llVolumeDetect Wierdness

I have some strange issues with a prim that has been set to true on volume detection. If I collide with it then it raises the appropriate event (collision_start). However, if I touch it while i am collinding with the object, it throws the collision_start event AGAIN. The touch event just has simple llOwnerSay functions. Any idea why this occurs?

It's a bug that I reported over a year ago. I can't remember the specifics but I think it will throw that collision_start if you click anything while standing in it. Also Av's teleporting won't throw collision_end.

It ruined one of my projects way back when.. was a really neat project to, but I've had lots of really great ideas ruined by a bug or just something they left out of lsl..

What other things would cause no colision end or no collision start?

Both of these problems can be worked around, with extra state and timers.

Ok now i am experiencing another bug where the person enteres the volume detect prim and sits down. For some reason when they get up and leave it does not throw a collision_end event.

Oy. Vey.

Hard work for unsatisfactory behavior.

Joy.

- C

Just about any real-time control system has problems with extra and missing input from sensors. LSL seems needlessly sloppy about it, though, given it's not actually simulating real sensors.

What I'd do would be to have the collision_start kick off an llSensorRepeat(), put all the work in sensor()/no_sensor(), and have that turn off when the avatar left the detection zone. That way you're not using CPU time when idle but you're avoiding all the missed state transition and many of the queue issues.