Research consolidation: authenticate PR279's complete parent inventory (no new kappa)

docs/research/pr-ledger.md

@@ -28,7 +28,7 @@ This frozen round covers #543/#545/#555 and related construction, research and t
 1. Follow the new Linux CI for the published #442/#484/#509/#516/#520 batch; the [scoped review](community-formal-20261012-review.md) records completed local checks.
 2. Follow the explicit construction-family dispositions and remaining obligations; do not repeat completed foundation replays without a source/input change.
 3. Review #525’s checkpoint/ceiling and #529/#532’s distinct construction mechanisms; retain parallel credit.
-4. Reconcile #210/#279 updates with their earlier incorporated heads; audit #236’s source-package integration separately from its already published review receipts.
+4. Reconcile #210's remaining ancestral paths and audit #236's source-package integration separately from its published review receipts. For #279, the [commit-authenticated parent inventory](../../research/pr279-parent-closure/README.md) closes the 116-path source-coverage question left by batch970 using #828's retained sources and replay; acceptance and parent disposition remain separate maintainer decisions.
 5. Continue finite-instance connections for formal interfaces; abstract theorem builds do not instantiate all current physical constructions.
 
 ## Starting PR inventory

research/pr279-parent-closure/OVERLAP.json

@@ -0,0 +1,63 @@
+{
+  "date": "2026-10-13",
+  "base": "1b59b624d21924f8e3537cdc1ac52d3be6eda2fe",
+  "queries": [
+    "gh pr list --state open --limit 200",
+    "gh pr list --state all --search 279 --limit 40",
+    "gh pr list --state all --search PR279 --limit 40",
+    "discussion 1166 latest 30 comments; PR279 live head"
+  ],
+  "explicit_pr279_results": [
+    {
+      "headRefOid": "f063aab8d6e514188e53db57218515df9cdd186e",
+      "number": 280,
+      "state": "OPEN",
+      "title": "Conditional \u03ba = 0.000724189082265949: gen4 retiming, 440 early restorations and 60 dirty entrances",
+      "url": "https://github.com/CrocSwap/integer-mult-bounds/pull/280"
+    },
+    {
+      "headRefOid": "e2fae6a92775b8f609d917a9cc93b31e2559dd5f",
+      "number": 372,
+      "state": "OPEN",
+      "title": "Refine reviewed gen4 envelopes: conditional kappa = 0.000723571593879006127174",
+      "url": "https://github.com/CrocSwap/integer-mult-bounds/pull/372"
+    },
+    {
+      "headRefOid": "2e43537c0e8ed1c9f9ef21cab0ef9cddb5d79976",
+      "number": 828,
+      "state": "MERGED",
+      "title": "Research consolidation: reconcile PR279 target-prefix sources and replay",
+      "url": "https://github.com/CrocSwap/integer-mult-bounds/pull/828"
+    },
+    {
+      "headRefOid": "6fe78cbe3a5499acabdeefd7f91ea710a20f6f8a",
+      "number": 393,
+      "state": "CLOSED",
+      "title": "Deep retiming",
+      "url": "https://github.com/CrocSwap/integer-mult-bounds/pull/393"
+    },
+    {
+      "headRefOid": "f6092eac77d965d6dc197cf796b0a3580382c0ef",
+      "number": 526,
+      "state": "OPEN",
+      "title": "[Lean] Bound automatic-width counting by 12n+4",
+      "url": "https://github.com/CrocSwap/integer-mult-bounds/pull/526"
+    },
+    {
+      "headRefOid": "222fb22c0ae2d13c25674cb8c18d168d8518cef5",
+      "number": 930,
+      "state": "MERGED",
+      "title": "Research consolidation: reconcile and replay PR210's parity/fusion update (no new kappa)",
+      "url": "https://github.com/CrocSwap/integer-mult-bounds/pull/930"
+    }
+  ],
+  "parent_refresh": {
+    "headRefOid": "913ccc48a0bc1173709ea92280b96252256b7754",
+    "state": "OPEN",
+    "updatedAt": "2026-10-10T05:21:07Z"
+  },
+  "claim_urls": [
+    "https://github.com/CrocSwap/integer-mult-bounds/discussions/1166#discussioncomment-18908144"
+  ],
+  "distinction": "PR828 already reconstructs/replays the update; this contribution authenticates both original Git commits/27 trees, every matched parent path/blob, and the repository-wide increment. No new mathematical replay or disposition claim. PR116 and selected PR1042 provenance audits concern different heads."
+}

research/pr279-parent-closure/README.md

@@ -0,0 +1,132 @@
+# PR279: commit-authenticated parent inventory closure
+
+**Research consolidation; no new bound or mathematical replay.** The batch970
+[companion disposition](../community-batch970-audit/COMPANION-REVIEWS.json)
+explicitly leaves reconciliation of PR279's original **116-file parent
+inventory** open after incorporating #828. This contribution supplies an offline
+Git-object proof that resolves that source-identity question, without changing
+PR279's disposition or the selected result.
+
+## Finding and distinction from #828
+
+[mihir-s-05's #828](https://github.com/CrocSwap/integer-mult-bounds/pull/828)
+already reconstructs all 115 later package files and the unchanged workflow,
+checks their declared Git blobs/SHA-256 digests, and retains a nine-stage replay.
+It must receive credit for that reconstruction and execution: neither is new
+here. Its checker verifies the twelve unresolved path/blob pairs and the count
+of 104 previously matching paths, but does not authenticate the latter path/blob
+pairs to a repository commit. Its declared commit IDs and file inventory do not
+contain Git tree/commit payloads from which that relationship can be checked.
+
+The new checker authenticates both raw commits and their relevant Git trees,
+then binds the **actual retained bytes** to their complete package subtrees and
+workflow entries. It also compares the entire two commit roots, recursively
+opening every differing tree and skipping only equal authenticated object IDs.
+Consequently the delta census is repository-wide, not just a search for changes
+inside a presumed package root. Exact file modes are retained in that census.
+
+The result distinguishes three boundaries:
+
+| Boundary | Exact coverage |
+|---|---|
+| Earlier reviewed package | 113 files authenticated to `4d8c31eec7ae618645cb058101cb4e72a97268d3` |
+| Later parent inventory | 115 package files plus one unchanged workflow authenticated to `913ccc48a0bc1173709ea92280b96252256b7754` |
+| Repository-wide increment between those commits | Ten modified files and two added files, exactly #828's twelve-file archive |
+
+The 116 paths are exactly the union of the original comparison's 104 matched
+paths and twelve unresolved paths, with **every path and Git blob** checked,
+not only the counts. The candidate commit's direct parent is the reviewed head.
+No additional repository change exists between those two authenticated trees.
+This is not a claim that every unrelated inherited repository file was replayed.
+
+The remaining batch970 instruction to reconcile the original parent inventory
+is therefore satisfied at this pinned head. Whether the source and already
+recorded scoped review warrant accepting or closing PR279 is a separate
+maintainer decision. The historical review and central disposition remain
+unchanged; the original source packages, archives and selected frontier are not
+modified.
+
+## Reproduce
+
+Python 3.11+ and the repository's retained source/evidence files suffice. No
+network, Git installation, `.git` directory, SymPy, or mathematical replay is
+needed by the verification commands:
+
+```sh
+python3 -B research/pr279-parent-closure/verify.py
+python3 -I -O -B research/pr279-parent-closure/test_verify.py
+```
+
+The first emits `PASS_PR279_PARENT_COMMIT_CLOSURE`, `parent_paths: 116`,
+`retained_paths: 104`, `companion_paths: 12`, and repository changes
+`added: 2, modified: 10`. It reports nine **retained** replay stages and
+`fresh_mathematical_execution: false`. The second exercises the positive case
+and 17 negative controls: altered historical/companion/workflow bytes;
+omitted/extra files; a later file substituted into the reviewed base; duplicated,
+omitted or false-blob inventory entries at unchanged cardinality; swapped
+coverage partitions; missing/misclassified updates; forged/substituted commits;
+forged/missing trees; and an unrelated authentic tree object.
+
+`TREE-PROOF.json` stores raw Git object payloads in base64. Git's exact
+`type + space + decimal-length + NUL + payload` SHA-1 identity is recomputed
+before use. Relevant tree leaves are bound to the retained bytes; every provided
+tree object must be used. The fixed source head is the trust anchor, not a
+mutable API response or an unauthenticated replacement file list. As with Git
+source pinning elsewhere, this relies on the integrity of the pinned Git SHA-1
+identities; it does not certify authorship or establish a signature trust chain.
+
+To independently recapture the proof from fetched original Git objects, write
+to a **new** output path (the collector refuses overwrite):
+
+```sh
+git fetch origin pull/279/head
+python3 -B research/pr279-parent-closure/capture.py --output /var/tmp/pr279-tree-proof-new.json
+```
+
+The collector reads the fixed heads, never `FETCH_HEAD`, and changes no source
+checkout. The result is deterministic. The proof verifier does not invoke or
+trust this collector. No compressed source artifact is regenerated.
+
+## Pins, bounded process and evidence reuse
+
+- Research base: `1b59b624d21924f8e3537cdc1ac52d3be6eda2fe`.
+- Fresh OS-random assignment seed: **97**, research consolidation.
+- Reviewed PR279 head: `4d8c31eec7ae618645cb058101cb4e72a97268d3`.
+- Later PR279 head: `913ccc48a0bc1173709ea92280b96252256b7754`.
+- #828 head: `2e43537c0e8ed1c9f9ef21cab0ef9cddb5d79976`.
+- The inherited #828 checker, source inventory and evidence inventory SHA-256
+  pins are explicit in `verify.py`; its unmodified source and evidence checks
+  are reused. No replacement verifier or rerun of its nine mathematical stages
+  is claimed.
+- Local claim: [discussion 1166 comment](https://github.com/CrocSwap/integer-mult-bounds/discussions/1166#discussioncomment-18908144).
+
+Three candidates were compared: (1) this complete PR279 parent inventory;
+(2) PR210's broader 128 unresolved ancestral paths; (3) PR236's source-package
+integration, distinct from its already published review receipts. The bounded,
+explicitly recorded PR279 gap was selected instead of a newest-frontier audit.
+The stopping rule was closure or a concrete unresolved path for the two fixed
+heads and original 116-path inventory, not a search for a stronger exponent.
+There is no failed search or impossibility theorem here.
+
+Nearest related work: #828 supplies the source/replay foundation; #1007 concerns
+a different PR116 evidence-only head; #1142 concerns the selected PR1042 replay
+capsule, not this historical parent commit; #1132 distinguishes R480/R492
+receipts; #1110 reconciles PR225 pricing. The current #1094 conditional selected
+value at the research base is `836330098538459/500000000000000000`; this work
+neither changes it nor claims historical PR279 values as current. PR279's
+all-size/compiler, chart, restoration, routing, prime-supply, precision/recovery,
+complex symbolic and analytic-transfer assumptions remain inherited. This is
+not a formal multiplication theorem, new acceptance decision or hosted CI run.
+
+## Attribution and license
+
+Prepared by **Adam Gall (@adamgall)** with substantial **OpenAI Codex assistance**
+in research, implementation, source comparison, controls and writing.
+**mihir-s-05** authored #828's retained reconstruction and replay. **Rohan Arun**
+authored PR279's descent/integration, with the original Anthropic Claude
+assistance disclosed in its commit; **eumemic** supplied the PR268 target-prefix
+transform; **DreamingOfClouds** supplied the gen4 word/package. **Douglas Colkitt**
+supplied the review and reconciliation context. Henry Grant, Jacob Sussman,
+Evan McKinney, Chafik Boukhalfa, Avi Eisenberg and the full source lineage retain
+their original credit and notices. No independent human peer review is claimed.
+New work is Apache-2.0; all original licenses and source bytes remain intact.

research/pr279-parent-closure/RESULTS.json

@@ -0,0 +1,40 @@
+{
+  "base_commit": "1b59b624d21924f8e3537cdc1ac52d3be6eda2fe",
+  "assignment_seed": 97,
+  "track": "research consolidation",
+  "python": "Python 3.14.4",
+  "command": "python3 -B research/pr279-parent-closure/verify.py && python3 -I -O -B research/pr279-parent-closure/test_verify.py",
+  "exit_code": 0,
+  "verification_stdout": {
+    "authenticated_commits": 2,
+    "authenticated_trees": 27,
+    "candidate_head": "913ccc48a0bc1173709ea92280b96252256b7754",
+    "candidate_package_tree": "4c7079b88c9f6ab5b597c50073782d039a24ecc8",
+    "candidate_tree": "f419e848c935cfa7b666be39797edb245b2a9e91",
+    "companion_paths": 12,
+    "fresh_mathematical_execution": false,
+    "parent_paths": 116,
+    "repository_changes": {"added": 2, "modified": 10},
+    "retained_paths": 104,
+    "retained_replay": {
+      "fresh_execution": false,
+      "retained_stages": ["banks", "bit", "complex", "finite", "math", "primes", "raw", "scalar", "virtual"],
+      "scope": "Checks stored evidence; use --replay for a fresh run.",
+      "status": "PASS_RETAINED_REPLAY_BINDING"
+    },
+    "reviewed_head": "4d8c31eec7ae618645cb058101cb4e72a97268d3",
+    "reviewed_package_tree": "70d8599cc7f840251a4298105ada23dd2730a3c7",
+    "reviewed_tree": "9620d20482a4ef68dddeb0d081ec64025b8ec59e",
+    "scope": "Source identity and retained evidence only; parent disposition remains a maintainer decision.",
+    "status": "PASS_PR279_PARENT_COMMIT_CLOSURE"
+  },
+  "tests": {
+    "positive_cases": 1,
+    "negative_controls": 17,
+    "summary": "Ran 18 tests in 17.764s\n\nOK",
+    "optimized_isolated_interpreter": true
+  },
+  "not_executed": ["new mathematical replay", "selected frontier cold replay", "Lean", "hosted CI"],
+  "retained_mathematical_evidence": "research/pr279-target-prefix-reconciliation/EVIDENCE.json",
+  "claim_url": "https://github.com/CrocSwap/integer-mult-bounds/discussions/1166#discussioncomment-18908144"
+}