Combining llRot2 Fwd/Left/Up

Bear with me on this description. I know exactly what I want, but I tend to confuse...

Basically, I have a child prim who's position and rotation need to change at once, and local to the position/rotation of the object itself. I can do position movement and rotation at once using llSetPrimitiveParams() with PRIM_POSITION and PRIM_ROTATION.

However, to make each of those changes local, I've had to do some messy things. Anyways I got rotations to work just fine, but using the llRot2 Fwd/Left/Up functions in combination for positioning is giving me the hardest time. When I used just llRot2Up on a test, it worked well. When I add the other two, my results seemed random. Perhaps I shouldn't be adding them? I don't know.. but I know if I want my final product unbreakable and rotated as a whole while still able to function.. I need to use these llRot2 functions.

Please, I know I could settle for simple local position movements but I want the best. I can do those simple movement on my own, but I know if the object gets rotated certain ways, it will break. I just need help with llRot2 functions in combination. Thank you very much.

Hmm, I'm not sure exactly what you're trying to do (code snippets might help!)

But won't llSetRot(llEuler2Rot(<x,y,z>) suffice? I find that works for most of my needs when trying to rotate a certain amount on each axis.

I find it much easier to use the vector * rotation operator when I want to specify arbitrary vectors that are relative to the local (rotated) axis of the object.

llRot2Fwd(llGetRot()) is equivelant to <1, 0, 0> * llGetRot(). Im not exactly sure what you want to do though - if you want the object to move so that its position is +1 meter on every axis, try <1, 1, 1> * llGetRot(). Or if you simply want the object to go in that direction but only travel one meter in total use llVecNorm(<1, 1, 1> * llGetRot().
==Chris

Okay,

            llSetPrimitiveParams([
                        PRIM_ROTATION, llEuler2Rot(<270,67.5,292.5>*DEG_TO_RAD)/( (ZERO_ROTATION / llGetLocalRot()) * llGetRot()),
                        PRIM_POSITION, (llRot2Fwd(llGetLocalRot())*(-0.03))+(llRot2Left(llGetLocalRot())*(-0.013))+(llRot2Up(llGetLocalRot())*0.158)
            ]);

            llSleep(3);

            llSetPrimitiveParams([
                        PRIM_ROTATION, llEuler2Rot(<180,0,0>*DEG_TO_RAD)/( (ZERO_ROTATION / llGetLocalRot()) * llGetRot()),
                        PRIM_POSITION, (llRot2Fwd(llGetLocalRot())*0)+(llRot2Left(llGetLocalRot())*0)+(llRot2Up(llGetLocalRot())*0.115)
            ]);

im multiplying the llRot2 functions by offsets from the root prim. The angles are also offsets from the root prim. That's what seemed logical to me, but I know I did something wrong.

This is what I'm rotating:

http://pegasus.cc.ucf.edu/~an162020/rot_slices.gif

Thats three pics of the same movement, to give a better idea. Each pizza slice shape there is the same prim at each end position it would be after a rotation. This is a tapered, flat, and cut cylinder. So, the center of it is the tip of it. But I want it to seem like its rotating about the middle of its curved edge. I've had this model for a long time, unlinked. It's a total mess full of listeners and hinges. It broke easily as well. So.. I hope that makes things more clear.

I changed all the llRot2 functions to a <x,y,z>*llGetLocalRot() but i get exactly the same problem. : /

Dividing my displacement vectors like this:

(<x,y,z> / ( (ZERO_ROTATION / llGetLocalRot()) * llGetRot()))

did the trick! I have no idea what the logic is behind that.. looks like there might be another way to write that.. but wow what a relief. I've been on this forever...