Dynamic slicing of aspect-oriented programs

Abstract

Program slicing is a decomposition technique which has many applications in various software engineering activities such as program debugging, testing, maintenance etc. Aspect-oriented programming (AOP) is a new programming paradigm that enables modular implementation of cross-cutting concerns such as exception handling, security, synchronization, logging etc. The unique features of AOP such as join-point, advice, aspect, introduction etc. pose difficulties for slicing of AOPs. We propose a dynamic slicing algorithm for aspect-oriented programs. Our algorithm uses a dependence-based representation called Dynamic Aspect-Oriented Dependence Graph (DADG) as the intermediate program representation. The DADG is an arc-classified digraph which represents various dynamic dependences between the statements of the aspect-oriented program. We have used a trace file to store the execution history of the program. We have developed a tool called Dynamic Depenedence Slicing Tool (DDST) to implement our algorithm. We have tested our algorithm on many programs for 40-50 runs. The resulting dynamic slice is precise as we create a node in the DADG for each occurrence of the statement in the execution trace.