Alexander Kovelenov

Thomas Fabini wrote:

If that’s something which could be roughly build with puzzles, the type of navigation used in the VR Demo would greatly benefit from this.

Hi Thomas,
Here is how we do it. You can use this snippet as a basis for your own puzzle or with “exec script”:

_handleCollisions() {
    if (!this._enableCollisions || this.collisionMeshes.length == 0)
        return;

    this._raycaster.ray.origin.copy(this.object.position);
    this._raycaster.far = this.storyHeight;

    const intersections = this._raycaster.intersectObjects(this.collisionMeshes, false);
    if (intersections.length > 0) {

        const intersection = intersections[0];
        this.object.position.y = intersection.point.y + this.gazeLevel;
        this._onCollisionMesh = true;

    } else if (this._onCollisionMesh) {

        let minDistance = Infinity;

        for (const mesh of this.collisionMeshes) {
            if (!mesh.geometry.boundsTree) {
                console.warn('v3d.FirstPersonControls: no bounding tree found on ' + mesh.name);
                continue;
            }

            this._toMeshSpaceMatrix.copy(mesh.matrixWorld).invert();

            this._lastIntersection.copy(this.object.position);
            this._lastIntersection.y -= this.gazeLevel;
            this._lastIntersection.applyMatrix4(this._toMeshSpaceMatrix);

            const target = this._ppTarget;
            mesh.geometry.boundsTree.closestPointToPoint(this._lastIntersection, target);

            if (target.distance < minDistance) {
                this._nearestPoint.copy(target.point);
                this._nearestPoint.applyMatrix4(mesh.matrixWorld);
                minDistance = target.distance;
            }
        }

        if (isFinite(minDistance)) {
            this.object.position.copy(this._nearestPoint);
            this.object.position.y += this.gazeLevel;
        } else {
            // NOTE: fallback if no bounding trees present on collision meshes
            this.object.position.copy(this._lastPosition);
        }
    }
}

This new code is from Verge3D preview that uses closestPointToPoint() method to push the camera back to the collision map. You’ll need to modify it to handle the “camera control object” instead.

• This reply was modified 1 week, 4 days ago by Alexander Kovelenov.

Thank you for bringing this up. We’ve already uploaded the fix to WordPress repository

I see no critical errors in the console you provided. I believe there is a logical issue in the puzzles scenario itself. There are some hints that might help in debugging it:

1. Use more print to console to find our the possible reason of this misbehavior.

2. Temporary disable blocks that are not relevant to this particular functionality. This will help you in narrowing it down.

Hi,
There are no specific puzzles required for making it work. The most probable cause of this misbehavior might be specifying a wrong value for the “query selector” puzzle, hence it does not return anything. You might use the print to console puzzle to verify that the proper element is returned by the query selector.

Hi,
Unfortunately, it’s not possible to disable the propagation of scaling through the bone hierarchy in Verge3D. You might try to create additional “scalable” bones to address this problem.

Hi,
Sure, you can modify the source puzzles of this app a to make multiple objects work. Just use some variable as an “active object” and move that instead of directly referencing the object by the name.

Hi, the code in the “exec script” puzzle does not seem to be a valid JavaScript… Are you trying to get some value of an HTML element?