Re: TAI and qmail

On Sat, Jul 13, 2002 at 04:01:19AM +0200, Jan Pisacka wrote:
> So, to summarize: qmail computes time labels directly from a system time 
> expressed as the number of seconds elapsed from an epoch, without taking 
> care what epoch it is based on. Without patching, concurrent use of 
> clockspeed and qmail is not a good idea. If I am making any logical error, 
> please, correct me.

Yes, qmail isn't dealing with leapseconds. It's not so clear why no one
noticed it earlier.

-Alex

> On Thu, 11 Jul 2002, Paul Jarc wrote:
> 
> > Alex Stevens <[email protected]> wrote:
> > > I can't see a reason for this date more than another, or a reason
> > > why qmail timestamps ignore how the system treats leapseconds.
> > 
> > Well, it ensures that you can interpret the produced string (i.e.,
> > find the original time_t value) without needing to know how $TZ or
> > /etc/localtime might have been set when the string was generated.  But
> > you lose the knowledge of how the rest of the system interprets the
> > same time_t value, so that value is of limited, uh, value.
> > 
> > 
> > paul
> > 
> 

-- 
__________________
[email protected]
------------------

"Yeah, I'm drunk all right, and you're crazy. Tomorrow I'll
be sober, but you'll still be crazy for the rest of your life."
-W.C. Fields