Effects of different flow insensitive points-to analyses on DEF/USE sets

Points-to analysis is required as a preliminary step for many code analyses used in program understanding and maintenance. Different flow insensitive points-to analyses have been proposed in literature. They are extremely appealing with respect to the flow sensitive counterparts because of their high efficiency. Their output consisting of a set of points-to pairs which hold for the given program, can be integrated with other analyses, such as reaching definitions or call graph computation. In any case, their role is to resolve dereference chains into the referenced locations. Two variants of flow insensitive points-to analysis are considered. The resulting points-to pairs are used to determine the locations that are defined or used by every program statement. When a definition or a use exploits a pointer to access the defined or used location, the preliminary computed points-to pairs allow one to find the accessed locations. The effect of different choices of points-to analyses on DEF/USE sets is discussed, together with the consequences on the efficiency of the computation