Automating The Detection and Resolution of Build Conflicts in Software Merge for Java Programs

Version control systems (VCS) like Git provide an efficient environment for collaborative software development. However, a major challenge of using such systems is the conflicts that occur when developers try to merge branches. This research focuses on build conflicts--a category of higher-order software merge conflicts. The goal of our research is to develop and implement methodologies that automate the detection and resolution process of build conflicts. Our research consists of three pieces. First, we designed and implemented a graph-based pattern matching approach, BUCOND (Build Conflict Detector), to detect conflicts via static analysis. Our evaluation shows that BUCOND accurately identified build conflicts on both synthetic and real-world datasets; it demonstrated great applicability in scenarios where compiler-based tools are inapplicable. Second, we created a hybrid program transformation approach, BuCoR (Build Conflict Resolver), to opportunistically resolve conflicts. We evaluated BuCoR in real-world merging scenarios, observing its great capabilities of generating syntactically correct resolutions and mimicking human developers' resolution practices. Third, we explored a new approach of resolving conflicts using LLMs, and empirically compared the usage of different LLMs as well as diverse prompt designs. We demonstrate that, when properly guided, LLMs can provide highly accurate conflict resolutions. Our research will help developers detect conflicts more efficiently and resolve conflicts with higher effectiveness as well as rigor. Through addressing issues caused by merge conflicts, it will help improve programmer productivity and software reliability.

Name thesis:degree_name

Doctor of Philosophy

Level thesis:degree_level

doctoral

Discipline thesis:degree_discipline

Computer Science & Applications

Department dc:contributor.department

Computer Science and#38; Applications

Grantor dc:publisher

Virginia Tech

Year dc:date.issued

2026

Author dc:creator

• Towqir, Sheikh Shadab

Chair dc:contributor.committeechair

• Meng, Na

Committee members dc:contributor.committeemember

• Gulzar, Muhammad Ali
• Servant Cortes, Francisco Javier
• Tilevich, Eli
• He, Fei

• Software Merge
• Build Conflict
• Static Analysis
• Pattern Matching
• Program Transformation
• Large Language Models (LLMs)

Statement dc:rights

• Creative Commons Attribution 4.0 International

Licence dc:rights.uri

• http://creativecommons.org/licenses/by/4.0/

Language dc:language.iso

en

Dc Identifier Other

vt_gsexam:45596

OAI identifier oai:identifier

oai:vtechworks.lib.vt.edu:10919/140818