Re: [gpsd-dev] Please review Miroslav's attempted fix.

[Eric S. Raymond]

Miroslav Lichvar <address@hidden>:
> On Fri, Aug 22, 2014 at 06:14:40AM -0400, Eric S. Raymond wrote:
> > Miroslav Lichvar <address@hidden>:
> > > With the patch, do you see any "PPS: no current GPS seconds:"
> > > messages? It would be useful to see the full debug output.
> > 
> > I don't know whether that message comes up from gpsd, because I noticed the
> > bug and verified the fix with gpsmon. With your patch in place there were
> > no PPS-event bars in the gpsmon display; when I reverted it, they were
> > restored.
> 
> The PPS samples are probably just failing in one of the checks, the
> debug output would show that.

It seems not.  Re-applying the patch, I see this sort of message:

gpsd:INFO: PPS hooks called with accepted 1408732340.000434890 offset 
-0.000434890
gpsd:PROG: PPS edge accepted 1408732340.000434890 offset -0.000434890
gpsd:PROG: PPS edge: 1, cycle: 1000028 uSec, duration:   99644 uSec @ 
1408732340.100079685

So it may be that the breakage is confined to gpsmon. Now checking this by
telnet...yes, I get PPS notifications in the JSON.  A representative one:

{"class":"PPS","device":"/dev/ttyUSB0","real_sec":1408732721, 
"real_nsec":0,"clock_sec":1408732720,"clock_nsec":999753418}

> > Do you have a PPS-capable device you can test with?
> 
> I have a Garmin 18x and it works for me. It sends the NMEA messages
> about 0.5 seconds after the pulse. I suspect your GPS may be sending
> them too close to the pulse and the pulses are rejected as too old
> or from the future.

I now think it's just the gpsmon part of the patch that's slightly wrong.
If that can be fixed, we can re-integrate it.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>