Re: TAI and qmail

On Thu, 11 Jul 2002, Paul Jarc wrote:

> Jan Pisacka <[email protected]> wrote:
> > On the other hand, on machine_2 (TAI), the timestamp in the "Received:
> > header is famous 22 seconds ahead of all the other timestamps.
> 
> It seems that qmail-queue contains its own implementation of gmtime()
> (datetime_tai()) which makes no use of $TZ or /etc/localtime - it
> always assumes the system clock does not count leap seconds.

Paul, thank you for the hint. Unfortunately, I am not a c-programmer so it
is quite difficult for me to check the sources. However, if this behaviour
has not been observed before, I suppose that the shortcoming is not in the
original DJB qmail-1.03 version. I am using a patched version by Bruce
Guenter, [email protected], which has an attractive feature for me - it is a
source RPM - see http://untroubled.org/qmail+patches/current/

I glimpsed at the patches but didn't notice anything suspicious. Obviously 
because, as I'v already confessed,  I am not a programmer ;-). 

Maybe I should rather try to bother the people at the Qmail mailing list, 
but they might have no interest in libtai? 

Jan