This dataset is designed to be used in evaluation studies of regression test prioritization techniques. It includes 20 open-source Java projects from GitHub and over 100,000 logs of real-world build logs from TravisCI. The projects span a wide range with regard to size, number of contributors, and maturity of open-source Java projects available on GitHub. Futher, the dataset includes the results of baseline approaches to ease the comparison of new techniques applied to the dataset. A readme file with a more detailed description of the structure of the dataset is included. For even more information see the corresponding MSR 2020 publication. Versions: 2020-09-23 (version 1.1) Updated archived `deeplearning4j` repository with a fork that contains all of the original commits. Repository at the original GitHub location had been replaced. Defect identified by Daniel Elsner (Technische Universität München). Renamed root folder from MSR2 to rtp-torrent 2020-05-25 (version 1.0) Initial release