Re: Unexpected logging from gpsd

[Mick Durkin]

Hi Gary,

Yup, it seems ok now. Glad I wasn't going nuts...

You said:-

> Installing into /etc usually requires you do it as root.  In your case
> it still worked:

I was in my normal user directory as I normally work on the code as that user, 
only changing to root
to run the actual install phase. This time however, I ran the whole show as 
root,
but still within my normal user tree. That's why the install worked correctly.

I attach the same format of logs as last time in case they are of any use. I 
also
added checks for any libgp* files. I only searched under /usr, as lib and lib64 
are symlinks
to the /usr tree on this distro.

Thanks for the prompt attention (as always!).

BR

Mick Durkin


On 11/10/2021 19:04, Gary E. Miller wrote:
> Yo Mick!
>
> On Mon, 11 Oct 2021 11:25:12 +0100
> Mick Durkin <mickdurkinuk@gmail.com> wrote:
>
> > I removed gpsd with "scons uninstall" and then deleted the entire
> > gpsd directory.
>
> That will not remove any gpsd installed another way.  That needs manual
> chacking.  Can you look under /lib* and /usr/lib* for any leftover
> gpsd libs?  And /usr/include for leftover header files?
>
> > I then pulled a fresh copy from the repository and did a full "scons
> > && scons check && scons udev-install". This went without any problems
> > that I could see apart from some warnings about missing stuff that
> > seems to be related to graphical things (this is a text-only server).
>
> I note this;
> $ cd gpsdscons && scons check && scons udev-install
>
> Installing into /etc usually requires you do it as root.  In your case
> it still worked:
>
>      mkdir -p /lib/udev/rules.d
>      cp /home/mick.durkin/gpsd/gpsd-3.23.2~dev/gpsd.rules 
> /lib/udev/rules.d/25-gpsd.rules
>
> > I then started gpsd with my normal "gpsd -n /dev/ttyS0" and ran the
> > gpsdebuginfo script, logging these steps to the attached
> > "gpsdebuginfo.log".
>
> Odd, you are on Intel, but a Raspberry Pi option got turned on:
>
>      # Magic Hat enabled.
>
> But, unrelated.
>
> > The bottom of this log also includes a couple of short snapshots of
> > "tail -10 /var/log/messages" which show the output I originally found
> > still appearing.
>
> Yeah, weird.  Just the messages from the PPS thread.  Never seen/heard
> anything like that.
>
> > Below you said "All the time?  I thought you said only for PPS
> > stuff?", what I meant was that I see these PPS logs all the time. I
> > do not see any output from the non-thread parts of gpsd.
>
> Yeah, I see that now.  And "gpsd -n /dev/ttyS0" is running, just as
> you said.
>
> > The thread specific output is something I have not worked with as
> > either it is more recent than the versions of gpsd I have worked with
> > or it never came to my attention before.
>
> It is at least 15 years old.  I still find corners of the code that I
> did not know existed.
>
> > I certainly never set
> > anything deliberately to trigger these thread related logs.
>
> Nor is there anything that you can do to only trigger PPS logs.  OTOH,
> something in your syslog rules could be selecting just the PPS stuff.
> Maybe run "gpsd -nN /dev/ttyS0" to rul that out.
>
> > If there is any other information or logs I can supply, please let me
> > know.
>
> Uh, oh.  Looking at ppsthread_hook() and my recent changes, it looks
> like I moved a redundant check out of gpsd_vlog(), only it was not
> redundant.  Not sure why I can't duplicate your problem, but I see
> my mistake.
>
> Fix pushed to git head, thanks for your help, please test.
>
> 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

gpsdebuginfo.log
Description: Text document

install.log
Description: Text document