Characterizing Issue Management in Runtime Systems

Modern programming languages like Java require runtime systems to support the implementation and deployment of software applications in diverse computing platforms and operating systems. These runtime systems are normally developed in GitHub-hosted repositories based on close collaboration between large software companies (e.g., IBM, Microsoft) and OSS developers. However, despite their popularity and broad usage; to the best of our knowledge, these repositories have never been studied. We report an empirical study of around 118K issues from 34 runtime system repos in GitHub. We found that issues regarding enhancement, test failure and bug are mostly posted on runtime system repositories and solution related discussion are mostly present on issue discussion. 82.69% issues in the runtime system repositories have been resolved and 0.69% issues are ignored; median of issue close rate, ignore rate and addressing time in these repositories are 76.1%, 2.2% and 58 days respectively. 82.65% issues are tagged with labels while only 28.30% issues have designated assignees and 90.65% issues contain at least one comment; also presence of these features in an issue report can affect issue closure. Based on the findings, we offer six recommendat