A Study of Debugging Sequence on Multi-Bug Program

Huitong Liu, Wei Liu, Ying Zhou, Xinyu Li, Wenqing Cheng

Abstract: Debugging behavior measurement is a traditional field. There are many studies on debugging behaviors, but most of the measurement materials are programs with a single bug, and there are rarely multiple bugs in one program. In a multi-bug test scenario, unlike a single bug, we need to consider the connection between multiple bugs, i.e. the relationship between the order or performance of bug fixation and the code in which bugs are located. There is little basic knowledge about the research based on multiple bugs at present. In this paper, we tentatively perform a measurement study based on debugging behaviors in a multi-bug program. We concentrate on discovery and solution sequences and temporal performance of bugs, as well as the relationship between complexity and debugging behaviors for different bugs. Our results show that debugging behaviors have some correlations with the bug complexity if the code indentation is cited to describe the complexity of bugs. Our work can be used as a reference for the future in-depth study of code complexity and debugging behaviors.