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
New attribute runs start with an empty interval · case 01
New attribute runs start with an empty interval.
New attribute runs start with an empty interval · case 02
New attribute runs start with an empty interval.
New attribute runs start with an empty interval · case 03
New attribute runs start with an empty interval.
New attribute runs start with an empty interval · case 04
New attribute runs start with an empty interval.
New attribute runs start with an empty interval · case 05
New attribute runs start with an empty interval.
Unexposed inline objects enter accessible text · case 01
Unexposed inline objects enter accessible text.
Unexposed inline objects enter accessible text · case 02
Unexposed inline objects enter accessible text.
Unexposed inline objects enter accessible text · case 03
Unexposed inline objects enter accessible text.
Unexposed inline objects enter accessible text · case 04
Unexposed inline objects enter accessible text.
Unexposed inline objects enter accessible text · case 05
Unexposed inline objects enter accessible text.
Decorative text tokens consume accessibility offsets · case 01
Decorative text tokens consume accessibility offsets.
Decorative text tokens consume accessibility offsets · case 02
Decorative text tokens consume accessibility offsets.
Decorative text tokens consume accessibility offsets · case 03
Decorative text tokens consume accessibility offsets.
Decorative text tokens consume accessibility offsets · case 04
Decorative text tokens consume accessibility offsets.
Decorative text tokens consume accessibility offsets · case 05
Decorative text tokens consume accessibility offsets.
Literal text is trimmed while constructing object offsets · case 01
Literal text is trimmed while constructing object offsets.
Literal text is trimmed while constructing object offsets · case 02
Literal text is trimmed while constructing object offsets.
Literal text is trimmed while constructing object offsets · case 03
Literal text is trimmed while constructing object offsets.
Literal text is trimmed while constructing object offsets · case 04
Literal text is trimmed while constructing object offsets.
Literal text is trimmed while constructing object offsets · case 05
Literal text is trimmed while constructing object offsets.
Accessible line break tokens disappear from text · case 01
Accessible line break tokens disappear from text.
Accessible line break tokens disappear from text · case 02
Accessible line break tokens disappear from text.
Accessible line break tokens disappear from text · case 03
Accessible line break tokens disappear from text.
Accessible line break tokens disappear from text · case 04
Accessible line break tokens disappear from text.
Accessible line break tokens disappear from text · case 05
Accessible line break tokens disappear from text.
Object alternatives replace markers even when disabled · case 01
Object alternatives replace markers even when disabled.
Object alternatives replace markers even when disabled · case 02
Object alternatives replace markers even when disabled.
Object alternatives replace markers even when disabled · case 03
Object alternatives replace markers even when disabled.
Object alternatives replace markers even when disabled · case 04
Object alternatives replace markers even when disabled.
Object alternatives replace markers even when disabled · case 05
Object alternatives replace markers even when disabled.
Empty alternative erases an embedded object position · case 01
Empty alternative erases an embedded object position.
Empty alternative erases an embedded object position · case 02
Empty alternative erases an embedded object position.
Empty alternative erases an embedded object position · case 03
Empty alternative erases an embedded object position.
Empty alternative erases an embedded object position · case 04
Empty alternative erases an embedded object position.
Empty alternative erases an embedded object position · case 05
Empty alternative erases an embedded object position.
Object starts are measured after the replacement is appended · case 01
Object starts are measured after the replacement is appended.
Object starts are measured after the replacement is appended · case 02
Object starts are measured after the replacement is appended.
Object starts are measured after the replacement is appended · case 03
Object starts are measured after the replacement is appended.
Object starts are measured after the replacement is appended · case 04
Object starts are measured after the replacement is appended.
Object starts are measured after the replacement is appended · case 05
Object starts are measured after the replacement is appended.
Alternative text spans retain a one-character object extent · case 01
Alternative text spans retain a one-character object extent.
Alternative text spans retain a one-character object extent · case 02
Alternative text spans retain a one-character object extent.
Alternative text spans retain a one-character object extent · case 03
Alternative text spans retain a one-character object extent.
Alternative text spans retain a one-character object extent · case 04
Alternative text spans retain a one-character object extent.
Alternative text spans retain a one-character object extent · case 05
Alternative text spans retain a one-character object extent.
Inline object IDs use filtered output ordinals · case 01
Inline object IDs use filtered output ordinals.
Inline object IDs use filtered output ordinals · case 02
Inline object IDs use filtered output ordinals.
Inline object IDs use filtered output ordinals · case 03
Inline object IDs use filtered output ordinals.
Inline object IDs use filtered output ordinals · case 04
Inline object IDs use filtered output ordinals.
Inline object IDs use filtered output ordinals · case 05
Inline object IDs use filtered output ordinals.
Object mappings lose authored alternative text · case 01
Object mappings lose authored alternative text.
Object mappings lose authored alternative text · case 02
Object mappings lose authored alternative text.
Object mappings lose authored alternative text · case 03
Object mappings lose authored alternative text.
Object mappings lose authored alternative text · case 04
Object mappings lose authored alternative text.
Object mappings lose authored alternative text · case 05
Object mappings lose authored alternative text.
Disabled hyperlinks remain reachable through text interfaces · case 01
Disabled hyperlinks remain reachable through text interfaces.
Disabled hyperlinks remain reachable through text interfaces · case 02
Disabled hyperlinks remain reachable through text interfaces.
Disabled hyperlinks remain reachable through text interfaces · case 03
Disabled hyperlinks remain reachable through text interfaces.
Disabled hyperlinks remain reachable through text interfaces · case 04
Disabled hyperlinks remain reachable through text interfaces.
Disabled hyperlinks remain reachable through text interfaces · case 05
Disabled hyperlinks remain reachable through text interfaces.
Empty hyperlink targets are represented as valid links · case 01
Empty hyperlink targets are represented as valid links.
Empty hyperlink targets are represented as valid links · case 02
Empty hyperlink targets are represented as valid links.
Empty hyperlink targets are represented as valid links · case 03
Empty hyperlink targets are represented as valid links.
Empty hyperlink targets are represented as valid links · case 04
Empty hyperlink targets are represented as valid links.
Empty hyperlink targets are represented as valid links · case 05
Empty hyperlink targets are represented as valid links.
Hyperlink offsets begin before document text · case 01
Hyperlink offsets begin before document text.
Hyperlink offsets begin before document text · case 02
Hyperlink offsets begin before document text.
Hyperlink offsets begin before document text · case 03
Hyperlink offsets begin before document text.
Hyperlink offsets begin before document text · case 04
Hyperlink offsets begin before document text.
Hyperlink offsets begin before document text · case 05
Hyperlink offsets begin before document text.
Hyperlink spans exceed document extent · case 01
Hyperlink spans exceed document extent.
Hyperlink spans exceed document extent · case 02
Hyperlink spans exceed document extent.
Hyperlink spans exceed document extent · case 03
Hyperlink spans exceed document extent.
Hyperlink spans exceed document extent · case 04
Hyperlink spans exceed document extent.
Hyperlink spans exceed document extent · case 05
Hyperlink spans exceed document extent.
Zero-length hyperlink spans pollute indexed link counts · case 01
Zero-length hyperlink spans pollute indexed link counts.
Zero-length hyperlink spans pollute indexed link counts · case 02
Zero-length hyperlink spans pollute indexed link counts.
Zero-length hyperlink spans pollute indexed link counts · case 03
Zero-length hyperlink spans pollute indexed link counts.
Zero-length hyperlink spans pollute indexed link counts · case 04
Zero-length hyperlink spans pollute indexed link counts.
Zero-length hyperlink spans pollute indexed link counts · case 05
Zero-length hyperlink spans pollute indexed link counts.
Hyperlink indices follow backend delivery order · case 01
Hyperlink indices follow backend delivery order.
Hyperlink indices follow backend delivery order · case 02
Hyperlink indices follow backend delivery order.
Hyperlink indices follow backend delivery order · case 03
Hyperlink indices follow backend delivery order.
Hyperlink indices follow backend delivery order · case 04
Hyperlink indices follow backend delivery order.
Hyperlink indices follow backend delivery order · case 05
Hyperlink indices follow backend delivery order.
A wide hyperlink masks a narrower coincident text link · case 01
A wide hyperlink masks a narrower coincident text link.
A wide hyperlink masks a narrower coincident text link · case 02
A wide hyperlink masks a narrower coincident text link.
A wide hyperlink masks a narrower coincident text link · case 03
A wide hyperlink masks a narrower coincident text link.
A wide hyperlink masks a narrower coincident text link · case 04
A wide hyperlink masks a narrower coincident text link.
A wide hyperlink masks a narrower coincident text link · case 05
A wide hyperlink masks a narrower coincident text link.
Repeated hyperlink IDs yield multiple indexed descriptors · case 01
Repeated hyperlink IDs yield multiple indexed descriptors.
Repeated hyperlink IDs yield multiple indexed descriptors · case 02
Repeated hyperlink IDs yield multiple indexed descriptors.
Repeated hyperlink IDs yield multiple indexed descriptors · case 03
Repeated hyperlink IDs yield multiple indexed descriptors.
Repeated hyperlink IDs yield multiple indexed descriptors · case 04
Repeated hyperlink IDs yield multiple indexed descriptors.
Repeated hyperlink IDs yield multiple indexed descriptors · case 05
Repeated hyperlink IDs yield multiple indexed descriptors.
Text position at a link end still resolves to that link · case 01
Text position at a link end still resolves to that link.
Text position at a link end still resolves to that link · case 02
Text position at a link end still resolves to that link.
Text position at a link end still resolves to that link · case 03
Text position at a link end still resolves to that link.
Text position at a link end still resolves to that link · case 04
Text position at a link end still resolves to that link.
Text position at a link end still resolves to that link · case 05
Text position at a link end still resolves to that link.
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 ↗