Effective and efficient localization of multiple faults using value replacement

Abstract

We previously presented a fault localization technique called Value Replacement that repeatedly alters the state of an executing program to locate a faulty statement [9]. The technique searches for program statements involving values that can be altered during runtime to cause the incorrect output of a failing run to become correct. We showed that highly effective fault localization results could be achieved by the technique on programs containing single faults. In the current work, we generalize Value Replacement so that it can also perform effectively in the presence of multiple faults. We improve scalability by describing two techniques that significantly improve the efficiency of Value Replacement. In our experimental study, our generalized technique effectively isolates multiple simultaneous faults in time on the order of minutes in each case, whereas in [9], the technique had sometimes required time on the order of hours to isolate only single faults.

Probabilistic Fault Localisation

“…Their approach is designed for the simultaneous fault localisation of sets of multiple faults, and were only scalable to our additional experiments. Other heavyweight techniques are similarly unscalable [30,[36][37][38][39][40][41], which emphasises the importance of developing lightweight techniques such as pfl/sbfl.…”

Using likely invariants for automated software fault localization

“…Execution perturbations [25,46] have been used to find root causes for small programs. We do not believe that these techniques alone will scale to larger programs.…”

Applying enhanced fault localization technology to Monte Carlo simulations

“…They identify suspicious source code by associating program statements (Jones and Harrold 2005), variable states (Huang et al 2007), or predicates (Liblit 2004) with these failures. They often depend on repeatable behavior (Cleve and Zeller 2005, Jeffrey et al 2008, Jeffrey et al 2009), assuming that programs are deterministic. Further, they tend to focus on integral rather than floating-point data types (Nainar et al 2010, Liblit 2004, Arumuga Nainar et al 2007, Liblit et al 2003, Liblit et al 2005, Zhang et al 2006b).…”