FreeBSD handles leapsecond correctly

Matthias Andree matthias.andree at
Mon Jan 2 13:53:07 PST 2006

On Mon, 02 Jan 2006, Poul-Henning Kamp wrote:

> And then, feel free come back and tell me and the rest of the world
> how easy this problem can be solved correctly.

I made no statement to the extent this problem were _easy_ to solve, and
I am aware that any "solution" goes beyond FreeBSD's scope.

> I assumed this to be self-evident, but I guess I need to lower
> the bar in future emails.

The actual point was that the definition of "correct" depends on the
scope, and that drifted through the discussion.

"Correct in POSIX:" yes, "correct" as "doing the right thing[tm] for all
fields of application": clearly no.

"Handles leapsecond according to IEEE Std 1003.1-2001" (or whichever you
chose) avoids the need to disambiguate.

BTW, thanks for providing evidence that your system was made aware of
the leap second, and processed it in accordance with POSIX -- this is a
data point for "leap seconds do work".

Matthias Andree

More information about the freebsd-current mailing list