Scope ExecObjects with Tokens

The original PrepareFor* methods of ArrayHandle returned an object to be used in the execution environment on a particular device that pointed to data in the array. The pointer to the data was contingent on the state of the ArrayHandle not changing. The assumption was that the calling code would next use the returned execution environment object and would not further change the ArrayHandle until done with the execution environment object.

This assumption is broken if multiple threads are running in the control environment. After one thread has called PrepareFor* and is in the process of using the resulting execution object and another thread attempts to write to or otherwise change the same array.

Because we cannot use the object itself to manage its own resources, we use a proxy object we are calling a Token. The Token object manages the scope of the return execution object. As long as the Token is still in scope, the execution object will remain valid. When the Token is destroyed (or DetachFromAll is called on it), then the execution object is no longer protected.

When a Token is attached to an ArrayHandle to protect an execution object, it's read or write mode is recorded. Multiple Tokens can be attached to read the ArrayHandle at the same time. However, only one Token can be used to write to the ArrayHandle.