Re: [gpsd-dev] Leap Second - end of June

[Eric S. Raymond]

Hal Murray <address@hidden>:
> 
> IERS has announced that there will be a leap second at the end of June.
> 
> GPS satellites started announcing it around the start of Jan 22 UTC.  (Or at 
> least that's when I first saw it.  But who knows what sort of filtering or 
> voting is going on inside the receivers.)
> 
> The leap-second pending info seems to confuse some GPS receivers.  They 
> report times that are early by a second.  I have log data from one that was 
> early for a couple of days and then recovered.
> 
> Please keep your eyes open for glitches like that and/or don't bang your head 
> against the wall trying to find the bug in your code if your GPS receiver has 
> that bug.  We should probably collect a list.

Thanks for the heads-up; that was useful.

This means we need to be prepared to do a release on July 1st.
 
> Eric:  Would gpsd get confused if it was using 2 receivers and they differed 
> by a second?

Not normally.  Time handling is independent per-device except in the following
circumstances:

(1) Two devices are reporting 4-digit years but a third is not.  When
a device reports 4-digit years, the century is copied into the global
session context.  In theory, therefore, it is possible during the two
seconds around a century change that the first two devices could
report different centuries and time from the third one could get 
disambiguated to the wrong one.

(2) There's an analogous potential problem in the two seconds around
a leap-second-offset change.

> There is another screwup that has been observed.  The HP GPSDOs say the leap 
> will happen at the end of March rather than the end of June.

Bletch...
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>