Skip to content
Defexts edited this page Jun 13, 2019 · 4 revisions

Defexts

Defexts is a family of Java Virtual Machine (JVM) language specific datasets which contain real-world bugs within real-world projects. Bug benchmarks are critical in furthering software engineering research in many topics, such as bug localization, (automated) program repair, and software maintenance. Numerous bug benchmarks exist for the most popular programming languages (Java, C++, Python, etc) but several programming languages lack similar mature bug benchmarks. Defexts is an attempt to create and facilitate the usage of new benchmarks towards these underrepresented languages, notably JVM languages.

With respect to its projects, Defexts currently exclusively contains:

  • public projects hosted on GitHub
  • projects compatible with Maven 3.3.9 or Gradle 4.8

Every Defexts entry contains:

  • Two git branches
    • One branch corresponding to the fixed version of the project
    • One branch corresponding to the buggy version of the project
  • Supplemental project information (contained in a references.csv file)

Defexts currently contains three bug datasets, DefextsKotlin, DefextsGroovy, and DefextsScala

DefextsKotlin

  • Focused towards the Kotlin programming language
  • Contains 225 bugs from 152 distinct Kotlin projects

DefextsGroovy

  • Focused towards the Groovy programming language
  • Contains 301 bugs from 170 distinct Groovy projects

DefextsScala

  • Focused towards the Scala programming language
  • Contains 128 bugs from 91 distinct Groovy projects

View the Installation wiki page to view how to install + run Defexts on your local machine