UniLoc: Unified Fault Localization of Continuous Integration Failures. This site contains dataset of 700 build failures in CI/CD environment and corresponding fixes to resolve the build failures. We extended this build fix dataset from TravisTorrent [1].
row - Row ID of Initial Passed Build. This row value can be used to join with TravisTorrent dataset.
f1row- Row ID of Build Fail introducing commit. This value can be used to join with TravisTorrent dataset.
f2row- Row ID of last failed commit. This value can be used to join with TravisTorrent dataset.
gh_project_name-Project Name.
git_branch-Git Branch Name.
git_commit-Initial passed commit id.
git_failintro_commit- Commit that changes build status passed to fail.
git_lastfail_commit-Last failed commit before the fix.
git_fix_commit-Commit that fixed the build failure. Change between git_lastfail_commit and git_fix_commit is the actual build fix change.
bl_log-Build Failure log. This log file is generated at git_lastfail_commit build.
failpart_log-Fail portion of the large log stored in bl_log column.
passpart_log- In failed log some portion might still be successful. This part contains those pass part.
failpart_sim-Performed filtering even on failpart_log to remove noises. This column contain fail log text after noise removal and used as IR query.
fail_inducing_filelist-List of files modified during Passed to Fail Status Change
fix_filelist- List of files modified to change last Fail status at git_lastfail_commit to Pass status and used as ground truth for our tool.
totalfile_count-Total file count of the project.
ev_diffdep_pos-TopN for without ranking optimization(V3)
ev_diffdep_mrr-MRR for without ranking optimization(V3)
ev_diffdep_map-MAP for without ranking optimization(V3)
ev_diffboost_pos-TopN for without search space optimization(V2)
ev_diffboost_mrr-MRR for without search space optimization(V2)
ev_diffboost_map-MAP for without search space optimization(V2)
ev_fulllogboost_pos-TopN for without query optimization(V1)
ev_fulllogboost_mrr-MRR for without query optimization(V1)
ev_fulllogboost_map-MAP for without query optimization(V1)
ev_baselinebluir_pos-TopN for BLuIR baseline
ev_baselinebluir_mrr-MRR for BLuIR baseline
ev_baselinebluir_map-MAP for BLuIR baseline
ev_baseline1_pos-TopN for baseline1(file name mentioned in log)
ev_baseline1_mrr-MRR for baseline1(file name mentioned in log)
ev_baseline1_map-MAP for baseline1(file name mentioned in log)
ev_uniloc_pos-TopN for proposed approach
ev_uniloc_mrr-MRR for proposed approach
ev_uniloc_map-MAP for proposed approach
dt_fail_type-Build fail type. Test Fail or Other Fail
dt_dataset_type-Training or Evaluation row
Will Add Later
[1] Beller M, Gousios G, Zaidman A. (2017) TravisTorrent: Synthesizing Travis CI and GitHub for Full-Stack Research on Continuous Integration