Repository navigation
chore(r3-debt-paydown): velocity-tripwire script + first 2026-05-02 measurement (FIRED 25.33:1) - #1534
Merged
Conversation
briansrls
marked this pull request as ready for review
May 2, 2026 17:08
This was referenced May 2, 2026
Closed
briansrls
added a commit
that referenced
this pull request
May 7, 2026
… rule (#2101) * WIP: R3 Debt-Paydown Mgr — standing program through R3 close * docs(r3): cite specific Director-ratification comment URL in §7.6.2 Per Director micro-suggestion on #2101 review — match #1534 source-column URL discipline so future readers can navigate to rationale without searching. Debt receipt: no debt introduced; no debt retired (review-absorption edit). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
May 7, 2026
* WIP: R3 Debt-Paydown Mgr — standing program through R3 close * docs(r3): cite specific Director-ratification comment URL in §7.6.2 Per Director micro-suggestion on #2101 review — match #1534 source-column URL discipline so future readers can navigate to rationale without searching. Debt receipt: no debt introduced; no debt retired (review-absorption edit). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): §7.6.1 log intra-cadence tripwire reading 1.56:1 (466 PRs) Adds a 2026-05-07 intra-cadence re-check row to the readings log. Same 7-day window as Wave 3, advanced ~3.5h. D-side grew 47 → 54 while I-side held at 84; ratio improved from 1.79:1 to 1.56:1. Confirms continued dissolution-side accrual — no breach. Debt receipt: no debt introduced; no debt retired (readings-log entry). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
May 9, 2026
…ng briefs (Phase 1+2) (#2437) * WIP: R3 Debt-Paydown Mgr — standing program through R3 close * docs(r3): cite specific Director-ratification comment URL in §7.6.2 Per Director micro-suggestion on #2101 review — match #1534 source-column URL discipline so future readers can navigate to rationale without searching. Debt receipt: no debt introduced; no debt retired (review-absorption edit). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): §7.6.1 log intra-cadence tripwire reading 1.56:1 (466 PRs) Adds a 2026-05-07 intra-cadence re-check row to the readings log. Same 7-day window as Wave 3, advanced ~3.5h. D-side grew 47 → 54 while I-side held at 84; ratio improved from 1.79:1 to 1.56:1. Confirms continued dissolution-side accrual — no breach. Debt receipt: no debt introduced; no debt retired (readings-log entry). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * WIP: R3 Debt-Paydown Mgr — standing program through R3 close * docs(r3): r3-debt-sweep §1 Class A-G populated + Class P added — Phase 1 organize cycle Per Brian operator directive 2026-05-09 ~20:30Z (lane scope expanded to all v3 debt; active posture), Debt-Paydown Mgr (gentle-newt-665) Phase 1 organize cycle replaces 7 [Mgr canvas populates rows] placeholders with initial substantive rows seeded from: - docs/audit/v3-comprehensive-debt-audit-2026-05-09.md 10-source inventory (PR #2436) - ROADMAP ## Tracked debts (10 sections; ~40 still-open rows) - r3-program-plan §10.3 ratifications + §1.8 ledger gate references - 4 stranded bug-fix briefs (PR #2373: u128 mirror / FieldProject dual-authority / resolve_producer_opt / CallGraph) - 6 gpt-5-5-pro novel findings (PR #2358 §8 meta-finding cycle) - SG-0 census untracked entries (113 of 159) + drift-class anti-patterns Class A-G total ~83 rows; new Class P (past-R3 / R4+ ambition parking lot) ~8 rows per Brian directive. Mgr re-organize cycle continues; initial pass not exhaustive. Phase 2 dispatch list surfacing to PM #846 follows in separate comment. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): 6 worker briefs for gpt-5-5-pro novel findings F1-F6 Authored under Mgr standing authority per PM confirmation at #846 c#4413701937 (Brian operator directive 2026-05-09 active posture). Mirrors PR #2373 brief template. Briefs: - F1 MissingEmissionPath typed-axes substrate (Class C; HIGH; Substrate Mgr) — substrate stringifies axes that Rust mirror correctly types - F2 ShapeATarget closed enum vs LanguageSpec extensibility (Class F; HIGH; Grounding Mgr) — closed Rust enum negates thesis extensibility claim - F3 Map<String,Bool> as set (Class F; MEDIUM; Substrate) — Set<A> declared but bypassed; set_has helper - F4 PartitionResult anonymous bypass (Class G; LOW-MEDIUM; Substrate) — named type declared but consumer uses anonymous return - F5 ComposedEffect illegal product (Class C; HIGH; Substrate) — idempotent + breaking_operation contradictory pair representable - F6 derive_op_effect string parser at structural boundary (Class C; HIGH; Substrate) — string-keyed dispatch over typed HttpMethod/Path carriers Each brief: problem statement with file:line citations + 1-2 fix options with PM/Mgr recommendation + receipt criteria + cross-cutting constraints. Cross-references sweep-doc Class A-G rows for traceability. Phase 2 dispatch under Brian-directive active posture. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): SG-0 census per-entry classification (Phase 2(3) deliverable) Per PM dispatch at #846 c#4413701937 (Brian directive 2026-05-09 active posture; Debt-Paydown Mgr standing authority). scripts/classify-sg0-census.py: heuristic regex classification of preceding-comment-block + path against Class A-G/P framework. Output: docs/audit/sg0-census-classification-2026-05-09.md — per-entry table + summary. Initial pass: 159 entries (104 TEST + 53 NON_TEST + 2 FRAGMENTS) — 115 untagged (no preceding dissolution-trigger comment), matching audit's "113 of 159" estimate. Heuristic class breakdown (initial): - A 0, B 4, C 1, D 3, E 7, F 0, G 143, STRUCTURAL 1 Coarse-grained — heuristic sweeps Class E (Tier3/v2-retirement/T-LensProducer) reliably; substrate-gap (Class A) and ontology (Class F) need Mgr-eye review on the 115 untagged entries. Initial pass is dispatchable input, not ratified classification. Mgr next-action documented in output §"Mgr next-action": review heuristics → confirm-or-correct → fold per-entry classifications back into r3-debt-sweep-2026-05-06.md §1 Class A-G rows. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): one-source SG-0 untagged-count derivation (codex BLOCKING fix sha 03a8830) Codex flagged: hard-coded SG-0 untagged-count appeared in both sweep doc Class G row + classification doc Summary, creating drift risk. Fix per "one checkable value" discipline: - sweep doc Class G row 16: drop hard-coded "159 entries; 113 untagged"; cite classification doc §Summary as live source. - classification doc preamble: explicit "single source for SG-0 untagged-count" framing; regenerated by `scripts/classify-sg0-census.py` from `sg0_census_test.rs`. Other consumers of the count (audit prose at PR #2436) are owned by PM; my PR's contribution is making my surfaces non-duplicating. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): supersede-disambig marker on TC3 D4 brief line 121 (CI R4-carve discipline fix) CI scripts/check-r4-carve-dissolution-discipline.sh fail-closed on docs/briefs/r3-evaluator-tc3-d4-eval-step-producer-worker.md:121 — the line read "(α)/(β) novel-substrate-introduction explicitly carved to R4+ via Class P partition" without a supersession marker. Disambiguating note added inline: "(Note: this is a novel R4+ partition for (α)/(β); does NOT supersede or invoke the DISSOLVED prior C1/C2/C3 carves per Director carve-promotion-IN-R3 at gunbc#846 #issuecomment-4412330468.)" The clarification is structurally accurate — Class P partition is a NEW R4+ routing for specific Pattern-A items, distinct from the prior C1/C2/C3 carves which Director DISSOLVED + reclassified R3-load-bearing within Cluster F (per gunbc#846 #issuecomment-4412330468). Marker "supersede" + "DISSOLVED" both present satisfies the discipline. Surfaced fix scope justification: PR #2437 introduces Class P partition; the brief at line 121 forward-references it. Cleanest fix is on the citing brief, not the discipline script (preserves single-Mgr-ownership of the Cluster F discipline). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): refine SG-0 classifier with path-based heuristics (better Class A/C/D/E surfacing) Per PM observation at gunbc#846 c#4413720143 — Class A=0/F=0 from comment-only heuristic was the substrate-gap and ontology-duplication semantic gap; path-based rules can fire on canonical compiler-Rust-mirror locations. Added PATH_RULES tier (fires when CLASS_RULES miss): - src/v3/compiler/src/dag.rs / lower.rs / infer.rs / emit*.rs / value.rs / eval/ → Class C - src/v3/compiler/src/parse / tokenize / test_runner → Class A (substrate-gap parser/grammar) - src/v3/compiler/src/bin/regen_ / bootstrap → Class D - src/v3/compiler/benches → STRUCTURAL - src/v3/lenses → Class E (T-LensProducer) - src/v3/grounding_pilot / grounding_cross_target / spec / std → Class C Class breakdown delta: - A: 0 → 1 - C: 1 → 16 - D: 3 → 12 - E: 7 → 9 - G: 143 → 115 (default drops as path-based rules catch real classes) 115 still-G are entries that don't fit canonical compiler patterns — test fixtures, integration boundaries, ad-hoc scaffolds; legitimate Mgr-eye-review cohort. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
May 9, 2026
* WIP: R3 Debt-Paydown Mgr — standing program through R3 close * docs(r3): cite specific Director-ratification comment URL in §7.6.2 Per Director micro-suggestion on #2101 review — match #1534 source-column URL discipline so future readers can navigate to rationale without searching. Debt receipt: no debt introduced; no debt retired (review-absorption edit). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): §7.6.1 log intra-cadence tripwire reading 1.56:1 (466 PRs) Adds a 2026-05-07 intra-cadence re-check row to the readings log. Same 7-day window as Wave 3, advanced ~3.5h. D-side grew 47 → 54 while I-side held at 84; ratio improved from 1.79:1 to 1.56:1. Confirms continued dissolution-side accrual — no breach. Debt receipt: no debt introduced; no debt retired (readings-log entry). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * WIP: R3 Debt-Paydown Mgr — standing program through R3 close * docs(r3): r3-debt-sweep §1 Class A-G populated + Class P added — Phase 1 organize cycle Per Brian operator directive 2026-05-09 ~20:30Z (lane scope expanded to all v3 debt; active posture), Debt-Paydown Mgr (gentle-newt-665) Phase 1 organize cycle replaces 7 [Mgr canvas populates rows] placeholders with initial substantive rows seeded from: - docs/audit/v3-comprehensive-debt-audit-2026-05-09.md 10-source inventory (PR #2436) - ROADMAP ## Tracked debts (10 sections; ~40 still-open rows) - r3-program-plan §10.3 ratifications + §1.8 ledger gate references - 4 stranded bug-fix briefs (PR #2373: u128 mirror / FieldProject dual-authority / resolve_producer_opt / CallGraph) - 6 gpt-5-5-pro novel findings (PR #2358 §8 meta-finding cycle) - SG-0 census untracked entries (113 of 159) + drift-class anti-patterns Class A-G total ~83 rows; new Class P (past-R3 / R4+ ambition parking lot) ~8 rows per Brian directive. Mgr re-organize cycle continues; initial pass not exhaustive. Phase 2 dispatch list surfacing to PM #846 follows in separate comment. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): 6 worker briefs for gpt-5-5-pro novel findings F1-F6 Authored under Mgr standing authority per PM confirmation at #846 c#4413701937 (Brian operator directive 2026-05-09 active posture). Mirrors PR #2373 brief template. Briefs: - F1 MissingEmissionPath typed-axes substrate (Class C; HIGH; Substrate Mgr) — substrate stringifies axes that Rust mirror correctly types - F2 ShapeATarget closed enum vs LanguageSpec extensibility (Class F; HIGH; Grounding Mgr) — closed Rust enum negates thesis extensibility claim - F3 Map<String,Bool> as set (Class F; MEDIUM; Substrate) — Set<A> declared but bypassed; set_has helper - F4 PartitionResult anonymous bypass (Class G; LOW-MEDIUM; Substrate) — named type declared but consumer uses anonymous return - F5 ComposedEffect illegal product (Class C; HIGH; Substrate) — idempotent + breaking_operation contradictory pair representable - F6 derive_op_effect string parser at structural boundary (Class C; HIGH; Substrate) — string-keyed dispatch over typed HttpMethod/Path carriers Each brief: problem statement with file:line citations + 1-2 fix options with PM/Mgr recommendation + receipt criteria + cross-cutting constraints. Cross-references sweep-doc Class A-G rows for traceability. Phase 2 dispatch under Brian-directive active posture. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): SG-0 census per-entry classification (Phase 2(3) deliverable) Per PM dispatch at #846 c#4413701937 (Brian directive 2026-05-09 active posture; Debt-Paydown Mgr standing authority). scripts/classify-sg0-census.py: heuristic regex classification of preceding-comment-block + path against Class A-G/P framework. Output: docs/audit/sg0-census-classification-2026-05-09.md — per-entry table + summary. Initial pass: 159 entries (104 TEST + 53 NON_TEST + 2 FRAGMENTS) — 115 untagged (no preceding dissolution-trigger comment), matching audit's "113 of 159" estimate. Heuristic class breakdown (initial): - A 0, B 4, C 1, D 3, E 7, F 0, G 143, STRUCTURAL 1 Coarse-grained — heuristic sweeps Class E (Tier3/v2-retirement/T-LensProducer) reliably; substrate-gap (Class A) and ontology (Class F) need Mgr-eye review on the 115 untagged entries. Initial pass is dispatchable input, not ratified classification. Mgr next-action documented in output §"Mgr next-action": review heuristics → confirm-or-correct → fold per-entry classifications back into r3-debt-sweep-2026-05-06.md §1 Class A-G rows. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): one-source SG-0 untagged-count derivation (codex BLOCKING fix sha 03a8830) Codex flagged: hard-coded SG-0 untagged-count appeared in both sweep doc Class G row + classification doc Summary, creating drift risk. Fix per "one checkable value" discipline: - sweep doc Class G row 16: drop hard-coded "159 entries; 113 untagged"; cite classification doc §Summary as live source. - classification doc preamble: explicit "single source for SG-0 untagged-count" framing; regenerated by `scripts/classify-sg0-census.py` from `sg0_census_test.rs`. Other consumers of the count (audit prose at PR #2436) are owned by PM; my PR's contribution is making my surfaces non-duplicating. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): supersede-disambig marker on TC3 D4 brief line 121 (CI R4-carve discipline fix) CI scripts/check-r4-carve-dissolution-discipline.sh fail-closed on docs/briefs/r3-evaluator-tc3-d4-eval-step-producer-worker.md:121 — the line read "(α)/(β) novel-substrate-introduction explicitly carved to R4+ via Class P partition" without a supersession marker. Disambiguating note added inline: "(Note: this is a novel R4+ partition for (α)/(β); does NOT supersede or invoke the DISSOLVED prior C1/C2/C3 carves per Director carve-promotion-IN-R3 at gunbc#846 #issuecomment-4412330468.)" The clarification is structurally accurate — Class P partition is a NEW R4+ routing for specific Pattern-A items, distinct from the prior C1/C2/C3 carves which Director DISSOLVED + reclassified R3-load-bearing within Cluster F (per gunbc#846 #issuecomment-4412330468). Marker "supersede" + "DISSOLVED" both present satisfies the discipline. Surfaced fix scope justification: PR #2437 introduces Class P partition; the brief at line 121 forward-references it. Cleanest fix is on the citing brief, not the discipline script (preserves single-Mgr-ownership of the Cluster F discipline). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): refine SG-0 classifier with path-based heuristics (better Class A/C/D/E surfacing) Per PM observation at gunbc#846 c#4413720143 — Class A=0/F=0 from comment-only heuristic was the substrate-gap and ontology-duplication semantic gap; path-based rules can fire on canonical compiler-Rust-mirror locations. Added PATH_RULES tier (fires when CLASS_RULES miss): - src/v3/compiler/src/dag.rs / lower.rs / infer.rs / emit*.rs / value.rs / eval/ → Class C - src/v3/compiler/src/parse / tokenize / test_runner → Class A (substrate-gap parser/grammar) - src/v3/compiler/src/bin/regen_ / bootstrap → Class D - src/v3/compiler/benches → STRUCTURAL - src/v3/lenses → Class E (T-LensProducer) - src/v3/grounding_pilot / grounding_cross_target / spec / std → Class C Class breakdown delta: - A: 0 → 1 - C: 1 → 16 - D: 3 → 12 - E: 7 → 9 - G: 143 → 115 (default drops as path-based rules catch real classes) 115 still-G are entries that don't fit canonical compiler patterns — test fixtures, integration boundaries, ad-hoc scaffolds; legitimate Mgr-eye-review cohort. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): novel-findings F1-F6 dispatch tracker (Phase 3 Task A) Per Director Phase 3 re-task at #828 → gentle-newt-665 inbox c#4413891498. Operator-authorized 2026-05-09 ~20:30Z (lane scope all v3 debt; active posture). Per-brief dispatch-path classification for the 6 gpt-5-5-pro novel-finding worker briefs from PR #2437. Output: - Path (a) Mgr-self-author retirement: 0 - Path (b) cross-Mgr surface: 6 (Substrate-owned 5 + Grounding-owned 1; PB v2-retirement coord on F3/F5/F6) - Path (c) operator-tier ratification: 0 F6 substrate-prereq audited dep-clear (HttpMethod carrier exists at std/types.dag). Cross-Mgr dispatch surfaces already posted at gunbc#2068 c#4413712338 + gunbc#2063 c#4413712632 in Phase 2 cycle; this tracker provides the option-(c) tightening evidence per PR #2361 §3. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
May 9, 2026
…ation) (#2456) * WIP: R3 Debt-Paydown Mgr — standing program through R3 close * docs(r3): cite specific Director-ratification comment URL in §7.6.2 Per Director micro-suggestion on #2101 review — match #1534 source-column URL discipline so future readers can navigate to rationale without searching. Debt receipt: no debt introduced; no debt retired (review-absorption edit). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): §7.6.1 log intra-cadence tripwire reading 1.56:1 (466 PRs) Adds a 2026-05-07 intra-cadence re-check row to the readings log. Same 7-day window as Wave 3, advanced ~3.5h. D-side grew 47 → 54 while I-side held at 84; ratio improved from 1.79:1 to 1.56:1. Confirms continued dissolution-side accrual — no breach. Debt receipt: no debt introduced; no debt retired (readings-log entry). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * WIP: R3 Debt-Paydown Mgr — standing program through R3 close * docs(r3): r3-debt-sweep §1 Class A-G populated + Class P added — Phase 1 organize cycle Per Brian operator directive 2026-05-09 ~20:30Z (lane scope expanded to all v3 debt; active posture), Debt-Paydown Mgr (gentle-newt-665) Phase 1 organize cycle replaces 7 [Mgr canvas populates rows] placeholders with initial substantive rows seeded from: - docs/audit/v3-comprehensive-debt-audit-2026-05-09.md 10-source inventory (PR #2436) - ROADMAP ## Tracked debts (10 sections; ~40 still-open rows) - r3-program-plan §10.3 ratifications + §1.8 ledger gate references - 4 stranded bug-fix briefs (PR #2373: u128 mirror / FieldProject dual-authority / resolve_producer_opt / CallGraph) - 6 gpt-5-5-pro novel findings (PR #2358 §8 meta-finding cycle) - SG-0 census untracked entries (113 of 159) + drift-class anti-patterns Class A-G total ~83 rows; new Class P (past-R3 / R4+ ambition parking lot) ~8 rows per Brian directive. Mgr re-organize cycle continues; initial pass not exhaustive. Phase 2 dispatch list surfacing to PM #846 follows in separate comment. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): 6 worker briefs for gpt-5-5-pro novel findings F1-F6 Authored under Mgr standing authority per PM confirmation at #846 c#4413701937 (Brian operator directive 2026-05-09 active posture). Mirrors PR #2373 brief template. Briefs: - F1 MissingEmissionPath typed-axes substrate (Class C; HIGH; Substrate Mgr) — substrate stringifies axes that Rust mirror correctly types - F2 ShapeATarget closed enum vs LanguageSpec extensibility (Class F; HIGH; Grounding Mgr) — closed Rust enum negates thesis extensibility claim - F3 Map<String,Bool> as set (Class F; MEDIUM; Substrate) — Set<A> declared but bypassed; set_has helper - F4 PartitionResult anonymous bypass (Class G; LOW-MEDIUM; Substrate) — named type declared but consumer uses anonymous return - F5 ComposedEffect illegal product (Class C; HIGH; Substrate) — idempotent + breaking_operation contradictory pair representable - F6 derive_op_effect string parser at structural boundary (Class C; HIGH; Substrate) — string-keyed dispatch over typed HttpMethod/Path carriers Each brief: problem statement with file:line citations + 1-2 fix options with PM/Mgr recommendation + receipt criteria + cross-cutting constraints. Cross-references sweep-doc Class A-G rows for traceability. Phase 2 dispatch under Brian-directive active posture. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): SG-0 census per-entry classification (Phase 2(3) deliverable) Per PM dispatch at #846 c#4413701937 (Brian directive 2026-05-09 active posture; Debt-Paydown Mgr standing authority). scripts/classify-sg0-census.py: heuristic regex classification of preceding-comment-block + path against Class A-G/P framework. Output: docs/audit/sg0-census-classification-2026-05-09.md — per-entry table + summary. Initial pass: 159 entries (104 TEST + 53 NON_TEST + 2 FRAGMENTS) — 115 untagged (no preceding dissolution-trigger comment), matching audit's "113 of 159" estimate. Heuristic class breakdown (initial): - A 0, B 4, C 1, D 3, E 7, F 0, G 143, STRUCTURAL 1 Coarse-grained — heuristic sweeps Class E (Tier3/v2-retirement/T-LensProducer) reliably; substrate-gap (Class A) and ontology (Class F) need Mgr-eye review on the 115 untagged entries. Initial pass is dispatchable input, not ratified classification. Mgr next-action documented in output §"Mgr next-action": review heuristics → confirm-or-correct → fold per-entry classifications back into r3-debt-sweep-2026-05-06.md §1 Class A-G rows. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): one-source SG-0 untagged-count derivation (codex BLOCKING fix sha 03a8830) Codex flagged: hard-coded SG-0 untagged-count appeared in both sweep doc Class G row + classification doc Summary, creating drift risk. Fix per "one checkable value" discipline: - sweep doc Class G row 16: drop hard-coded "159 entries; 113 untagged"; cite classification doc §Summary as live source. - classification doc preamble: explicit "single source for SG-0 untagged-count" framing; regenerated by `scripts/classify-sg0-census.py` from `sg0_census_test.rs`. Other consumers of the count (audit prose at PR #2436) are owned by PM; my PR's contribution is making my surfaces non-duplicating. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): supersede-disambig marker on TC3 D4 brief line 121 (CI R4-carve discipline fix) CI scripts/check-r4-carve-dissolution-discipline.sh fail-closed on docs/briefs/r3-evaluator-tc3-d4-eval-step-producer-worker.md:121 — the line read "(α)/(β) novel-substrate-introduction explicitly carved to R4+ via Class P partition" without a supersession marker. Disambiguating note added inline: "(Note: this is a novel R4+ partition for (α)/(β); does NOT supersede or invoke the DISSOLVED prior C1/C2/C3 carves per Director carve-promotion-IN-R3 at gunbc#846 #issuecomment-4412330468.)" The clarification is structurally accurate — Class P partition is a NEW R4+ routing for specific Pattern-A items, distinct from the prior C1/C2/C3 carves which Director DISSOLVED + reclassified R3-load-bearing within Cluster F (per gunbc#846 #issuecomment-4412330468). Marker "supersede" + "DISSOLVED" both present satisfies the discipline. Surfaced fix scope justification: PR #2437 introduces Class P partition; the brief at line 121 forward-references it. Cleanest fix is on the citing brief, not the discipline script (preserves single-Mgr-ownership of the Cluster F discipline). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): refine SG-0 classifier with path-based heuristics (better Class A/C/D/E surfacing) Per PM observation at gunbc#846 c#4413720143 — Class A=0/F=0 from comment-only heuristic was the substrate-gap and ontology-duplication semantic gap; path-based rules can fire on canonical compiler-Rust-mirror locations. Added PATH_RULES tier (fires when CLASS_RULES miss): - src/v3/compiler/src/dag.rs / lower.rs / infer.rs / emit*.rs / value.rs / eval/ → Class C - src/v3/compiler/src/parse / tokenize / test_runner → Class A (substrate-gap parser/grammar) - src/v3/compiler/src/bin/regen_ / bootstrap → Class D - src/v3/compiler/benches → STRUCTURAL - src/v3/lenses → Class E (T-LensProducer) - src/v3/grounding_pilot / grounding_cross_target / spec / std → Class C Class breakdown delta: - A: 0 → 1 - C: 1 → 16 - D: 3 → 12 - E: 7 → 9 - G: 143 → 115 (default drops as path-based rules catch real classes) 115 still-G are entries that don't fit canonical compiler patterns — test fixtures, integration boundaries, ad-hoc scaffolds; legitimate Mgr-eye-review cohort. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): novel-findings F1-F6 dispatch tracker (Phase 3 Task A) Per Director Phase 3 re-task at #828 → gentle-newt-665 inbox c#4413891498. Operator-authorized 2026-05-09 ~20:30Z (lane scope all v3 debt; active posture). Per-brief dispatch-path classification for the 6 gpt-5-5-pro novel-finding worker briefs from PR #2437. Output: - Path (a) Mgr-self-author retirement: 0 - Path (b) cross-Mgr surface: 6 (Substrate-owned 5 + Grounding-owned 1; PB v2-retirement coord on F3/F5/F6) - Path (c) operator-tier ratification: 0 F6 substrate-prereq audited dep-clear (HttpMethod carrier exists at std/types.dag). Cross-Mgr dispatch surfaces already posted at gunbc#2068 c#4413712338 + gunbc#2063 c#4413712632 in Phase 2 cycle; this tracker provides the option-(c) tightening evidence per PR #2361 §3. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(audit): broaden dispatch tracker to 10 briefs (PM scope-consolidation 2026-05-09) Per PM re-task at #846 → gentle-newt-665 inbox c#4413913108 (operator directive 2026-05-09 ~23:15Z: orient ALL pending v3 debt work-items through Debt-Paydown Mgr as central tracker). Changes: - Renamed: r3-novel-findings-dispatch-tracker-2026-05-09.md → r3-debt-dispatch-tracker-2026-05-09.md - Scope broadened: 6 (F1-F6 novel findings) → 10 (+ 4 P0 bug-fix briefs from PR #2373) - Added Capacity column reflecting cross-Mgr workload state per PM observation: - Substrate (warm-wolf-698): WORKING heavy (9 of 10 briefs in this lane; no escalation pressure) - Grounding (sunny-koi-893): IDLE with capacity (1 brief: F2) - PB: coord on F3/F5/F6 v2-retirement timing - F-resolve-producer-opt reclassified: Verification → Substrate per brief header (PM scope-consolidation) Triage outcome: 10 of 10 path (b) cross-Mgr surface; 0 (a) Mgr-self-author; 0 (c) operator-tier ratification. Velocity expectation: Substrate heavy-queue → ~2 briefs/cycle absorption rate; cycle-1 = u128 + FieldProject. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
briansrls
pushed a commit
that referenced
this pull request
Jun 11, 2026
Context-scoped MutationCounters on the copy-on-update primitives (map_insert/merge/list_push/concat/set ops: calls + entries copied — the triangular/quadratic receipt), thread-local flatten counters on the free_monoid_to_vec chokepoint (fires inside Value::eq, so no ctx; two fixed-size integers, not a cache), and a sharing-aware retained-value byte accounting walk (per-variant counts/bytes, visited-set dedup) on InterpContext. claim_batch prints the report under GUNBC_INTERP_STATS=1. Receipt on the real v4 gate workload (7 affected_testgen witnesses, all PASS): native mutation primitives 0 calls — v4 collections are closure chains + FreeMonoid trees, so the host copy cost rides the flatten chokepoint, confirming ctrl PR #1534's layer analysis. Retained: 21.8 MB / 219K allocations (Record 12.7MB with 48K sharing hits; 31K un-interned Strings) — the M-D interning/positional-record receipt baseline. Read-only tooling per ctrl#1533 phase 0; no semantic change to any evaluation path (counters + opt-in report only). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
Jun 11, 2026
* WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * feat(interp): phase-0 memory measurement harness (ctrl#1533) Context-scoped MutationCounters on the copy-on-update primitives (map_insert/merge/list_push/concat/set ops: calls + entries copied — the triangular/quadratic receipt), thread-local flatten counters on the free_monoid_to_vec chokepoint (fires inside Value::eq, so no ctx; two fixed-size integers, not a cache), and a sharing-aware retained-value byte accounting walk (per-variant counts/bytes, visited-set dedup) on InterpContext. claim_batch prints the report under GUNBC_INTERP_STATS=1. Receipt on the real v4 gate workload (7 affected_testgen witnesses, all PASS): native mutation primitives 0 calls — v4 collections are closure chains + FreeMonoid trees, so the host copy cost rides the flatten chokepoint, confirming ctrl PR #1534's layer analysis. Retained: 21.8 MB / 219K allocations (Record 12.7MB with 48K sharing hits; 31K un-interned Strings) — the M-D interning/positional-record receipt baseline. Read-only tooling per ctrl#1533 phase 0; no semantic change to any evaluation path (counters + opt-in report only). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * WIP: memory usage * fix(interp): one copy-work definition per primitive; delta-scope fm_flatten in claim_batch Review #28288 (cursor/composer-2.5) on #4653: the copy-work counters mixed definitions across dispatch paths, breaking the single-receipt property (P2). - One definition per counter, by operation semantics, not dispatch path: add-one ops (map_insert/list_push/set_insert) count the receiver's pre-existing entries; merge ops (map_merge/list_concat/set_union) count both operands' entries. Method .concat/.append/.push buckets by what the arg IS (collection -> concat, atomic -> push); binop + and builtins agree. - set_union gets its own row (was folded into set_insert — same P2 class). - builtin map_merge instrumented (was counted on the method path only). - claim_batch fm_flatten row is now a delta sampled across the witness loop, matching the context-scoped counters next to it. Tests pin the unified semantics: concat counts both operands (9 for the 3⊕1 then 4⊕1 chain), atomic append lands in list_push with receiver-only copy-work, and neither bucket leaks into the other. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Brian Searls <briansearls1@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
Jun 11, 2026
…ntity laws (M-C) with executable witnesses (#4660) * WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * feat(interp): phase-0 memory measurement harness (ctrl#1533) Context-scoped MutationCounters on the copy-on-update primitives (map_insert/merge/list_push/concat/set ops: calls + entries copied — the triangular/quadratic receipt), thread-local flatten counters on the free_monoid_to_vec chokepoint (fires inside Value::eq, so no ctx; two fixed-size integers, not a cache), and a sharing-aware retained-value byte accounting walk (per-variant counts/bytes, visited-set dedup) on InterpContext. claim_batch prints the report under GUNBC_INTERP_STATS=1. Receipt on the real v4 gate workload (7 affected_testgen witnesses, all PASS): native mutation primitives 0 calls — v4 collections are closure chains + FreeMonoid trees, so the host copy cost rides the flatten chokepoint, confirming ctrl PR #1534's layer analysis. Retained: 21.8 MB / 219K allocations (Record 12.7MB with 48K sharing hits; 31K un-interned Strings) — the M-D interning/positional-record receipt baseline. Read-only tooling per ctrl#1533 phase 0; no semantic change to any evaluation path (counters + opt-in report only). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * WIP: memory usage * fix(interp): one copy-work definition per primitive; delta-scope fm_flatten in claim_batch Review #28288 (cursor/composer-2.5) on #4653: the copy-work counters mixed definitions across dispatch paths, breaking the single-receipt property (P2). - One definition per counter, by operation semantics, not dispatch path: add-one ops (map_insert/list_push/set_insert) count the receiver's pre-existing entries; merge ops (map_merge/list_concat/set_union) count both operands' entries. Method .concat/.append/.push buckets by what the arg IS (collection -> concat, atomic -> push); binop + and builtins agree. - set_union gets its own row (was folded into set_insert — same P2 class). - builtin map_merge instrumented (was counted on the method path only). - claim_batch fm_flatten row is now a delta sampled across the witness loop, matching the context-scoped counters next to it. Tests pin the unified semantics: concat counts both operands (9 for the 3⊕1 then 4⊕1 chain), atomic append lands in list_push with receiver-only copy-work, and neither bucket leaks into the other. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * WIP: memory usage * WIP: memory usage * review: explicit collection imports + free-function concat in value_carrier_laws; strengthen insert-order witness to 3 keys Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Brian Searls <briansearls1@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
Jun 11, 2026
…iveContract + M-D name-interning declarations (#4666) * WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * feat(interp): phase-0 memory measurement harness (ctrl#1533) Context-scoped MutationCounters on the copy-on-update primitives (map_insert/merge/list_push/concat/set ops: calls + entries copied — the triangular/quadratic receipt), thread-local flatten counters on the free_monoid_to_vec chokepoint (fires inside Value::eq, so no ctx; two fixed-size integers, not a cache), and a sharing-aware retained-value byte accounting walk (per-variant counts/bytes, visited-set dedup) on InterpContext. claim_batch prints the report under GUNBC_INTERP_STATS=1. Receipt on the real v4 gate workload (7 affected_testgen witnesses, all PASS): native mutation primitives 0 calls — v4 collections are closure chains + FreeMonoid trees, so the host copy cost rides the flatten chokepoint, confirming ctrl PR #1534's layer analysis. Retained: 21.8 MB / 219K allocations (Record 12.7MB with 48K sharing hits; 31K un-interned Strings) — the M-D interning/positional-record receipt baseline. Read-only tooling per ctrl#1533 phase 0; no semantic change to any evaluation path (counters + opt-in report only). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * WIP: memory usage * fix(interp): one copy-work definition per primitive; delta-scope fm_flatten in claim_batch Review #28288 (cursor/composer-2.5) on #4653: the copy-work counters mixed definitions across dispatch paths, breaking the single-receipt property (P2). - One definition per counter, by operation semantics, not dispatch path: add-one ops (map_insert/list_push/set_insert) count the receiver's pre-existing entries; merge ops (map_merge/list_concat/set_union) count both operands' entries. Method .concat/.append/.push buckets by what the arg IS (collection -> concat, atomic -> push); binop + and builtins agree. - set_union gets its own row (was folded into set_insert — same P2 class). - builtin map_merge instrumented (was counted on the method path only). - claim_batch fm_flatten row is now a delta sampled across the witness loop, matching the context-scoped counters next to it. Tests pin the unified semantics: concat counts both operands (9 for the 3⊕1 then 4⊕1 chain), atomic append lands in list_push with receiver-only copy-work, and neither bucket leaks into the other. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * WIP: memory usage * WIP: memory usage * review: explicit collection imports + free-function concat in value_carrier_laws; strengthen insert-order witness to 3 keys Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * model: ctrl#1533 phase-1 follow-up — M-B carrier-cost facts on PrimitiveContract + M-D name-interning declarations M-B: PrimitiveContract gains carrier_cost: CarrierCostSensitivity. 12 carrier-sensitive primitives declare both arms (ephemeral = copy-before- update, what the v2 interpreter does today; persistent = the declared M-A carrier cost); 49 carrier-insensitive primitives marked explicitly. M-D: name-interning facts at the v4.std.value_carrier authority — type/ variant/field names are references into the resolved graph's declared name set; positional-record layout derives from this fact (phase 3 implements). New witness name_interning_covers_all_domains_holds (8/8 green). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Brian Searls <briansearls1@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
Jun 11, 2026
…B (persistent updates, copied-counters collapse to 0) (#4672) * WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * WIP: memory usage * feat(interp): phase-0 memory measurement harness (ctrl#1533) Context-scoped MutationCounters on the copy-on-update primitives (map_insert/merge/list_push/concat/set ops: calls + entries copied — the triangular/quadratic receipt), thread-local flatten counters on the free_monoid_to_vec chokepoint (fires inside Value::eq, so no ctx; two fixed-size integers, not a cache), and a sharing-aware retained-value byte accounting walk (per-variant counts/bytes, visited-set dedup) on InterpContext. claim_batch prints the report under GUNBC_INTERP_STATS=1. Receipt on the real v4 gate workload (7 affected_testgen witnesses, all PASS): native mutation primitives 0 calls — v4 collections are closure chains + FreeMonoid trees, so the host copy cost rides the flatten chokepoint, confirming ctrl PR #1534's layer analysis. Retained: 21.8 MB / 219K allocations (Record 12.7MB with 48K sharing hits; 31K un-interned Strings) — the M-D interning/positional-record receipt baseline. Read-only tooling per ctrl#1533 phase 0; no semantic change to any evaluation path (counters + opt-in report only). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * WIP: memory usage * fix(interp): one copy-work definition per primitive; delta-scope fm_flatten in claim_batch Review #28288 (cursor/composer-2.5) on #4653: the copy-work counters mixed definitions across dispatch paths, breaking the single-receipt property (P2). - One definition per counter, by operation semantics, not dispatch path: add-one ops (map_insert/list_push/set_insert) count the receiver's pre-existing entries; merge ops (map_merge/list_concat/set_union) count both operands' entries. Method .concat/.append/.push buckets by what the arg IS (collection -> concat, atomic -> push); binop + and builtins agree. - set_union gets its own row (was folded into set_insert — same P2 class). - builtin map_merge instrumented (was counted on the method path only). - claim_batch fm_flatten row is now a delta sampled across the witness loop, matching the context-scoped counters next to it. Tests pin the unified semantics: concat counts both operands (9 for the 3⊕1 then 4⊕1 chain), atomic append lands in list_push with receiver-only copy-work, and neither bucket leaks into the other. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * WIP: memory usage * WIP: memory usage * review: explicit collection imports + free-function concat in value_carrier_laws; strengthen insert-order witness to 3 keys Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * model: ctrl#1533 phase-1 follow-up — M-B carrier-cost facts on PrimitiveContract + M-D name-interning declarations M-B: PrimitiveContract gains carrier_cost: CarrierCostSensitivity. 12 carrier-sensitive primitives declare both arms (ephemeral = copy-before- update, what the v2 interpreter does today; persistent = the declared M-A carrier cost); 49 carrier-insensitive primitives marked explicitly. M-D: name-interning facts at the v4.std.value_carrier authority — type/ variant/field names are references into the resolved graph's declared name set; positional-record layout derives from this fact (phase 3 implements). New witness name_interning_covers_all_domains_holds (8/8 green). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * ctrl#1533 phase 2: swap Value::Map/List host carriers to im-rc HAMT/RRB Value::Map now carries Rc<im_rc::HashMap> (HAMT, matches map_carrier ^hash_array_mapped_trie) and Value::List carries Rc<im_rc::Vector> (RRB, matches list_carrier ^relaxed_radix_balanced_tree). Rc wrapper kept: value_rc_identity memo keys and accounting dedup rely on Rc::as_ptr. map insert/merge and list push/concat no longer copy entries; the *_entries_copied / *_items_copied counters now count only FreeMonoid chain flattens — the copied-counter collapse is the phase-2 receipt. List ops route native List and Empty/Cons chains through one value_to_list_carrier bridge (B1 alias transparency preserved). New witnesses: list/map prior-version persistence in value_carrier_laws.dag; host-side at-scale probes (n=500, past inline chunks) in value_carrier_swap_test.rs. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * fix at-scale probes: use DSL fold(list, init:, f: fn(){}) syntax Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: Brian Searls <briansearls1@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Slice 1C measurement-only PR for the R3 Debt-Paydown cadence.
Tripwire reading: FIRED.
2026-04-20→2026-05-02876325.33:12026-05-01..2026-05-02: verified introduction-side PRs by reading diffs, not just titlesWhat changed
scripts/r3-debt-velocity.sh: hermetic bash measurement script for the ROADMAP post-merge debt sectionsdocs/debt/velocity-tripwire-2026-05-02.md: first tripwire report for the 2026-04-20 → 2026-05-02 windowWhy
Per coordination brief #1518, the standing R3 Debt-Paydown program needs a measurable cadence for introduction:dissolution ratio reporting against
INVARIANTS.md§P5(c). This PR enables future receipts without touching tracked debt rows itself.Authority
INVARIANTS.md§P5(c), velocity tripwire thresholdReceipt
Disposition #3: No debt touched.
This PR adds reporting infrastructure only. It does not gate CI and does not alter any ROADMAP debt rows.