Re: [gpsd-users] GL* strings and gpspipe

[Gary E. Miller]

Yo John!

On Tue, 15 Jan 2019 15:24:44 +0000
John Klug <address@hidden> wrote:

> It is not pointless to do both gpspipe and gpsd, as we have
> potentially multiple applications using gpsd, including ntpd.  We
> would prefer our customers use gpsd, as it provides the most
> flexibility.  gpspipe is a stop-gap.

I'm not sure you got the drift of my argument.  Of course gpspipe and
gpsd are wonderful, so use them as intended.

> The simultated gps strings from gpspipe do not seem to include all
> the strings written natively by the U-Blox chip when it is writing
> NMEA,

An impossibility.  Either gpsd is creating simulated NMEA, or passing
on the GPS NMEA, but not both at the same time.  If you want the
u-blox NMEA, then turn on the NMEA.

> and we are worried about compatibility for existing customers
> when we add gpsd to their system.

Good luck with that, gpsd is its own thing, unlike any other thing.
And since we have no idea what you are replacing it is hard to
be accomodating.

>  We will try to make it easy for
> them to shut off gpsd and run without gpsd, but they will lose other
> functionality, which they never had before anyway.

Yes, let gpsd be gpsd.

> The patch for leap second for the u-blox driver and ntpd I posted
> some time ago on this mailing list.
> 
> http://git.multitech.net/cgi-bin/cgit.cgi/meta-mlinux.git/tree/recipes-navigation/gpsd/gpsd-3.16/0006-ubxtimelps.patch

Sorry, I must have missed or dropped it when it was sent.  A lot of changes
since 3.16.

Some things wrong with that patch.  At a minimum the #ifdef has to
go: #ifdef UBLOXTIMELS_ENABLE

gpsd is going to be slowly removing as many #ifdef as possible, Those
made send in 1996 to save code and compute cycles, but now adds
complexity and makes many things harder.

The code in ubx_msg_nav_timels() has merit., but needs work.

For example:

+    if(valid_time_to_ls_event) {
+      char *src="Unknown";

We know that some compilers will not tolerate defining variables like that.

One of the other additions, UBX_RXM_SFRBX, is already in gpsd.

That code added another initialization string, but the gpsd u-blox init
strings are already too long and not reliable.  So other work needs
to be done before adding that new init.

Also the code formatting does not comply to the gpsd code standards.

If you can submit just the ubx_msg_nav_timels() change first, based on
git head, then I would be interested in applying that.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
        address@hidden  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can’t measure it, you can’t improve it." - Lord Kelvin

pgpMkY29Iv3Xd.pgp
Description: OpenPGP digital signature