Research And Application Of Self-admitted Refactorings In The Open Source Software Ecosystem

Nowadays,open source software has become an indispensable part of software field.In the background of big data,the code of software is becoming larger and larger,which makes code quality maintenance(maintainability,readability,etc.)increasingly hard.In open source code development system represented by Github,developers commit code changes with commit comments,which is a common action.In the commit message,the committer may explicitly claim that the commit is a refactoring with the intention of quality improvement.We defined such commit as a self-admitted refactorings(SAR).Since Self-admitted refactorings are unremarkable in the code of the commit information,however,there is little knowledge about the SAR phenomenon,and the impact of SAR on software projects is not clear.In this paper,we performed an exploratory study on SAR by using the assessment of code smells.We investigated SAR in several aspects:code quality of SAR,number of modified source files involved in SAR,number of code smells introduced in SAR and their priorities,number of committers performing SAR,whether there is refactorings in SAR.In the case study,we used 3 large open source projects(Fastjson,Junit4 and Commons-IO)as cases,we conducted an open source software PMDlet which is associated with our experimental process and can be used to collect data,compare the PMD reports.The following conclusions are drawn from the case study of this paper:(1)there is a tendency to improve code quality in the overall SAR versions,and SAR is a signal of improving code quality.(2)Dataflow Anomaly Analysis is the most frequent and reduced code smell type in SAR,and the code smell types affected in each project are not the same.(3)In the SAR version,the severity of most code smell is "Warning High".(4)In SAR,the number of modified source files is not significantly correlated with the number of newly introduced code smell.(5)SAR is not a common behavior,and SAR developers account for a small proportion of all developers.(6)In the project of this paper,the occurrence frequency of SAR is higher in the early and middle period of software development lifecycle.(7)There is not a fact that refactorings must be appeared in SAR versions.The research of SAR phenomenon will help developers to be familiar with the code situation and enhance the collaborative development of open source software,which greatly deepens people's understanding of the SAR.