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
Mix gains per source · case 01
The sum of source gains is applied to the sum of signals, introducing cross terms.
Mix gains per source · case 02
The sum of source gains is applied to the sum of signals, introducing cross terms.
Mix gains per source · case 03
The sum of source gains is applied to the sum of signals, introducing cross terms.
Mix gains per source · case 04
The sum of source gains is applied to the sum of signals, introducing cross terms.
Mix gains per source · case 05
The sum of source gains is applied to the sum of signals, introducing cross terms.
Gain ramp reaches final frame · case 01
Dividing by sample count never reaches the requested final gain.
Gain ramp reaches final frame · case 02
Dividing by sample count never reaches the requested final gain.
Gain ramp reaches final frame · case 03
Dividing by sample count never reaches the requested final gain.
Gain ramp reaches final frame · case 04
Dividing by sample count never reaches the requested final gain.
Gain ramp reaches final frame · case 05
Dividing by sample count never reaches the requested final gain.
Crossfade complementary gains · case 01
Applying the same fade gain to both sources changes total gain and source weighting.
Crossfade complementary gains · case 02
Applying the same fade gain to both sources changes total gain and source weighting.
Crossfade complementary gains · case 03
Applying the same fade gain to both sources changes total gain and source weighting.
Crossfade complementary gains · case 04
Applying the same fade gain to both sources changes total gain and source weighting.
Crossfade complementary gains · case 05
Applying the same fade gain to both sources changes total gain and source weighting.
Channel gains independent · case 01
A single gain replaces separately configured channel volumes.
Channel gains independent · case 02
A single gain replaces separately configured channel volumes.
Channel gains independent · case 03
A single gain replaces separately configured channel volumes.
Channel gains independent · case 04
A single gain replaces separately configured channel volumes.
Channel gains independent · case 05
A single gain replaces separately configured channel volumes.
Mute preserves duration · case 01
Dropping muted samples advances the playback timeline instead of silencing it.
Mute preserves duration · case 02
Dropping muted samples advances the playback timeline instead of silencing it.
Mute preserves duration · case 03
Dropping muted samples advances the playback timeline instead of silencing it.
Mute preserves duration · case 04
Dropping muted samples advances the playback timeline instead of silencing it.
Mute preserves duration · case 05
Dropping muted samples advances the playback timeline instead of silencing it.
Channel solo versus mute · case 01
Solo state is ignored when selecting audible tracks.
Channel solo versus mute · case 02
Solo state is ignored when selecting audible tracks.
Channel solo versus mute · case 03
Solo state is ignored when selecting audible tracks.
Channel solo versus mute · case 04
Solo state is ignored when selecting audible tracks.
Channel solo versus mute · case 05
Solo state is ignored when selecting audible tracks.
Pan linear complement · case 01
The pan law doubles total amplitude by omitting normalization.
Pan linear complement · case 02
The pan law doubles total amplitude by omitting normalization.
Pan linear complement · case 03
The pan law doubles total amplitude by omitting normalization.
Pan linear complement · case 04
The pan law doubles total amplitude by omitting normalization.
Pan linear complement · case 05
The pan law doubles total amplitude by omitting normalization.
Mix bus clear each block · case 01
An uncleared mix bus accumulates audio from the prior callback.
Mix bus clear each block · case 02
An uncleared mix bus accumulates audio from the prior callback.
Mix bus clear each block · case 03
An uncleared mix bus accumulates audio from the prior callback.
Mix bus clear each block · case 04
An uncleared mix bus accumulates audio from the prior callback.
Mix bus clear each block · case 05
An uncleared mix bus accumulates audio from the prior callback.
Route channel matrix · case 01
Multiplying row sums mixes each sample through unrelated gains.
Route channel matrix · case 02
Multiplying row sums mixes each sample through unrelated gains.
Route channel matrix · case 03
Multiplying row sums mixes each sample through unrelated gains.
Route channel matrix · case 04
Multiplying row sums mixes each sample through unrelated gains.
Route channel matrix · case 05
Multiplying row sums mixes each sample through unrelated gains.
Callback event half open window · case 01
A boundary event is scheduled in two adjacent callbacks.
Callback event half open window · case 02
A boundary event is scheduled in two adjacent callbacks.
Callback event half open window · case 03
A boundary event is scheduled in two adjacent callbacks.
Callback event half open window · case 04
A boundary event is scheduled in two adjacent callbacks.
Callback event half open window · case 05
A boundary event is scheduled in two adjacent callbacks.
Schedule insert stable time order · case 01
Appending an earlier event breaks chronological dispatch.
Schedule insert stable time order · case 02
Appending an earlier event breaks chronological dispatch.
Schedule insert stable time order · case 03
Appending an earlier event breaks chronological dispatch.
Schedule insert stable time order · case 04
Appending an earlier event breaks chronological dispatch.
Schedule insert stable time order · case 05
Appending an earlier event breaks chronological dispatch.
Playback pause freezes position · case 01
The transport clock advances while playback is paused.
Playback pause freezes position · case 02
The transport clock advances while playback is paused.
Playback pause freezes position · case 03
The transport clock advances while playback is paused.
Playback pause freezes position · case 04
The transport clock advances while playback is paused.
Playback pause freezes position · case 05
The transport clock advances while playback is paused.
Loop playback wrap offset · case 01
Modulo uses the loop length without restoring the nonzero loop origin.
Loop playback wrap offset · case 02
Modulo uses the loop length without restoring the nonzero loop origin.
Loop playback wrap offset · case 03
Modulo uses the loop length without restoring the nonzero loop origin.
Loop playback wrap offset · case 04
Modulo uses the loop length without restoring the nonzero loop origin.
Loop playback wrap offset · case 05
Modulo uses the loop length without restoring the nonzero loop origin.
Seek clamps to frame duration · case 01
Negative seeks remain before the stream beginning.
Seek clamps to frame duration · case 02
Negative seeks remain before the stream beginning.
Seek clamps to frame duration · case 03
Negative seeks remain before the stream beginning.
Seek clamps to frame duration · case 04
Negative seeks remain before the stream beginning.
Seek clamps to frame duration · case 05
Negative seeks remain before the stream beginning.
Stream end stops after tail · case 01
The last callback claims requested frames beyond stream end.
Stream end stops after tail · case 02
The last callback claims requested frames beyond stream end.
Stream end stops after tail · case 03
The last callback claims requested frames beyond stream end.
Stream end stops after tail · case 04
The last callback claims requested frames beyond stream end.
Stream end stops after tail · case 05
The last callback claims requested frames beyond stream end.
Latency position subtracts queued · case 01
Submitted position is exposed as audible position before queued audio plays.
Latency position subtracts queued · case 02
Submitted position is exposed as audible position before queued audio plays.
Latency position subtracts queued · case 03
Submitted position is exposed as audible position before queued audio plays.
Latency position subtracts queued · case 04
Submitted position is exposed as audible position before queued audio plays.
Latency position subtracts queued · case 05
Submitted position is exposed as audible position before queued audio plays.
Dequeue audio preserves tail · case 01
A partial dequeue discards the rest of the queued audio.
Dequeue audio preserves tail · case 02
A partial dequeue discards the rest of the queued audio.
Dequeue audio preserves tail · case 03
A partial dequeue discards the rest of the queued audio.
Dequeue audio preserves tail · case 04
A partial dequeue discards the rest of the queued audio.
Dequeue audio preserves tail · case 05
A partial dequeue discards the rest of the queued audio.
Write ring wraparound · case 01
A contiguous write discards the portion crossing the ring end.
Write ring wraparound · case 02
A contiguous write discards the portion crossing the ring end.
Write ring wraparound · case 03
A contiguous write discards the portion crossing the ring end.
Write ring wraparound · case 04
A contiguous write discards the portion crossing the ring end.
Write ring wraparound · case 05
A contiguous write discards the portion crossing the ring end.
Read ring logical order · case 01
A slice stops at the physical buffer end instead of wrapping.
Read ring logical order · case 02
A slice stops at the physical buffer end instead of wrapping.
Read ring logical order · case 03
A slice stops at the physical buffer end instead of wrapping.
Read ring logical order · case 04
A slice stops at the physical buffer end instead of wrapping.
Read ring logical order · case 05
A slice stops at the physical buffer end instead of wrapping.
Callback cancellation removes only target · case 01
Cancelling one scheduled sound clears unrelated scheduled work.
Callback cancellation removes only target · case 02
Cancelling one scheduled sound clears unrelated scheduled work.
Callback cancellation removes only target · case 03
Cancelling one scheduled sound clears unrelated scheduled work.
Callback cancellation removes only target · case 04
Cancelling one scheduled sound clears unrelated scheduled work.
Callback cancellation removes only target · case 05
Cancelling one scheduled sound clears unrelated scheduled work.
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 ↗