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
Q15 FIR pushes new samples onto the wrong end of the delay line · case 01
The impulse response comes out reversed for asymmetric taps.
Q15 FIR pushes new samples onto the wrong end of the delay line · case 02
The impulse response comes out reversed for asymmetric taps.
Q15 FIR pushes new samples onto the wrong end of the delay line · case 03
The impulse response comes out reversed for asymmetric taps.
Q15 FIR pushes new samples onto the wrong end of the delay line · case 04
The impulse response comes out reversed for asymmetric taps.
Q15 FIR pushes new samples onto the wrong end of the delay line · case 05
The impulse response comes out reversed for asymmetric taps.
Notch design places the zeros at half the tone frequency · case 01
A 50 Hz notch at 1 kHz sampling attenuates 25 Hz instead.
Notch design places the zeros at half the tone frequency · case 02
A 50 Hz notch at 1 kHz sampling attenuates 25 Hz instead.
Notch design places the zeros at half the tone frequency · case 03
A 50 Hz notch at 1 kHz sampling attenuates 25 Hz instead.
Notch design places the zeros at half the tone frequency · case 04
A 50 Hz notch at 1 kHz sampling attenuates 25 Hz instead.
Notch design places the zeros at half the tone frequency · case 05
A 50 Hz notch at 1 kHz sampling attenuates 25 Hz instead.
Notch design uses r instead of r^2 for a2 · case 01
The poles do not sit at radius r; the notch is far wider than specified.
Notch design uses r instead of r^2 for a2 · case 02
The poles do not sit at radius r; the notch is far wider than specified.
Notch design uses r instead of r^2 for a2 · case 03
The poles do not sit at radius r; the notch is far wider than specified.
Notch design uses r instead of r^2 for a2 · case 04
The poles do not sit at radius r; the notch is far wider than specified.
Notch design uses r instead of r^2 for a2 · case 05
The poles do not sit at radius r; the notch is far wider than specified.
Notch design places the zeros at pi - w0 · case 01
The notch appears at fs/2 - f0.
Notch design places the zeros at pi - w0 · case 02
The notch appears at fs/2 - f0.
Notch design places the zeros at pi - w0 · case 03
The notch appears at fs/2 - f0.
Notch design places the zeros at pi - w0 · case 04
The notch appears at fs/2 - f0.
Notch design places the zeros at pi - w0 · case 05
The notch appears at fs/2 - f0.
Notch design inverts the DC gain correction · case 01
The passband gain is squared away from unity instead of corrected.
Notch design inverts the DC gain correction · case 02
The passband gain is squared away from unity instead of corrected.
Notch design inverts the DC gain correction · case 03
The passband gain is squared away from unity instead of corrected.
Notch design inverts the DC gain correction · case 04
The passband gain is squared away from unity instead of corrected.
Notch design inverts the DC gain correction · case 05
The passband gain is squared away from unity instead of corrected.
Allpass adds the feedback term · case 01
The magnitude response is no longer flat; the section boosts low frequencies.
Allpass adds the feedback term · case 02
The magnitude response is no longer flat; the section boosts low frequencies.
Allpass adds the feedback term · case 03
The magnitude response is no longer flat; the section boosts low frequencies.
Allpass adds the feedback term · case 04
The magnitude response is no longer flat; the section boosts low frequencies.
Allpass adds the feedback term · case 05
The magnitude response is no longer flat; the section boosts low frequencies.
Allpass passes the input without the a coefficient · case 01
The section is not allpass; its DC gain differs from 1.
Allpass passes the input without the a coefficient · case 02
The section is not allpass; its DC gain differs from 1.
Allpass passes the input without the a coefficient · case 03
The section is not allpass; its DC gain differs from 1.
Allpass passes the input without the a coefficient · case 04
The section is not allpass; its DC gain differs from 1.
Allpass passes the input without the a coefficient · case 05
The section is not allpass; its DC gain differs from 1.
Allpass swaps the input and output delay slots · case 01
The impulse response differs from the allpass kernel after the first sample.
Allpass swaps the input and output delay slots · case 02
The impulse response differs from the allpass kernel after the first sample.
Allpass swaps the input and output delay slots · case 03
The impulse response differs from the allpass kernel after the first sample.
Allpass swaps the input and output delay slots · case 04
The impulse response differs from the allpass kernel after the first sample.
Allpass swaps the input and output delay slots · case 05
The impulse response differs from the allpass kernel after the first sample.
TON credits the interval before the rung went true · case 01
The on-delay finishes one scan early because the scan that energized the rung already contributes elapsed time.
TON credits the interval before the rung went true · case 02
The on-delay finishes one scan early because the scan that energized the rung already contributes elapsed time.
TON credits the interval before the rung went true · case 03
The on-delay finishes one scan early because the scan that energized the rung already contributes elapsed time.
TON credits the interval before the rung went true · case 04
The on-delay finishes one scan early because the scan that energized the rung already contributes elapsed time.
TON credits the interval before the rung went true · case 05
The on-delay finishes one scan early because the scan that energized the rung already contributes elapsed time.
TON accumulator runs past its preset · case 01
After the delay completes the ACC value keeps growing on every true scan instead of holding at PRE.
TON accumulator runs past its preset · case 02
After the delay completes the ACC value keeps growing on every true scan instead of holding at PRE.
TON accumulator runs past its preset · case 03
After the delay completes the ACC value keeps growing on every true scan instead of holding at PRE.
TON accumulator runs past its preset · case 04
After the delay completes the ACC value keeps growing on every true scan instead of holding at PRE.
TON accumulator runs past its preset · case 05
After the delay completes the ACC value keeps growing on every true scan instead of holding at PRE.
TON behaves retentively when its rung drops · case 01
A second energization completes too soon because time accumulated before an interruption is kept.
TON behaves retentively when its rung drops · case 02
A second energization completes too soon because time accumulated before an interruption is kept.
TON behaves retentively when its rung drops · case 03
A second energization completes too soon because time accumulated before an interruption is kept.
TON behaves retentively when its rung drops · case 04
A second energization completes too soon because time accumulated before an interruption is kept.
TON behaves retentively when its rung drops · case 05
A second energization completes too soon because time accumulated before an interruption is kept.
TON done bit waits for ACC to exceed a clamped preset · case 01
A running timer never reports done even though ACC equals PRE.
TON done bit waits for ACC to exceed a clamped preset · case 02
A running timer never reports done even though ACC equals PRE.
TON done bit waits for ACC to exceed a clamped preset · case 03
A running timer never reports done even though ACC equals PRE.
TON done bit waits for ACC to exceed a clamped preset · case 04
A running timer never reports done even though ACC equals PRE.
TON done bit waits for ACC to exceed a clamped preset · case 05
A running timer never reports done even though ACC equals PRE.
TON timing bit stays on after the delay completes · case 01
TT and DN are both true on the same scan once the timer has finished.
TON timing bit stays on after the delay completes · case 02
TT and DN are both true on the same scan once the timer has finished.
TON timing bit stays on after the delay completes · case 03
TT and DN are both true on the same scan once the timer has finished.
TON timing bit stays on after the delay completes · case 04
TT and DN are both true on the same scan once the timer has finished.
TON timing bit stays on after the delay completes · case 05
TT and DN are both true on the same scan once the timer has finished.
Zero-preset TON reports done with a false rung · case 01
A timer with a zero preset shows DN set on scans where its rung is false.
Zero-preset TON reports done with a false rung · case 02
A timer with a zero preset shows DN set on scans where its rung is false.
Zero-preset TON reports done with a false rung · case 03
A timer with a zero preset shows DN set on scans where its rung is false.
Zero-preset TON reports done with a false rung · case 04
A timer with a zero preset shows DN set on scans where its rung is false.
Zero-preset TON reports done with a false rung · case 05
A timer with a zero preset shows DN set on scans where its rung is false.
TOF output energizes at power-up with a false rung · case 01
After a restart the off-delay output is on for a full preset even though its rung was never true.
TOF output energizes at power-up with a false rung · case 02
After a restart the off-delay output is on for a full preset even though its rung was never true.
TOF output energizes at power-up with a false rung · case 03
After a restart the off-delay output is on for a full preset even though its rung was never true.
TOF output energizes at power-up with a false rung · case 04
After a restart the off-delay output is on for a full preset even though its rung was never true.
TOF output energizes at power-up with a false rung · case 05
After a restart the off-delay output is on for a full preset even though its rung was never true.
TOF keeps stale delay time when its rung returns · case 01
An off-delay that is re-energized mid-delay then expires early on the next drop.
TOF keeps stale delay time when its rung returns · case 02
An off-delay that is re-energized mid-delay then expires early on the next drop.
TOF keeps stale delay time when its rung returns · case 03
An off-delay that is re-energized mid-delay then expires early on the next drop.
TOF keeps stale delay time when its rung returns · case 04
An off-delay that is re-energized mid-delay then expires early on the next drop.
TOF keeps stale delay time when its rung returns · case 05
An off-delay that is re-energized mid-delay then expires early on the next drop.
TOF credits the scan on which the rung dropped · case 01
The output turns off one scan early because the de-energizing scan already contributes time.
TOF credits the scan on which the rung dropped · case 02
The output turns off one scan early because the de-energizing scan already contributes time.
TOF credits the scan on which the rung dropped · case 03
The output turns off one scan early because the de-energizing scan already contributes time.
TOF credits the scan on which the rung dropped · case 04
The output turns off one scan early because the de-energizing scan already contributes time.
TOF credits the scan on which the rung dropped · case 05
The output turns off one scan early because the de-energizing scan already contributes time.
TOF never expires once ACC is clamped · case 01
The off-delay output stays on forever after the rung drops.
TOF never expires once ACC is clamped · case 02
The off-delay output stays on forever after the rung drops.
TOF never expires once ACC is clamped · case 03
The off-delay output stays on forever after the rung drops.
TOF never expires once ACC is clamped · case 04
The off-delay output stays on forever after the rung drops.
TOF never expires once ACC is clamped · case 05
The off-delay output stays on forever after the rung drops.
TOF timing bit reports an expired timer as timing · case 01
TT is true on false-rung scans after the off-delay has already finished, and at power-up.
TOF timing bit reports an expired timer as timing · case 02
TT is true on false-rung scans after the off-delay has already finished, and at power-up.
TOF timing bit reports an expired timer as timing · case 03
TT is true on false-rung scans after the off-delay has already finished, and at power-up.
TOF timing bit reports an expired timer as timing · case 04
TT is true on false-rung scans after the off-delay has already finished, and at power-up.
TOF timing bit reports an expired timer as timing · case 05
TT is true on false-rung scans after the off-delay has already finished, and at power-up.
RTO discards accumulated time when its rung goes false · case 01
A retentive timer loses its partial accumulation across a stop and restarts from zero.
RTO discards accumulated time when its rung goes false · case 02
A retentive timer loses its partial accumulation across a stop and restarts from zero.
RTO discards accumulated time when its rung goes false · case 03
A retentive timer loses its partial accumulation across a stop and restarts from zero.
RTO discards accumulated time when its rung goes false · case 04
A retentive timer loses its partial accumulation across a stop and restarts from zero.
RTO discards accumulated time when its rung goes false · case 05
A retentive timer loses its partial accumulation across a stop and restarts from zero.
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 ↗