llRotLookAt This

[Midi Nova]

 

Another rotation Mystry,

 

The below snipit  I am trying to get to work from different attachment points, namely the Skull and both Foramrs. What I can't figure out is the rotation math needed to get from local atacment to region. I have spent a few days on this so i figured it time to ask the question out there to you rotation experts.

 

How do you get a directional rotation such that it causes  +y of an attachemnt to point at the choosen region position?

 

 

       vector regionPos; // Some Region Position
	   
       vector agentPos = llGetRootPosition(); // Agent Position in region
       rotation agentRot = llGetRootRotation(); // Agent rotation in region    
	   
       vector localPos = llGetLocalPos(); // Attachment Position relative to attachment point	   
       rotation  localRot = llGetLocalRot(); // Attachment Rotation relative to attachment point

       

      
       llRotLookAt( llRotBetween( < 0.0 , 1.0 , 0.0 > , llVecNorm( ( regionPos - agentPos ) / agentRot ) ), 1.0, 0.4 );

 

[ObviousAltIsObvious]

sorry, you can't really do this in LSL, the simulator does not have the necessary information about joint rotations. that is all done on the viewer side.

for more details see http://wiki.secondlife.com/wiki/LlGetRot#Caveats and http://wiki.secondlife.com/wiki/Rotation#Single_or_Root_Prims_vs_Linked_Prims_vs_Attachments

[Midi Nova]

What I posed will work from the pelvis, I know the viewer handles the animation.

the degree of accuracy im looking for is the same as what it does on the pelvis, but just attached to a different attachment point.

[ObviousAltIsObvious]

it will appear to work from the pelvis most of the time, only because many animations do not try to change it. it is not completely reliable. run a spinning, flying, laying animation and you will see the problem. the avatar root point is the only one that wil not be affected by animations at all.