FAILURE MAP

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 ↗
100840Executable case variants
20168Distinct failure mechanisms
302520Executed implementations
20168Open-access cases

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

Python · Standard library
REFERENCEFAILURE MECHANISMDOMAINACCESS
FA-16101

ULP distance returns signed separation · case 01

ULP distance returns signed separation.

Floating-point arithmetic● Open access↗
FA-16102

ULP distance returns signed separation · case 02

ULP distance returns signed separation.

Floating-point arithmetic◈ Members↗
FA-16103

ULP distance returns signed separation · case 03

ULP distance returns signed separation.

Floating-point arithmetic◈ Members↗
FA-16104

ULP distance returns signed separation · case 04

ULP distance returns signed separation.

Floating-point arithmetic◈ Members↗
FA-16105

ULP distance returns signed separation · case 05

ULP distance returns signed separation.

Floating-point arithmetic◈ Members↗
FA-16106

Subtracting one erases a tiny square-root increment · case 01

Subtracting one erases a tiny square-root increment.

Floating-point arithmetic● Open access↗
FA-16107

Subtracting one erases a tiny square-root increment · case 02

Subtracting one erases a tiny square-root increment.

Floating-point arithmetic◈ Members↗
FA-16108

Subtracting one erases a tiny square-root increment · case 03

Subtracting one erases a tiny square-root increment.

Floating-point arithmetic◈ Members↗
FA-16109

Subtracting one erases a tiny square-root increment · case 04

Subtracting one erases a tiny square-root increment.

Floating-point arithmetic◈ Members↗
FA-16110

Subtracting one erases a tiny square-root increment · case 05

Subtracting one erases a tiny square-root increment.

Floating-point arithmetic◈ Members↗
FA-16111

Square-root increment silently clamps invalid radicands · case 01

Square-root increment silently clamps invalid radicands.

Floating-point arithmetic● Open access↗
FA-16112

Square-root increment silently clamps invalid radicands · case 02

Square-root increment silently clamps invalid radicands.

Floating-point arithmetic◈ Members↗
FA-16113

Square-root increment silently clamps invalid radicands · case 03

Square-root increment silently clamps invalid radicands.

Floating-point arithmetic◈ Members↗
FA-16114

Square-root increment silently clamps invalid radicands · case 04

Square-root increment silently clamps invalid radicands.

Floating-point arithmetic◈ Members↗
FA-16115

Square-root increment silently clamps invalid radicands · case 05

Square-root increment silently clamps invalid radicands.

Floating-point arithmetic◈ Members↗
FA-16116

Square-root endpoint is assigned positive one · case 01

Square-root endpoint is assigned positive one.

Floating-point arithmetic● Open access↗
FA-16117

Square-root endpoint is assigned positive one · case 02

Square-root endpoint is assigned positive one.

Floating-point arithmetic◈ Members↗
FA-16118

Square-root endpoint is assigned positive one · case 03

Square-root endpoint is assigned positive one.

Floating-point arithmetic◈ Members↗
FA-16119

Square-root endpoint is assigned positive one · case 04

Square-root endpoint is assigned positive one.

Floating-point arithmetic◈ Members↗
FA-16120

Square-root endpoint is assigned positive one · case 05

Square-root endpoint is assigned positive one.

Floating-point arithmetic◈ Members↗
FA-16121

Square-root increment loses the sign of exact negative zero · case 01

Square-root increment loses the sign of exact negative zero.

Floating-point arithmetic● Open access↗
FA-16122

Square-root increment loses the sign of exact negative zero · case 02

Square-root increment loses the sign of exact negative zero.

Floating-point arithmetic◈ Members↗
FA-16123

Square-root increment loses the sign of exact negative zero · case 03

Square-root increment loses the sign of exact negative zero.

Floating-point arithmetic◈ Members↗
FA-16124

Square-root increment loses the sign of exact negative zero · case 04

Square-root increment loses the sign of exact negative zero.

Floating-point arithmetic◈ Members↗
FA-16125

Square-root increment loses the sign of exact negative zero · case 05

Square-root increment loses the sign of exact negative zero.

Floating-point arithmetic◈ Members↗
FA-16126

Rationalized square-root increment uses the wrong conjugate · case 01

Rationalized square-root increment uses the wrong conjugate.

Floating-point arithmetic● Open access↗
FA-16127

Rationalized square-root increment uses the wrong conjugate · case 02

Rationalized square-root increment uses the wrong conjugate.

Floating-point arithmetic◈ Members↗
FA-16128

Rationalized square-root increment uses the wrong conjugate · case 03

Rationalized square-root increment uses the wrong conjugate.

Floating-point arithmetic◈ Members↗
FA-16129

Rationalized square-root increment uses the wrong conjugate · case 04

Rationalized square-root increment uses the wrong conjugate.

Floating-point arithmetic◈ Members↗
FA-16130

Rationalized square-root increment uses the wrong conjugate · case 05

Rationalized square-root increment uses the wrong conjugate.

Floating-point arithmetic◈ Members↗
FA-16131

Complementary log exponentials subtract near unity · case 01

Complementary log exponentials subtract near unity.

Floating-point arithmetic● Open access↗
FA-16132

Complementary log exponentials subtract near unity · case 02

Complementary log exponentials subtract near unity.

Floating-point arithmetic◈ Members↗
FA-16133

Complementary log exponentials subtract near unity · case 03

Complementary log exponentials subtract near unity.

Floating-point arithmetic◈ Members↗
FA-16134

Complementary log exponentials subtract near unity · case 04

Complementary log exponentials subtract near unity.

Floating-point arithmetic◈ Members↗
FA-16135

Complementary log exponentials subtract near unity · case 05

Complementary log exponentials subtract near unity.

Floating-point arithmetic◈ Members↗
FA-16136

Complementary log exponentials lose the distant tail · case 01

Complementary log exponentials lose the distant tail.

Floating-point arithmetic● Open access↗
FA-16137

Complementary log exponentials lose the distant tail · case 02

Complementary log exponentials lose the distant tail.

Floating-point arithmetic◈ Members↗
FA-16138

Complementary log exponentials lose the distant tail · case 03

Complementary log exponentials lose the distant tail.

Floating-point arithmetic◈ Members↗
FA-16139

Complementary log exponentials lose the distant tail · case 04

Complementary log exponentials lose the distant tail.

Floating-point arithmetic◈ Members↗
FA-16140

Complementary log exponentials lose the distant tail · case 05

Complementary log exponentials lose the distant tail.

Floating-point arithmetic◈ Members↗
FA-16141

Complementary log exponentials select unstable branches · case 01

Complementary log exponentials select unstable branches.

Floating-point arithmetic● Open access↗
FA-16142

Complementary log exponentials select unstable branches · case 02

Complementary log exponentials select unstable branches.

Floating-point arithmetic◈ Members↗
FA-16143

Complementary log exponentials select unstable branches · case 03

Complementary log exponentials select unstable branches.

Floating-point arithmetic◈ Members↗
FA-16144

Complementary log exponentials select unstable branches · case 04

Complementary log exponentials select unstable branches.

Floating-point arithmetic◈ Members↗
FA-16145

Complementary log exponentials select unstable branches · case 05

Complementary log exponentials select unstable branches.

Floating-point arithmetic◈ Members↗
FA-16146

Complementary logarithm assigns a finite value at zero · case 01

Complementary logarithm assigns a finite value at zero.

Floating-point arithmetic● Open access↗
FA-16147

Complementary logarithm assigns a finite value at zero · case 02

Complementary logarithm assigns a finite value at zero.

Floating-point arithmetic◈ Members↗
FA-16148

Complementary logarithm assigns a finite value at zero · case 03

Complementary logarithm assigns a finite value at zero.

Floating-point arithmetic◈ Members↗
FA-16149

Complementary logarithm assigns a finite value at zero · case 04

Complementary logarithm assigns a finite value at zero.

Floating-point arithmetic◈ Members↗
FA-16150

Complementary logarithm assigns a finite value at zero · case 05

Complementary logarithm assigns a finite value at zero.

Floating-point arithmetic◈ Members↗
FA-16151

Complementary logarithm reflects positive input into its domain · case 01

Complementary logarithm reflects positive input into its domain.

Floating-point arithmetic● Open access↗
FA-16152

Complementary logarithm reflects positive input into its domain · case 02

Complementary logarithm reflects positive input into its domain.

Floating-point arithmetic◈ Members↗
FA-16153

Complementary logarithm reflects positive input into its domain · case 03

Complementary logarithm reflects positive input into its domain.

Floating-point arithmetic◈ Members↗
FA-16154

Complementary logarithm reflects positive input into its domain · case 04

Complementary logarithm reflects positive input into its domain.

Floating-point arithmetic◈ Members↗
FA-16155

Complementary logarithm reflects positive input into its domain · case 05

Complementary logarithm reflects positive input into its domain.

Floating-point arithmetic◈ Members↗
FA-16156

Softplus materializes a large positive exponential · case 01

Softplus materializes a large positive exponential.

Floating-point arithmetic● Open access↗
FA-16157

Softplus materializes a large positive exponential · case 02

Softplus materializes a large positive exponential.

Floating-point arithmetic◈ Members↗
FA-16158

Softplus materializes a large positive exponential · case 03

Softplus materializes a large positive exponential.

Floating-point arithmetic◈ Members↗
FA-16159

Softplus materializes a large positive exponential · case 04

Softplus materializes a large positive exponential.

Floating-point arithmetic◈ Members↗
FA-16160

Softplus materializes a large positive exponential · case 05

Softplus materializes a large positive exponential.

Floating-point arithmetic◈ Members↗
FA-16161

Softplus adds one before taking the logarithm · case 01

Softplus adds one before taking the logarithm.

Floating-point arithmetic● Open access↗
FA-16162

Softplus adds one before taking the logarithm · case 02

Softplus adds one before taking the logarithm.

Floating-point arithmetic◈ Members↗
FA-16163

Softplus adds one before taking the logarithm · case 03

Softplus adds one before taking the logarithm.

Floating-point arithmetic◈ Members↗
FA-16164

Softplus adds one before taking the logarithm · case 04

Softplus adds one before taking the logarithm.

Floating-point arithmetic◈ Members↗
FA-16165

Softplus adds one before taking the logarithm · case 05

Softplus adds one before taking the logarithm.

Floating-point arithmetic◈ Members↗
FA-16166

Softplus evaluates the wrong stabilized branch · case 01

Softplus evaluates the wrong stabilized branch.

Floating-point arithmetic● Open access↗
FA-16167

Softplus evaluates the wrong stabilized branch · case 02

Softplus evaluates the wrong stabilized branch.

Floating-point arithmetic◈ Members↗
FA-16168

Softplus evaluates the wrong stabilized branch · case 03

Softplus evaluates the wrong stabilized branch.

Floating-point arithmetic◈ Members↗
FA-16169

Softplus evaluates the wrong stabilized branch · case 04

Softplus evaluates the wrong stabilized branch.

Floating-point arithmetic◈ Members↗
FA-16170

Softplus evaluates the wrong stabilized branch · case 05

Softplus evaluates the wrong stabilized branch.

Floating-point arithmetic◈ Members↗
FA-16171

Positive softplus drops its linear asymptote · case 01

Positive softplus drops its linear asymptote.

Floating-point arithmetic● Open access↗
FA-16172

Positive softplus drops its linear asymptote · case 02

Positive softplus drops its linear asymptote.

Floating-point arithmetic◈ Members↗
FA-16173

Positive softplus drops its linear asymptote · case 03

Positive softplus drops its linear asymptote.

Floating-point arithmetic◈ Members↗
FA-16174

Positive softplus drops its linear asymptote · case 04

Positive softplus drops its linear asymptote.

Floating-point arithmetic◈ Members↗
FA-16175

Positive softplus drops its linear asymptote · case 05

Positive softplus drops its linear asymptote.

Floating-point arithmetic◈ Members↗
FA-16176

Negative softplus uses a reflected exponential · case 01

Negative softplus uses a reflected exponential.

Floating-point arithmetic● Open access↗
FA-16177

Negative softplus uses a reflected exponential · case 02

Negative softplus uses a reflected exponential.

Floating-point arithmetic◈ Members↗
FA-16178

Negative softplus uses a reflected exponential · case 03

Negative softplus uses a reflected exponential.

Floating-point arithmetic◈ Members↗
FA-16179

Negative softplus uses a reflected exponential · case 04

Negative softplus uses a reflected exponential.

Floating-point arithmetic◈ Members↗
FA-16180

Negative softplus uses a reflected exponential · case 05

Negative softplus uses a reflected exponential.

Floating-point arithmetic◈ Members↗
FA-16181

Sigmoid exponentiates the large magnitude of a negative input · case 01

Sigmoid exponentiates the large magnitude of a negative input.

Floating-point arithmetic● Open access↗
FA-16182

Sigmoid exponentiates the large magnitude of a negative input · case 02

Sigmoid exponentiates the large magnitude of a negative input.

Floating-point arithmetic◈ Members↗
FA-16183

Sigmoid exponentiates the large magnitude of a negative input · case 03

Sigmoid exponentiates the large magnitude of a negative input.

Floating-point arithmetic◈ Members↗
FA-16184

Sigmoid exponentiates the large magnitude of a negative input · case 04

Sigmoid exponentiates the large magnitude of a negative input.

Floating-point arithmetic◈ Members↗
FA-16185

Sigmoid exponentiates the large magnitude of a negative input · case 05

Sigmoid exponentiates the large magnitude of a negative input.

Floating-point arithmetic◈ Members↗
FA-16186

Sigmoid computes its negative tail as a complement · case 01

Sigmoid computes its negative tail as a complement.

Floating-point arithmetic● Open access↗
FA-16187

Sigmoid computes its negative tail as a complement · case 02

Sigmoid computes its negative tail as a complement.

Floating-point arithmetic◈ Members↗
FA-16188

Sigmoid computes its negative tail as a complement · case 03

Sigmoid computes its negative tail as a complement.

Floating-point arithmetic◈ Members↗
FA-16189

Sigmoid computes its negative tail as a complement · case 04

Sigmoid computes its negative tail as a complement.

Floating-point arithmetic◈ Members↗
FA-16190

Sigmoid computes its negative tail as a complement · case 05

Sigmoid computes its negative tail as a complement.

Floating-point arithmetic◈ Members↗
FA-16191

Sigmoid reuses the negative-branch numerator on positive inputs · case 01

Sigmoid reuses the negative-branch numerator on positive inputs.

Floating-point arithmetic● Open access↗
FA-16192

Sigmoid reuses the negative-branch numerator on positive inputs · case 02

Sigmoid reuses the negative-branch numerator on positive inputs.

Floating-point arithmetic◈ Members↗
FA-16193

Sigmoid reuses the negative-branch numerator on positive inputs · case 03

Sigmoid reuses the negative-branch numerator on positive inputs.

Floating-point arithmetic◈ Members↗
FA-16194

Sigmoid reuses the negative-branch numerator on positive inputs · case 04

Sigmoid reuses the negative-branch numerator on positive inputs.

Floating-point arithmetic◈ Members↗
FA-16195

Sigmoid reuses the negative-branch numerator on positive inputs · case 05

Sigmoid reuses the negative-branch numerator on positive inputs.

Floating-point arithmetic◈ Members↗
FA-16196

Sigmoid omits one from the negative-branch normalizer · case 01

Sigmoid omits one from the negative-branch normalizer.

Floating-point arithmetic● Open access↗
FA-16197

Sigmoid omits one from the negative-branch normalizer · case 02

Sigmoid omits one from the negative-branch normalizer.

Floating-point arithmetic◈ Members↗
FA-16198

Sigmoid omits one from the negative-branch normalizer · case 03

Sigmoid omits one from the negative-branch normalizer.

Floating-point arithmetic◈ Members↗
FA-16199

Sigmoid omits one from the negative-branch normalizer · case 04

Sigmoid omits one from the negative-branch normalizer.

Floating-point arithmetic◈ Members↗
FA-16200

Sigmoid omits one from the negative-branch normalizer · case 05

Sigmoid omits one from the negative-branch normalizer.

Floating-point arithmetic◈ Members↗

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 ↗