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
Reduced density matrix conjugates the row amplitude · case 01
For 0.6|0> + 0.8i|1> the off-diagonal entry rho[0][1] is +0.48i instead of -0.48i (the transpose).
Reduced density matrix conjugates the row amplitude · case 02
For 0.6|0> + 0.8i|1> the off-diagonal entry rho[0][1] is +0.48i instead of -0.48i (the transpose).
Reduced density matrix conjugates the row amplitude · case 03
For 0.6|0> + 0.8i|1> the off-diagonal entry rho[0][1] is +0.48i instead of -0.48i (the transpose).
Reduced density matrix conjugates the row amplitude · case 04
For 0.6|0> + 0.8i|1> the off-diagonal entry rho[0][1] is +0.48i instead of -0.48i (the transpose).
Reduced density matrix conjugates the row amplitude · case 05
For 0.6|0> + 0.8i|1> the off-diagonal entry rho[0][1] is +0.48i instead of -0.48i (the transpose).
Reduced density matrix reads the kept bit big-endian · case 01
Keeping qubit 0 of a product state returns the other qubit reduced state.
Reduced density matrix reads the kept bit big-endian · case 02
Keeping qubit 0 of a product state returns the other qubit reduced state.
Reduced density matrix reads the kept bit big-endian · case 03
Keeping qubit 0 of a product state returns the other qubit reduced state.
Reduced density matrix reads the kept bit big-endian · case 04
Keeping qubit 0 of a product state returns the other qubit reduced state.
Reduced density matrix reads the kept bit big-endian · case 05
Keeping qubit 0 of a product state returns the other qubit reduced state.
Reduced density matrix is not trace-normalized · case 01
An unnormalized GHZ input yields a reduced matrix with trace 8.
Reduced density matrix is not trace-normalized · case 02
An unnormalized GHZ input yields a reduced matrix with trace 8.
Reduced density matrix is not trace-normalized · case 03
An unnormalized GHZ input yields a reduced matrix with trace 8.
Reduced density matrix is not trace-normalized · case 04
An unnormalized GHZ input yields a reduced matrix with trace 8.
Reduced density matrix is not trace-normalized · case 05
An unnormalized GHZ input yields a reduced matrix with trace 8.
Reduced density purity ignores off-diagonal entries · case 01
A pure product qubit such as |+> reports purity 0.5 instead of 1.
Reduced density purity ignores off-diagonal entries · case 02
A pure product qubit such as |+> reports purity 0.5 instead of 1.
Reduced density purity ignores off-diagonal entries · case 03
A pure product qubit such as |+> reports purity 0.5 instead of 1.
Reduced density purity ignores off-diagonal entries · case 04
A pure product qubit such as |+> reports purity 0.5 instead of 1.
Reduced density purity ignores off-diagonal entries · case 05
A pure product qubit such as |+> reports purity 0.5 instead of 1.
Shot sampler assigns boundary uniforms to the earlier outcome · case 01
u = 0 selects an outcome with zero weight, and u exactly on a cumulative edge lands in the lower bucket.
Shot sampler assigns boundary uniforms to the earlier outcome · case 02
u = 0 selects an outcome with zero weight, and u exactly on a cumulative edge lands in the lower bucket.
Shot sampler assigns boundary uniforms to the earlier outcome · case 03
u = 0 selects an outcome with zero weight, and u exactly on a cumulative edge lands in the lower bucket.
Shot sampler assigns boundary uniforms to the earlier outcome · case 04
u = 0 selects an outcome with zero weight, and u exactly on a cumulative edge lands in the lower bucket.
Shot sampler assigns boundary uniforms to the earlier outcome · case 05
u = 0 selects an outcome with zero weight, and u exactly on a cumulative edge lands in the lower bucket.
Shot sampler writes qubit 0 as the leftmost character · case 01
An outcome on basis index 1 of two qubits is counted as "10" instead of "01".
Shot sampler writes qubit 0 as the leftmost character · case 02
An outcome on basis index 1 of two qubits is counted as "10" instead of "01".
Shot sampler writes qubit 0 as the leftmost character · case 03
An outcome on basis index 1 of two qubits is counted as "10" instead of "01".
Shot sampler writes qubit 0 as the leftmost character · case 04
An outcome on basis index 1 of two qubits is counted as "10" instead of "01".
Shot sampler writes qubit 0 as the leftmost character · case 05
An outcome on basis index 1 of two qubits is counted as "10" instead of "01".
Shot sampler accepts u = 1 · case 01
A uniform of exactly 1 is silently mapped to the last index, even when its weight is zero, instead of being rejected.
Shot sampler accepts u = 1 · case 02
A uniform of exactly 1 is silently mapped to the last index, even when its weight is zero, instead of being rejected.
Shot sampler accepts u = 1 · case 03
A uniform of exactly 1 is silently mapped to the last index, even when its weight is zero, instead of being rejected.
Shot sampler accepts u = 1 · case 04
A uniform of exactly 1 is silently mapped to the last index, even when its weight is zero, instead of being rejected.
Shot sampler accepts u = 1 · case 05
A uniform of exactly 1 is silently mapped to the last index, even when its weight is zero, instead of being rejected.
Shot sampler accepts negative weights · case 01
Weights [2, -1] produce counts instead of the "negative-weight" error.
Shot sampler accepts negative weights · case 02
Weights [2, -1] produce counts instead of the "negative-weight" error.
Shot sampler accepts negative weights · case 03
Weights [2, -1] produce counts instead of the "negative-weight" error.
Shot sampler accepts negative weights · case 04
Weights [2, -1] produce counts instead of the "negative-weight" error.
Shot sampler accepts negative weights · case 05
Weights [2, -1] produce counts instead of the "negative-weight" error.
Shot sampler checks list emptiness instead of total weight · case 01
All-zero weights produce counts on the last outcome instead of "no-support".
Shot sampler checks list emptiness instead of total weight · case 02
All-zero weights produce counts on the last outcome instead of "no-support".
Shot sampler checks list emptiness instead of total weight · case 03
All-zero weights produce counts on the last outcome instead of "no-support".
Shot sampler checks list emptiness instead of total weight · case 04
All-zero weights produce counts on the last outcome instead of "no-support".
Shot sampler checks list emptiness instead of total weight · case 05
All-zero weights produce counts on the last outcome instead of "no-support".
QFT controlled-phase angles are halved · case 01
QFT of |1> on two qubits produces phases at multiples of 45 degrees instead of 90 degrees.
QFT controlled-phase angles are halved · case 02
QFT of |1> on two qubits produces phases at multiples of 45 degrees instead of 90 degrees.
QFT controlled-phase angles are halved · case 03
QFT of |1> on two qubits produces phases at multiples of 45 degrees instead of 90 degrees.
QFT controlled-phase angles are halved · case 04
QFT of |1> on two qubits produces phases at multiples of 45 degrees instead of 90 degrees.
QFT controlled-phase angles are halved · case 05
QFT of |1> on two qubits produces phases at multiples of 45 degrees instead of 90 degrees.
QFT omits the final bit-reversal swaps · case 01
QFT|1> on three qubits returns amplitudes whose frequencies are indexed in bit-reversed order.
QFT omits the final bit-reversal swaps · case 02
QFT|1> on three qubits returns amplitudes whose frequencies are indexed in bit-reversed order.
QFT omits the final bit-reversal swaps · case 03
QFT|1> on three qubits returns amplitudes whose frequencies are indexed in bit-reversed order.
QFT omits the final bit-reversal swaps · case 04
QFT|1> on three qubits returns amplitudes whose frequencies are indexed in bit-reversed order.
QFT omits the final bit-reversal swaps · case 05
QFT|1> on three qubits returns amplitudes whose frequencies are indexed in bit-reversed order.
Inverse QFT reverses the gates but keeps phase signs · case 01
The inverse QFT applied to |j> is not the conjugate spectrum; round trips do not return to the input.
Inverse QFT reverses the gates but keeps phase signs · case 02
The inverse QFT applied to |j> is not the conjugate spectrum; round trips do not return to the input.
Inverse QFT reverses the gates but keeps phase signs · case 03
The inverse QFT applied to |j> is not the conjugate spectrum; round trips do not return to the input.
Inverse QFT reverses the gates but keeps phase signs · case 04
The inverse QFT applied to |j> is not the conjugate spectrum; round trips do not return to the input.
Inverse QFT reverses the gates but keeps phase signs · case 05
The inverse QFT applied to |j> is not the conjugate spectrum; round trips do not return to the input.
QFT adds a controlled phase of a qubit with itself · case 01
Every qubit gets an extra Z after its Hadamard, negating half of the spectrum.
QFT adds a controlled phase of a qubit with itself · case 02
Every qubit gets an extra Z after its Hadamard, negating half of the spectrum.
QFT adds a controlled phase of a qubit with itself · case 03
Every qubit gets an extra Z after its Hadamard, negating half of the spectrum.
QFT adds a controlled phase of a qubit with itself · case 04
Every qubit gets an extra Z after its Hadamard, negating half of the spectrum.
QFT adds a controlled phase of a qubit with itself · case 05
Every qubit gets an extra Z after its Hadamard, negating half of the spectrum.
Dynamic circuit condition reads the register big-endian · case 01
A condition on register value 2 fires when clbit 0 is set instead of clbit 1.
Dynamic circuit condition reads the register big-endian · case 02
A condition on register value 2 fires when clbit 0 is set instead of clbit 1.
Dynamic circuit condition reads the register big-endian · case 03
A condition on register value 2 fires when clbit 0 is set instead of clbit 1.
Dynamic circuit condition reads the register big-endian · case 04
A condition on register value 2 fires when clbit 0 is set instead of clbit 1.
Dynamic circuit condition reads the register big-endian · case 05
A condition on register value 2 fires when clbit 0 is set instead of clbit 1.
Dynamic circuit reset flips the qubit · case 01
Resetting a qubit that is already |0> leaves it in |1>.
Dynamic circuit reset flips the qubit · case 02
Resetting a qubit that is already |0> leaves it in |1>.
Dynamic circuit reset flips the qubit · case 03
Resetting a qubit that is already |0> leaves it in |1>.
Dynamic circuit reset flips the qubit · case 04
Resetting a qubit that is already |0> leaves it in |1>.
Dynamic circuit reset flips the qubit · case 05
Resetting a qubit that is already |0> leaves it in |1>.
Dynamic circuit Toffoli fires on a single control · case 01
ccx with only its first control set flips the target.
Dynamic circuit Toffoli fires on a single control · case 02
ccx with only its first control set flips the target.
Dynamic circuit Toffoli fires on a single control · case 03
ccx with only its first control set flips the target.
Dynamic circuit Toffoli fires on a single control · case 04
ccx with only its first control set flips the target.
Dynamic circuit Toffoli fires on a single control · case 05
ccx with only its first control set flips the target.
Dynamic circuit measure swaps qubit and clbit operands · case 01
measure 0 -> c1 copies qubit 1 into clbit 0.
Dynamic circuit measure swaps qubit and clbit operands · case 02
measure 0 -> c1 copies qubit 1 into clbit 0.
Dynamic circuit measure swaps qubit and clbit operands · case 03
measure 0 -> c1 copies qubit 1 into clbit 0.
Dynamic circuit measure swaps qubit and clbit operands · case 04
measure 0 -> c1 copies qubit 1 into clbit 0.
Dynamic circuit measure swaps qubit and clbit operands · case 05
measure 0 -> c1 copies qubit 1 into clbit 0.
Dynamic circuit prints registers with index 0 leftmost · case 01
A circuit that excites qubit 0 of three reports "100" instead of "001".
Dynamic circuit prints registers with index 0 leftmost · case 02
A circuit that excites qubit 0 of three reports "100" instead of "001".
Dynamic circuit prints registers with index 0 leftmost · case 03
A circuit that excites qubit 0 of three reports "100" instead of "001".
Dynamic circuit prints registers with index 0 leftmost · case 04
A circuit that excites qubit 0 of three reports "100" instead of "001".
Dynamic circuit prints registers with index 0 leftmost · case 05
A circuit that excites qubit 0 of three reports "100" instead of "001".
Dynamic circuit swap overwrites one operand first · case 01
swap after exciting qubit 0 leaves both qubits excited.
Dynamic circuit swap overwrites one operand first · case 02
swap after exciting qubit 0 leaves both qubits excited.
Dynamic circuit swap overwrites one operand first · case 03
swap after exciting qubit 0 leaves both qubits excited.
Dynamic circuit swap overwrites one operand first · case 04
swap after exciting qubit 0 leaves both qubits excited.
Dynamic circuit swap overwrites one operand first · case 05
swap after exciting qubit 0 leaves both qubits excited.
Grover planner takes asin of the marked fraction instead of its square root · case 01
For N=1024, M=1 the planner asks for about 804 iterations instead of 25.
Grover planner takes asin of the marked fraction instead of its square root · case 02
For N=1024, M=1 the planner asks for about 804 iterations instead of 25.
Grover planner takes asin of the marked fraction instead of its square root · case 03
For N=1024, M=1 the planner asks for about 804 iterations instead of 25.
Grover planner takes asin of the marked fraction instead of its square root · case 04
For N=1024, M=1 the planner asks for about 804 iterations instead of 25.
Grover planner takes asin of the marked fraction instead of its square root · case 05
For N=1024, M=1 the planner asks for about 804 iterations instead of 25.
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 ↗