Daily Arxiv

This is a page that curates AI-related papers published worldwide.
All content here is summarized using Google Gemini and operated on a non-profit basis.
Copyright for each paper belongs to the authors and their institutions; please make sure to credit the source when sharing.

CFaults: Model-Based Diagnosis for Fault Localization in C Programs with Multiple Test Cases

Created by
  • Haebom

Author

Pedro Orvalho, Mikol a\v{s} Janota, Vasco Manquinho

Outline

This paper proposes CFaults, a novel fault localization technique for C programs containing multiple faults. CFaults leverages model-based diagnosis (MBD) with multiple observations to aggregate all failing test cases into a unified MaxSAT formula. This ensures consistency across all observations and simplifies the fault localization process. Experimental results on two C program benchmark sets, TCAS and C-Pack-IPAs, demonstrate that CFaults is faster than other FBFL approaches, such as BugAssist and SNIPER, and generates only a minimal subset of diagnostics without generating redundant diagnostics.

Takeaways, Limitations

•
Takeaways:
◦
Enables efficient and accurate error localization for C programs with multiple errors.
◦
It improves efficiency by generating faster speed and minimum subset diagnosis than existing FBFL methods.
◦
Ensures consistency across all failing test cases.
•
Limitations:
◦
It is only applicable to C programs, and its applicability to other programming languages requires further study.
◦
The size of the benchmark set used may be limited, and validation on a more diverse and larger dataset is needed.
◦
Depending on the complexity of generating and solving the MaxSAT formula, performance degradation may occur.
👍
Made with Slashpage