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
Dictionary clear codes retain stale learned phrases · case 01
Dictionary clear codes retain stale learned phrases.
Dictionary clear codes retain stale learned phrases · case 02
Dictionary clear codes retain stale learned phrases.
Dictionary clear codes retain stale learned phrases · case 03
Dictionary clear codes retain stale learned phrases.
Dictionary clear codes retain stale learned phrases · case 04
Dictionary clear codes retain stale learned phrases.
Dictionary clear codes retain stale learned phrases · case 05
Dictionary clear codes retain stale learned phrases.
Canonical Huffman codes advance without shifting at length changes · case 01
Canonical Huffman codes advance without shifting at length changes.
Canonical Huffman codes advance without shifting at length changes · case 02
Canonical Huffman codes advance without shifting at length changes.
Canonical Huffman codes advance without shifting at length changes · case 03
Canonical Huffman codes advance without shifting at length changes.
Canonical Huffman codes advance without shifting at length changes · case 04
Canonical Huffman codes advance without shifting at length changes.
Canonical Huffman codes advance without shifting at length changes · case 05
Canonical Huffman codes advance without shifting at length changes.
Bit reservoir loses codes that cross byte boundaries · case 01
Bit reservoir loses codes that cross byte boundaries.
Bit reservoir loses codes that cross byte boundaries · case 02
Bit reservoir loses codes that cross byte boundaries.
Bit reservoir loses codes that cross byte boundaries · case 03
Bit reservoir loses codes that cross byte boundaries.
Bit reservoir loses codes that cross byte boundaries · case 04
Bit reservoir loses codes that cross byte boundaries.
Bit reservoir loses codes that cross byte boundaries · case 05
Bit reservoir loses codes that cross byte boundaries.
Padding bits are decoded as extra compressed symbols · case 01
Padding bits are decoded as extra compressed symbols.
Padding bits are decoded as extra compressed symbols · case 02
Padding bits are decoded as extra compressed symbols.
Padding bits are decoded as extra compressed symbols · case 03
Padding bits are decoded as extra compressed symbols.
Padding bits are decoded as extra compressed symbols · case 04
Padding bits are decoded as extra compressed symbols.
Padding bits are decoded as extra compressed symbols · case 05
Padding bits are decoded as extra compressed symbols.
Delta decompression carries a predictor across independent blocks · case 01
Delta decompression carries a predictor across independent blocks.
Delta decompression carries a predictor across independent blocks · case 02
Delta decompression carries a predictor across independent blocks.
Delta decompression carries a predictor across independent blocks · case 03
Delta decompression carries a predictor across independent blocks.
Delta decompression carries a predictor across independent blocks · case 04
Delta decompression carries a predictor across independent blocks.
Delta decompression carries a predictor across independent blocks · case 05
Delta decompression carries a predictor across independent blocks.
Burrows-Wheeler inversion ignores the primary row · case 01
Burrows-Wheeler inversion ignores the primary row.
Burrows-Wheeler inversion ignores the primary row · case 02
Burrows-Wheeler inversion ignores the primary row.
Burrows-Wheeler inversion ignores the primary row · case 03
Burrows-Wheeler inversion ignores the primary row.
Burrows-Wheeler inversion ignores the primary row · case 04
Burrows-Wheeler inversion ignores the primary row.
Burrows-Wheeler inversion ignores the primary row · case 05
Burrows-Wheeler inversion ignores the primary row.
Move-to-front decoder updates the wrong position · case 01
Move-to-front decoder updates the wrong position.
Move-to-front decoder updates the wrong position · case 02
Move-to-front decoder updates the wrong position.
Move-to-front decoder updates the wrong position · case 03
Move-to-front decoder updates the wrong position.
Move-to-front decoder updates the wrong position · case 04
Move-to-front decoder updates the wrong position.
Move-to-front decoder updates the wrong position · case 05
Move-to-front decoder updates the wrong position.
Bounded run packets silently discard overflow · case 01
Bounded run packets silently discard overflow.
Bounded run packets silently discard overflow · case 02
Bounded run packets silently discard overflow.
Bounded run packets silently discard overflow · case 03
Bounded run packets silently discard overflow.
Bounded run packets silently discard overflow · case 04
Bounded run packets silently discard overflow.
Bounded run packets silently discard overflow · case 05
Bounded run packets silently discard overflow.
Combining attacks bypasses per-hit armor · case 01
Combining attacks bypasses per-hit armor.
Combining attacks bypasses per-hit armor · case 02
Combining attacks bypasses per-hit armor.
Combining attacks bypasses per-hit armor · case 03
Combining attacks bypasses per-hit armor.
Combining attacks bypasses per-hit armor · case 04
Combining attacks bypasses per-hit armor.
Combining attacks bypasses per-hit armor · case 05
Combining attacks bypasses per-hit armor.
Sequential combat suppresses a defeated fighter’s committed strike · case 01
Sequential combat suppresses a defeated fighter’s committed strike.
Sequential combat suppresses a defeated fighter’s committed strike · case 02
Sequential combat suppresses a defeated fighter’s committed strike.
Sequential combat suppresses a defeated fighter’s committed strike · case 03
Sequential combat suppresses a defeated fighter’s committed strike.
Sequential combat suppresses a defeated fighter’s committed strike · case 04
Sequential combat suppresses a defeated fighter’s committed strike.
Sequential combat suppresses a defeated fighter’s committed strike · case 05
Sequential combat suppresses a defeated fighter’s committed strike.
Returning a rook restores a forfeited castling right · case 01
Returning a rook restores a forfeited castling right.
Returning a rook restores a forfeited castling right · case 02
Returning a rook restores a forfeited castling right.
Returning a rook restores a forfeited castling right · case 03
Returning a rook restores a forfeited castling right.
Returning a rook restores a forfeited castling right · case 04
Returning a rook restores a forfeited castling right.
Returning a rook restores a forfeited castling right · case 05
Returning a rook restores a forfeited castling right.
Reapplying poison extends a nonstacking effect additively · case 01
Reapplying poison extends a nonstacking effect additively.
Reapplying poison extends a nonstacking effect additively · case 02
Reapplying poison extends a nonstacking effect additively.
Reapplying poison extends a nonstacking effect additively · case 03
Reapplying poison extends a nonstacking effect additively.
Reapplying poison extends a nonstacking effect additively · case 04
Reapplying poison extends a nonstacking effect additively.
Reapplying poison extends a nonstacking effect additively · case 05
Reapplying poison extends a nonstacking effect additively.
A queued bonus turn schedules an eliminated player · case 01
A queued bonus turn schedules an eliminated player.
A queued bonus turn schedules an eliminated player · case 02
A queued bonus turn schedules an eliminated player.
A queued bonus turn schedules an eliminated player · case 03
A queued bonus turn schedules an eliminated player.
A queued bonus turn schedules an eliminated player · case 04
A queued bonus turn schedules an eliminated player.
A queued bonus turn schedules an eliminated player · case 05
A queued bonus turn schedules an eliminated player.
A full winning board is classified as a draw · case 01
A full winning board is classified as a draw.
A full winning board is classified as a draw · case 02
A full winning board is classified as a draw.
A full winning board is classified as a draw · case 03
A full winning board is classified as a draw.
A full winning board is classified as a draw · case 04
A full winning board is classified as a draw.
A full winning board is classified as a draw · case 05
A full winning board is classified as a draw.
Multiple aces force an avoidable blackjack bust · case 01
Multiple aces force an avoidable blackjack bust.
Multiple aces force an avoidable blackjack bust · case 02
Multiple aces force an avoidable blackjack bust.
Multiple aces force an avoidable blackjack bust · case 03
Multiple aces force an avoidable blackjack bust.
Multiple aces force an avoidable blackjack bust · case 04
Multiple aces force an avoidable blackjack bust.
Multiple aces force an avoidable blackjack bust · case 05
Multiple aces force an avoidable blackjack bust.
A pawn promotes before reaching its final rank · case 01
A pawn promotes before reaching its final rank.
A pawn promotes before reaching its final rank · case 02
A pawn promotes before reaching its final rank.
A pawn promotes before reaching its final rank · case 03
A pawn promotes before reaching its final rank.
A pawn promotes before reaching its final rank · case 04
A pawn promotes before reaching its final rank.
A pawn promotes before reaching its final rank · case 05
A pawn promotes before reaching its final rank.
An exploding die consumes rolls belonging to the next die · case 01
An exploding die consumes rolls belonging to the next die.
An exploding die consumes rolls belonging to the next die · case 02
An exploding die consumes rolls belonging to the next die.
An exploding die consumes rolls belonging to the next die · case 03
An exploding die consumes rolls belonging to the next die.
An exploding die consumes rolls belonging to the next die · case 04
An exploding die consumes rolls belonging to the next die.
An exploding die consumes rolls belonging to the next die · case 05
An exploding die consumes rolls belonging to the next die.
A defeated flag carrier scores on entering home · case 01
A defeated flag carrier scores on entering home.
A defeated flag carrier scores on entering home · case 02
A defeated flag carrier scores on entering home.
A defeated flag carrier scores on entering home · case 03
A defeated flag carrier scores on entering home.
A defeated flag carrier scores on entering home · case 04
A defeated flag carrier scores on entering home.
A defeated flag carrier scores on entering home · case 05
A defeated flag carrier scores on entering home.
Strong components collapse one-way reachability into equivalence · case 01
Strong components collapse one-way reachability into equivalence.
Strong components collapse one-way reachability into equivalence · case 02
Strong components collapse one-way reachability into equivalence.
Strong components collapse one-way reachability into equivalence · case 03
Strong components collapse one-way reachability into equivalence.
Strong components collapse one-way reachability into equivalence · case 04
Strong components collapse one-way reachability into equivalence.
Strong components collapse one-way reachability into equivalence · case 05
Strong components collapse one-way reachability into equivalence.
Parallel edges are all removed when testing a single bridge · case 01
Parallel edges are all removed when testing a single bridge.
Parallel edges are all removed when testing a single bridge · case 02
Parallel edges are all removed when testing a single bridge.
Parallel edges are all removed when testing a single bridge · case 03
Parallel edges are all removed when testing a single bridge.
Parallel edges are all removed when testing a single bridge · case 04
Parallel edges are all removed when testing a single bridge.
Parallel edges are all removed when testing a single bridge · case 05
Parallel edges are all removed when testing a single bridge.
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 ↗