Regenerate docs data from saved full test262 transcript

Closed chris opened this 11 days ago ·

chris 11 days ago

Scope

Save a full parent test262 transcript from the current runner.
Regenerate docs/.data/test262.json from that transcript with scripts/gen-test262-data.ts.
Restore per-suite docs rows only from parsed transcript evidence.
Keep docs/DOCS_ACCURACY_PLAN.md aligned with the actual workflow.

Exit criteria

docs/.data/test262.json top-level totals and suite rows come from a saved run transcript.
docs build passes.
README and docs do not contain hand-entered stale conformance numbers.

Notes

Current verified configured result: VALID 48368/48368, NEGATIVE 4669/4669, host-fail 0, skipped 140.
If no transcript is available, per-suite rows must remain empty rather than fabricated.

chris 11 days ago

Implemented in ec91da9. Committed docs/.data/test262-run-2026-07-04.txt and regenerated docs/.data/test262.json from it with bun run docs:data -- --from docs/.data/test262-run-2026-07-04.txt. Final saved run: VALID 48369/48369, NEGATIVE 4669/4669, parse/runtime/host failures 0, skipped 139. Fixed the parser to preserve full suite paths, including RegExp shard names with colons; JSON now has 282 evidence-backed suite rows.