On Tue, Apr 4, 2017 at 12:02 PM, Chris Woodbury via tz <tz@iana.org> wrote:
I note, in the event of a (highly unlikely) negative leap
second, the LEAP directive generated is for 23:59:59 which,
in the case of a negative leap second, doesn't exist.  I
reviewed the C code and, it seems to me, the transaction
added should actually be for the second before, 23:59:58,
as the transaction itself causes 23:59:59 to be skippedaltogether.  (Kinda hard to skip an event if you're in it.)

The "Leap" line references the second that should be removed, so 23:59:59 (stationary) is correct.