• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.20 no.6
• /
• pp.33-39
• /
• 2020

Developers spend considerable time manually repairing code that was not built during project construction. If the build fails, it is necessary to understand the failed execution, identify the cause of the failure, and then implement the solution. Build tools such as Gradle have been developed to reduce this effort and automate project construction. However, build tools still do not solve many errors, requiring developers to try to solve build errors. In this study, we propose a solution recommendation method to increase the success rate of Gradle build and reduce the effort required to resolve errors. We provide a way to collect build errors and a way to transition from build error messages to successful builds. In particular, 296 build error messages collected from Github's Java project are classified as solutions, and 89% show that the solution can be applied.