Change search
ReferencesLink to record
Permanent link

Direct link
Using mutation to design tests for aspect-oriented models
University of Skövde, School of Informatics. University of Skövde, The Informatics Research Centre. (Distributed Real-Time Systems (DRTS))
George Mason University, Fairfax VA, USA.
Swedish Institute of Computer Science, Kista, Sweden.
University of Skövde, School of Informatics. University of Skövde, The Informatics Research Centre. (Distributed Real-Time Systems (DRTS))
Show others and affiliations
2017 (English)In: Information and Software Technology, ISSN 0950-5849, E-ISSN 1873-6025, Vol. 81, 112-130 p.Article in journal (Refereed) PublishedText
Abstract [en]

Abstract Context: Testing for properties such as robustness or security is complicated because their concerns are often repeated in many locations and muddled with the normal code. Such “cross-cutting concerns” include things like interrupt events, exception handling, and security protocols. Aspect-oriented (AO) modeling allows developers to model the cross-cutting behavior independently of the normal behavior, thus supporting model-based testing of cross-cutting concerns. However, mutation operators defined for AO programs (source code) are usually not applicable to AO models (AOMs) and operators defined for models do not target the AO features. Objective: We present a method to design abstract tests at the aspect-oriented model level. We define mutation operators for aspect-oriented models and evaluate the generated mutants for an example system. Method: AOMs are mutated with novel operators that specifically target the AO modeling features. Test traces killing these mutant models are then generated. The generated and selected traces are abstract tests that can be transformed to concrete black-box tests and run on the implementation level, to evaluate the behavior of the woven cross-cutting concerns (combined aspect and base models). Results: This paper is a significant extension of our paper at Mutation 2015. We present a complete fault model, additional mutation operators, and a thorough analysis of the mutants generated for an example system. Conclusions: The analysis shows that some mutants are stillborn (syntactically illegal) but none is equivalent (exhibiting the same behavior as the original model). Additionally, our AOM-specific mutation operators can be combined with pre-existing operators to mutate code or models without any overlap.

Place, publisher, year, edition, pages
Elsevier, 2017. Vol. 81, 112-130 p.
Keyword [en]
Model-based testing, Aspect-oriented model, Mutation testing
National Category
Computer Science
URN: urn:nbn:se:his:diva-12767DOI: 10.1016/j.infsof.2016.04.007ISI: 000387634200009ScopusID: 2-s2.0-84963813590OAI: oai:DiVA.org:his-12767DiVA: diva2:951614
Available from: 2016-08-09 Created: 2016-08-09 Last updated: 2016-12-01Bibliographically approved

Open Access in DiVA

No full text

Other links

Publisher's full textScopus

Search in DiVA

By author/editor
Lindström, BirgittaAndler, Sten F.
By organisation
School of InformaticsThe Informatics Research Centre
In the same journal
Information and Software Technology
Computer Science

Search outside of DiVA

GoogleGoogle Scholar

Altmetric score

Total: 214 hits
ReferencesLink to record
Permanent link

Direct link