Scenario-based testing using symbolic animation of B models

摘要

This article presents a model-based test generation technique, from user-defined scenarios, for behavioral models expressed as B machines. Scenarios are expressed using a customized formalism, based on regular expressions, that makes it possible to describe sequences of operation calls possibly reaching specific states of the system. A symbolic animation engine, simulating the execution of a model using constraint logic programming, is then exploited to play the unfolded scenarios on the model and to instantiate the test cases, providing the expected results used to establish the conformance verdict. This approach is tool supported by a research prototype and has been successfully applied in an industrial context of a smart card applet. This tool is extended by a scenario generator, which automatically generates testing strategies for exercising user-defined properties, written using specific patterns. Copyright © 2012 John Wiley & Sons, Ltd.