A New Approach for Concurrent Program Slicing

Abstract

AbstractRegarding the progress made in model analysis, more complex models, and consequently more complex programs can now be analyzed. However, this remains a difficult task in particular for concurrent programs which induce a lot of combinatory. Another way to reduce this complexity is to use program decomposition. Program decomposition technics extract a part of a given program while preserving the behavior of the original program w.r.t. a specified property. Quasar analyzes concurrent Ada programs, using program slicing as decomposition technic. The program slicer is built using the ASIS tools, that provides syntactic and semantic informations on an Ada source code. These informations can be considered as the “semantic and syntactic graph” mapping an Ada program. This allows to save building the graphs used by traditional program slicing technics and thus to design a simpler and more evolutive algorithm.This paper presents Yasnost, the program slicer used by Quasar, describes the method used to slice concurrent Ada programs and illustrates with two significant examples how concurrent programs analysis can take advantage of program slicing for reducing the analyzed program complexity.KeywordsDependence GraphData DependenceOriginal ProgramConcurrent ProgramTransitivity GraphThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.