Understand the failure.
Verify the repair.
Small, reproducible software failures. The broken implementation, the fix that didn’t work, and the one that passed—preserved together.
Explore the cases ↓How results are verified ↗WHAT THE ARCHIVE CONTAINS
100840 executable cases. 20168 are open.
Every case records the implementation that fails, the fix that did not work, and the repair that passed its checks—with recorded outputs and source hashes. This release adds 100840 cases across 20168 failure mechanisms and 254 domains.
The open tier gives you the failure and the unsuccessful fix for one case in every mechanism. The remaining 80672 cases, 5 variants per mechanism, are member-only: the verified repair, its recorded checks, and the full fixture suite are held in the member archive. Read the methodology ↗
A RECORD OF WHAT WENT WRONG
Browse the archive / 100840
Full outer join marks equal-key residual failures as matched · case 01
Full outer join marks equal-key residual failures as matched.
Full outer join marks equal-key residual failures as matched · case 02
Full outer join marks equal-key residual failures as matched.
Full outer join marks equal-key residual failures as matched · case 03
Full outer join marks equal-key residual failures as matched.
Full outer join marks equal-key residual failures as matched · case 04
Full outer join marks equal-key residual failures as matched.
Full outer join marks equal-key residual failures as matched · case 05
Full outer join marks equal-key residual failures as matched.
Full outer join stores right IDs but tests physical positions · case 01
Full outer join stores right IDs but tests physical positions.
Full outer join stores right IDs but tests physical positions · case 02
Full outer join stores right IDs but tests physical positions.
Full outer join stores right IDs but tests physical positions · case 03
Full outer join stores right IDs but tests physical positions.
Full outer join stores right IDs but tests physical positions · case 04
Full outer join stores right IDs but tests physical positions.
Full outer join stores right IDs but tests physical positions · case 05
Full outer join stores right IDs but tests physical positions.
Full outer join carries a successful match into the next left row · case 01
Full outer join carries a successful match into the next left row.
Full outer join carries a successful match into the next left row · case 02
Full outer join carries a successful match into the next left row.
Full outer join carries a successful match into the next left row · case 03
Full outer join carries a successful match into the next left row.
Full outer join carries a successful match into the next left row · case 04
Full outer join carries a successful match into the next left row.
Full outer join carries a successful match into the next left row · case 05
Full outer join carries a successful match into the next left row.
Full outer join omits unmatched left placeholders · case 01
Full outer join omits unmatched left placeholders.
Full outer join omits unmatched left placeholders · case 02
Full outer join omits unmatched left placeholders.
Full outer join omits unmatched left placeholders · case 03
Full outer join omits unmatched left placeholders.
Full outer join omits unmatched left placeholders · case 04
Full outer join omits unmatched left placeholders.
Full outer join omits unmatched left placeholders · case 05
Full outer join omits unmatched left placeholders.
Full outer join emits already matched right rows again · case 01
Full outer join emits already matched right rows again.
Full outer join emits already matched right rows again · case 02
Full outer join emits already matched right rows again.
Full outer join emits already matched right rows again · case 03
Full outer join emits already matched right rows again.
Full outer join emits already matched right rows again · case 04
Full outer join emits already matched right rows again.
Full outer join emits already matched right rows again · case 05
Full outer join emits already matched right rows again.
Full outer join treats missing keys as equal matches · case 01
Full outer join treats missing keys as equal matches.
Full outer join treats missing keys as equal matches · case 02
Full outer join treats missing keys as equal matches.
Full outer join treats missing keys as equal matches · case 03
Full outer join treats missing keys as equal matches.
Full outer join treats missing keys as equal matches · case 04
Full outer join treats missing keys as equal matches.
Full outer join treats missing keys as equal matches · case 05
Full outer join treats missing keys as equal matches.
A resume cursor stops at an empty batch instead of normalizing · case 01
A resume cursor stops at an empty batch instead of normalizing.
A resume cursor stops at an empty batch instead of normalizing · case 02
A resume cursor stops at an empty batch instead of normalizing.
A resume cursor stops at an empty batch instead of normalizing · case 03
A resume cursor stops at an empty batch instead of normalizing.
A resume cursor stops at an empty batch instead of normalizing · case 04
A resume cursor stops at an empty batch instead of normalizing.
A resume cursor stops at an empty batch instead of normalizing · case 05
A resume cursor stops at an empty batch instead of normalizing.
Cross-batch resume carries the old row offset into the next batch · case 01
Cross-batch resume carries the old row offset into the next batch.
Cross-batch resume carries the old row offset into the next batch · case 02
Cross-batch resume carries the old row offset into the next batch.
Cross-batch resume carries the old row offset into the next batch · case 03
Cross-batch resume carries the old row offset into the next batch.
Cross-batch resume carries the old row offset into the next batch · case 04
Cross-batch resume carries the old row offset into the next batch.
Cross-batch resume carries the old row offset into the next batch · case 05
Cross-batch resume carries the old row offset into the next batch.
Fetch decrements quota per batch rather than per emitted row · case 01
Fetch decrements quota per batch rather than per emitted row.
Fetch decrements quota per batch rather than per emitted row · case 02
Fetch decrements quota per batch rather than per emitted row.
Fetch decrements quota per batch rather than per emitted row · case 03
Fetch decrements quota per batch rather than per emitted row.
Fetch decrements quota per batch rather than per emitted row · case 04
Fetch decrements quota per batch rather than per emitted row.
Fetch decrements quota per batch rather than per emitted row · case 05
Fetch decrements quota per batch rather than per emitted row.
Returned resume cursor points back at the last emitted row · case 01
Returned resume cursor points back at the last emitted row.
Returned resume cursor points back at the last emitted row · case 02
Returned resume cursor points back at the last emitted row.
Returned resume cursor points back at the last emitted row · case 03
Returned resume cursor points back at the last emitted row.
Returned resume cursor points back at the last emitted row · case 04
Returned resume cursor points back at the last emitted row.
Returned resume cursor points back at the last emitted row · case 05
Returned resume cursor points back at the last emitted row.
Batch fetch reads the beginning of a partially consumed batch · case 01
Batch fetch reads the beginning of a partially consumed batch.
Batch fetch reads the beginning of a partially consumed batch · case 02
Batch fetch reads the beginning of a partially consumed batch.
Batch fetch reads the beginning of a partially consumed batch · case 03
Batch fetch reads the beginning of a partially consumed batch.
Batch fetch reads the beginning of a partially consumed batch · case 04
Batch fetch reads the beginning of a partially consumed batch.
Batch fetch reads the beginning of a partially consumed batch · case 05
Batch fetch reads the beginning of a partially consumed batch.
Fetch drains all remaining rows despite its output quota · case 01
Fetch drains all remaining rows despite its output quota.
Fetch drains all remaining rows despite its output quota · case 02
Fetch drains all remaining rows despite its output quota.
Fetch drains all remaining rows despite its output quota · case 03
Fetch drains all remaining rows despite its output quota.
Fetch drains all remaining rows despite its output quota · case 04
Fetch drains all remaining rows despite its output quota.
Fetch drains all remaining rows despite its output quota · case 05
Fetch drains all remaining rows despite its output quota.
Constant vector materialization ignores scalar validity · case 01
Constant vector materialization ignores scalar validity.
Constant vector materialization ignores scalar validity · case 02
Constant vector materialization ignores scalar validity.
Constant vector materialization ignores scalar validity · case 03
Constant vector materialization ignores scalar validity.
Constant vector materialization ignores scalar validity · case 04
Constant vector materialization ignores scalar validity.
Constant vector materialization ignores scalar validity · case 05
Constant vector materialization ignores scalar validity.
Constant vector materialization ignores row validity · case 01
Constant vector materialization ignores row validity.
Constant vector materialization ignores row validity · case 02
Constant vector materialization ignores row validity.
Constant vector materialization ignores row validity · case 03
Constant vector materialization ignores row validity.
Constant vector materialization ignores row validity · case 04
Constant vector materialization ignores row validity.
Constant vector materialization ignores row validity · case 05
Constant vector materialization ignores row validity.
Constant vector treats zero as a null scalar · case 01
Constant vector treats zero as a null scalar.
Constant vector treats zero as a null scalar · case 02
Constant vector treats zero as a null scalar.
Constant vector treats zero as a null scalar · case 03
Constant vector treats zero as a null scalar.
Constant vector treats zero as a null scalar · case 04
Constant vector treats zero as a null scalar.
Constant vector treats zero as a null scalar · case 05
Constant vector treats zero as a null scalar.
Constant vector materializes the first selection-length rows · case 01
Constant vector materializes the first selection-length rows.
Constant vector materializes the first selection-length rows · case 02
Constant vector materializes the first selection-length rows.
Constant vector materializes the first selection-length rows · case 03
Constant vector materializes the first selection-length rows.
Constant vector materializes the first selection-length rows · case 04
Constant vector materializes the first selection-length rows.
Constant vector materializes the first selection-length rows · case 05
Constant vector materializes the first selection-length rows.
Constant vector removes repeated gathered rows · case 01
Constant vector removes repeated gathered rows.
Constant vector removes repeated gathered rows · case 02
Constant vector removes repeated gathered rows.
Constant vector removes repeated gathered rows · case 03
Constant vector removes repeated gathered rows.
Constant vector removes repeated gathered rows · case 04
Constant vector removes repeated gathered rows.
Constant vector removes repeated gathered rows · case 05
Constant vector removes repeated gathered rows.
Constant vector loses source row ordinals in output · case 01
Constant vector loses source row ordinals in output.
Constant vector loses source row ordinals in output · case 02
Constant vector loses source row ordinals in output.
Constant vector loses source row ordinals in output · case 03
Constant vector loses source row ordinals in output.
Constant vector loses source row ordinals in output · case 04
Constant vector loses source row ordinals in output.
Constant vector loses source row ordinals in output · case 05
Constant vector loses source row ordinals in output.
Scatter writes lanes disabled by the update selection mask · case 01
Scatter writes lanes disabled by the update selection mask.
Scatter writes lanes disabled by the update selection mask · case 02
Scatter writes lanes disabled by the update selection mask.
Scatter writes lanes disabled by the update selection mask · case 03
Scatter writes lanes disabled by the update selection mask.
Scatter writes lanes disabled by the update selection mask · case 04
Scatter writes lanes disabled by the update selection mask.
Scatter writes lanes disabled by the update selection mask · case 05
Scatter writes lanes disabled by the update selection mask.
Scatter uses update position instead of target address · case 01
Scatter uses update position instead of target address.
Scatter uses update position instead of target address · case 02
Scatter uses update position instead of target address.
Scatter uses update position instead of target address · case 03
Scatter uses update position instead of target address.
Scatter uses update position instead of target address · case 04
Scatter uses update position instead of target address.
Scatter uses update position instead of target address · case 05
Scatter uses update position instead of target address.
INSPECTABLE BY DESIGN
Every result has a runnable source.
Runnable implementations with recorded outputs, source hashes, and explicit contracts. Related variants share a failure mechanism and belong together in evaluation splits.
Read the methodology ↗