Constraint-Based Testing (CBT) is the process of generating test cases against a testing objective by using constraint solving techniques. In CBT, testing objectives are given under the form of properties to be satisfied by program's input/output. Whenever the program or the properties contain disjunctions or multiplications between variables, CBT faces the problem of solving non-linear constraint systems. Currently, existing CBT tools tackle this problem by exploiting a finite-domains constraint solver. But, solving a non-linear constraint system over finite domains is NP hard and CBT tools fail to handle properly most properties to be tested. In this paper, we present a CBT approach where a finite domain constraint solver is enhanced by Dynamic Linear Relaxations (DLRs). DLRs are based on linear abstractions derived during the constraint solving process. They dramatically increase the solving capabilities of the solver in the presence of non-linear constraints without compromising the completeness or soundness of the overall CBT process. We implemented DLRs within the CBT tool TAUPO that generates test data for programs written in C. The approach has been validated on difficult non-linear properties over a few (academic) C programs.

Automatic test data generation leads to the identification of input values on which a selected path or a selected branch is executed within a program (path-oriented vs. goal-oriented methods). In both cases, several approaches based on constraint solving exist, but in the presence of pointer variables only path-oriented methods have been proposed. This paper proposes to extend an existing goal-oriented test data generation technique to deal with multi-level pointer variables. The approach exploits the results of an intraprocedural flow-sensitive points-to analysis to automatically generate goal-oriented test data at the unit testing level. Implementation is in progress and a few examples are presented.

The current trend in debugging and testing is to cross-check information collected during several executions. Jones et al., for example, propose to use the instruction coverage of passing and failing runs in order to visualize suspicious statements. This seems promising but lacks a formal justification. In this paper, we show that the method of Jones et al. can be re-interpreted as a data mining procedure. More particularly, they define an indicator which characterizes association rules between data. With this formal framework we are able to explain intrinsic limitations of the above indicator.

Constraint-Based Test data generation (CBT) exploits constraint satisfaction techniques to generate test data able to kill a given mutant or to reach a selected branch in a program. When pointer variables are present in the program, aliasing problems may arise and may lead to the failure of current CBT approaches. In our work, we propose an overall CBT method that exploits the results of an intraprocedural points-to analysis and provides two specific constraint combinators for automatically generating test data able to reach a selected branch. Our approach correctly handles multi-levels stack-directed pointers that are mainly used in real-time control systems. The method has been fully implemented in the test data generation tool INKA and first experiences in applying it to a variety of existing programs tend to show the interest of the approach.