[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gpsd-dev] Where to put tests for: $GPVTG

From: Gary E. Miller
Subject: Re: [gpsd-dev] Where to put tests for: $GPVTG
Date: Sun, 29 Apr 2018 12:37:55 -0700

Yo teyrana!

On Sat, 28 Apr 2018 23:45:14 -0400
teyrana <address@hidden> wrote:

> (in order vvvvv )
> > Use the built in regression scheme.  Add a GPS capture file in
> > tests/daemon that has the VTG commands.
> >  
> > >   a) modify every device-specific test case which contains VTG
> > > messages? b) emplace a specific test somewhere else, that's not
> > > obvious to me...  
> >
> > Nope just add a new test file.  
> Okay, *before* I add a new test-file, a whole bunch of existing tests
> fail.  (because they contain previously-ignored VTG messages).

And that is a GOOD THING!

> So this leaves us with a few options:
> a) perhaps the patch's existing behavior is flawed, and it's output
> *should* match those existing .chk files


> b) the existing behavior for VTG is incomplete, and we merely need to
> update the *.log.chk files.


> c) update both existing and new .chk files?


If you send in your patch, and new log file, I'll check it out.  If your
patch looks good, I'll take care of updating the existing regressions.

This sort of thing is not uncommon.

> I've run the `scons check` target, with the draft patch, and while it
> completes (at least) there are numerous failures.

As expected from any patch that gets more, or better, data from old
log files.

> From your comments, it sounds like either (b) or (c). *ALSO* we
> probably need to adjust the behavior of what VTG messages affect, but
> I think we'll have to adjust existing .log.chk files. (Unless I
> misunderstand)

Two different issues.  First we validate you are extracting good
data from VTG and sending it along in a good way.  Second, use the
existing logs to validate the improvement.

My guess is the new VTG data goes into the exsting TPV JSON.

> > > 2. I notice you guys also don't have a test *.log.chk for our
> > > brand (Hemisphere). Would it be worth while to add it?  
> >
> > Yup, more tests the better. Put the new raw data in tests/daemon,
> > then use regres-driver -b to create the new check file.
> >
> >  
> Awesome! I'll include it, then.

Looking forward to it.

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

Attachment: pgpRVU6XcMHRU.pgp
Description: OpenPGP digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]