Adding Contextual Guidance to the Automated Search for Probabilistic Test Profiles - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2014

Adding Contextual Guidance to the Automated Search for Probabilistic Test Profiles

Résumé

Statistical testing is a probabilistic approach to test data generation that has been demonstrated to be very effective at revealing faults. Its premise is to compensate for the imperfect connection between coverage criteria and the faults to be revealed by exercising each coverage element several times with different random data. The cornerstone of the approach is the often complex task of determining a suitable input profile, and recent work has shown that automated metaheuristic search can be a practical method of synthesising such profiles. The starting point of this paper is the hypothesis that, for some software, the existing grammar-based representation used by the search algorithm fails to capture important relationships between input arguments and this can limit the fault-revealing power of the synthesised profiles. We provide evidence in support of this hypothesis, and propose a solution in which the user provides some basic contextual knowledge to guide the search. Empirical results for two case studies are promising: knowledge gained by a very straightforward review of the software-under-test is sufficient to dramatically increase the efficacy of the profiles synthesised by search.
Fichier principal
Vignette du fichier
ICST2014.pdf (458.67 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01953483 , version 1 (13-12-2018)

Identifiants

  • HAL Id : hal-01953483 , version 1

Citer

Hélène Waeselynck, Simon Poulding. Adding Contextual Guidance to the Automated Search for Probabilistic Test Profiles. 7th IEEE International Conference on Software Testing, Verification and Validation (ICST 2014), Mar 2014, Cleveland, United States. ⟨hal-01953483⟩
27 Consultations
73 Téléchargements

Partager

Gmail Facebook X LinkedIn More