Conditional termination of loops over heap-allocated data

Static analysis which takes into account the values of data stored in the heapis considered complex and computationally intractable in practice. Thus, moststatic analyzers do not keep track of object fields nor of array contents, i.e.,they are heap-insensitive. In this article, we propose locality conditions forsoundly tracking heap-allocated data in Java (bytecode) programs, by meansof ghost non-heap allocated variables. This way, heap-insensitive analysis overthe transformed program can infer information on the original heap-allocateddata without sacrificing efficiency. If the locality conditions cannot be provenunconditionally, we seek to generate aliasing preconditions which, when theyhold in the initial state, guarantee the termination of the program. Experimen-tal results show that we greatly improve the accuracy w.r.t. a heap-insensitiveanalysis while the overhead introduced is reasonable.