7.1 KiB
Public mixed read: runtime evidence review
Read-only evidence reconciliation completed. Final source is locally committed bab6f96a5b05b37f2b5d67a430aed0f67d783e73. Parent executed the independently authored data-only checker once, exit0; this reviewer inspected the resulting immutable readback and recorded logs without running the checker again. No compiler, test or CLI was executed by this reviewer.
Final result and binding
Checker SHA256 132729db90aab52ce9fc1e225dbed8dc53338372bd7f0212efd847131efde919; PUBLIC-READ-RUNTIME-READBACK.json SHA256 8cf8a233b3ab3f0383251fe627d0cb443d0248371313661dcdcef4235e74c4ee. Baseline is the accepted ordinary packet ZIP 879ec5a2198898a8a6d792d163710ca1d6ca83f8e7773ee093dcc93456c77378, with its explicit0c07/7b06 component provenance retained.
All562 declared final inputs match immutable Git and current source bytes. Final receipts bind unchanged source/head, empty post-run porcelain status, exact argv, input/log/launcher hashes and actual artifacts. Environment retains jobs1/incremental0, established toolchain/target settings, and absent stack/lifecycle/Rustdoc overrides. No test-name renaming or failure-body normalization is needed for the old library inventory.
- Core listing/full agree at299 actual passes: all298 prior named passes preserved plus the one real public-constructor bridge. Focused public_read passes exactly that bridge.
- Normal-library mixed_source8_read integration passes all ten independently authored external cases, with identical actual list/run names and artifact.
- Seven ordinary integration pairs pass65 total: source_project33, identity3, nested12, source_files8, r3_migration5, checked_save_3, selected source7 migration1. Exact prior name sets agree; these were rerun at bab6 rather than reused across the production Parser change.
- All three rustdoc cases actually pass under the exact reconciliation below. The recorder's name-mismatch failure remains preserved.
- Both ordinary non-test CLI commands pass with the same independently authored complete preview bytes, no write, and unchanged original directory/files. This is ordinary CLI regression evidence, not a mixed CLI feature.
- Strict core all-target Clippy exits0. Format exits1 with exact inherited bytes/hash
c1c6b12861c8117ce6ffc6d6742ebf92618aafb10160ec6f99d370ca72ebab2d; it is not reported as passing.
Core executable:43,348,480 bytes, SHA256 01a051033bb53512a2747c2ef87e366e9b189c14df060d24284db437d39ff96a. Non-test CLI executable:22,534,656 bytes, SHA256 10302f9a0c17e82b57ab6498e2d76c47d6f68388888855b89397660825caac5c. Current bytes were checked by the successful readback. Full raw log SHA256 159e94ce159fcde7314abf45383419fa1199519e54fe2a223b222f25e2e81adb; external raw log 45c09bf0a65b71f9e23aa96d4cdcfbf7fe8f04ee841c8e38f8fd479b5e4a2bfd; Clippy raw log 4061235dd23ba634e9275fc8237a9ccd9f78540980ff740accdf68760c96bb79.
External mixed-source consumer executable:20,148,736 bytes, SHA256 2b8c1f7ca484ce21f94da2d0b0fee7d082474244ebe1f212acad868d8564da22. Full core receipt SHA256 e54b77e90bf965f155a148440449df04929ba1ba063c190a2d29093a1093d058. Readback contains24 final Cargo gate receipts, three preserved historical Cargo attempts, and two actual CLI receipts.
Preserved attempts and finite stack scope
Initial ddf823a non-test integration listing exited101 at E0063 in semantic_source.rs191: the shared Parser initializer omitted mixed outside cfg(test). It produced no artifact, listing or test outcome. Initial source review missed that initializer. The raw failure, unchanged-input manifest, captured launcher and both recorder verification failures remain;8555e4 corrected only the missing initializer's cfg plus note.
At8555e4 the external listing contained ten names and all ten actually passed. Compiler warnings identified unused Budget::limits and the redundant language self import; bab6 gates follow the reviewed test-only observer/import cleanup, without changing API, policy or expected models/diagnostics. The old8555 artifacts are recorded historical identities, not falsely equated to final overwritten binaries.
Final full results include all eleven explicitly named baseline core stack controls in RUNTIME-PLAN.md: ten private analysis controls and the ordinary deep-declaration/positions control. The actual non-test source_nested_updates suite also passes deep_structural_writeback_runs_on_a_small_thread_stack. These support their finite exercised worker/cleanup scopes. No universal headroom bound, full-pipeline setup qualification, JSON depth128 switch or Linux evidence is inferred.
Actual rustdoc spelling discrepancy
Inspected public-read-doc-list-bab6f96-1.log, SHA256 d4b5207694f0bcbf061cc230057c2015adac6bb0754b5727c825e5eaa0d8378f, and public-read-doc-suite-bab6f96-1.log, SHA256 c93d7bfc96884dcbfb54aa2a36ba2294fd48bab036fb676931c54b7b8fd806d2, with the original suite receipt. Cargo exits0; all three actual cases pass in groups1+2 with zero failures/ignored/filtered cases. The recorder nevertheless records Doc runtime names must exactly match the prior nonzero listing and therefore refuses its own verification. That evaluator verdict is retained, not rewritten as a successful recorder run.
Rustdoc lists CheckedMixed at source lines36 and50 without a suffix, then executes them with exactly - compile fail appended. The positive check_mixed_bundle example at line87 has identical names in both records. Independent checker reconciliation therefore maps only these two complete actual names:
crates\oa-core\src\language\analysis_source\mixed\project.rs - language::analysis_source::mixed::project::CheckedMixed (line 36) - compile fail→ the same complete name without the suffix.crates\oa-core\src\language\analysis_source\mixed\project.rs - language::analysis_source::mixed::project::CheckedMixed (line 50) - compile fail→ the same complete name without the suffix.
No generic suffix stripping, path normalization, name exclusion or extra tolerated recorder failure is introduced. The exception binds exact source bab6f96 and raw suite-log hash; it requires the exact three listing names, exact three runtime names/statuses, unchanged positive example, two compile-fail cases, the precise1+2 successful group summaries, Cargo exit0 and the sole preserved recorder failure. The checker output records raw names, explicit mapping and original verifier failures separately.
This is an evaluator spelling correction, not a product/test change or permission to rerun and overwrite historical evidence. Original launchers, listing, log and receipt remain unchanged. The completed independent readback accepts actual successful rustdoc execution under this exact exception while retaining the unsuccessful recorder verdict.
Final disposition: evidence supports the selected borrowed, explicit-ID complete-bundle Rust read API and preserved ordinary behavior. No unresolved runtime evidence discrepancy remains. Public editing/formatting/save/history adoption, mixed CLI discovery, universal allocation guarantees, depth128 qualification and user-deferred Linux work remain outside this checkpoint. Source and evidence are local-only until publication is independently confirmed; parent owns integration and publication.