[time-nuts] Leap-second happens when...??

[M. Warner Losh]

In message: <4956D398.6060401 at rubidium.dyndns.org>
            Magnus Danielson <magnus at rubidium.dyndns.org> writes:
: 
: > : The first preference on which month to insert them is first to the June 
: > : and December.
: > : 
: > : The second preference is on Mars and September.
: > 
: > I assume you mean 'May' here.
: 
: No. Read on. (spelt March incorrectly. sorry)
: 
: >  The standard additionally allows for
: > leap seconds at the end of any month.  However, there's a lot of
: > software out there that will be dodgy for the Secondary times, and not
: > work at all for the others.
: 
: TF.460-4 says:
: 
: "A positive or negative leap-second should be the last second of a UTC 
: month, but first preference should be given to the end of December and 
: June, and second preference to the end of March and September."
: 
: There is an incorrect interpretation that only December and June are 
: allowed. Another incorrect interpretation is that only March, June, 
: September and December is allowed. Each month is allowed. There is only 
: a preference towards the others. The flagging of upcoming leap second 
: occuring in December already in July is ambiguous. Ah well.
: 
: Anyway, the TF.460-4 is very clear on it. May is allowed, but not preferred.

Yes.  But there's very little software that actually implements the
secondary months, let alone the others.  There's been cases where the
GPS constellation turned on leap indicator before the end of September
and the HP Oscilloquartz GPSDO clocks glitched at the end of September
due to a bug in the firmware...

Warner