Re: ✘3.23~rc1 -- gpxlogger

[Fred Wright]

On Tue, 3 Aug 2021, Gary E. Miller wrote:
> On Mon, 2 Aug 2021 20:07:45 -0700 (PDT)
> Fred Wright <fw@fwright.net> wrote:
>
> > On Mon, 2 Aug 2021, Gary E. Miller wrote:
> >
> > > I pushed another fix to the gpxlogger issues.  Turns out this bugs
> > > was there since day one.
> > >
> > > Please test.
> >
> > What I see now is:
> >
> > Running with no args and no gpsd running exits 1 with no message.
>
> Good.
>
> But, the message is in the syslog.  All gpxlogger messages go into the
> syslof.

If it's not in daemon mode, messages should go to the terminal.  Syslog is 
for daemons et al.  And it's questionable that a nonprivileged program 
should be able to write to syslog at all.

A subtlety with that is that "daemon mode" in this context should mean 
actually running as a daemon, not just the command flag.  In particular, 
if it's unable to daemonize, the error to that effect should go to the 
terminal (by default, anyway).

I notice that the manpage says that -d should be accompanied by -f, though 
-f isn't directly documented.  Though it wouldn't be unreasonable for -d 
without -f to go to syslog (if possible).

> > Running with no args and gpsd running with no receivers exits 1 with
> > no message.
>
> Good, also syslog.

Aside from where the message goes, I'd expect the behavior in this case 
should be to output the header and wait, in case a receiver appears later. 
But it should of course be quittable while waiting.

> > Running with no args and gpsd running with good fixes exits 1 with no
> > message.
>
> Odd, works for me.  You do not get a valid gpx file?

Nope.  This is with no args at all.

> > Running against an external gpsd works.
>
> Good.

Fred Wright