Sensor range capped by sim border??

Hi all!

I got a strange problem. I live at a sim border, and it appears that an object with a sensor can only "sense" as far as it is away from the sim border.

llSensorRepeat("", NULL_KEY, ACTIVE | PASSIVE, 40, PI, 1);

As you see the range is set for 40 meters.
Now imagine my object rezzed before you. To its left side is the sim border, say, 4 meters away. Now to its right side is another object, which i want to find with the sensor.
This works IF it is not farther away from my sensor than the sim border on the other side.

So... I'd say the radius of the sensor is capped to the distance of the sim border.
Any opinions on this?

Thw Wiki page on this (http://rpgstats.com/wiki/index.php?title=LlSensorRepeat) days that with llSensorRepeat() I can even detect objects in another sim, so I don't see why this should happen.

Thanks,
Stukas

It's been my experience that sim borders are almost like event horizons for sensors, particle targets, and llListen().

I've observed that there's a good 5 second delay between when an object crosses a sim border and when a particle effect targeting that object will find it. also, there is a significant delay between someone saying something across a border and llListen being able to pick it up. This is Easily verifiable if you have two SL clients running in the same room: have each avatar on different sides of the border and have one talk. there will be a good 5-10 seconds between typing the message on one computer and seeing it on the other.

Sensors behave in the same way, there's a large delay between what it's looking for and the data being returned. it has been my experience that fast repeating sensors (1 or 2 seconds) will never pick up an object across the border, because the data being pulled from the sensor is getting to the script after another sensor call is given, thus causing the script to ignore the "I found it" being returned from the sensor. slower sensors (around 10-15 seconds) will usually (not always) be able to return a "I found something" but all information about what it found (Position, rotation, velocity) will most likely be inaccurate if the object was moving in any meaningful way.

So any sensor call that happens to follow too fast invalidates the previous sensor event?
And we don't know how long they will take in advance..
Daaamn that's stupid! *lol

But thanks for the answer, it helped.

- Stukas

I know you are just joking, but it is not stupid. It is asynchonous distributed computing. Guaranteeing order/sequence/delivery/response time in a system that runs on many separate computers is a very difficult problem. There are solutions for specific situations, but not in general, and that's the problem in SL...we don't know ahead of time what people will try to do.

Oh, no it's not stupid that we don't know how long it will take, but it's stupid that the later call invalidates the former event.
That's what I meant

It's not like this for HTTP or Dataserver calls. Not that I can tell anyway.

- Stukas