kirancodes.me
To Proof Maintenance & Beyond!
ICSE 2019

Could I have a stack trace to examine the dependency conflict issue?

Ying Wang, Ming Wen, Rongxin Wu, Zhenwei Liu, Shin Hwei Tan, Zhiliang Zhu, Hai Yu, Shing-Chi Cheung

DOI 10.1109/icse.2019.00068dblp

Abstract

Intensive use of libraries in Java projects brings potential risk of dependency conflicts, which occur when a project directly or indirectly depends on multiple versions of the same library or class. When this happens, JVM loads one version and shadows the others. Runtime exceptions can occur when methods in the shadowed versions are referenced. Although project management tools such as Maven are able to give warnings of potential dependency conflicts when a project is built, developers often ask for crashing stack traces before examining these warnings. It motivates us to develop Riddle, an automated approach that generates tests and collects crashing stack traces for projects subject to risk of dependency conflicts. Riddle, built on top of Asm and Evosuite, combines condition mutation, search strategies and condition restoration. We applied Riddle on 19 real-world Java projects with duplicate libraries or classes. We reported 20 identified dependency conflicts including their induced crashing stack traces and the details of generated tests. Among them, 15 conflicts were confirmed by developers as real issues, and 10 were readily fixed. The evaluation results demonstrate the effectiveness and usefulness of Riddle.

BibTeX
@inproceedings{Wang-al:ICSE19,
  author    = {Ying Wang and
               Ming Wen and
               Rongxin Wu and
               Zhenwei Liu and
               Shin Hwei Tan and
               Zhiliang Zhu and
               Hai Yu and
               Shing{-}Chi Cheung},
  title     = {Could I have a stack trace to examine the dependency conflict issue?},
  booktitle = {ICSE},
  pages     = {572--583},
  publisher = {{IEEE} / {ACM}},
  year      = {2019},
}

Related papers