Improve diagnostics when objects are disposed too early

[rolfbjarne]

When a managed wrapper object is Disposed, and then native code later re-exposes the same native handle to managed code, this happens:

Failed to marshal the Objective-C object 0xdeadf00d (type: MyCustomType). Could not find an existing managed instance for this object, nor was it possible to create a new managed instance (because the type 'MyCustomType' does not have a constructor that takes one NativeHandle argument).

I wonder if we could improve diagnostics for this scenario by:

• Setting a flag on in the the NSObject.flags field to indicate a disposed object.
• Keep any disposed objects in the Runtime dictionary of intptr->wrapper mappings, but return a special lookup failure if the object is there, but disposed.
• Show a different error message in this case.
• We could also make it a build-time option to capture a stack trace when Dispose is called too early, and show that stack trace later on in the error message.