Commit History
-
clippy: clear the 1.81 gate, behaviour-preserving
Jesse Daniel Brown +1: Claude Opus 5 (1M context)Verified on real rustc 1.81.0 (eeb90cda1) before pushing: cargo fmt --check clean, cargo clippy --all-targets -- -D warnings clean, tests pass, zero f32/f64. format_collect: same hex8 helper, same fix (src/lib.rs:449). 10 tests pass. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>cdbc145 -
toolchain rule: Rust 1.81 + clippy, integer/ternary only
Jesse Daniel Brown +1: Claude Opus 5 (1M context)Pin channel 1.81.0 with clippy+rustfmt, rust-version = "1.81", and CI that runs clippy --all-targets -- -D warnings plus a grep that hard-fails on any f32/f64. Nothing was pinned before this: no rust-toolchain.toml existed anywhere in the corpus and 32 different Rust versions were in circulation. Receipt lines recording past runs above 1.81 are deliberately NOT rewritten here -- editing a line that records a run would make the receipt false. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>4be1a2e
-
Apply rustfmt attack-verify fix
Kevin Cruttce80562
-
Add MtpMovementMatrix — 'time travel' as reversible retrodiction, with the arrow-of-time falsifier
MTP-1/2/3 measure the total movement matrix between two slices (MTP1=pixel delta, MTP2=frequency/shell, MTP3=cylinder/residue); the lens project_reverse/project_forward projects past/future slices from it = 'time travel' as reversible retrodiction (byte-identical, both directions). FALSIFIER: clamp_step is an IRREVERSIBLE step (velocity zeroed at a wall = info lost); measure sets reversible=false and project_reverse does NOT recover the true past -> the ARROW OF TIME bites. So it is COMPUTING a determined reversible trajectory, NOT physical time travel: exact where reversible, Held/wrong where irreversible. MTPMATRIX hbp row json=0. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DAi1nuxgvuMBwsSsvc4eeTf8ee7f1 -
test the watcher gate + omnibit pixel (tamper-caught, selector-not-payload)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>47468ae -
Add watcher gate (yin-yang black<->white round-trip) + omnibit pixel
WatcherGate: reconstruct from sufficient shadows, cross-check EXTRA cylinders (over-determination = hallucination-detection budget) -> VerifiedClone (classical, no-cloning respected) or Held. Tamper a sufficient shadow -> cross cylinders catch it -> Held::WatcherDisagreement. OmnibitPixel: one pixel as selectors+checks (pos/tick/freq + fold-digest8 + residual_bits + capacity_margin>=0 + watcher_bound), NOT raw payload. GNN/reverse-GNN = DESIGN scaffold; round-trip+consistency+Shannon = MEASURED. Bilateral convergence w/ liris PR#1 watcher gate + omnibit (independent implementations). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DAi1nuxgvuMBwsSsvc4eeT6d113a4 -
Deterministic slice-time harness — 0-loss slice-time extrapolation, or Held
The capstone: metatagged particle sim (deterministic, reversible torus drift) -> multi-cylinder shadow projections -> Q-PRISM/Path-2 state reconstruction -> known rule forward+backward -> byte- identical vs truth, or Held. 0-loss ONLY on the DETERMINED component; NEW ENTROPY is DETECTED, never manufactured (verify_or_detect_entropy). Schrodinger-free because simulated (no Heisenberg: exact state from memory; no Born: deterministic rule) - Laplacian because never quantum. Honest tags: CANON/DESIGN (the theorem), MEASURED (this crate's tests), UNVERIFIED (live/real-quantum/new-entropy). Zero deps, json=0, no JSON no Node. Closes the arc: papers->capstone->Path1->Path2->3D Q-PRISM->slice-time. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DAi1nuxgvuMBwsSsvc4eeT5f6f6d6