DIVERSITY is a multi-purpose customizable platform based on symbolic execution. DIVERSITY has been designed for the purpose of managing the diversity of different semantics, but also the diversity of possible analyses based on symbolic execution. In this paper, we show how the input language of DIVERSITY can be used to encode the semantics of UML scenarios which include timing constraints expressed with the VSL language (standardized in the UML profile for embedded systems MARTE). We apply symbolic execution on practical scenarios of a system-on-chip example3 in order to select test behaviors using an advanced exploration strategy implemented in DIVERSITY.
Interaction models describe the exchange of messages between the different components of distributed systems. This paper presents an approach for checking the validity of multi-traces against interaction models. A multi-trace is a collection of traces (sequences of emissions and receptions), each representing a local view of the same global execution of the distributed system. We formally prove our approach, study its complexity, and implement it in a prototype tool.
Modern industrial systems are often large and distributed. Consequently, building the test harness for them can be technically challenging. A compositional approach attempts to overcome this problem by partitioning the system into smaller parts easier to test separately. And in particular, compositionality helps to avoid as much as possible testing the whole monolithic system thanks to mathematical results which relate the global correctness of the system to the correctness of its constituent parts. In this paper, we present a compositionality result for model-based testing in the setting of the conformance relation tioco which is dedicated to timed systems. We show how to exploit this result in practice by extending a previously defined symbolic testing framework.
Interactions are graphical models representing communication flows between actors. Well-known interaction languages include UML Sequence Diagrams or Message Sequence Charts. Even though interactions allow for concise and intuitive specifications, their use remains limited in formal verification, partly because the subsets of formalized languages often lack expressiveness. As many verification methods, such as model-checking or runtime verification, are routinely available for finite automata, we propose a new approach to generate finite automata from an expressive interaction language with operators such as the concurrent region. Our approach leverages an operational semantics to compute derivatives of an interaction and assimilate them to states of a finite automata. In addition, we use term rewriting to merge states on-the-fly so as to obtain small automata without relying on costly a-posteriori minimization techniques.