Wrong rotation in getAbsoluteTransformation()

[nathanf534]

Code: Select all

ISceneNode* laser=smgr->addMeshSceneNode(laserMesh,0,0,location+direction*offset,rotation,vector3df(1,1,10),false);

matrix4 mat = matrix4(laser->getAbsoluteTransformation());
vector3df r1=mat.getRotationDegrees();
cout<<r1.X<<","<<r1.Y<<","<<r1.Z<<endl;
mat.setRotationDegrees(laser->getRotation());
vector3df r2=mat.getRotationDegrees();
cout<<r2.X<<","<<r2.Y<<","<<r2.Z<<endl;

output:

Code: Select all

2.39263,0.489813,0.156511
22.6769,0.489813,0.156511

Unless I don't understand something, both outputs should be the same.
I'm using Irrlicht 1.6 due to some rendering issues in 1.7 that I can't find the source of. If this was fixed in 1.7, then sorry.

[Eigen]

You should call updateAbsolutePosition() on the node before getAbsoluteTransformation(). Also, why the new keyword for the matrix?

[nathanf534]

I already tried updateAbsolutePosition(), it doesn't change anything, and that only updates the position, I don't believe it does anything to the rotation.
And I just tried the new keyword to see if it would do anything different, Ive tried it without it too.

edit: I edited the first post so it doesn't use new anymore.

[Eigen]

Even without the updateAbsolutePosition(), the code you provided works for me. I'm using 1.7.1. Hmm ..

Wasn't there a bug with rotation of a matrix that had been scaled or something? that was fixed in version 1.7 I think. If you scale your node uniformly (1,1,1) what does it say then?

edit: In the change log I found this under 1.7:
- getRotation now handles matrices with scaling as well
Maybe that was the bug?

[nathanf534]

ok, that was probably it, I was using something with a scale (1,10,1)