maybe if you gave the dts a loscol the normal static shapes type will have that effect too??
I've never tested LOSCol w/a camera that's attached to a vehicle. Is that what LOSCol is for? I want to try that.
Here's my version of Tendon's export.py that I edited: Maybe I did something wrong?
As far as I can tell... probably not? If you have the object centers for the collision shapes all set to 0 0 0 there shouldn't be any problems.
But if you have the object centers set to the bounds center of each seperate object...
bpy.ops.transform.translate(value=dupe.location*-1)
That line will set each object's transform to 0 0 0. So you'd end up with all of the objects stacked on top of each other.
And this line:
textplace.write("\t\tposition = \"" + "%.6f" % object.location[0] + " " + "%.6f" % object.location[1] + " " + "%.6f" % object.location[2] + "\";\n")
Is going to use the position of every 60th object. So the offset will be weird.
Maybe you could try exporting the objects in groups smaller than 60. Or you could make a .blend specifically for testing your version of the script.
Better- before spending all of this time making the exporter work, try exporting a group of the shapes as a single dts manually, to see if the collision is ok.
I tested exporting 2 cubes seperated by 4 units as a single dts. The cube faces that faced each other were invalid, and I could drive into the cubes through most of the faces. :c