We need to provide more details about what is going on at build time so that users can register additional processing if necessary. In particular, being able to register processors that infer native configurations for custom use case is important.