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
Iterative removal of backtracking spikes: repeat until stable · case 01
Nested spikes are only partly removed.
Iterative removal of backtracking spikes: repeat until stable · case 02
Nested spikes are only partly removed.
Iterative removal of backtracking spikes: repeat until stable · case 03
Nested spikes are only partly removed.
Iterative removal of backtracking spikes: repeat until stable · case 04
Nested spikes are only partly removed.
Iterative removal of backtracking spikes: repeat until stable · case 05
Nested spikes are only partly removed.
Iterative removal of backtracking spikes: spike predicate · case 01
Straight-through vertices are deleted as if they were spikes.
Iterative removal of backtracking spikes: spike predicate · case 02
Straight-through vertices are deleted as if they were spikes.
Iterative removal of backtracking spikes: spike predicate · case 03
Straight-through vertices are deleted as if they were spikes.
Iterative removal of backtracking spikes: spike predicate · case 04
Straight-through vertices are deleted as if they were spikes.
Iterative removal of backtracking spikes: spike predicate · case 05
Straight-through vertices are deleted as if they were spikes.
Iterative removal of backtracking spikes: collapse threshold · case 01
Rings reduced to a segment are returned as polygons.
Iterative removal of backtracking spikes: collapse threshold · case 02
Rings reduced to a segment are returned as polygons.
Iterative removal of backtracking spikes: collapse threshold · case 03
Rings reduced to a segment are returned as polygons.
Iterative removal of backtracking spikes: collapse threshold · case 04
Rings reduced to a segment are returned as polygons.
Iterative removal of backtracking spikes: collapse threshold · case 05
Rings reduced to a segment are returned as polygons.
Iterative removal of backtracking spikes: closing vertex · case 01
Cleaned rings come back unclosed.
Iterative removal of backtracking spikes: closing vertex · case 02
Cleaned rings come back unclosed.
Iterative removal of backtracking spikes: closing vertex · case 03
Cleaned rings come back unclosed.
Iterative removal of backtracking spikes: closing vertex · case 04
Cleaned rings come back unclosed.
Iterative removal of backtracking spikes: closing vertex · case 05
Cleaned rings come back unclosed.
Missing node (T-junction) detection in a coverage: shared endpoint exclusion · case 01
Properly shared vertices are reported as T-junctions.
Missing node (T-junction) detection in a coverage: shared endpoint exclusion · case 02
Properly shared vertices are reported as T-junctions.
Missing node (T-junction) detection in a coverage: shared endpoint exclusion · case 03
Properly shared vertices are reported as T-junctions.
Missing node (T-junction) detection in a coverage: shared endpoint exclusion · case 04
Properly shared vertices are reported as T-junctions.
Missing node (T-junction) detection in a coverage: shared endpoint exclusion · case 05
Properly shared vertices are reported as T-junctions.
Missing node (T-junction) detection in a coverage: distinct vertex walk · case 01
The first vertex of a ring is reported twice.
Missing node (T-junction) detection in a coverage: distinct vertex walk · case 02
The first vertex of a ring is reported twice.
Missing node (T-junction) detection in a coverage: distinct vertex walk · case 03
The first vertex of a ring is reported twice.
Missing node (T-junction) detection in a coverage: distinct vertex walk · case 04
The first vertex of a ring is reported twice.
Missing node (T-junction) detection in a coverage: distinct vertex walk · case 05
The first vertex of a ring is reported twice.
Missing node (T-junction) detection in a coverage: collinearity requirement · case 01
Vertices merely inside the bounding box of a diagonal edge are reported.
Missing node (T-junction) detection in a coverage: collinearity requirement · case 02
Vertices merely inside the bounding box of a diagonal edge are reported.
Missing node (T-junction) detection in a coverage: collinearity requirement · case 03
Vertices merely inside the bounding box of a diagonal edge are reported.
Missing node (T-junction) detection in a coverage: collinearity requirement · case 04
Vertices merely inside the bounding box of a diagonal edge are reported.
Missing node (T-junction) detection in a coverage: collinearity requirement · case 05
Vertices merely inside the bounding box of a diagonal edge are reported.
Missing node (T-junction) detection in a coverage: report ordering · case 01
Report order depends on polygon traversal order.
Missing node (T-junction) detection in a coverage: report ordering · case 02
Report order depends on polygon traversal order.
Missing node (T-junction) detection in a coverage: report ordering · case 03
Report order depends on polygon traversal order.
Missing node (T-junction) detection in a coverage: report ordering · case 04
Report order depends on polygon traversal order.
Missing node (T-junction) detection in a coverage: report ordering · case 05
Report order depends on polygon traversal order.
Heat index regression: regression gate average · case 01
Near 80 degF the reported heat index jumps between the simple and regression formulas at the wrong temperatures.
Heat index regression: regression gate average · case 02
Near 80 degF the reported heat index jumps between the simple and regression formulas at the wrong temperatures.
Heat index regression: regression gate average · case 03
Near 80 degF the reported heat index jumps between the simple and regression formulas at the wrong temperatures.
Heat index regression: regression gate average · case 04
Near 80 degF the reported heat index jumps between the simple and regression formulas at the wrong temperatures.
Heat index regression: regression gate average · case 05
Near 80 degF the reported heat index jumps between the simple and regression formulas at the wrong temperatures.
Heat index regression: dry adjustment direction · case 01
Very dry desert readings come out hotter than the regression instead of cooler.
Heat index regression: dry adjustment direction · case 02
Very dry desert readings come out hotter than the regression instead of cooler.
Heat index regression: dry adjustment direction · case 03
Very dry desert readings come out hotter than the regression instead of cooler.
Heat index regression: dry adjustment direction · case 04
Very dry desert readings come out hotter than the regression instead of cooler.
Heat index regression: dry adjustment direction · case 05
Very dry desert readings come out hotter than the regression instead of cooler.
Heat index regression: humid adjustment sign · case 01
Saturated air at 80 to 86 degF reads cooler than the regression.
Heat index regression: humid adjustment sign · case 02
Saturated air at 80 to 86 degF reads cooler than the regression.
Heat index regression: humid adjustment sign · case 03
Saturated air at 80 to 86 degF reads cooler than the regression.
Heat index regression: humid adjustment sign · case 04
Saturated air at 80 to 86 degF reads cooler than the regression.
Heat index regression: humid adjustment sign · case 05
Saturated air at 80 to 86 degF reads cooler than the regression.
Heat index regression: humid adjustment window · case 01
Tropical 88 to 97 degF readings with RH above 85 are pulled down by a negative humid correction.
Heat index regression: humid adjustment window · case 02
Tropical 88 to 97 degF readings with RH above 85 are pulled down by a negative humid correction.
Heat index regression: humid adjustment window · case 03
Tropical 88 to 97 degF readings with RH above 85 are pulled down by a negative humid correction.
Heat index regression: humid adjustment window · case 04
Tropical 88 to 97 degF readings with RH above 85 are pulled down by a negative humid correction.
Heat index regression: humid adjustment window · case 05
Tropical 88 to 97 degF readings with RH above 85 are pulled down by a negative humid correction.
Heat index regression: simple estimate base offset · case 01
Mild-temperature heat index values are tens of degrees too high.
Heat index regression: simple estimate base offset · case 02
Mild-temperature heat index values are tens of degrees too high.
Heat index regression: simple estimate base offset · case 03
Mild-temperature heat index values are tens of degrees too high.
Heat index regression: simple estimate base offset · case 04
Mild-temperature heat index values are tens of degrees too high.
Heat index regression: simple estimate base offset · case 05
Mild-temperature heat index values are tens of degrees too high.
Wind chill from km/h anemometer: anemometer unit conversion · case 01
Wind chill is several degrees too cold for every windy reading.
Wind chill from km/h anemometer: anemometer unit conversion · case 02
Wind chill is several degrees too cold for every windy reading.
Wind chill from km/h anemometer: anemometer unit conversion · case 03
Wind chill is several degrees too cold for every windy reading.
Wind chill from km/h anemometer: anemometer unit conversion · case 04
Wind chill is several degrees too cold for every windy reading.
Wind chill from km/h anemometer: anemometer unit conversion · case 05
Wind chill is several degrees too cold for every windy reading.
Wind chill from km/h anemometer: warm cutoff inclusivity · case 01
At exactly 50 degF the index returns the air temperature despite qualifying wind.
Wind chill from km/h anemometer: warm cutoff inclusivity · case 02
At exactly 50 degF the index returns the air temperature despite qualifying wind.
Wind chill from km/h anemometer: warm cutoff inclusivity · case 03
At exactly 50 degF the index returns the air temperature despite qualifying wind.
Wind chill from km/h anemometer: warm cutoff inclusivity · case 04
At exactly 50 degF the index returns the air temperature despite qualifying wind.
Wind chill from km/h anemometer: warm cutoff inclusivity · case 05
At exactly 50 degF the index returns the air temperature despite qualifying wind.
Wind chill from km/h anemometer: calm wind gate units · case 01
Winds between 3 and 4.8 km/h produce a wind chill although they are below 3 mph.
Wind chill from km/h anemometer: calm wind gate units · case 02
Winds between 3 and 4.8 km/h produce a wind chill although they are below 3 mph.
Wind chill from km/h anemometer: calm wind gate units · case 03
Winds between 3 and 4.8 km/h produce a wind chill although they are below 3 mph.
Wind chill from km/h anemometer: calm wind gate units · case 04
Winds between 3 and 4.8 km/h produce a wind chill although they are below 3 mph.
Wind chill from km/h anemometer: calm wind gate units · case 05
Winds between 3 and 4.8 km/h produce a wind chill although they are below 3 mph.
Wind chill from km/h anemometer: temperature-wind interaction term · case 01
Wind chill no longer converges as wind changes; values are off by several degrees.
Wind chill from km/h anemometer: temperature-wind interaction term · case 02
Wind chill no longer converges as wind changes; values are off by several degrees.
Wind chill from km/h anemometer: temperature-wind interaction term · case 03
Wind chill no longer converges as wind changes; values are off by several degrees.
Wind chill from km/h anemometer: temperature-wind interaction term · case 04
Wind chill no longer converges as wind changes; values are off by several degrees.
Wind chill from km/h anemometer: temperature-wind interaction term · case 05
Wind chill no longer converges as wind changes; values are off by several degrees.
Magnus dew point with ice branch: freezing branch boundary · case 01
Readings at exactly 0 degC are computed with ice constants.
Magnus dew point with ice branch: freezing branch boundary · case 02
Readings at exactly 0 degC are computed with ice constants.
Magnus dew point with ice branch: freezing branch boundary · case 03
Readings at exactly 0 degC are computed with ice constants.
Magnus dew point with ice branch: freezing branch boundary · case 04
Readings at exactly 0 degC are computed with ice constants.
Magnus dew point with ice branch: freezing branch boundary · case 05
Readings at exactly 0 degC are computed with ice constants.
Magnus dew point with ice branch: relative humidity fraction · case 01
Dew points exceed the air temperature for every humid reading.
Magnus dew point with ice branch: relative humidity fraction · case 02
Dew points exceed the air temperature for every humid reading.
Magnus dew point with ice branch: relative humidity fraction · case 03
Dew points exceed the air temperature for every humid reading.
Magnus dew point with ice branch: relative humidity fraction · case 04
Dew points exceed the air temperature for every humid reading.
Magnus dew point with ice branch: relative humidity fraction · case 05
Dew points exceed the air temperature for every humid reading.
Magnus dew point with ice branch: inverse Magnus solve · case 01
Dew points collapse toward zero regardless of humidity.
Magnus dew point with ice branch: inverse Magnus solve · case 02
Dew points collapse toward zero regardless of humidity.
Magnus dew point with ice branch: inverse Magnus solve · case 03
Dew points collapse toward zero regardless of humidity.
Magnus dew point with ice branch: inverse Magnus solve · case 04
Dew points collapse toward zero regardless of humidity.
Magnus dew point with ice branch: inverse Magnus solve · case 05
Dew points collapse toward zero regardless of humidity.
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 ↗