Many software engineering tools such as program slicers must perform data-flow analysis in order to extract necessary information from the program source. These tools typically borrow much of their implementation from optimizing compilers. However, since these tools are expected to analyze programs in their entirety, rather than functions in isolation, the time and space performance of the data-flow analyses are of major concern. We present techniques that reduce the time and space required to perform data-flow analysis of large programs. We have used these techniques to implement an efficient program slicing tool for C programs and have computed slices of programs with more than 100,000 lines of code.

Pre-2018 CSE ID: CS2001-0665

Recently, a few whole-program static slicers for the C programming language have been developed, permitting a variety of hypotheses about time--precision tradeoffs in program analysis for software engineering to be tested. This paper reports an initial investigation into these claims through GrammaTech's CodeSurfer and UCSD's Sprite research prototype, which represent two very different approaches in the program analysis tool design space. First, it was found that algorithmic superiority tended to provide large improvements in relative precision in select cases. Second, a number of non-algorithmic design choices had a substantial and sometimes unintuitive influence on slice results. Third, considerable expertise and time was required to discern the reasons why a particular statement appeared in a slice, diminishing the slice's probable usefulness. These results provide recommendations for future tool design.

Pre-2018 CSE ID: CS2001-0668