cvs commit: src/usr.bin/vmstat vmstat.c src/usr.bin/w w.c

M. Warner Losh imp at bsdimp.com
Fri Oct 21 09:03:58 PDT 2005


In message: <30677.1129909322 at critter.freebsd.dk>
            "Poul-Henning Kamp" <phk at phk.freebsd.dk> writes:
: In message <20051021.093234.116607170.imp at bsdimp.com>, "M. Warner Losh" writes:
: 
: >It is a problem.  If I boot a system today, the authors of the
: >software still cannot know the example that I gave.  Since there's no
: >leap second table by default, the system may get the answer wrong.
: >That's what is so evil about leap seconds.  You can't plan more than 6
: >months into the future.
: 
: It's worse than 6 months, just because IERS says there will or wont
: be a leap second in 6 months time doesn't mean that they can't change
: their mind later on, technically it is only one month ahead.
: 
: Fortunately, this is not relevant with the current geophysics.

True.  The international standard on leap seconds does say they can
happen at the end of any month, with June/Dec being primary and
Mar/Sept being secondary.  If current models are accurate, we'll need
more than two leap seconds a year in the next century or so.  We'll
need more than 12 a century or so after that.  And more than one a day
in a few thousand years.  Clearly a better way to sychronize time will
be needed at some point.  I vote steering the earth so that its
rotation is restored to its true length :-)

Still haven't mentioned that deep and abiding love for leap seconds...

Warner


More information about the cvs-src mailing list