Thèse soutenue

Langages applicatifs et machines abstraites pour la couverture de code structurelle
FR
Accès à la thèse
Auteur / Autrice : Philippe Wang
Direction : Emmanuel Chailloux
Type : Thèse de doctorat
Discipline(s) : Informatique
Date : Soutenance en 2012
Etablissement(s) : Paris 6

Résumé

FR  |  
EN

This thesis presents a study on structural code coverage fora language of the ML family, in response to an industrial need insafety-critical software domain to develop tools. In this context, MLappears as a particularly rich and high-level language with a highdegree of expressiveness. Its use is a progress but also raises issueswhen trying to apply classical safety-critical software engineeringprocesses. In particular, the two notions of condition and decision,as well as coverage criteria associated with them, rapidly become verycomplex. The first contribution of this thesis answers the question ofwhat conditions and decisions mean for a language of the ML family, bygiving several formal definitions. Then, we present a formalisedtechnique for structural code coverage which rewrites the source codeto produce traces at run-time. We name it the intrusiveinstrumentation. We also formalise another technique which does notrewrite the source code, which allows to use the same binary for bothtesting activities and production. This second technique is called nonintrusive and consists in generating at compile-time the informationneeded to match the machine code back to the source code. Otherinformation are also generated for the execution environment to recordspecific traces that we need to generate a coverage report involvingBoolean measures. Finally, we compare these two techniques bothformally and practically, but also in terms of implementation.