A flexible architecture for pointcut-advice language implementations

Current implementations for aspect-oriented programming languages map the aspect-oriented concepts of source programs to object-oriented bytecode. This hinders execution environments with dedicated support for such concepts in applying their optimizations, as they have to recover the original aspect definition from bytecode. To address this representational gap we propose an architecture for implementations of pointcut-advice languages where aspect-oriented concepts are preserved as first-class entities. In this architecture, compilers generate a model of the crosscutting which is executed by virtual machines. In this paper we discuss a meta-model for aspect-oriented concepts and a virtual machine-integrated weaver for the meta-model. As a proof of concept, we also provide an instantiation of the meta-model with the AspectJ language and an AspectJ compiler complying with the proposed architecture. We also discuss how preexisting high-performance optimizations of aspect-oriented concepts benefit from the proposed architecture