6.0 KiB
Independent entry-producer runtime readback
Source 29417130c71f70c4eaefe8ed4c879efba78eb72c passes the specified bounded regression readback. No checker exception or relaxation was needed. This is not a wholly passing suite:16 inherited tests fail and workspace format exits1. No compiler, test binary, debugger or launcher was executed by this reviewer.
Readback identity and gates
Executed only the prepared read-only inspect_runtime.py, using bundled Python -I -B with arguments 29417130c71f70c4eaefe8ed4c879efba78eb72c 2941713 6. Checker SHA256 792b2eff0a9288a73bad9495b05ca06927ca0717a176dce5913206ffa9731080 was unchanged. Exclusive result ENTRY-PRODUCER-RUNTIME-READBACK.json SHA256 f45c86d42a1d89c4cf2b22342f2c41dcd4514622f0bd2367101f219367d1c66c; checker exit0.
| Gate | Exit | Raw evidence |
|---|---|---|
| Focused | 0 | Six entry-producer tests passed,275 filtered |
| Full | 101 | 281 unique results:265 passed,16 failed, zero ignored |
| List | 0 | 281 unique names, exactly full runtime inventory |
| Strict core all-target Clippy | 0 | Locked/offline -D warnings completed |
| Workspace format | 1 | Identical prior raw output, SHA256 c1c6b12861c8117ce6ffc6d6742ebf92618aafb10160ec6f99d370ca72ebab2d |
All275 predecessor outcomes are unchanged. Exactly six new names occur, all passed, and focused is exactly that new set. All16 inherited failure bodies are identical after numeric thread-PID replacement only. There is no backtrace-hint exception: the97 baseline already contains the previously documented hint removal. All ten existing stack-named controls in the final full raw log passed.
The checker binds baseline source 97a8703adc75467a7f8e75f5c6cb00ca542b707c full/list/fmt logs and receipts byte-for-byte to the locally committed text-producer ZIP, SHA256 252ea94e7c7f500a2afb3d0fa47c43a051f39d9c844328337289eea0c463becd. This review makes no remote publication claim.
All five receipts bind the clean current source,557 declared inputs, raw Git blob versus disk equality, unchanged inputs through each command, exact commands/environment and the same captured launcher. Manifest SHA256 3453c4b85aa27e79fa602e491e333ddeb42914228bb6e3dda8efbabbbe14f02c. Final focused/full/list record the same actual current executable,42,898,944 bytes, SHA256 adb0bd907f614c512bc3883dfe9a88ca0e4268638d85f4899ad07a5c1d3ec4e9; checker reread those bytes. Individual log/receipt pins are preserved in the readback JSON. Full raw log SHA256 47fef6f46d136be0e53f48f8ff03a96be73635fc4c3f57350c1e641ad44b01c4.
Launcher review
Read actual captured launcher SHA256 43a25f27b417cc66f81ec1714c57eb079a9923995bdadbe79440b00836ac73fc, without executing it. It hardpins2941713, checks clean tracked source and raw declared files before launch, captures exact helper bytes, records unfiltered stdout/stderr and actual process return code, checks source/input/helper identity afterward, and records executable hashes from actual Cargo Running lines. The readback separately re-parses results and rejects duplicate/missing names and incomplete summaries rather than trusting the recorder's summary alone.
Commands remain locked/offline oa-core library focused/full/list, locked/offline core all-target Clippy, and workspace format. Jobs1/incremental0 and baseline environment match; RUST_MIN_STACK, RUSTFLAGS, custom RUSTC, RUSTUP_TOOLCHAIN and lifecycle override are absent. This establishes recorded declared-input/artifact concordance, not independent compiler provenance or completeness of every undeclared/untracked dependency.
Original compile failure retained
Initial source a629ae01d49bb3cb482a047eb41832ad4246a5c6 focused attempt exited101 after26.156 seconds, before any tests. Raw E0277 identifies entry_producer.rs:241 unwrap_err requiring Debug for successful Vec<(Uuid, EditDeclaration)>. Receipt has zero outcomes and no artifacts. Independently hashed raw log cfd534fb67405666ba7a86aa3aa3164655d803b00d4e80f09893d1e312c2a58a, receipt 0490608967cc91ac5e666453304a657a01743508d3d98aa9b3e469482b3629f0; log matches receipt pin.
The previously reviewed2941713 repair replaces only that extraction with explicit Err/Ok matching, retaining both quota refusal assertions; production unchanged, implementation note updated. Final test SHA256 ebdcd69eae66103a7ed596356e5f629099399d944b5659ace54edc9f558e7bc9. Original failed evidence is not replaced by later successful execution.
Stack observations and scope
Read parent STACK-OBSERVATIONS and raw prior/final stack, active and projector JSON disassemblies. Final selected reservations match the stated values: parse_impl12888, parse_with_profile216, parse280, accepted1608, indexed1608, actual worker7368, Parser::unit12456. Their selected changes versus97 sum2048 bytes. Typed/numeric/Pratt selected reservations remain unchanged; binding2520 plus prefix11720 is14240, and continuation13416 is an alternative, not simultaneously added.
Projector records expand from nine baseline to17 final specializations. Raw mixed projector changes616→1032 (+416). Ordinary old body584 becomes wrapper80 plus producer1000=1080 selected reservation bytes; outer wrapper remains72. Specialized token/value/fields are120/536/216; new quota-consume120; injected quota test has a separate1000-byte specialization. These immediate prologue reservations exclude pushes and other runtime overhead; neither their sum nor record count is total live stack/headroom. Projection occurs after Parser::unit returns, so combining both phases as a single measured chain would be false.
All old stack controls passed in this full run, but the six new entry tests run on ordinary test-harness stacks; this does not establish a new128KiB strong-mode pipeline guarantee. Prior extra process repetitions, debugger observations and historical failures retain their own source/scope. No additional runtime, retry or stack override was used for this review.
Admission is limited to these finite retained-entry evaluators and unchanged baseline disposition. It does not establish all-heap accounting, accepted rewritten output, public constructor authority, GUI behavior or unresolved scratch-domain closure.