Stewart Andreason wrote:
Jim Wilson wrote:
There are a couple bugs or at least they were there a week or so ago. One is
just a mapping typo where latitude goes into both latitude and altitude. The
other is under linux the fg-root and fg-scenery parameters don't get saved and
passed on to fgfs (no prefs.set done) unless you hit previous to display
page[0] (the function that processes page[0] saves those strings).
Ah, is this why fgfs get stuck in a freezing loop, when I give latitude and longitude parameters at the command line?
(Tries a few more runs)
Ah! that's it. If I give a --lat that is less than the ground elevation, it freezes in a loop. and ignores the --altitude parameter.
I believe I reported this Jan.11, but had not figured out the exact conditions that triggered it.
Thanks Jim.
A serious bug.
I can explain the bug to you. If you specify a lon/lat that lies on the *exact* border between two tiles (i.e. --lat=90 --lon=45) then at startup the ground intersection code can fail. This means that the scenery subsystem never returns a valid groud elevation. Now the problem is that the flight dynamics model *needs* to know the ground elevation before it can position the aircraft. Complicating the matter is that when the FDM is first initialized the tiled scenery loader may not have the current tile loaded yet. So the FDM doesn't know if it's in a dead lock state or if it just needs to wait a bit longer for the threaded tile loader to catch up.
The solution would be to make the ground intersection code more robust to this boundary condition, but I believe that might be burried deep in plib.
Regards,
Curt.
-- Curtis Olson http://www.flightgear.org/~curt HumanFIRST Program http://www.humanfirst.umn.edu/ FlightGear Project http://www.flightgear.org Unique text: 2f585eeea02e2c79d7b1d8c4963bae2d
_______________________________________________ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
