• Summary
• Related Documents
  • Active Bibliography
  • Co-citation
• Version History

Abstract:

Program slicing is a potentially beneficial analysis for aiding program understanding. However, slices of even small programs are often too large to be generally useful. Imprecise pointer analyses have been suggested as one cause of this problem. In this paper, we explore using dynamic points-to information as an alternative to static information to improve precision. Since dynamic points-to data reflects the values of pointers during actual execution, the dynamic points-to sets are generally much smaller, and therefore the sizes of the resulting slices should also be smaller. Our experiments showed that the smaller points-to sets generally led to significantly smaller data-flow sets in the analysis. However, because of many direct data dependences in the programs, the improved precision of pointer data did generally not translate into significantly reduced slices. Nevertheless, for programs with many calls through function pointers, the more precise dynamic pointer information produced a significantly smaller call graph, resulting in considerably better slices. The implications of our results are twofold. First, to improve the quality of slices in general, other techniques than improved pointer analysis will be necessary. Second, for applications with an imprecise call graph, an optimistic approximation of the true graph is likely to yield considerable improvement in slice size.

Citations