Automatic Testing of Program Slicers

Abstract

Program slicing is a technique to extract the part of a program (the slice) that influences or is influenced by a set of variables at a given point (the slicing criterion). Computing minimal slices is undecidable in the general case, and obtaining the minimal slice of a given program is normally computationally prohibitive even for very small programs. Therefore, no matter what program slicer we use, in general, we cannot be sure that our slices are minimal. This is probably the fundamental reason why no benchmark collection of minimal program slices exists. In this work, we present a method to automatically produce quasi-minimal slices. Using our method, we have produced a suite of quasi-minimal slices for Erlang that we have later manually proved they are minimal. We explain the process of constructing the suite, the methodology and tools that were used, and the results obtained. The suite comes with a collection of Erlang benchmarks together with different slicing criteria and the associated minimal slices.

Analysis Techniques for Software Maintenance

“…The chapter presents the notion of quasi-minimal slice and it describes a methodology to compute them and to use them to compare the performance of different program slicers. Most of the content of this chapter has been extracted from [152].…”

“…In order to compare both implementations we have chosen Bencher [152], a program slicing benchmark suite for Erlang. It contains 18 Erlang programs, in which a variety of structures and program constructs can be found.…”

“…Moreover, in contrast to them, where only variable nodes can be used as slicing criteria, in our work (and also previously in [188]), any AST node can be the slicing criterion. This increases the slicing capabilities and is particularly useful, e.g., for testing program slicers [152].…”

“…In many cases, deleting only one line/node results in a syntax error; and several lines/nodes must be deleted to make the slice compilable. The combinatorial explosion of deleting sets of lines/nodes can significantly increase the slicing time due to its exponential cost [152].…”

“…• Chapter 6 exposed a methodology to compute quasi-minimal slices, which are minimal slices for a specific set of program inputs [152]. The chapter described a model that produces quasi-minimal slices by (optionally) using a group of static program slicers, and implementing the tree-ORBS slicing algorithm and a set of automatically generated test cases that attempt to maximise branch coverage.…”