Modular Change Impact Analysis for Configurable Software
Florian Angerer, Herbert Prähofer, Paul Grünbacher · 2016
Slicing-based change impact analysis is an important and established technique to assess the effects of modifications to source code. Program slicing is usually done by performing a graph reachability analysis on a system dependence graph representing the control and data flow dependencies between statements. However, analyzing large-scale software systems can lead to performance issues, resulting in huge dependence graphs and long analysis times. In this paper we present an approach that exploits the modularity of large-scale systems to first perform program analysis for individual modules, and later compose the pre-computed analysis results. However, partitioning a system dependence graph is not straightforward, as it carries presence conditions representing variability. We thus use placeholders that are resolved when composing the pre-computed SDG modules during configuration-aware program analysis. Our approach is particularly useful in the context of product lines, when product variants are derived by composing modules depending on specific customer requirements. We present preliminary results of an evaluation based on a product line from an industrial partner.