SZZ revisited: verifying when changes induce fixes

ABSTRACT

Automatically identifying commits that induce fixes is an important task, as it enables researchers to quickly and efficiently validate many types of software engineering analyses, such as software metrics or models for predicting faulty components. Previous work on SZZ, an algorithm designed by Sliwerski et al and improved upon by Kim et al, provides a process for automatically identifying the fix-inducing predecessor lines to lines that are changed in a bug-fixing commit. However, as of yet no one has verified that the fix-inducing lines identified by SZZ are in fact responsible for introducing the fixed bug. Also, the SZZ algorithm relies on annotation graphs, which are imprecise in the face of large blocks of modified code, for back-tracking through previous revisions to the fix-inducing change.

In this work we outline several improvements to the SZZ algorithm: First, we replace annotation graphs with line-number maps that track unique source lines as they change over the lifetime of the software; and second, we use DiffJ, a Java syntax-aware diff tool, to ignore comments and formatting changes in the source. Finally, we begin verifying how often a fix-inducing change identified by SZZ is the true source of a bug.

REFERENCES

• 1: Gerardo Canfora , Luigi Cerulo , Massimiliano Di Penta, Identifying Changed Source Code Lines from Version Repositories, Proceedings of the Fourth International Workshop on Mining Software Repositories, p.14, May 20-26, 2007 [doi>10.1109/MSR.2007.14]
• 2: Sunghun Kim , Thomas Zimmermann , Kai Pan , E. James Jr. Whitehead, Automatic Identification of Bug-Introducing Changes, Proceedings of the 21st IEEE/ACM International Conference on Automated Software Engineering, p.81-90, September 18-22, 2006 [doi>10.1109/ASE.2006.23]
• 3: J. Pace. A tool which compares java files based on content. http://www.incava.org/projects/java/diffj, 2007.
• 4: Jacek Śliwerski , Thomas Zimmermann , Andreas Zeller, When do changes induce fixes?, ACM SIGSOFT Software Engineering Notes, v.30 n.4, July 2005
• 5: Chadd C. Williams , Jaime W. Spacco, Branching and merging in the repository, Proceedings of the 2008 international working conference on Mining software repositories, May 10-11, 2008, Leipzig, Germany [doi>10.1145/1370750.1370754]
• 6: Thomas Zimmermann , Sunghun Kim , Andreas Zeller , E. James Whitehead, Jr., Mining version archives for co-changed lines, Proceedings of the 2006 international workshop on Mining software repositories, May 22-23, 2006, Shanghai, China [doi>10.1145/1137983.1138001]

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.4 INFORMATION SYSTEMS APPLICATIONS

Additional Classification:
  D. Software
  D.2 SOFTWARE ENGINEERING
      D.2.8 Metrics
          Subjects: Performance measures; Complexity measures