Continuous Conflict Prediction during Collaborative Software Development: A step-before Continuous Integration

Abstract

Concurrent activities of collaborative developers over shared project repositories might lead to direct and indirect conflicts. Software Configuration Management systems are designed to capture direct or merge conflicts which arise due to concurrent editing of same shared artifact. However, inconsistencies caused owing to indirect conflicts which arise because of concurrent editing of related artifacts might enter the codebase, since SCM systems have limited capabilities to capture these. Although, Continuous Integration process which is deployed to build the entire codebase with every commit, is quite effective in capturing several type of inconsistencies. However, still few categories of behavioral semantic inconsistencies might evade the build process and penetrate into codebase. In this paper, we propose the Continuous Conflict Prediction Framework which describes a cyclic, real-time, continuous process for conflict prediction which is executed during the process of code creation by collaborative developers. This framework entails a critical conflict-prediction and awareness-generation process which helps in capturing conflicts during development process itself and hence minimizes the number of conflicts entering the project codebase. The proposed framework is realized through implementation of the tool named Collaboration Over GitHub.