A Large Language Model Approach to Identify Flakiness in C++ Projects

TL;DR

This paper presents an approach using large language models to identify flaky tests in C++ projects, improving debugging efficiency by accurately classifying flaky tests at the code level.

Contribution

It introduces a fine-tuned LLM-based method for detecting flaky tests in C++ and Java, with comprehensive evaluation and practical recommendations.

Findings

Mistral-7b outperforms other models on all metrics

Models perform comparably on C++ and Java datasets

LLMs demonstrate high accuracy in classifying flaky tests

Abstract

The role of regression testing in software testing is crucial as it ensures that any new modifications do not disrupt the existing functionality and behaviour of the software system. The desired outcome is for regression tests to yield identical results without any modifications made to the system being tested. In practice, however, the presence of Flaky Tests introduces non-deterministic behaviour and undermines the reliability of regression testing results. In this paper, we propose an LLM-based approach for identifying the root cause of flaky tests in C++ projects at the code level, with the intention of assisting developers in debugging and resolving them more efficiently. We compile a comprehensive collection of C++ project flaky tests sourced from GitHub repositories. We fine-tune Mistral-7b, Llama2-7b and CodeLlama-7b models on the C++ dataset and an existing Java dataset and evaluate the performance in terms of precision, recall, accuracy, and F1 score. We assess the performance of the models across various datasets and offer recommendations for both research and industry applications. The results indicate that our models exhibit varying performance on the C++ dataset, while their performance is comparable to that of the Java dataset. The Mistral-7b surpasses the other two models regarding all metrics, achieving a score of 1. Our results demonstrate the exceptional capability of LLMs to accurately classify flakiness in C++ and Java projects, providing a promising approach to enhance the efficiency of debugging flaky tests in practice.