Toward Efficient Package Maintenance: An Empirical Study of Patch Sharing across Four Linux Distributions
As a cornerstone of cloud computing and artificial intelligence, the Linux operating system has given rise to many popular distributions (distros), such as Fedora and Debian. A distro includes a rich collection of software packages derived from third-party open-source projects (upstream). Due to the nature of reusing upstream, some packages originate from the same upstream projects, noted as homologous packages. Maintaining packages is a core but tedious task for distros, which relies on continuously developing code patches. The maintenance of homologous packages could benefit from cross-distro collaboration. However, their shared origin may also lead to redundant efforts and delayed reuse of valuable maintenance work performed by other distros. To measure potential inefficiencies and explore possible solutions, this paper conducts an empirical study of patch sharing among homologous packages across four representative distros spanning three distinct families: Debian, Ubuntu, Fedora, and openEuler. We find that homologous packages are prevalent across distros, even across different distro families. Specifically, there are 13,873 homologous packages shared between Debian and Fedora, accounting for 40.3% and 56.1% of all packages in Debian and Fedora, respectively. For homologous packages with the same upstream version, we found that 19.5% of the patches are shared between Fedora and Debian, and 46.7% between Fedora and openEuler. However, these shared patches often exhibit remarkable propagation latency, with a median lag of 530.0 days between Fedora and Debian and 169.0 days between Fedora and openEuler. These results highlight the importance and necessity of cross-distro-family patch recommendations. We identified 17 shared Fedora-exclusive patches and contributed them to openEuler, among which 15 patches were successfully merged, demonstrating the potential of cross-distro-family patch recommendation and inspiring further research in this area.
Wed 15 AprDisplayed time zone: Brasilia, Distrito Federal, Brazil change
16:00 - 17:30 | Analytics 2SE In Practice (SEIP) / Research Track at Oceania I Chair(s): Emerson Murphy-Hill Microsoft | ||
16:00 15mTalk | Energy-Efficient Software Development: A Multi-dimensional Empirical Analysis of Stack Overflow Research Track Bihui Jin University of Waterloo, Heng Li Polytechnique Montréal, Pengyu Nie University of Waterloo, Ying Zou Queen's University, Kingston, Ontario Pre-print | ||
16:15 15mTalk | Towards Supporting Open Source Library Maintainers with Community-Based Analytics Research Track | ||
16:30 15mTalk | The Cost vs the Benefit of Adding an Extra Code Reviewer to Mitigate Developer Turnover through Reviewer Recommenders Research Track Mohammadali Sefidi Esfahani Concordia University, Fahimeh Hajari Concordia University, Peter Rigby Concordia University; Meta | ||
16:45 15mTalk | Toward Efficient Package Maintenance: An Empirical Study of Patch Sharing across Four Linux Distributions Research Track Jian Peng Institute of Software, Chinese Academy of Sciences, Jiaxin Zhu Institute of Software at Chinese Academy of Sciences, Yuwei Zhang Institute of Software Chinese Academy of Sciences, Wei Chen Institute of Software at Chinese Academy of Sciences, Guoquan Wu Institute of Software at Chinese Academy of Sciences; University of Chinese Academy of Sciences; University of Chinese Academy of Sciences Nanjing College; China Southern Power Grid, Wei Wang Institute of Software at Chinese Academy of Sciences, Jun Wei Institute of Software at Chinese Academy of Sciences; University of Chinese Academy of Sciences | ||
17:00 15mTalk | Evolving Trends, Patterns, and Hidden Pitfalls: Unveiling JavaScript Feature Usage in the Wild Research Track Dawei Chen Xi'an Jiaotong University, Wuxia Jin Xi'an Jiaotong University, Hui Guo Xi'an Jiaotong University, Guanlin Qiao Xi'an Jiaotong University, Peng Di Ant Group & UNSW Sydney, Ting Liu Xi'an Jiaotong University | ||
17:15 15mTalk | What’s DAT? Three Case Studies of Measuring Software Development Productivity at Meta With Diff Authoring Time SE In Practice (SEIP) Moritz Beller Meta Platforms, Inc., USA, Amanda Park Meta Platforms, Inc., Karim Nakad Meta, Akshay Patel Meta Platforms, Inc., Sarita Mohanty Meta Platforms, Inc., Ford Garberson Meta Platforms, Inc., Henri Verroken Meta Platforms, Inc., Andrew Kennedy Meta Platforms, Inc., Ian G. Malone Meta Platforms, Inc., Vaishali Garg Meta Platforms, Inc., Pavel Avgustinov Meta Platforms, Inc. Pre-print File Attached | ||