Asymptotic Inference with Response-Adaptive Treatment Allocation Designs