On Friday, April 13, 2007 at 01:25:58 +0059, Jason McIntyre wrote:
>On Thu, Apr 12, 2007 at 10:34:25PM +0200, Maurice Janssen wrote:
>> Hi,
>> 
>> The manpage for rdate(8) uses the -c option in the examples at the
>> bottom (leap second correction), but the given host (ptbtime1.ptb.de)
>> doesn't need this.  In fact, I've never come across a time server that
>> needed -c, but I suppose there are some servers out there that need it.
>> 
>> Anyway, I think it's better to skip the -c option in the examples.
>> 
>> Maurice
>
>why? if you need -c, you have it. if you don't, it won;t do any harm to
>specify it. as i understand it, -nc is a fair combination.

At first, I thought that -c is needed to correct for replies from broken
servers (hence my remark that ptbtime1.ptb.de doesn't need it).
As Matthew pointed out, it is needed when using a zone under
/usr/share/zoneinfo/right/ and using TAI instead of UTC for the system
clock.
Perhaps this can be added to the manpage.  Something like:

-c  Correct leap seconds.  Only needed when the system clock is set to
    TAI instead of the default UTC to avoid steps during leap second
    insertions.  In this case, a timezone under
    /usr/share/zoneinfo/right/ should be used to display the correct
    local time.

I'm not sure about the last sentence.  It has nothing to do with rdate,
but makes the whole thing clearer IMHO.

Maurice

Reply via email to