ArrayIndexOutOfBoundsException at GenericVertexController.updateMesh

[raft]

any idea why this happens ?

Code Select Expand

Exception in thread "main" java.lang.ArrayIndexOutOfBoundsException: 8
at com.threed.jpct.GenericVertexController.updateMesh(Unknown Source)
at com.threed.jpct.Mesh.applyVertexController(Unknown Source)
..

[EgonOlsen]

When does this happen? Do you have a test case?

[raft]

found it i guess. it happens when an empty Object3D is created with new constructor:

Code Select Expand

Object3D object = new Object3D(new float[0], new float[0], new int[0], TextureManager.TEXTURE_NOTFOUND);
object.build();

object.getMesh().setVertexController(..);
object.getMesh().applyVertexController();


an empty object created with old constructor throws same exception with different index (1)

Code Select Expand

Object3D object = new Object3D(0);


hope this helps,
r a f t

[EgonOlsen]

I'll look into it. But what's the point of using an IVertexController on an empty object anyway?

[raft]

But what's the point of using an IVertexController on an empty object anyway?

no point, the collada exporter just exported so..

[EgonOlsen]

Try this jar: http://www.jpct.net/download/beta/jpct.jar. The problem was, that the controller implementation didn't require that there is a bounding box but it assumed that at least the space for it is reserved, which isn't true to "empty" objects like these. I hope this is fixed now. This release also fixes the rotate?()-methods.

[raft]

empty objects created with new constructor still throws same exception

Code Select Expand

new Object3D(new float[0], new float[0], new int[0], TextureManager.TEXTURE_NOTFOUND);

[EgonOlsen]

Argh...my test case didn't call build(), which obviously makes a difference. I'll fix this tomorrow.

[EgonOlsen]

I've updated the jar. Please give it a try.

[raft]

it seems ok now, thanks