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
Retired accessibility proxies accept new pins · case 01
Retired accessibility proxies accept new pins.
Retired accessibility proxies accept new pins · case 02
Retired accessibility proxies accept new pins.
Retired accessibility proxies accept new pins · case 03
Retired accessibility proxies accept new pins.
Retired accessibility proxies accept new pins · case 04
Retired accessibility proxies accept new pins.
Retired accessibility proxies accept new pins · case 05
Retired accessibility proxies accept new pins.
Extra proxy release underflows reference count · case 01
Extra proxy release underflows reference count.
Extra proxy release underflows reference count · case 02
Extra proxy release underflows reference count.
Extra proxy release underflows reference count · case 03
Extra proxy release underflows reference count.
Extra proxy release underflows reference count · case 04
Extra proxy release underflows reference count.
Extra proxy release underflows reference count · case 05
Extra proxy release underflows reference count.
Earlier proxy snapshots change after later mutation · case 01
Earlier proxy snapshots change after later mutation.
Earlier proxy snapshots change after later mutation · case 02
Earlier proxy snapshots change after later mutation.
Earlier proxy snapshots change after later mutation · case 03
Earlier proxy snapshots change after later mutation.
Earlier proxy snapshots change after later mutation · case 04
Earlier proxy snapshots change after later mutation.
Earlier proxy snapshots change after later mutation · case 05
Earlier proxy snapshots change after later mutation.
New proxy generations inherit an old pin count · case 01
New proxy generations inherit an old pin count.
New proxy generations inherit an old pin count · case 02
New proxy generations inherit an old pin count.
New proxy generations inherit an old pin count · case 03
New proxy generations inherit an old pin count.
New proxy generations inherit an old pin count · case 04
New proxy generations inherit an old pin count.
New proxy generations inherit an old pin count · case 05
New proxy generations inherit an old pin count.
Unexposed headings enter document navigation metadata · case 01
Unexposed headings enter document navigation metadata.
Unexposed headings enter document navigation metadata · case 02
Unexposed headings enter document navigation metadata.
Unexposed headings enter document navigation metadata · case 03
Unexposed headings enter document navigation metadata.
Unexposed headings enter document navigation metadata · case 04
Unexposed headings enter document navigation metadata.
Unexposed headings enter document navigation metadata · case 05
Unexposed headings enter document navigation metadata.
Headings from embedded documents pollute the parent outline · case 01
Headings from embedded documents pollute the parent outline.
Headings from embedded documents pollute the parent outline · case 02
Headings from embedded documents pollute the parent outline.
Headings from embedded documents pollute the parent outline · case 03
Headings from embedded documents pollute the parent outline.
Headings from embedded documents pollute the parent outline · case 04
Headings from embedded documents pollute the parent outline.
Headings from embedded documents pollute the parent outline · case 05
Headings from embedded documents pollute the parent outline.
Boolean attributes are interpreted as level one headings · case 01
Boolean attributes are interpreted as level one headings.
Boolean attributes are interpreted as level one headings · case 02
Boolean attributes are interpreted as level one headings.
Boolean attributes are interpreted as level one headings · case 03
Boolean attributes are interpreted as level one headings.
Boolean attributes are interpreted as level one headings · case 04
Boolean attributes are interpreted as level one headings.
Boolean attributes are interpreted as level one headings · case 05
Boolean attributes are interpreted as level one headings.
Out-of-domain heading levels enter the outline · case 01
Out-of-domain heading levels enter the outline.
Out-of-domain heading levels enter the outline · case 02
Out-of-domain heading levels enter the outline.
Out-of-domain heading levels enter the outline · case 03
Out-of-domain heading levels enter the outline.
Out-of-domain heading levels enter the outline · case 04
Out-of-domain heading levels enter the outline.
Out-of-domain heading levels enter the outline · case 05
Out-of-domain heading levels enter the outline.
Heading whitespace-only names create empty navigation entries · case 01
Heading whitespace-only names create empty navigation entries.
Heading whitespace-only names create empty navigation entries · case 02
Heading whitespace-only names create empty navigation entries.
Heading whitespace-only names create empty navigation entries · case 03
Heading whitespace-only names create empty navigation entries.
Heading whitespace-only names create empty navigation entries · case 04
Heading whitespace-only names create empty navigation entries.
Heading whitespace-only names create empty navigation entries · case 05
Heading whitespace-only names create empty navigation entries.
Nameless headings become structural ancestors · case 01
Nameless headings become structural ancestors.
Nameless headings become structural ancestors · case 02
Nameless headings become structural ancestors.
Nameless headings become structural ancestors · case 03
Nameless headings become structural ancestors.
Nameless headings become structural ancestors · case 04
Nameless headings become structural ancestors.
Nameless headings become structural ancestors · case 05
Nameless headings become structural ancestors.
Later duplicate heading IDs overwrite earlier document positions · case 01
Later duplicate heading IDs overwrite earlier document positions.
Later duplicate heading IDs overwrite earlier document positions · case 02
Later duplicate heading IDs overwrite earlier document positions.
Later duplicate heading IDs overwrite earlier document positions · case 03
Later duplicate heading IDs overwrite earlier document positions.
Later duplicate heading IDs overwrite earlier document positions · case 04
Later duplicate heading IDs overwrite earlier document positions.
Later duplicate heading IDs overwrite earlier document positions · case 05
Later duplicate heading IDs overwrite earlier document positions.
Outline ancestry follows transport arrival order · case 01
Outline ancestry follows transport arrival order.
Outline ancestry follows transport arrival order · case 02
Outline ancestry follows transport arrival order.
Outline ancestry follows transport arrival order · case 03
Outline ancestry follows transport arrival order.
Outline ancestry follows transport arrival order · case 04
Outline ancestry follows transport arrival order.
Outline ancestry follows transport arrival order · case 05
Outline ancestry follows transport arrival order.
Interactive image loses all semantic representation · case 01
Interactive image loses all semantic representation.
Interactive image loses all semantic representation · case 02
Interactive image loses all semantic representation.
Interactive image loses all semantic representation · case 03
Interactive image loses all semantic representation.
Interactive image loses all semantic representation · case 04
Interactive image loses all semantic representation.
Interactive image loses all semantic representation · case 05
Interactive image loses all semantic representation.
Caption text loses priority to tooltip text · case 01
Caption text loses priority to tooltip text.
Caption text loses priority to tooltip text · case 02
Caption text loses priority to tooltip text.
Caption text loses priority to tooltip text · case 03
Caption text loses priority to tooltip text.
Caption text loses priority to tooltip text · case 04
Caption text loses priority to tooltip text.
Caption text loses priority to tooltip text · case 05
Caption text loses priority to tooltip text.
Images without an alt fail to expose a title fallback · case 01
Images without an alt fail to expose a title fallback.
Images without an alt fail to expose a title fallback · case 02
Images without an alt fail to expose a title fallback.
Images without an alt fail to expose a title fallback · case 03
Images without an alt fail to expose a title fallback.
Images without an alt fail to expose a title fallback · case 04
Images without an alt fail to expose a title fallback.
Images without an alt fail to expose a title fallback · case 05
Images without an alt fail to expose a title fallback.
Image recognition labels bypass consent · case 01
Image recognition labels bypass consent.
Image recognition labels bypass consent · case 02
Image recognition labels bypass consent.
Image recognition labels bypass consent · case 03
Image recognition labels bypass consent.
Image recognition labels bypass consent · case 04
Image recognition labels bypass consent.
Image recognition labels bypass consent · case 05
Image recognition labels bypass consent.
OCR from an unfinished image is exposed · case 01
OCR from an unfinished image is exposed.
OCR from an unfinished image is exposed · case 02
OCR from an unfinished image is exposed.
OCR from an unfinished image is exposed · case 03
OCR from an unfinished image is exposed.
OCR from an unfinished image is exposed · case 04
OCR from an unfinished image is exposed.
OCR from an unfinished image is exposed · case 05
OCR from an unfinished image is exposed.
Recognition from a failed image is retained · case 01
Recognition from a failed image is retained.
Recognition from a failed image is retained · case 02
Recognition from a failed image is retained.
Recognition from a failed image is retained · case 03
Recognition from a failed image is retained.
Recognition from a failed image is retained · case 04
Recognition from a failed image is retained.
Recognition from a failed image is retained · case 05
Recognition from a failed image is retained.
Machine inferred alternatives are labeled as authored · case 01
Machine inferred alternatives are labeled as authored.
Machine inferred alternatives are labeled as authored · case 02
Machine inferred alternatives are labeled as authored.
Machine inferred alternatives are labeled as authored · case 03
Machine inferred alternatives are labeled as authored.
Machine inferred alternatives are labeled as authored · case 04
Machine inferred alternatives are labeled as authored.
Machine inferred alternatives are labeled as authored · case 05
Machine inferred alternatives are labeled as authored.
Broken loaded images report a successful ready state · case 01
Broken loaded images report a successful ready state.
Broken loaded images report a successful ready state · case 02
Broken loaded images report a successful ready state.
Broken loaded images report a successful ready state · case 03
Broken loaded images report a successful ready state.
Broken loaded images report a successful ready state · case 04
Broken loaded images report a successful ready state.
Broken loaded images report a successful ready state · case 05
Broken loaded images report a successful ready state.
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 ↗