English

A Case for Dynamic Reverse-code Generation to Debug Non-deterministic Programs

Programming Languages 2013-09-23 v1 Software Engineering

Abstract

Backtracking (i.e., reverse execution) helps the user of a debugger to naturally think backwards along the execution path of a program, and thinking backwards makes it easy to locate the origin of a bug. So far backtracking has been implemented mostly by state saving or by checkpointing. These implementations, however, inherently do not scale. Meanwhile, a more recent backtracking method based on reverse-code generation seems promising because executing reverse code can restore the previous states of a program without state saving. In the literature, there can be found two methods that generate reverse code: (a) static reverse-code generation that pre-generates reverse code through static analysis before starting a debugging session, and (b) dynamic reverse-code generation that generates reverse code by applying dynamic analysis on the fly during a debugging session. In particular, we espoused the latter one in our previous work to accommodate non-determinism of a program caused by e.g., multi-threading. To demonstrate the usefulness of our dynamic reverse-code generation, this article presents a case study of various backtracking methods including ours. We compare the memory usage of various backtracking methods in a simple but nontrivial example, a bounded-buffer program. In the case of non-deterministic programs such as this bounded-buffer program, our dynamic reverse-code generation outperforms the existing backtracking methods in terms of memory efficiency.

Keywords

Cite

@article{arxiv.1309.5152,
  title  = {A Case for Dynamic Reverse-code Generation to Debug Non-deterministic Programs},
  author = {Jooyong Yi},
  journal= {arXiv preprint arXiv:1309.5152},
  year   = {2013}
}

Comments

In Proceedings Festschrift for Dave Schmidt, arXiv:1309.4557

R2 v1 2026-06-22T01:30:42.666Z