[ntp:questions] Purpose of a leap second file?
David J Taylor
david-taylor at blueyonder.co.uk.invalid
Thu Feb 16 15:13:41 UTC 2012
"Dave Hart" <hart at ntp.org> wrote in message
news:CAMbSiYDGA1NK8KJLWfg9bDL+ih0CCNDJuJK1-A_GfDJfvbry_Q at mail.gmail.com...
> On Thu, Feb 16, 2012 at 07:25, David J Taylor
> <david-taylor at blueyonder.co.uk.invalid> wrote:
>> Ah, I found the instructions about using the file:
> I've updated that section to give a few more details.
Thanks, more information oftem helps!
>> I had the impression, falsely, from the way people were talking that it
>> an OS thing, and not an NTP thing.
> It's handled by both ntpd and the OS in the common situation of a
> kernel with precision timekeeping extensions. ntpd notifies the
> kernel loop discipline which takes care of implementing the insertion.
OK - that's what I have on the FreeBSD system.
>> Still don't know if Windows supports it.
> Windows doesn't have those extensions. When the kernel loop
> discipline is unavailable or disabled, the daemon loop discipline
> implements insertion by stepping the clock backward one second on
> POSIX systems. On Windows, the clock is run at half speed for two
Yes, but what I don't know if whether the two types of GPS receiver I have
will tell about the leap second, or whether the message from the Internet
servers will be enough. I'll find out in June.
>> I also see that entry format is
>> different for the current entry compared to all previous, so I hope the
>> white-space count isn't important....
> I'm not sure what you're referring to. Here's the prior entry and the
> 3439756800 34 # 1 Jan 2009
> 3550089600 35 # 1 Jul 2012
Well, that's interesting, and thanks for checking. Using two different
editors (Notepad++ and Programmer's File Editor, PFE) I get different
results! PFE shows a consistent tab separated field all the way down the
file, but Notepad++ shows a space before the tab on the last three entries
(2014 & 2015). The tabs are being expanded differently, sigh!
I'd also found a file: leap-seconds.3535228800 which didn't have those
future, commented out, entries.
> You can verify the leapsecond file is loaded by seeing tai=34 among
> the output of ntpq -crv
> Dave Hart
I will check that shortly. Yes, looking good:
tai=34, leapsec=201207010000, expire=201212280000
More information about the questions