Detection of Interferences in Aspect-Oriented Programs Using Executable Assertions

Abstract

Aspect-oriented programming (AOP) is a technique that promotes separation of concerns. Unfortunately, it still suffers from well-known composition issues, in particular from undesirable interferences when multiple concerns are applied at the same join point. In this paper we propose an approach to detect interferences side effect using executable assertions. The assertions are inserted in the aspect chain to detect various types of interferences. The implementation is based on the AIRIA resolver construct, recently introduced to better control conflicting aspects in AspectJ. Resolvers add observation points that were lacking in AspectJ. We propose to take advantage of this to implement automated detection of interferences at execution time. We study the feasibility of this approach and demonstrate it on artificial examples.