[gpsd-dev] bug: Two different NMEA time stamps per second

[Hal Murray]

address@hidden said:
> Hmm, it just occured to me, it is your SHM0 that is sending 2 updates a
> second?  Was it the SHM0?  If so, that is not a problem, as long as the time
> stamps look right, it just means the NMEA is reporting 2x a second.  The
> only thing we care about is the SHM1. 

I think whatever gpsd is doing is currently buggy.  At least relative to what 
ntpd expects.

With the default Venus setup, it sends GPGGA and GPRMC (and others).  Both 
have time and gpsmon shows two time updates.

The problem is that you are giving two different offsets per second so which 
one ntpd sees depends on when in the second it happens to poll the SHM 
segment.  I don't care which one you use.  I can work with either one.  They 
both need fudging.

You can use both as long as they give the same answer.  It might be 
interesting to see if both point to the same place if you subtract off the 
delay due to the NMEA sentences you have seen already this burst.

This may all be tangled up with the start-too-early problem.   When I look at 
an Adafruit HAT or a GR-601, both send GPGGA and GPRMC but I only see 1 time 
update per second on gpsmon.  (Both have a .00 seconds fractional part.)



-- 
These are my opinions.  I hate spam.