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
Annotations intersecting an edit are silently retained · case 01
Annotations intersecting an edit are silently retained.
Annotations intersecting an edit are silently retained · case 02
Annotations intersecting an edit are silently retained.
Annotations intersecting an edit are silently retained · case 03
Annotations intersecting an edit are silently retained.
Annotations intersecting an edit are silently retained · case 04
Annotations intersecting an edit are silently retained.
Annotations intersecting an edit are silently retained · case 05
Annotations intersecting an edit are silently retained.
Leading whitespace trim leaves annotation offsets stale · case 01
Leading whitespace trim leaves annotation offsets stale.
Leading whitespace trim leaves annotation offsets stale · case 02
Leading whitespace trim leaves annotation offsets stale.
Leading whitespace trim leaves annotation offsets stale · case 03
Leading whitespace trim leaves annotation offsets stale.
Leading whitespace trim leaves annotation offsets stale · case 04
Leading whitespace trim leaves annotation offsets stale.
Leading whitespace trim leaves annotation offsets stale · case 05
Leading whitespace trim leaves annotation offsets stale.
Sentence partition retains untrimmed cross-boundary spans · case 01
Sentence partition retains untrimmed cross-boundary spans.
Sentence partition retains untrimmed cross-boundary spans · case 02
Sentence partition retains untrimmed cross-boundary spans.
Sentence partition retains untrimmed cross-boundary spans · case 03
Sentence partition retains untrimmed cross-boundary spans.
Sentence partition retains untrimmed cross-boundary spans · case 04
Sentence partition retains untrimmed cross-boundary spans.
Sentence partition retains untrimmed cross-boundary spans · case 05
Sentence partition retains untrimmed cross-boundary spans.
Repeated mention alignment always selects the first occurrence · case 01
Repeated mention alignment always selects the first occurrence.
Repeated mention alignment always selects the first occurrence · case 02
Repeated mention alignment always selects the first occurrence.
Repeated mention alignment always selects the first occurrence · case 03
Repeated mention alignment always selects the first occurrence.
Repeated mention alignment always selects the first occurrence · case 04
Repeated mention alignment always selects the first occurrence.
Repeated mention alignment always selects the first occurrence · case 05
Repeated mention alignment always selects the first occurrence.
Tab expansion leaves annotation endpoints on old columns · case 01
Tab expansion leaves annotation endpoints on old columns.
Tab expansion leaves annotation endpoints on old columns · case 02
Tab expansion leaves annotation endpoints on old columns.
Tab expansion leaves annotation endpoints on old columns · case 03
Tab expansion leaves annotation endpoints on old columns.
Tab expansion leaves annotation endpoints on old columns · case 04
Tab expansion leaves annotation endpoints on old columns.
Tab expansion leaves annotation endpoints on old columns · case 05
Tab expansion leaves annotation endpoints on old columns.
Discontinuous annotation fragments merge across a gap · case 01
Discontinuous annotation fragments merge across a gap.
Discontinuous annotation fragments merge across a gap · case 02
Discontinuous annotation fragments merge across a gap.
Discontinuous annotation fragments merge across a gap · case 03
Discontinuous annotation fragments merge across a gap.
Discontinuous annotation fragments merge across a gap · case 04
Discontinuous annotation fragments merge across a gap.
Discontinuous annotation fragments merge across a gap · case 05
Discontinuous annotation fragments merge across a gap.
Token coverage ratio counts only fully contained tokens · case 01
Token coverage ratio counts only fully contained tokens.
Token coverage ratio counts only fully contained tokens · case 02
Token coverage ratio counts only fully contained tokens.
Token coverage ratio counts only fully contained tokens · case 03
Token coverage ratio counts only fully contained tokens.
Token coverage ratio counts only fully contained tokens · case 04
Token coverage ratio counts only fully contained tokens.
Token coverage ratio counts only fully contained tokens · case 05
Token coverage ratio counts only fully contained tokens.
Serialized end offsets remain exclusive · case 01
Serialized end offsets remain exclusive.
Serialized end offsets remain exclusive · case 02
Serialized end offsets remain exclusive.
Serialized end offsets remain exclusive · case 03
Serialized end offsets remain exclusive.
Serialized end offsets remain exclusive · case 04
Serialized end offsets remain exclusive.
Serialized end offsets remain exclusive · case 05
Serialized end offsets remain exclusive.
Dataset deduplication: case variants survive text-key deduplication · case 01
case variants survive text-key deduplication.
Dataset deduplication: case variants survive text-key deduplication · case 02
case variants survive text-key deduplication.
Dataset deduplication: case variants survive text-key deduplication · case 03
case variants survive text-key deduplication.
Dataset deduplication: case variants survive text-key deduplication · case 04
case variants survive text-key deduplication.
Dataset deduplication: case variants survive text-key deduplication · case 05
case variants survive text-key deduplication.
Dataset deduplication: canonically equivalent text receives separate keys · case 01
canonically equivalent text receives separate keys.
Dataset deduplication: canonically equivalent text receives separate keys · case 02
canonically equivalent text receives separate keys.
Dataset deduplication: canonically equivalent text receives separate keys · case 03
canonically equivalent text receives separate keys.
Dataset deduplication: canonically equivalent text receives separate keys · case 04
canonically equivalent text receives separate keys.
Dataset deduplication: canonically equivalent text receives separate keys · case 05
canonically equivalent text receives separate keys.
Dataset deduplication: formatting whitespace creates duplicate records · case 01
formatting whitespace creates duplicate records.
Dataset deduplication: formatting whitespace creates duplicate records · case 02
formatting whitespace creates duplicate records.
Dataset deduplication: formatting whitespace creates duplicate records · case 03
formatting whitespace creates duplicate records.
Dataset deduplication: formatting whitespace creates duplicate records · case 04
formatting whitespace creates duplicate records.
Dataset deduplication: formatting whitespace creates duplicate records · case 05
formatting whitespace creates duplicate records.
Dataset deduplication: token permutations are treated as distinct documents · case 01
token permutations are treated as distinct documents.
Dataset deduplication: token permutations are treated as distinct documents · case 02
token permutations are treated as distinct documents.
Dataset deduplication: token permutations are treated as distinct documents · case 03
token permutations are treated as distinct documents.
Dataset deduplication: token permutations are treated as distinct documents · case 04
token permutations are treated as distinct documents.
Dataset deduplication: token permutations are treated as distinct documents · case 05
token permutations are treated as distinct documents.
Dataset deduplication: token multiplicity is discarded from bag identity · case 01
token multiplicity is discarded from bag identity.
Dataset deduplication: token multiplicity is discarded from bag identity · case 02
token multiplicity is discarded from bag identity.
Dataset deduplication: token multiplicity is discarded from bag identity · case 03
token multiplicity is discarded from bag identity.
Dataset deduplication: token multiplicity is discarded from bag identity · case 04
token multiplicity is discarded from bag identity.
Dataset deduplication: token multiplicity is discarded from bag identity · case 05
token multiplicity is discarded from bag identity.
Dataset deduplication: later duplicate replaces the first source record · case 01
later duplicate replaces the first source record.
Dataset deduplication: later duplicate replaces the first source record · case 02
later duplicate replaces the first source record.
Dataset deduplication: later duplicate replaces the first source record · case 03
later duplicate replaces the first source record.
Dataset deduplication: later duplicate replaces the first source record · case 04
later duplicate replaces the first source record.
Dataset deduplication: later duplicate replaces the first source record · case 05
later duplicate replaces the first source record.
Dataset deduplication: deduplication retains a lower-quality example · case 01
deduplication retains a lower-quality example.
Dataset deduplication: deduplication retains a lower-quality example · case 02
deduplication retains a lower-quality example.
Dataset deduplication: deduplication retains a lower-quality example · case 03
deduplication retains a lower-quality example.
Dataset deduplication: deduplication retains a lower-quality example · case 04
deduplication retains a lower-quality example.
Dataset deduplication: deduplication retains a lower-quality example · case 05
deduplication retains a lower-quality example.
Dataset deduplication: deleted record is resurrected by older duplicate · case 01
deleted record is resurrected by older duplicate.
Dataset deduplication: deleted record is resurrected by older duplicate · case 02
deleted record is resurrected by older duplicate.
Dataset deduplication: deleted record is resurrected by older duplicate · case 03
deleted record is resurrected by older duplicate.
Dataset deduplication: deleted record is resurrected by older duplicate · case 04
deleted record is resurrected by older duplicate.
Dataset deduplication: deleted record is resurrected by older duplicate · case 05
deleted record is resurrected by older duplicate.
Dataset deduplication: pair exactly at overlap threshold is rejected · case 01
pair exactly at overlap threshold is rejected.
Dataset deduplication: pair exactly at overlap threshold is rejected · case 02
pair exactly at overlap threshold is rejected.
Dataset deduplication: pair exactly at overlap threshold is rejected · case 03
pair exactly at overlap threshold is rejected.
Dataset deduplication: pair exactly at overlap threshold is rejected · case 04
pair exactly at overlap threshold is rejected.
Dataset deduplication: pair exactly at overlap threshold is rejected · case 05
pair exactly at overlap threshold is rejected.
Dataset deduplication: near-duplicate score ignores unmatched token union · case 01
near-duplicate score ignores unmatched token union.
Dataset deduplication: near-duplicate score ignores unmatched token union · case 02
near-duplicate score ignores unmatched token union.
Dataset deduplication: near-duplicate score ignores unmatched token union · case 03
near-duplicate score ignores unmatched token union.
Dataset deduplication: near-duplicate score ignores unmatched token union · case 04
near-duplicate score ignores unmatched token union.
Dataset deduplication: near-duplicate score ignores unmatched token union · case 05
near-duplicate score ignores unmatched token union.
Dataset deduplication: near-duplicate chain is split into separate groups · case 01
near-duplicate chain is split into separate groups.
Dataset deduplication: near-duplicate chain is split into separate groups · case 02
near-duplicate chain is split into separate groups.
Dataset deduplication: near-duplicate chain is split into separate groups · case 03
near-duplicate chain is split into separate groups.
Dataset deduplication: near-duplicate chain is split into separate groups · case 04
near-duplicate chain is split into separate groups.
Dataset deduplication: near-duplicate chain is split into separate groups · case 05
near-duplicate chain is split into separate groups.
Dataset deduplication: different-length fingerprints appear identical · case 01
different-length fingerprints appear identical.
Dataset deduplication: different-length fingerprints appear identical · case 02
different-length fingerprints appear identical.
Dataset deduplication: different-length fingerprints appear identical · case 03
different-length fingerprints appear identical.
Dataset deduplication: different-length fingerprints appear identical · case 04
different-length fingerprints appear identical.
Dataset deduplication: different-length fingerprints appear identical · case 05
different-length fingerprints appear identical.
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 ↗