
2017-01-01T00:00:01+0000 2016-12-31T16:00:01-0800
2017-01-01T00:00:02+0000 2016-12-31T16:00:02-0800
2017-01-01T00:00:03+0000 2016-12-31T16:00:03-0800
https://developers.google.com/time/smear#othersmears
http://support.ntp.org/bin/view/Support/ConfiguringNTP#Section_6.14.
$ fgrep leap /etc/ntp.conf
leapfile /usr/share/zoneinfo/leap-seconds.list
$ fgrep Jan\ 2017 /usr/share/zoneinfo/leap-seconds.list
3692217600      37      # 1 Jan 2017
$
My gosh, some might be off up to a second on the start of their
New Year's celebration.
Ah well, at least we'll get our clocks better aligned to solar time again.
And we don't have the nasty mess of the past, where the time period for
the length of the second changed, by definition.
At least the sun dials aren't too far off.
https://en.wikipedia.org/wiki/Leap_second
https://en.wikipedia.org/wiki/Unix_time#Leap_seconds
https://en.wikipedia.org/wiki/Network_Time_Protocol
https://en.wikipedia.org/wiki/Network_Time_Protocol#Leap_seconds
ntpd(1) (notably also -x option)