From 30ebbb88d8e83cf58e4781338da46d47f30cc243 Mon Sep 17 00:00:00 2001 From: Claude Date: Mon, 2 Nov 2026 21:50:26 +0000 Subject: [PATCH 1/2] board: post-merge record for #1932, #1933 and ruff #135/#136 Co-Authored-By: Claude Opus 5.5 Claude-Session: https://claude.ai/code/session_01X6y3drwKSE2zSgoexheLFX --- .claude/board/LATEST_STATE.md | 9 +++++++++ .claude/board/PR_ARC_INVENTORY.md | 31 +++++++++++++++++++++++++++++++ 2 files changed, 40 insertions(+) diff --git a/.claude/board/LATEST_STATE.md b/.claude/board/LATEST_STATE.md index 28c856295..077d89eb1 100644 --- a/.claude/board/LATEST_STATE.md +++ b/.claude/board/LATEST_STATE.md @@ -1,3 +1,12 @@ +## 2026-11-02 — main after #1932 and #1933 (merged `d1b3ea59f`), ruff #135/#136 + +- Codegraph coverage gate: + - `DeadFunction` is `Partial` unless the residuals, dynamic calls and partial crates all show no gap. + - `?` placeholder symbols are refused. + - A residuals-only change is picked up by refresh. +- A10 probes (test-local): the F1/F3/F4/F6 outcomes are recorded, with the admission kinds Given/Observed/Derived/Measured. No production admission writer and no CE64 change. +- NNUE is parked for the XLOG triangle optimizer: the exact-degree fold beats the estimate argmin, with regret 1.000 against 13.1. + ## 2026-11-02 — main after #1929 (merged `4fadef69d`) — plateau - Codegraph reads `.rows.bin` (written by ruff `ruff_charon`, ruff #133/#134). An unchanged refresh reads 64 bytes (2.58 µs). A changed crate still interns every symbol (2.5 ms against a 1 ms gate). This is a plateau; the interner step (B3) is open. diff --git a/.claude/board/PR_ARC_INVENTORY.md b/.claude/board/PR_ARC_INVENTORY.md index 1b82cb55d..504c6c7c4 100644 --- a/.claude/board/PR_ARC_INVENTORY.md +++ b/.claude/board/PR_ARC_INVENTORY.md @@ -1,3 +1,34 @@ +## 2026-11-02 — lance-graph PR #1933 (merged `d1b3ea59f`, branch `claude/clone-repositories-71a5sw`) — A10 probes + admission probe, NNUE triangle baseline, S5 erratum, #1931 record + +- **Added:** + - `cognitive-shader-driver/tests/a10_epistemic_admission_probes.rs` (F1/F3/F6); + - `lance-graph-planner/tests/a10_admission_probe.rs` (test-local Given/Observed/Derived/Measured admission, plus F2/F4/F5); + - `xlog-relcomp-probe/examples/triangle_pairing_baselines.rs`; + - board entries for the SGR recovery gates, A10 admission + NNUE feasibility, and A10 outcomes + S6 + NNUE baselines. +- **Locked (measured):** + - F1: `raw5 = 0` is indistinguishable from an unwritten edge inside CE64. + - F3: `truth_code` and `learn` count a duplicated root. + - F4: a written certification survives the withdrawal of its support. + - NNUE is PARKED for the triangle optimizer: the exact-degree fold has regret 1.000 on 72/72. +- **Not done:** + - The admission probe is not production. + - The rule-identity carrier and the production admission boundary are named gaps. + - The second writer, `deepnsm_v2::wire`, needs an owner decision. + - The S6 fork change (xlog `ea924b3`) is compiled but not run. +- **Confidence:** each guard was disable-verified red-then-green; the triangle numbers were re-run independently. + +## 2026-11-02 — lance-graph PR #1932 (merged `3355be0b8`, branch `claude/sgr-coverage-gates`) — SGR recovery: residuals + coverage gate, `?::?` refusal; with ruff #135/#136 + +- **Added:** + - codegraph reads `.residuals.tsv` and gives a typed `Coverage`; + - `DeadFunction` runs `Partial` unless there is no known gap; + - a placeholder (`?`) symbol segment is refused at open and at refresh; + - refresh skips only when the rows and the residuals are both unchanged (codex P2, `ccb7cba44`). +- **ruff #135** (merged `9eb6e11e76`): Charon `?::?` identity residual; typed `.effects.tsv`, with assert kinds and DEREF_WRITE Arg/Local covered by the `derefw.rs` fixture. +- **ruff #136** (merged `667936eebb`): C++ resolution status (Resolved / Lookup / Dynamic) with namespaces kept; constructors resolve; a call through a returned function pointer is dynamic. +- **Locked:** the absence of a CALLS/REFERENCES row proves nothing unless coverage is complete. +- **Confidence:** each gate has a can-fire and a can-stay-silent test, all disable-verified. + ## 2026-11-02 — lance-graph PR #1931 (merged `ef6f34999`, branch `claude/clone-repositories-71a5sw`) — /coresearch Semantic Grammar Recovery map; #1930 post-merge record - **Added (board only):** `entries/2026-11-02-coresearch-semantic-grammar-recovery.md`. It contains: From 7c8b8fa4782aa46395ec20f5b649f36e73abf02d Mon Sep 17 00:00:00 2001 From: Claude Date: Mon, 2 Nov 2026 21:51:47 +0000 Subject: [PATCH 2/2] board: remove private-repo specifics from the NNUE second-application note The residue-ranking note named a private consumer repo, its exam and one of its refused concepts. It now describes the target and the gate without them; outcome labels stay on the consumer side. Co-Authored-By: Claude Opus 5.5 Claude-Session: https://claude.ai/code/session_01X6y3drwKSE2zSgoexheLFX --- .../2026-11-02-a10-probes-s6-magic-nnue-baselines.md | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/.claude/board/entries/2026-11-02-a10-probes-s6-magic-nnue-baselines.md b/.claude/board/entries/2026-11-02-a10-probes-s6-magic-nnue-baselines.md index 231a1543c..073e3a9b4 100644 --- a/.claude/board/entries/2026-11-02-a10-probes-s6-magic-nnue-baselines.md +++ b/.claude/board/entries/2026-11-02-a10-probes-s6-magic-nnue-baselines.md @@ -63,10 +63,10 @@ AdaWorldAPI/xlog `claude/magic-query-relative-demand` @ `ea924b3`. Second application (Ruff × ARM × SoC): - **Verdict: NEEDS-GROUND-TRUTH-FIRST.** -- **Candidate target:** ranking the next residue on the MedCare furnace exam. The deterministic baseline already exists: `rekey_exam` sorts by method count. +- **Candidate target:** ranking the next unresolved residue in a consumer transcode. The deterministic baseline already exists in ruff: `rekey_exam` sorts by method count. - **Not applicable yet:** - Ranking Concern and Scope splits: nothing generates candidate splits. - ARM: it has no `FeatureSpec` over ruff facts. - CB2Vec and NORU: absent from the workspace. -- **Gate before any trial:** a private outcome TSV in MedCare-rs and a measured precision@15 baseline. -- **Kill condition:** the ranker ranks the gate-refused `user_right` above a confirmed mint. +- **Gate before any trial:** an outcome table kept on the consumer side, and a measured precision@15 baseline. Outcome labels are not recorded here. +- **Kill condition:** the ranker ranks a residue the mint gate refused above a confirmed mint.