#89 Realtime tracking

gps support (6)
Fabian Affolter

Direct after enabling the realtime tracking ('Start Realtime Tracking' from a GPS layer) viking crashes. After the crash gpsd is still up and running.

$ gpsctl
gpsctl:SHOUT: /dev/ttyUSB3 identified as SiRF binary at 4800

I'm not sure if the crash is releated to the setting of the GPS receiver. The device is providing SiRF binary data. xgps/tangogps are working properly.


  • Rob Norris
    Rob Norris

    Which version of Viking (and Linux Distribution) are you using?

    How repeatable is this fault?

    I don't think the GPS receivers mode should matter.

    I don't use Viking's realtime tracking mode much, but after just trying it - the latest 1.3 version still works for me.

    If possible can you run Viking under a debugger and report the stack trace when the fault occurs (for all threads).

  • Rob Norris
    Rob Norris

    Oops didn't see the attached log. (NB I run a 64 bit system too.)

    Hmmm, I can crash it by *re-creating* a GPS layer - this may have something to do with the use of realloc...

    I keep you informed of my progress.

  • Rob Norris
    Rob Norris

    I guess you're 'fab' from:

    Since your backtrace crashes in the 'realloc' function this most likely means (I think) Viking has been compiled against GPSD API version 3 (probably GPSD release version 2.39). However the versions of other system libraries seem quite up-to-date.

    Viking is 'For But Not With' GPSD API v3. In that compiling with the old API methods has been maintained (in that that code seemed to work in the past has been kept) yet we (I) don't offer any support for it actually still working, since I refactored it to work with GPSD API v5 (and v4 too).

  • I used the latest release of viking under Fedora. I ran into this problem while updating the package for Fedora. Perhaps this issue occurred due to a mistake in the RPM build process or the update process on my system.

    • status: open --> closed