The construction of contextual def-use associations for object-oriented systems

Abstract

This paper describes a program representation and algorithms for realizing a novel structural testing methodology that not only focuses on addressing the complex features of object-oriented languages, but also incorporates the structure of object-oriented software into the approach. The testing methodology is based on the construction of contextual defuse associations, which provide context to each definition and use of an object. Testing based on contextual defuse associations can provide increased test coverage by identifying multiple unique contextual defuse associations for the same context-free association. Such a testing methodology promotes more thorough and focused testing of the manipulation of objects in object-oriented programs. This paper presents a technique for the construction of contextual defuse associations, as well as detailed examples illustrating their construction, an analysis of the cost of constructing contextual defuse associations with this approach, and a description of a prototype testing tool that shows how the theoretical contributions of this work can be useful for structural test coverage.

Dynamic data-flow testing

“…We can model data dependencies that arise from the nested state of classes extending definitions and uses with the information on the context of the method invocations, where the context of methods invocation is the chain of (possibly nested) method invocations that, from the class, leads to the definition or the use of the class state vari- Contextual data flow information is computed with an inter-procedural data flow analysis that propagate data flow equations over multi-class extension of the class control flow graph, which models the control flow between methods of different classes [SP03,DGP08]. The construction of the inter-procedural graph on objectoriented systems, which extensively use virtual call invocations instead of static ones (caused by polymorphism and in general dynamic binding), presents well-known challenges and impact on the applicability and on the precision of the analysis.…”

“…Control-flow criteria and intra class data flow testing techniques do not identify any element that characterize this complex interactions, being limited to control structures and intra-class interactions, respectively. Souter and Pollock, and Denaro et al [SP03,DGP08] have formalized the core object-oriented data flow testing concepts. Souter and Pollock noticed that to properly capture the inter procedural control flow relations between methods and classes, it is necessary to distinguish definitions and uses according to their invocation context, formalised the concept of contextual definition-use pair reported in Section 2.2.2, and proposed to use contextual definition use pairs to define thorough data flow testing techniques for object-oriented systems.…”

“…In particular, conservative implementations of data flow analysis both compute too many test objectives to be practical, and include infeasible test objectives in the computation, i.e., infeasible paths from definitions to uses of the same variable. Therefore, existing data flow testing techniques use strong approximations in their analyses to scale to large complex systems, for example excluding inter-procedural data relations and aliasing [HR94,SP03,DGP08,VMGF13].…”

Masked types for sound object initialization

“…However, Java cannot safely prevent reading from uninitialized fields. There has been work on interprocedural def-use analysis in the context of object-oriented languages [30,29], with varying cost and precision. This prior work detects initialization bugs on fields, but requires non-modular whole-program def-use analyses and is subject to the typically limited accuracy of whole-program alias/points-to analyses.…”

An Empirical Study of the Use of Frankl-Weyuker Data Flow Testing Criteria to Test BPEL Web Services

“…Souter and Pollack [24] propose a way to construct contextual def-use associations, in which each definition and use of an object is associated with contexts. The effect of environment information in pervasive contextaware software is studied in [16].…”