*please* *please* switch back to the posix conformant way. The leap seconds support currently breaks kerberos V5 (or at least preaks the test suite -- I *think* it breaks the protocol for real, I haven't had time to look closer.) *NO* other platform uses the "correct" leap seconds; as I put in my bug report, NetBSD tried for a while a couple of years ago and changed it back as well (since, really, we're a Unix! POSIX is a higher authority than IERS, for us...)