On Wed, Apr 19, 2017 at 07:18:29AM +0200, Patrik Fältström wrote:
I want as potential fixes which I think are doable:
- an updated POSIX definition where the variable which holds the
number of seconds since the epoch when converting back and forth to
date actually will hold the number of seconds since the epoch.
So we should... brigade the Open Group?
- a slot in the NTP protocol not used today include the number of leap
seconds so people get to know that.
FWIW I am a person that is against smear. I think a clock must be able
to give the correct time and that timestamps should be correct. Either
in seconds (and fractions of seconds) since the epoch or in correct
date and time UTC. Both are increasing all the time. And because of
that smear should not have to happen.
That said, I am not against a BCP explaining the issues due to for
example the broken POSIX specification (which makes it hard to "do the
right thing" in software).