Assumption-based Reasoning with Constraints for Diagnosing Program Errors

Abstract

Model checking is an efficient technique for detecting errors of a system. However diagnosing program errors is the most time consuming hard work. One of the major advantages of model checking is the production of a counterexample when a property violation is detected. The error trace produced by a model checker may exhibit the symptom related to the cause of errors. Thus, counterexamples can be enough and are indicative for the cause of violation of the property. We present an assumption-based approach to localize the cause of a property violation using reasoning with constraints. The assumption among the statements in counterexample is made to point out which statement(s) is (are) faulty. Some constraints will be introduced from the specifications of the program. Moreover, we transform the counterexample into a propositional logic formula which is then fed to a SAT solver or a theorem prover together with constraints. A calculus of reasoning with these constraints proceeds under certain assumptions. If the result is satisfiable, the assumption is correct, otherwise, the assumption is wrong and a new assumption should be proposed. Some examples are presented to support the applicability and effectiveness of our approach.