Game Unit with Multiple Child Objects - "Garage Scene"

Hello ladies & gents,

Let’s say I have a TANK unit (for in-game use) with the following object hierarchy:

Tank Body (GameObject)
|__
   Turret Mount (empty)
   |__
      Turret (GameObject)
      |__
         Gun Mount (empty)
         |__
            Gun (GameObject)

So, this Tank unit has three moving (and visible) parts to it; namely, Tank Body, Turret, and Gun. And I want the 3D representation of this tank shown in a garage-like setting where the player can pick and choose from an array of different parts. Of course, the player’s selection will be immediately be reflected to the 3D tank model on the same screen.

Now, I got it to a point where the tank is showing up just fine. But, choosing a different option for, say, the turret (or any other part of the tank) has proven to be more challenging than I expected.

The main problem is that when I display the newly selected mesh, the mesh from the previous selection remains in place. So, I used the Destroy method to get rid of the previous mesh. But then, I ran into this error - “MissingReferenceException: The object of type ‘GameObject’ has been destroyed but you are still trying to access it.”

And I have been trying to find a workaround to no avail. Is there a cleaner/smarter way to cycle through different meshes and display only one at any given time?

Just to mind you, if I am dealing with one-part mesh, it wouldn’t have been an issue. My problem is rather unique in the way that the Tank unit is structured in a hierarchical manner as already mentioned above.

Any insight/help will be much appreciated!

Well, i would suggest a setup like this:

Tank Body (GameObject)
|__
   Turret Mount (empty)
   |__
   |  Turret (GameObject)
   |__
      Gun Mount (empty)
      |__
         Gun (GameObject)

This way your mount-points don’t get destroyed (or attached) with the exchangeable object. You also shouldn’t move the mounted parts inside their mount-points. Just use the mount-point object as gun or turret. All motion should be applied exclusively to those objects. This way the tank works even without any visual part. This seperates the logical part from the visual part.

You can of course have dependencies to the mounted object. For example the mounted Gun object could have a gun script which handles the actual firing (rate-of-fire, reload, which bullet prefab should be used, …). This script would get exchanged with the gun object (which is probably a prefab).

The pivot of the mountable objects should always match their mount points, so you can just attach the object and set the localPosition and localEulerAngles to (0,0,0).