*BSD News Article 25792


Return to BSD News archive

Newsgroups: comp.os.386bsd.questions
Path: sserve!newshost.anu.edu.au!munnari.oz.au!constellation!paladin.american.edu!howland.reston.ans.net!pipex!uknet!bradford.ac.uk!t.d.g.sandford
From: t.d.g.sandford@bradford.ac.uk (Thomas Sandford)
Subject: Re: ntp, time, timezones
Message-ID: <1994Jan10.150422.26859@info.brad.ac.uk>
X-Newsreader: TIN [version 1.2 PL2]
References: <1994Jan7.160541.17677@allegra.att.com>
Date: Mon, 10 Jan 94 15:04:22 GMT
Lines: 25

Henning G. Schulzrinne (hgs@allegra.att.com) wrote:
: First, I now have (what appears to be) a working copy of xntp, version
: 3 up on netbsd 0.9. (The version on agate is ancient and does not
: support NTPv3, the current spec.) Indications of interest are welcome.

: In testing, I discovered that ctime() translates the second count
: differently on my Sun (SunOS 4.1.1) and NetBSD, with an offset of
: 14 seconds, apparently. E.g.
: On NetBSD:
:  gettimeofday()    time()    ctime()
:  757958070.532501  757958070 Fri Jan  7 15:54:16 1994
: On Sun:
:  757958070.501251  757958070 Fri Jan  7 10:54:30 1994

: This is rather strange, naturally. Anything to do with leap seconds?

Yes - it is the leap seconds that causes this problem. ntp takes account of 
leap seconds internally, so your zoneinfo file should NOT include them.
Unfortunately the default *bsd one does (up to about 1985 :-) ).

There is more information about this in the comp.protocols.time.ntp
newsgroup FAQ.
 
--
Thomas Sandford | t.d.g.sandford@bradford.ac.uk