How are Developers Treating License Inconsistency Issues? A Case Study on License Inconsistency Evolution in FOSS Projects

Abstract

Abstract.A license inconsistency is the presence of two or more source files that evolved from the same original file containing different licenses. In our previous study, we have shown that license inconsistencies do exist in open source projects and may lead to potential license violation problems. In this study, we try to find out whether the issues of license inconsistencies are properly solved by analyzing two versions of a FOSS distribution-Debian-and investigate the evolution patterns of license inconsistencies. Findings are: license inconsistencies occur mostly because the original copyright owner updated the license while the reusers were still using the old version of the source files with the old license; most license inconsistencies would disappear when the reusers synchronize their project from the upstream, while some would exist permanently if reusers decide not to synchronize anymore. Legally suspicious cases have not been found yet in those Debian distributions.

Sourcerer's Apprentice and the study of code snippet migration

“…Wu et al use file-level clone detection to track the evolution of a file with respect to code changes and Ninka to track license evolution [42]. Wu's approach is extended and applied to analyze license inconsistencies in large-scale F/LOSS projects [43]. We extend the application of Ninka for license inconsistency detection by checking the package-level files that developers use to license every file in the project in addition to checking the license of the individual files.…”