Re: release?

[Gary E. Miller]

Yo Fred!

On Sun, 4 Dec 2022 16:30:28 -0800 (PST)
Fred Wright <fw@fwright.net> wrote:

> On Wed, 30 Nov 2022, Gary E. Miller wrote:
> 
> > The AIS regression is fixed, and any AIS test failure will now cause
> > "scons check" to fail.
> >
> > Are there any outstadning regressions?
> >
> > There are a few small feature requests pensing::
> >    update RMC psuedonmea to version 4.10
> >    add depth to gpxlogger
> >
> > I also need to:
> >    go over the GitLab issues
> >    check pycodestyle, Codacy and Covertiy for warnings.
> >
> > Anything else on the short list to a gpsd release?  
> 
> There was a regression prior to the last release that was missed.  In 
> prerelease testing, I saw some effectively empty SKY reports go by,

Not really emnpty, intentionnally.

> but didn't look more closely since I was chasing something else.  I
> negelected to test xgps at the time, but later noticed that it had
> become very "flashy" (and not in a good way).  Looking right now, I
> don't see that behavior, though I'm not sure why; it could well be
> timing-related.

Not gonna worry about fixed issues.

> The offending SKY reports seem to be reporting DOP values without 
> reporting any satellites.

Yup.  Because some receivers report DOPs without satellites, and user
still want the DOP data.

Working on one of the last blocker to release that will make this more
so.

> Since DOP values are determined solely by 
> satellite geometry, this seems to be pretty pointless,

The receiver send them, users want them, so they get reported.

Users do not want gpsd to discard data from their receivers.

> and the
> effect, in at least some cases, is that the client clears its
> satellite list in response to the empty report, and then repopulates
> it when it receives the real report, but resulting in an intermittent
> view.  Sometimes the duty cycle is so low that the display is empty
> most of the time.

Yup.

> It's not even necessary to fire up a receiver to see the trouble,
> since the expectation of this behavior is now baked into the
> regression-test data.  I wrote a hack program to check for
> satellite-free SKY reports, and it appears that the culprit is commit
> bddd594aa:

"culprit" is not the word I would use.  I would say "improvement".

> I'll push the tool shortly.

Tool?

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

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

pgpUPYVvGKIBH.pgp
Description: OpenPGP digital signature