Refinando a Precisão da Detecção de Conflitos: Uma Análise do CSDiff com Abordagem Focalizada

Software development is increasingly complex, with developers simultaneously working on different parts of the source code to build, maintain, and enhance systems. However, this collaborative nature of development can lead to conflicts when multiple individuals attempt to simultaneously modify the same file. In this scenario, code merge tools play a crucial role in detecting and resolving these conflicts. One such tool is CSDiff, a conflict detection and resolution tool, an alternative to the traditional and widespread Diff3. CSDiff stands out by using customizable separators specific to each programming language to help in conflict resolution. In this article, we propose an improvement to the functionality of CSDiff focusing on reducing false positive and false negatives conflicts found when using the original tool. Through an analysis based on Python programs, we compare CSDiff with and without the proposed improvement; we assess the impact on reducing errors presented by the original version of the tool. The results indicate that the proposed improvement not only reduces the number of reported false positive conflicts, leading to a higher proportion of scenarios with correctly resolved conflicts, but also results in a decrease in false negatives when compared to the original tool.