Repository navigation
Add coercion-kind-mapped tests and audit runtime implementation - #37
Merged
Merged
Conversation
Audited classify_coercion() in coerce.rs against the executor's fan-in logic in execute.rs. All three CoercionKind variants (WrapScalar, OptionalToList, Widen) are correctly implemented by the runtime. Changes: - Add 4 coercion-kind-mapped tests in execute.rs that verify value shapes for each CoercionKind (the definitive correctness proof) - Improve build_coercion_coverage_tests in testgen codegen to assert target node execution, not just "DAG didn't crash" - Update TODO_hacks: mark coercion audit as resolved, update coverage test entry with partial fix status https://claude.ai/code/session_019RSYYEQsHBWb1TPrbUzSyf
The log.get(to_node).is_some() assertion was a half-measure — it didn't verify coercion value shapes. Revert to original _log binding. The real fix requires observing node *inputs* in generated tests, which is blocked on a design decision: either add an inputs field to LogEntry (memory/API tradeoff) or inject shape-assert nodes in testgen (IR complexity). Document both options and their sub-decisions in TODO_hacks. The 4 execute.rs coercion-kind-mapped tests remain as the definitive shape verification for each CoercionKind. https://claude.ai/code/session_019RSYYEQsHBWb1TPrbUzSyf
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 4e0cb4d69d
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Extract the per-edge fan-in collection logic into a standalone collect_fan_in() function, eliminating duplication between the main executor loop and build_node_inputs. Both call sites now delegate to the same function. Replace 4 integration-level coercion tests (which built full DAGs, ran the executor, and inspected the execution log via echo nodes) with 7 unit tests that call collect_fan_in() directly. Tests now verify coercion behavior at the level where it happens. https://claude.ai/code/session_019RSYYEQsHBWb1TPrbUzSyf
Full codebase scan for needless conversions, unnecessary allocations, suboptimal idioms, and correctness risks across lib/, core/, gunbc-dag/. Key findings: - MAJOR: swapped TCP timeout fields in executor.rs (§17.1/§10) - ~70 .expect() calls in Result-returning graph builders (§17.2/§12) - panic!() in Result-returning Executable impl (§17.6/§11) - ~100 push_str(&format!()) sites (§17.8, mechanical fix) - ~80 string literal allocations avoidable with Cow (§17.9) https://claude.ai/code/session_019RSYYEQsHBWb1TPrbUzSyf
Major: - Fix swapped TCP read/write timeout assignments in executor.rs Moderate: - Replace .expect() with ? error propagation in graph builders - Replace panic!() with Err(CliToolError::invariant()) in CLI dispatch - Replace is_none()+unwrap() with ok_or_else(?) in builder.rs - Introduce ParamType enum replacing stringly-typed type_id field Minor - Allocations: - push_str(&format!()) → write!() across ~30 sites - BlobHandleError::new accept impl Into<String> (avoid double alloc) - to_string_lossy().to_string() → into_owned() - sort()+dedup() → BTreeSet in registry and discovery_ops - O(n²) stage dedup → HashSet+Vec in GitLab CI renderer - Cow<'static, str> for FileHeader, Target, Category structs Minor - API: - &PathBuf → &Path in function signatures - Deduplicate execute_run / execute_run_with_path into shared helper - Deduplicate MapToGcpInputs / MapToGcpSecretInputs via extract_gcp_common_inputs() - Simplify gist subdag 12-variant match to wildcard with documentation Minor - Logic: - Remove dead expireTime statement in gcp-ops - Replace char collection with prev_char tracking in language/mod.rs - Document /root fallback assumption and Vec::contains constraint Net: 37 files changed, -60 lines https://claude.ai/code/session_019RSYYEQsHBWb1TPrbUzSyf
Analyzes how to bring the transport executor (currently a testing blind spot with zero TCP coverage) under the DAG/testgen umbrella. The swapped timeout bug (17.1) motivates the investigation. Evaluates 3 approaches: A) Full behavioral sub-DAGs (blocked by OS handle problem) B) Specification DAGs + code generation (feasible but heavy) C) Decomposed prepare/parse + behavioral tests (recommended) Recommends 4-phase incremental migration starting with immediate test coverage, then typed port decomposition to make field routing explicit at the DAG level. https://claude.ai/code/session_019RSYYEQsHBWb1TPrbUzSyf
Three compounding issues caused CI to hang indefinitely: 1. No timeout-minutes in generated GitHub Actions / GitLab CI YAML (GHA defaults to 360 min). Now all renderers emit timeout-minutes: 30. 2. execute_shell didn't set stdin to Stdio::null() for commands without stdin data — child inherited parent stdin, causing potential deadlocks. 3. ShellRequest had no timeout field — wait_with_output() blocks forever. Added timeout_ms with try_wait polling + kill-on-expiry in executor. All fixes are structural (central execute_shell, shared RenderConfig) so they apply across the entire codebase automatically. Includes TODO/TODO_ci_timeout_fermi.md for fermi-based timeout estimation. https://claude.ai/code/session_019RSYYEQsHBWb1TPrbUzSyf
Every shell command now gets a 5-minute timeout (FermiCost::S) by default instead of waiting forever. Callers needing longer can set explicit ShellRequest::timeout(ms). - Add FermiCost::timeout_ms() mapping (XS=30s, S=5m, M=10m, L=30m, XL=60m) - execute_shell uses DEFAULT_TIMEOUT_MS=300_000 when no explicit timeout - Update TODO doc to reflect implemented state, defer full per-op fermi https://claude.ai/code/session_019RSYYEQsHBWb1TPrbUzSyf
The default 5-min timeout (FermiCost::S) is too short for preflight steps that compile + run cargo binaries in CI with cold caches. The gunbc-testgen step was timing out at 300s in GitHub Actions. https://claude.ai/code/session_019RSYYEQsHBWb1TPrbUzSyf
briansrls
added a commit
that referenced
this pull request
May 9, 2026
Resolve lens_cost_target_realization_test.rs: keep R3 gate #37 ε-slice tests (symbolic_cost_of × realization cost × sequential) alongside main's RealizationCostTable integration tests and bootstrap_dag helper. Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
May 9, 2026
…population drift (#2333) * docs(briefs): Substrate bridge SourceSpan.file participation retirement worker brief Authored for gunbc#1958 Substrate-owned bridge slice. Targets audit-row #2 (kernel Bool patch BOOL_TYPES_FILE) + row #6 (pipeline authority PIPELINE_AUTHORITY_FILE) per r3-program-plan.md §5 line 353 scope-narrowing. Sibling #1959 closed as already-retired by PR #1272. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): S5 Variant-aware projection carrier canvas (#1947) Surfaces 3 carrier-shape options (α RestResponseProjection variant-tag / β Declaration variant_projection_metadata / γ free-standing CoproductProjection) for Director ratification before worker brief authoring. Mgr-tier recommendation = γ (DeclarationRef-keyed, avoids tag-string-as-identity bridge). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): SourceSpan.file brief — same-slice prerequisite + section anchor Director calibration (gunbc#2079 #issuecomment-...): 1. Promote ratchet-test cross-Mgr handoff from conditional Acceptance #4 to upfront same-slice BLOCKING prerequisite gate (per feedback_same_slice_dissolution_discipline). 2. Replace `r3-program-plan.md:353` line-cite with `§5 Y4 scope-clarification` section anchor (per feedback_section_anchors_over_line_numbers). Code-line anchors in bootstrap.rs left as-is (anchor sites worker navigates to). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): SourceSpan.file brief — umbrella stays Open per P2 ledger discipline Address BLOCKING inline review at line 46: bridge_source_span_file_participation_retired is an Open umbrella whose green predicate is "no production code path consults SourceSpan.file" per r3-structure.md:115. Partial retirement was explicitly rejected 2026-04-29 (Director acceptance #1130 / dispatch #1139). Changes: - Add Ledger-discipline preamble: umbrella row stays Open; receipt updates audit-packet enumeration, NOT bridge_ledger.dag. - Acceptance #3 reframed: do NOT mutate bridge_ledger.dag; mark rows #2 + #6 retired in the audit packet only. - Authority anchor updated: status=Open (not Proposed); add r3-structure.md:115 + bridge_ledger.dag:125-129 line refs. - Cross-Mgr section reframed: umbrella ratchet cannot flip on this PR alone; Verification's ledger-zero audit progress field is post-merge tracking, not a same-slice pre-merge blocker. Reconciles with BLOCKING finding (Director calibration #1 assumed umbrella ratchet could flip on partial retirement, which r3-structure.md:115 forbids). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): S5 CoproductProjection worker brief — γ ratified (#1947) Director ratification of option γ at gunbc#828 #issuecomment-4394369848. Free-standing CoproductProjection carrier in src/v3/std/, DeclarationRef-keyed, typed WireTagValue leaf, 4 same-slice acceptance gates. Surfaces 3 substrate observations for STOP-and-PING (DeclarationRef String alias; FieldRef does-not-exist-at-HEAD; tag_field String asymmetry) — worker must escalate, not silently work around. PB Mgr cross-Mgr ping at carrier landing (heads-up, not blocker). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): SourceSpan.file brief — resolve Acceptance #4 / Cross-Mgr contradiction openai-pro REQUEST_CHANGES at PR #2079 #issuecomment-... flagged Acceptance #4 ("ratchet test passes ... confirmed-present at HEAD before merge") contradicting the reframed Cross-Mgr section ("No same-slice ratchet-test gate applies; post-merge tracking only"). Both said different things about whether the umbrella ratchet is a pre-merge blocker. Resolution: Acceptance #4 reframed to explicitly state "No umbrella-ratchet pre-merge gate" — worker does NOT wait for Verification ratchet authoring; acceptance for this slice is the audit-packet receipt update in #3. Cross-Mgr handoff also updated to remove "ratchet authoring" from Verification's same-slice duties. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * WIP: R3 Substrate Mgr — lane through R3 close * docs(briefs): S5 brief — absorb Director pre-ratifications for 3 STOP-and-PING items Director pre-ratified dispositions at gunbc#828 #issuecomment-4394416049: 1. DeclarationRef alias: accept (b) + debt note (re-escalate only on same-slice break) 2. FieldRef: grep-decide between InputFieldRef-as-specialization vs rename 3. tag_field String asymmetry: typed introduction + debt note for migration Worker proceeds without re-pinging unless evidence forces escalation. STOP-criteria section narrowed accordingly. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): Q-Reification Gate A receipt — Option A (Dag-as-carrier) Director ratification at gunbc#828 #issuecomment-4394427814 + proposal merge at PR #2096 (commit fec869202): src/v3/std/substrate.dag::Dag IS the reflected program; no new substrate carrier required. ReflectedProgram<T> rejected. Gate A patches: 1. r3-program-plan.md §10.3: new Q-Reification row marked RATIFIED with PR #2096 merge link + Gate A receipt scope (this PR + #1960 closed-as-non-addition). 2. r3-v-pattern-a-tc1-v1-worker.md: 3 sites — status header (Q-Reification CLEARED, Branch B η non-vacuity remains), worker-pin gate, E6-G1.a/E3 producer dependency. Replaces ReflectedProgram<T> with consumer-wiring nuance: lens fold consumes Dag via .dag body authority through Evaluator. 3. r3-pr-e6-g1a-option3-static-lens-worker.md: 2 sites — same nuance: deferred work is consumer-wiring, NOT a separate carrier. 4. r3-pr-e8-w1-producer-contract-test-plan-worker.md: 1 site — fold-over-Dag reframe. Receipt of pass-by-construction: this PR adds NO new .dag declaration to src/v3/std/. The ratification is structurally a non-addition (Option A correctness proof per same-slice dissolution discipline). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): S5 — delete superseded canvas + name dissolution trigger for tag_field asymmetry openai-pro REQUEST_CHANGES at gunbc#2079: 2 BLOCKING findings (P2 single-authority + P5 dissolution trigger). Fixes: 1. Delete docs/briefs/r3-substrate-s5-variant-aware-projection-carrier-canvas.md (superseded by the γ-ratified worker brief in same PR; would otherwise leave two current-looking S5 status authorities post-merge — one saying ratification pending, one saying γ ratified). 2. Substrate observation #3 (tag_field String/FieldRef asymmetry) now carries a binding named dissolution trigger: when FieldRef exists as top-level carrier AND InputFieldRef is classified, migrate InternallyTaggedObject.tag_field via follow-on Substrate hygiene PR. Worker MUST add debt-paydown row to authoritative debt ledger before merging carrier-introduction PR. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): Q-Reification row — fix invariant cite (C-1 → P2) cursor review at gunbc#2079 #issuecomment-... flagged C-1 as mis-keyed: INVARIANTS.md C-1 is "missing args fail closed; no LitNull sentinels" (P3 fail-closed family), not parallel-representation/duplicate-authority. The correct cite for rejecting a parallel ReflectedProgram<T> alongside Dag is P2 single authority (INVARIANTS.md line 148: cost of change is proportional to how many files encode the same fact). Cite updated to "INVARIANTS.md P2 single authority" with inline gloss. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): S5 — name dissolution trigger for DeclarationRef alias debt openai-pro REQUEST_CHANGES at gunbc#2079: substrate observation #1 had desired endpoint ("future structural promotion of DeclarationRef") but no checkable dissolution trigger — same P5 gap that #3 (tag_field asymmetry) had been fixed for in commit 2601d7d26. Trigger now binding: promote DeclarationRef when EITHER (a) audit-row #14 (declaration_name_preference_rank / declaration_by_name rank-table) closes — dsl/std ↔ src/v3/std module convergence makes name-keyed identity unambiguous and structural module identity available, OR (b) any DeclarationRef-typed consumer surfaces a string-identity bridge per feedback_opaque_strings_attract_heuristics (heuristic patching, naming- convention dispatch, suffix/prefix matching). Worker MUST add debt-paydown row before merging carrier-introduction PR (same pattern as the tag_field debt requirement). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs: complete Q-Reification stale-cite sweep — e6-g1a brief :169 + Q-PAFS row codex BLOCKING at gunbc#2079: docs/briefs/r3-pr-e6-g1a-option3-static-lens-worker.md:169 still said TC1/V1 "waits for Q-Reification and the carrier landing", contradicting the same brief's lines 15-19 + 148-153 saying Dag IS the carrier and no separate carrier lands. Fixed: 1. e6-g1a brief line 167-170 paragraph: clarified V1 waits for consumer-wiring work (lens fold consuming Dag via .dag body authority through Evaluator), NOT for a carrier-introduction. 2. r3-program-plan.md:954 Q-PAFS row text was the same shape: "Resume after Q-Reification + ReflectedProgram<T>" — contradicted my new Q-Reification row in the same diff. Updated: Q-Reification STOP CLEARED 2026-05-07 (Option A); remaining hold = Branch B η non-vacuity only. Out-of-scope-for-this-PR: docs/briefs/r3-pr-e6-g1a-option3-feasibility-probe.md contains 4 stale cites but is not modified in this PR; stale-on-main can be swept in a follow-up if needed (single source of truth is the e6-g1a-static-lens worker brief, not the feasibility probe per Q-PAFS Path A acceptance). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): S5 — fix InputFieldRef mis-read; reframe observations #2 + #3 codex BLOCKING at gunbc#2079 line 22: my brief mis-read services.dag:28-36. The text actually says "No separate InputFieldRef carrier is introduced here, since ParamToken.name already carries the same shape and adding a wrapper would duplicate without strengthening the structural invariant" — i.e., InputFieldRef does NOT exist; the comment REJECTS the wrapper. Fixes: 1. Substrate observation #2 reframed: no FieldRef-shaped carrier exists at HEAD; services.dag:28-36 precedent argues against wrapper unless it strengthens structural invariant. New (a)/(b) STOP-and-PING: (a) follow services.dag precedent — wire_tag_field: String + key invariant on wire_tag_values + fixture-load fail-closed check; (b) introduce typed FieldRef — must justify per "duplicate without strengthening" test. 2. Carrier shape (line 19): wire_tag_field: FieldRef → {String|FieldRef} pending observation #2 resolution. 3. Substrate observation #3 reframed: InternallyTaggedObject asymmetry is conditional on path (b); under path (a) no asymmetry exists. Trigger correspondingly conditional. Removed stale "InputFieldRef classified" trigger clause. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): S5 — consolidate split per-variant maps into single keyed entry openai-pro REQUEST_CHANGES at gunbc#2079: CoproductProjection shape split per-variant data across two parallel maps (variant_field_projections + wire_tag_values), admitting illegal states where keysets drift (P2 boundary discipline / illegal-states-unrepresentable). Fix: introduce CoproductVariantProjection { field_projection, wire_tag_value } as the per-variant keyed value; CoproductProjection.variant_projections becomes Map<VariantId, CoproductVariantProjection>. Single keyed authority per variant. Director's binding constraint #2 enumerated the two fields separately but said "refine in implementation as ergonomics demand" — consolidation preserves the substantive constraints (typed WireTagValue leaf, structural per-variant projection) while enforcing keyset alignment by carrier shape. Empty-payload variants encoded via FieldProjection::Empty constructor (or analog), not via map-absence. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): S5 — fix stale split-map vocab in path (a) + scope STOP-and-PING umbrella openai-pro REQUEST_CHANGES at gunbc#2079: 2 BLOCKING findings. 1. Path (a) at line 43 still referenced "Map<VariantId, WireTagValue> key invariant on wire_tag_values" — that's the superseded split-map vocab; the consolidated shape is Map<VariantId, CoproductVariantProjection> on variant_projections. Path (a) now references the consolidated map; per-variant WireTagValue lives inside CoproductVariantProjection. 2. Pre-ratification umbrella at line 36 said "all 3 dispositions pre-ratified, proceed without re-pinging" — but observation #2 was re-opened after the codex InputFieldRef finding and explicitly says STOP-and-PING. Contradictory. Umbrella now scoped: observations #1 + #3 are pre-ratified; #2 is re-opened STOP-and-PING — worker MUST escalate before choosing path (a) vs (b). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): descent_execution_proof canvas — 4-residual coproduct-dissolution audit Director ratified split disposition at gunbc#828 #issuecomment-4394696074: descent_execution_proof STANDS ALONE (consumer-side termination-contract substrate function with fail-closed residual enumeration; folding into T-E-P-Producer-Broadening explicitly rejected — different concern axis). Canvas surfaces 3 carrier-shape options for the residual enumeration per Director's coproduct-dissolution audit suggestion: α: 4-variant coproduct verbatim from §10.3 row 966 (Missing | Unknown | Incomplete | NonStrict) β: 3-axis dimensional product (presence × completeness × strictness) γ (recommended): reuse DescentEvidence at termination.dag:14-17 for "absent/unknown" via EvidenceUnknown(DescentEvidence) payload-variant + separate EvidenceIncomplete — ratchets variant count 4 → 2 via dimensional folding while honoring services.dag "no parallel wrapper" precedent. Mgr recommendation γ; β rejected unless 3 axes provably compose orthogonally. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): SourceSpan brief Row #6 — derive from bootstrap_authority map, no new enum variant codex BLOCKING (post-merge of #2079, on commit 85ceb85a — same brief is now on main): my Row #6 disposition told the worker to introduce a new BootstrapAuthority::Pipeline variant ("extend the enum if needed"). That violates P2 single-authority — src/v3/std/bootstrap_authority.dag:90 already has "src/v3/compiler/pipeline.dag": CompilerAuthority, classifying pipeline.dag under the existing CompilerAuthority variant. The :14-17 comment is explicit: "the path itself is the BootstrapAuthoritySet map key; variants carry no duplicate path payload" — single authority, not extension-by-variant. Fix: Row #6 now instructs worker to derive the typed key from the existing bootstrap_authority-map witness (BootstrapAuthorityKey threading the existing CompilerAuthority classifier), refining the constructor surface if needed — NOT introducing a new enum variant. STOP-and-PING criteria updated to forbid new-variant resolution under any path. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): descent_execution_proof worker brief — γ ratified (2-variant residual) Director ratification at gunbc#828 #issuecomment-4395060514: - DescentResidual = EvidenceUnknown(DescentEvidence) | EvidenceIncomplete (4 → 2 dissolution: Missing+Unknown fold via DescentUnknown injection; NonStrict folds via NonIncreasing wrap; Incomplete retained — different concern axis from evidence-lattice) - DescentEvidence 3-variant lattice at termination.dag:14-17 confirmed as authoritative composition target - Type signature from §10.3 row 966 confirmed verbatim-binding Director-asked canvas-shape verification absorbed: worker STOPs if EvidenceIncomplete decomposes into payload-variants (timeout / depth-bound / evaluator-error-during-proof-construction); proceeds as 2-variant otherwise. 7 same-slice acceptance gates incl Evaluator E2 #1971 consumer wiring in same PR + §10.3 row 966 row-text refresh to cite γ-disposition. Worker pin: quick-koi-190 (pre-authorized per §10.3 row 966). Auto-spawn HOLD per L-sized threshold; surgical-recreate path ratified case-by-case if critical path blocked. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * WIP: R3 Substrate Mgr — lane through R3 close * docs(briefs): descent_execution_proof — narrow EvidenceUnknown payload via NonStrictEvidence subset; delete superseded canvas openai-pro REQUEST_CHANGES at gunbc#2105: 2 findings. 1. BLOCKING (LAYER MODEL / illegal states unrepresentable): worker brief shape `EvidenceUnknown(DescentEvidence)` admitted EvidenceUnknown(Strict) even though the canvas itself acknowledged Strict-isn't-a-residual as illegal. P2 requires API-level enforcement, not prose convention. Fix: introduce typed subset `NonStrictEvidence = NonIncreasing | DescentUnknown`; residual now `EvidenceUnknown(NonStrictEvidence)` — illegal-states-unrepresentable by construction. NonStrictEvidence lands in same file as DescentResidual; composes with existing 3-variant DescentEvidence via inhabitation, not re-definition. 2. NON-BLOCKING (live-state drift): canvas + worker brief both visible with "ratification needed" vs "ratified" status — same P2 single-authority shape as the S5 canvas/worker-brief co-existence at #2079. Fix: delete docs/briefs/r3-substrate-descent-execution-proof-canvas.md (worker brief frontmatter already names it as superseded; with canvas gone the live authority is unambiguous). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): descent_execution_proof — section-anchor cite for §10.3 row codex BLOCKING at gunbc#2105 line 47: my brief cited "§10.3 row 966" but the actual line is now 986 (after my Q-Reification row insert in PR #2079 shifted §10.3 by 20 lines). Reviewer's "no such section" claim is wrong on substance (file + section + row all exist) but the line drift IS real. Fix per feedback_section_anchors_over_line_numbers (Director calibration in gunbc#2079): replaced all "§10.3 row 966" with "§10.3 Q-EVAL-Descent- Termination-Contract row" — name-anchor instead of line-anchor, drift-immune. 5 occurrences cleaned (closure predicate, acceptance gate #3, gate #5, STOP criterion, worker pin justification). Stylistic "row row-text" repetition collapsed to "row text". Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): T-CostLens-Composition canvas — Lens<SymbolicCost> composition shape (#1957) Pre-staged per Director endorsement of T-CostLens-Composition canvas-shape authoring. Surfaces 3 composition options for the Lens<SymbolicCost> instance: α: two separate lenses, externally joined — REJECTED (violates §1.8 gate #39 no_coercion_cost_dimension by construction) β: single Lens<SymbolicCost> with composed witness in read — strong but requires Lens<C> carrier-shape refactor (target-context threading) γ (recommended): Lens<SymbolicCost> reads structural cost via algebra-fold + target-realization composed via existing Lookup<SymbolicCost> substrate at lookup.dag:48-60 — preserves generic Lens<C> carrier; satisfies all 4 §1.8 gates (#37-40) by construction; aligns with feedback_audit_adjacent_authority_first Adjacent substrate verified at HEAD: lens.dag:70-77 (Lens<C>), algebra.dag:12+ (SymbolicCost 7-variant + Semiring), lookup.dag:48-60 (Lookup<SymbolicCost> + MissingCost lens-boundary). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): T-Workflow-As-Data canvas — audit-receipt-honoring scope-narrowing Pre-staged per Director endorsement. Q-Workflow-As-Data-Carriers (§10.3 row 983) named 5 carriers; grep-verified at HEAD that 4 of 5 ALREADY EXIST in dsl/extdeps/github/actions.dag (218 lines). Only WorkflowSecret<Name> is wholly net-new substrate. Surfaces 3 options: α: maximalist 5-carrier introduction in dsl/std/workflow.dag — REJECTED (admits parallel-representation debt vs audit-receipt #1771 reuse-first directive) β (recommended): minimalist — only WorkflowSecret<Name> + Cron<Schedule> refinement net-new; lens consumes extdeps.github.actions directly. Honors feedback_audit_adjacent_authority_first. γ: like β + WorkflowObservationAnchor for typed lens-consumption-shape; natural ratchet from β if evidence accumulates. Sequencing: dispatch-ready post-T-LBP COMPLETE per §S4 design-schedule:95; brief authoring lands in advance per pre-staging discipline. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): T-CostLens-Composition worker brief — γ ratified; delete canvas Director ratification at gunbc#828 #issuecomment-4395691775: - γ option ratified (Lens<SymbolicCost> + Lookup<SymbolicCost> composition) - Lens<C> generic at lens.dag:70-77 confirmed authoritative (no refactor in scope) - symbolic_cost_dimension: AnalysisDimension<SymbolicCost> defers to separate Dimensions sub-lane Critical reframing absorbed: src/v3/lenses/cost.dag ALREADY EXISTS (status: STRUCTURALLY TERMINAL; BEHAVIORALLY PROXY). T-CostLens-Composition is behavioral-completion + target-realization-wiring, NOT P1 carrier introduction. Worker reads existing lens; advances PROXY → BEHAVIORALLY COMPLETE via Lookup<SymbolicCost> composition. 8 same-slice acceptance gates incl §1.8 #37-40 + #70 demonstration + lens status header refresh + §10.3 row text refresh. Out-of-scope (deferred per Director): symbolic_cost_dimension; Lens<C> generic refactor (STOP-and-PING if implementation reveals need). Canvas deleted per single-authority discipline (same precedent as S5 + descent_execution_proof canvas deletions). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * Delete superseded T-CostLens canvas (worker brief is single live authority) * docs(briefs): T-WAD Slice 1 worker brief — β ratified; delete canvas Director ratification at gunbc#828 #issuecomment-4395945465: 4 asks confirmed: 1. β ratified (minimalist reuse-first) 2. #1771 audit-receipt binding precedent confirmed 3. WorkflowSecret<Name> folds into dsl/extdeps/github/actions.dag (provider- specific scope; NOT new dsl/std/ file unless cross-provider evidence) 4. §1.8 gates #54 + #55 split into separate slices (slice 2 = timing-and-pattern; slice 3 = ci_workflow_modeled_as_dag) Slice 1 net-new substrate (only): - WorkflowSecret<Name> + SecretScope carriers - CronExpression + CronField (typed refinement of existing WorkflowTrigger::Schedule { cron: String } at actions.dag:43) Other 4 carriers from §10.3 row 983 reused as-is from extdeps.github.actions per audit-receipt #1771 directive. 6 same-slice acceptance gates incl no-parallel-representation grep + gate #53/#62/#63 advancement + bootstrap regen + clippy. Cross-provider STOP-and-PING per Director ask #3 caveat: worker greps adjacent provider work for WorkflowSecret-shape evidence; surfaces if found before finalizing fold-into-extdeps. Canvas deleted per single-authority discipline (S5 + descent_execution_proof + T-CostLens precedent). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): T-LAS Demo (#1952) complexity-contract compile-error worker brief Pre-staged execution brief per Director endorsement of T-LAS demos via direct worker brief path (no canvas — design-lock at docs/design-lens-application- surface.md specifies fixture shape verbatim at line 292). 7 same-slice acceptance gates: fixture program (O(n²) body + Enforce O(log n) budget) + TestClaim assertion + diagnostic structural validation + no regression on T-LBP cementing + bootstrap regen + clippy + §1.8 #92 advancement. Hard prerequisites STOP-and-PING: T-LAS Slice A landed (gates #88-#91) AND T-LBP complexity-lens BEHAVIORALLY COMPLETE (gate #79). Worker does not author against partial substrate. Sibling demos #1953 (CRDT cost) + #1954 (memory-peak cost) share the same hard-prerequisite + TestClaim shape pattern — could dispatch as 3 sequential PRs or single multi-demo PR (worker's call at dispatch). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): T-LAS Demos #1953 + #1954 — cost-basis sibling worker brief Pre-staged sibling brief covering both cost-basis demos (CRDT cost basis + memory-peak cost basis) per Director endorsement of T-LAS demos via direct worker-brief path. Single brief for two demos because they share: - Hard-prerequisite pattern (T-LAS Slice A + T-LBP cost-lens BEHAVIORALLY COMPLETE) - TestClaim + Diagnostic structural validation shape - apply_lens(cost, ...) Enforce mode Sibling of #1952 brief (complexity-contract compile-error). Worker's call at dispatch on multi-demo PR vs sequential PRs (Mgr ratification needed if going multi-demo). #1953 (CRDT) substrate per design §4.2 + cost-basis-declaration audit at docs/audit/t-user-authored-cost-basis-discipline-worked-examples.md. #1954 (memory-peak) substrate per design §4.3; STOP-and-PING if Dimension<SymbolicCost> substrate (deferred to Dimensions sub-lane per Director ratification at gunbc#828 #issuecomment-4395691775) not yet landed. 6 same-slice acceptance gates per demo + same STOP-and-PING discipline as #1952. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): T-CostLens — update Sub-issue from #1957 to #2141 (post-surgical-recreate) PM executed Director's surgical-recreate ratification (gunbc#828 #issuecomment-4397693699) at 17:54:43Z: closed #1957, created fresh #2141 to trigger pollAutoSpawn fresh-creation path. Worker fierce-ram-21 (#2153) spawned on #2141. 1-line brief update per Director's queued-action commitment: brief now references #2141 + cites surgical-recreate authority. #1957 closed-as-superseded. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): Q-Lens-Target-Context canvas (β-extended substrate-shape question) Director α-revised supersession at gunbc#828 #issuecomment-4400920572 elevated β-extended path (introduce first-precedent .dag-side fold-over-realization-rows + name active-target authority shape) from slice-tier to canvas-tier. Authored per Q-PAFS template: - Substrate-state at HEAD grep-verified (Lens<C>.read no target-context; zero .dag-side fold-over-realizations precedent; 15+ lens instances using generic Lens<C>) - Binary question: should .dag-side lenses receive target-context for target-keyed reading? - Options matrix: (i) LanguageSpec param to fold; (ii) per-target lens instances [parallel-representation debt]; (iii) global accessor [REJECTED global-state anti-pattern] - Mgr provisional preference: (i) - Cross-cutting: cost.dag load-bearing; emission_provenance.dag secondary; other 13+ lenses target-agnostic Framework discipline anchors per Director corrections: - feedback_same_slice_dissolution_discipline - feedback_parallel_representation_debt - feedback_abstraction_layering (sibling canvas) Sibling canvas (ε path): q-cost-composition-layering-canvas.md authoring follows. Both canvases together address deeper cross-cutting axis: does target-context belong .dag-side (β-extended) or emit-side (ε)? Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): Q-Cost-Composition-Layering canvas (ε substrate-shape question) Sibling canvas to q-lens-target-context-canvas.md. Director α-revised supersession at gunbc#828 #issuecomment-4400920572 elevated ε path (Rust-side cost-composition wiring; preserves cost.dag PROXY) from slice-tier to canvas-tier. Authored per Q-PAFS template: - Factoring claim: cost = (target-agnostic-shape SymbolicCost algebra) × (target-specific-values per-primitive realization-cost data loaded Rust-side) - Test against feedback_abstraction_layering (Director-named anchor at gunbc#828 c#4400921658): pure objective concepts × language-agnostic LanguageSpec × emit-side composition. Compliance test for each layer. - Pro factoring (ε structurally honest): SymbolicCost algebra is target- agnostic at HEAD; Rust-side composition is natural home; #38/#39 substrate- already-aligned per Slice 1 finding. - Con factoring (ε is parallel-representation debt): N parallel Rust-side compositions if multiple lenses need target-context; lens-framework abstraction stops at substrate boundary; future lenses hit same dilemma. Mgr provisional reading: factoring honest for COST specifically; NOT generalizable. ε right framing IF cost is the singular need + factoring holds. If N>1 target-context-needing lenses surface, β-extended (option (i) from sibling canvas) becomes the right path. Director ratification ask: 3-way choice (ε) Cost-specific Rust-side composition; preserves cost.dag PROXY (β-extended option (i)) Lens<C> refactor for LanguageSpec parameter (both sequenced) ε now for cost; β-extended later if N>1 Framework discipline anchors: - feedback_abstraction_layering (Director-named for this canvas) - feedback_parallel_representation_debt - feedback_same_slice_dissolution_discipline Canvas-pair complete; sibling canvas + this canvas address the deeper cross-cutting axis: target-context belongs .dag-side (β-extended) or emit-side (ε)? Director ratification across both informs T-CostLens follow-on slice + downstream lens architecture. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * WIP: R3 Substrate Mgr — lane through R3 close * WIP: R3 Substrate Mgr — lane through R3 close * docs: ε ratified canonical for cost; β-extended deferred to N=2 — gate updates Director ratification at gunb-ai/gunbc#2181 #issuecomment-... : - (ε) RATIFIED standalone for cost composition (canonical-not-transitional) - (β-extended option (i)) DEFERRED to N=2 trigger event (second lens with target-context need beyond cost) - ("both sequenced") REJECTED as bridge-with-named-dissolution anti-pattern Same-slice acceptance gates landed: 1. Q-Cost-Composition-Layering canvas: status updated from "ratification needed" → "Director-RATIFIED 2026-05-07 (ε standalone; canonical-not- transitional); PROPOSAL maturation pending" 2. Q-Lens-Target-Context canvas: status updated from "ratification needed" → "DRAFT/DEFERRED 2026-05-07; reopens on N=2 trigger event" (second lens with substantive target-context need; emission_provenance most likely candidate per cross-cutting analysis §3) 3. r3-program-plan.md gate-table rows #37 + #40 + #70: ε path ratified; close via Rust-side composition reading abstract SymbolicCost × per- primitive realization-cost in T-CostLens follow-on slice 4. r3-program-plan.md §10.3 T-CostLens-Composition row: ε ratified for #37/#40/#70; β-extended deferred per N=2 trigger 5. v3-lens-capability-register.md cost.dag row: ε disposition cited; #2175 continues as cross-cutting umbrella tracker Closed-system disposition per Director: if N=2 condition never fires, β-extended never fires — structurally correct under closed-system design. Framework discipline anchors honored: - feedback_abstraction_layering: ε respects layering (objective concepts pure; target-specific values flow at emit time per Layer-3 honesty test) - feedback_parallel_representation_debt: bounded to N=1; reopens at N=2 - feedback_same_slice_dissolution_discipline: ε ratified canonical, not transitional - feedback_construction_over_ratchets: substrate-shape question answered structurally - feedback_substrate_principle_audit: substrate-fact authored at canvas- tier; ratification follows P1 procedure Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): canvas inventory grounded in live src/v3/lenses/ registry codex BLOCKING at gunb-ai/gunbc#2181 (sha 00551a80): Q-Lens-Target-Context canvas inventory was "copied from expected lens set instead of live src/v3/lenses/ registry"; emission_provenance trigger references not grounded in live file content. Same shape of error as 3 prior BLOCKING reversals (Q-Reification, S5 DeclarationRef, T-CostLens merge-content). Substrate-state-grep is being treated as retrospective-correction rather than prerequisite. Tightening: this is the 4th occurrence; should be hard-prerequisite-discipline going forward. Fixes: 1. Q-Lens-Target-Context inventory section: replaced approximate "15+ lens instances" framing with exact `ls`-grep registry (15 .dag files listed by name), notes infer_helpers + lower_helpers are helper modules authoring shared structural fns rather than top-level lens instances. 2. emission_provenance.dag analysis grounded in HEAD file status header: STATUS = STRUCTURALLY TERMINAL; LENS-INSTANCE FIXTURE-BOUND; BEHAVIORALLY DEFERRED. `read` body is fail-closed Empty stub. Lens does NOT currently read target-context inside fold; producer-side instrumentation records target-specific entries consumed via standard framework. Reframed as "ungrounded at HEAD" rather than "POSSIBLY target-context need" — re-evaluates at producer-wiring landing. 3. Net finding: N=1 confirmed at HEAD (cost.dag only); N=2 is empirically open pending lens-completion cycles, NOT pre-claimable. Plus non-blocking improvement: Q-Cost-Composition-Layering line 9 stale `#issuecomment-...` placeholder replaced with concrete ratification comment id `#issuecomment-4401584012`. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): fix Director-ratification-ask typo + #37 gate-id clarification footnote cursor APPROVE_WITH_COMMENTS at gunb-ai/gunbc#2181 (sha 4209eb7f) flagged 2 NON-BLOCKING items + 1 exploratory: 1. (NON-BLOCKING) Q-Cost-Composition-Layering line 69: `## Directorratification ask` → `## Director ratification ask` (whitespace typo) 2. (NON-BLOCKING) Q-Cost-Composition-Layering line 9 placeholder `#issuecomment-...`: ALREADY ADDRESSED in commit `db88b1004` (replaced with `#issuecomment-4401584012`) 3. (Exploratory) Gate #37 id `cost_lens_reads_target_realization` framing implies .dag lens body performs realization read; per ε ratification it's Rust-side composition consumer. Added clarifying footnote noting gate-id retention + Rust-side closure path; rename candidate post-follow-on-slice landing. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): Q-Lens-Target-Context — absorb PM canvas-review concerns 1-3 Director relayed PM canvas review at gunb-ai/gunbc#2068 c#4401627536: Concern 1 (RESOLVED via grep-verify): empirical N=1 confirmation. PM ran `grep -lc 'TypeRealization|CallableRealization|MethodTemplateContract| target_realization|realization_cost|LanguageSpec' src/v3/lenses/*.dag`; only cost.dag has 3 hits, all 14 others have 0 refs (including emission_provenance.dag which I'd previously framed speculatively). Net-finding §3 updated with grep result inline + "N=1 empirically grounded" framing replacing speculative "secondary candidate" language. Empirical basis for DEFERRED β-extended disposition strengthened. Concern 2 (Option (ii) multiplier framing): per-target lens instances replicate the entire Lens<C> authoring surface — N×M × 4 (carrier + monoid sequential + branch + iterate) authoring overhead. Updated Con bullet to reflect the multiplier explicitly: 3 targets × 1 cost lens × 4 = 12 authored fns; grows to 36 if 3 lenses need target-context. Cites feedback_parallel_representation_debt as the P2 framing. Concern 3 (Option (i) threading cost quantification): replaced narrative "threading cost is real but manageable" with quantified breakdown — ~15 signature changes + 14 ignore-parameter patterns + 1 consumer + cementing- test revalidation. Helps future Director ratification at N=2 trigger event. All 3 amendments are docs-only refinements; ε ratification at gunbc#2181 c#4401584012 unchanged. Strengthens canvas as durable substrate-shape decision record. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): canvases — collapse stale ratification-ask sections post-ratification codex BLOCKING at gunb-ai/gunbc#2181 (sha be9a9c94): both canvases had stale "Director ratification ask" sections after status headers were updated to RATIFIED/DEFERRED. Per INVARIANTS P1 "Documentation Describes Live State" — canvas internally presented both settled and unsettled authority. Fixes: 1. Q-Cost-Composition-Layering line 69-76: "Director ratification ask" section collapsed to "Director ratification (RECEIVED 2026-05-07)" with ε ratified, β-extended deferred, "both sequenced" rejected. Three options recorded as historical with ratification cite. Eliminates the rejected "both, sequenced" line that conflicted with status-header "canonical-not- transitional" framing. 2. Q-Lens-Target-Context line 131-136: "Director ratification ask" section collapsed to "Director disposition (DEFERRED 2026-05-07; reopens on N=2 trigger event)". Original ratification asks recorded as FROZEN; revisit at N=2 trigger event with then-current substrate-state grep. Eliminates live-now-ratify framing that conflicted with status-header DEFERRED. Both canvases now single live authority — RATIFIED/DEFERRED dispositions, no internal ratification-ask drift. Future re-ratification (Q-Lens-Target- Context at N=2 trigger) refreshes options matrix at that point per substrate-state-honesty discipline. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): Q-Lens-Target-Context — separate live N=1 from reopen-only future N=2 openai-pro APPROVE_WITH_COMMENTS at gunb-ai/gunbc#2181 (sha be9a9c94): canvas line 127 mixed live N=1 authority with speculative N=2 candidate in one current-state cost count. Per P1 Documentation Describes Live State + P2 single-authority metadata: the live count at HEAD is `cost × 3 targets = 3 per-target instances` (× 4 authoring-multiplier per Option (ii) Con = 12 fns). The `emission_provenance × 3` figure is a reopen-only future scenario that materializes only if emission_provenance becomes the second real target-context lens at producer- wiring landing time. Updated to separate the two clearly: live count + reopen-only future scenario marked separately. Aligns with the grep-verified N=1 finding established earlier in canvas §3. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * WIP: R3 Substrate Mgr — lane through R3 close * docs(briefs): T-CostLens worker brief — ε supersession header (γ retained as context) PR #2181 merged at eff426de5 ratifies ε path canonical-not-transitional for cost composition. Brief was authored under γ scope (.dag-side Lookup<SymbolicCost> composition). Add binding ε supersession header at top; retain γ section as historical context per single-authority discipline. cost.dag stays PROXY under ε; composition is Rust-side (abstract SymbolicCost shape × per-primitive realization values). Authority: Director ratification at gunb-ai/gunbc#2181 #issuecomment-4401584012. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): Q-Complexity-Composition-Layering canvas — DRAFT (factoring tested, ε precedent does NOT apply) Director authorized canvas authoring at gunb-ai/gunbc#828 c#4402669133 as parallel structure to Q-Cost-Composition-Layering (ε RATIFIED). Substrate- grep on src/v3/lenses/complexity.dag at HEAD shows the factoring claim DIFFERS from cost-lens: - Cost-lens needed (target-agnostic shape × target-specific values) factoring → Rust-side composition (ε) - Complexity-lens has NO target-specific axis; output is structural property of DAG topology + algebra-instance composition; substrate- native fully-on-.dag-side completion Mgr provisional reading: ε precedent does NOT apply; complexity-lens BEHAVIORAL COMPLETION is direct slice-tier substrate authoring (carrier introduction follows SymbolicCost precedent + T-E-P P1 carrier consumption). Director ratification ask: Q1 (factoring finding correct), Q2 (slice-tier directly bypassing canvas), Q3 (fresh worker pin). Cross-Mgr: Verification Mgr (#2075) pinged on v2-oracle snapshot capture status (cross-cutting prerequisite for #1950/#1951 cementing dispatch). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): T-LBP complexity-lens substrate-completion worker brief (Q1/Q2/Q3 RATIFIED) Pre-authored brief for complexity-lens BEHAVIORAL COMPLETION substrate work per Director Q1/Q2/Q3 ratification at gunb-ai/gunbc#828 c#4402714255. Three deliverables: carrier introduction (ComplexityCost / WorkSpan / AsymptoticClass following SymbolicCost precedent) + lens widening (complexity.dag PROXY → COMPLETE) + T-E-P P1 carrier consumption (DescentEvidence / CallPattern / SubValueRelation for asymptotic classification of recursive-call behavior). Indirect-variant dependency surfaced as worker-grep concern at slice- authoring time (T-E-P P1 Slice 5+ pending; eager-bat-178 stream). Cementing test (#1950) is separate downstream brief; v2-oracle snapshot ownership pending Q4 cross-Mgr ratification. Worker pin: fresh-pool pick at dispatch time (NOT eager-bat-178 per Director Q3 framing; NOT fierce-ram-21 busy with cost-lens ε). Same-window-dispatch discipline applies post-canvas-merge (PR #2197). Pre-authored vs pre-known discipline applied per feedback_pre_known_vs_pre_authored_briefs.md memorialized 2026-05-08. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(briefs): rewrite complexity-lens brief consuming live design authority (codex BLOCKING fix) Codex BLOCKING (3 findings) at #2199 sha 7a8bb6dc: 1. Brief authored against missing canvas instead of reconciling with docs/design-complexity-lens-behavioral-completeness.md (which exists at HEAD with comprehensive substrate spec) 2. Producer coverage treated as optional corpus scope; should be hard T-E-P-Producer-Broadening prerequisite 3. Stale/non-in-repo planning authority cited; should be r3-structure.md row 146 (T-LBP slice 1) All three BLOCKING findings VALID — substrate-grep-before-authoring discipline failure on my part (feedback_substrate_grep_before_authoring already memorialized; violated own discipline). Rewrite delegates carrier shape, dimension wiring, and consumer rewrite to live design doc §§1.1-1.6 verbatim: - §1.1 SymbolicCost — already at algebra.dag; no change - §1.2 SizeVariable — display_name enrichment - §1.3 work_dimension + span_dimension — two AnalysisDimension instances - §1.4 AsymptoticClass + BoundedLattice instance - §1.5 Certainty (cost-aware composition; no lattice) - §1.6 cost_bound_to_symbolic projection - §1.7 ComplexitySummary record + lens widening T-E-P P1 hard prerequisite per design's authority discipline + r3- structure.md row 146; STOP-and-PING if T-E-P P1 not COMPLETE at slice authoring time. Authority cites updated to live r3-structure.md row 146. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): Q-PerfWithinBaseline canvas — DRAFT (canvas-tier P1 procedure per Q6 RATIFIED) Director Q6 RATIFIED canvas-tier-required at gunb-ai/gunbc#828 c#4403163639 for PerfWithinBaseline TestPredicate variant authoring. Three substantive substrate-shape questions resolved: Q1 baseline shape: (a) literal-Int rejected as strict-mirror-of-CostBounded defeats semantic load; (b) DeclarationRef-to-stored-data composes LensOutputEquals + data X: SymbolicCost precedent at HEAD; (c) runtime- fixture-injection over-authors. Mgr lean (b). Q2 ComparisonOp composition: (i) reuse existing ComparisonOp via test- runner-side resolution matches LensOutputEquals path; (ii) new relative- op-set introduces parallel-representation debt. Mgr lean (i). Q3 baseline-storage scope: (α) in-scope slice authors example data; (β) out-of-scope variant-only slice + PB consumer authors baseline data matches existing layering. Mgr lean (β). Combined Mgr lean: Q1(b) + Q2(i) + Q3(β) — variant authored as compositional extension; baseline-storage is PB consumer-tier work. Cross-Mgr: PB Mgr #2138 worker (crisp-swift-433) holds dispatch on #2204 land; T-Tier3-Dissolution #2085 R-4 prereq encoded. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): Q-PerfWithinBaseline canvas REVISED (supersedes wrong-shape original) Director disposition at gunb-ai/gunbc#828 c#4403265220 authorized supersession after BLOCKING at PR #2209 c#4403246233 verified VALID: original canvas conflated symbolic SymbolicCost (cost-lens substrate) with wall-clock measured median/p99 baseline (T-Tier3 perf gate substrate per docs/r3-structure.md:225 + :461 Director-locked 2026-04-28). Revised canvas frames as two-path ratification: - P1 author full perf-budget substrate in-R3 (multi-slice; pattern-5 genuinely-novel substrate-fact-introduction) - P2 explicitly post-R3 per Director-locked recommendation (zero in-R3 effort; structural close per r3-structure.md:461 'R3 deliverable is structural close; perf is downstream') Mgr lean: P2 — Director-locked recommendation explicit; trigger- condition ('someone authors perf-budget claim with concrete numbers and tooling') not met; R3 horizon constraint argues against multi- slice perf-budget substrate adding to 5-orthogonal-dependency picture. Original canvas at q-perf-within-baseline-canvas.md retained with SUPERSEDED-BY header per durable-decision-record discipline. 5th discipline-failure of 2026-05-08 cycle: substrate-grep verified substrate-precedent but missed consumer-side requirement at canonical authority doc; Director self-flagged symmetric two-axis verification gap; canvas-finding-taxonomy needs precondition 'consumer-side requirement grep-verified at canonical authority doc'. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): Q-PerfWithinBaseline canvas-revised — correct trigger-condition framing per Director c#4403297623 Director caught second-axis grep failure on revised canvas: my P2 reasoning #1 ("path-trigger condition for P1 is not met") was structurally wrong. Existing tooling at HEAD substantially meets the Director-locked 2026-04-28 trigger-condition ("someone authors the perf-budget claim with concrete numbers and tooling"): - docs/briefs/r3-pb-tier3-perf-budget-worker.md - docs/audit/c1-tier3-perf-budget-readiness-matrix.md - src/v3/compiler/benches/tier3_mirror_perf.rs - docs/audit/c1-tier3-baseline-capture-procedure.md - docs/audit/c1-r3-canonical-bench-host-decision-matrix.md Plus thresholds (≤2× median, ≤5× p99) + capture discipline (N=3 min, N=5 preferred) + canonical bench host option matrix. P1 is bridging existing tooling to substrate (compositional-extension), NOT multi-slice greenfield genuinely-novel pattern-5 as originally framed. Smaller scope than canvas-finding-taxonomy pattern 5 implies. Path-call genuinely depends on PB Mgr's R-3 (canonical CI bench host) + R-7 (tier3_baseline.json baseline-capture path) decisions per their audit response at c#4403059897. Path-call DEFERRED to cross-Mgr coordination outcome with PB Mgr (#2074); Substrate Mgr surfaces with corrected framing this turn. Sixth discipline-failure of cycle (mine, second-axis grep gap on existing-tooling state); same-class as Director's first-axis grep gap on consumer-side requirement at the prior turn. Memorialized as two-axis verification discipline anchor. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): Q-ValueBody-Drift-Resolution canvas — DRAFT (codegen-generation lean) D1 substrate-shape question per Q-ValueBody-Isomorphism RATIFIED at gunb-ai/gunbc#828 c#4403972737. Variant-inventory readiness input from PR #2218 (merged 2026-05-08) confirms drift: Rust 5 variants vs substrate 3 constructors; asymmetry on Scalar+List Rust-only + Map payload-parity-with-semantic-gap. Two-axis verification at HEAD: - Substrate-precedent: codegen tooling exists (regen_bootstrap_emit; 5 _generated.rs files; regen_bootstrap binary). Pattern-3 strict- mirror codegen extension applies; no pattern-5 P1 procedure. - Consumer-side requirement: V1 worker brief explicitly names mirror- parity-vs-codegen-generation as D1 path-pair. Mgr lean: (b) codegen-generation. Path (a) mirror-parity perpetuates parallel-representation debt; (b) dissolves the dual-taxonomy class by construction (substrate canonical, Rust codegens). Map dup-key gap handled via (ii) Rust-side post-codegen wrapping (substrate stays minimal). Director ratification ask: D1 (b) + D1-followup (ii). Carrier slice path post-ratification: extend regen_bootstrap_emit; add Scalar+List constructors to substrate; regen + remove hand-Rust enum; handle Map dup-key per ratified followup. Worker pin fresh-pool per same-window-dispatch. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(proposals): Q-ValueBody-Drift-Resolution canvas amend authority cite (codex BLOCKING fix) Director ratified (α) amendment-PR at gunb-ai/gunbc#828 c#4404398425 post-codex-BLOCKING at PR #2222 c#4404360699. Canvas originally cited regen_bootstrap_emit.rs as codegen authority for ValueBody runtime mirror; corrected to scripts/regen_runtime_mirrors.py which is the substrate→Rust runtime-mirror generator (reads substrate.dag, emits dag_scalar_generated.rs etc.). regen_bootstrap_emit.rs is a separate codegen system that generates bootstrap_generated.rs (bootstrap parser/lower compile snapshot). D1 (b) codegen-generation ratification UNCHANGED; only authority-path cites corrected. Carrier slice path at canvas §6-step plan now references correct runtime-mirror generator. Canvas inline notes mark amendment location for future-reader audit trail. 8th cycle-tier framing-failure (Mgr-tier; same fine-granularity gap parallel to Director-tier failure at c#4404256128). Two-axis verification discipline applied at insufficient granularity (verified codegen-tooling- exists but didn't disambiguate against bootstrap-vs-runtime-mirror systems). Discipline anchor refinement: when canvas cites a codegen system / tooling / authority path, identify the SPECIFIC generator/handler for the target-file-class. Don't accept 'tooling exists' at coarse granularity. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs: reconcile cross_target_coverage emission_path_projections post-population drift cross_target_coverage.dag:179 + sg0_census_test.rs:355 still described emission_path_projections as `(empty data state)` / `== []`, but the data block at cross_target_coverage.dag:186 is populated with 41 Phase-1 rows (Rust 13 / Python 16 / Go 12). Update both comment sites to reflect the populated state. P1 Modeling Faithfulness — comment-drift only, no behavior change. Surfaced by gentle-newt-665 R3 Debt-Paydown ROADMAP audit at gunb-ai/gunbc#2062; absorbed into Substrate lane (#1939) per L6 emission_path_projections carrier ownership. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix: pad cross_target_coverage.dag comment to preserve byte offsets Previous comment edit changed line 179 length 76→75, shifting SourceSpan byte offsets in bootstrap_generated.rs by 1 and breaking regen_bootstrap --verify in CI. Pad with one extra dash to restore original 76-char length; comment text intent unchanged. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * WIP: R3 Substrate Mgr — lane through R3 close * fix(canvas): clarify Source param doesn't carry subject identity (P2) Reviewer flagged on PR #2333 inline review: parametric form WorkflowObservationAnchor<BehaviorId, TimingMeasurement> would create redundant subject-identity authority if TimingMeasurement carries subject identity (P2 violation). Clarify Q-WAD-S2-Anchor (b) so Source parameter is observed-payload-only; subject identity is owned solely by Subject parameter. In worker's actual PR #2360 shape, TimingMeasurement is variant-typed report state (Observed { nanoseconds } | Missing | ...) and does NOT carry subject identity — concern is theoretical there but worth clarifying canvas language for clean Director ratification. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): correct fail-closed analysis for LensEnforcement.violates signature Reviewer flagged BLOCKING on PR #2333 inline review (canvas:90): LensEnforcement.violates signature is (Budget, Budget) -> Bool per src/v3/std/lens_application.dag:100 — does NOT see raw Output. Original canvas claim 'fail-closed: violates = Output != Observed' was structurally wrong. Updated: - (a) clarified: projection step (Output → Budget) must fabricate violating Budget for non-Observed variants (option (a)(i)) OR substrate-extend the violates signature (option (a)(ii)) which is canvas-tier and cascades across all lens consumers. - Added option (c) reflecting worker tidy-raven-610 PR #2360 shape: Output folded into TimingMeasurement carrier directly, Budget projection handles fabrication naturally. Status: shape ratification still pending Director per #828 c#4412018726. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): add 🟢 dissolution classification to TimingObservationOutcome variants Reviewer flagged BLOCKING on PR #2333 inline review (canvas:86): proposed sum has no 🟢/🟡/🔴 dissolution receipt per INVARIANTS P5 / modeling-discipline.md §coproduct dissolution. Worker could land unclassified coproduct. Added 🟢 GREEN (terminal) classification to both option (a) TimingObservationOutcome and option (c) folded TimingMeasurement — four variants exhaust externally-attested observation states; no richer-source-extraction path exists at the workflow-observation boundary. Worker .dag declarations MUST carry the 🟢 marker comment per modeling-discipline.md:132. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): correct Q-WAD-S2-Placement to src/v3/std/ (not dsl/std/) Reviewer flagged BLOCKING on PR #2333 inline review (canvas:108): recommending dsl/std/timing_lens.dag conflicts with src/v3/std/lens.dag:17-21 v3-only-carriers convention. Worker would land substrate in wrong layer. Earlier canvas claim 'T-LBP / T-CostLens / T-LAS lens carriers all live in dsl/std/' was factually wrong — they live in src/v3/std/ and src/v3/lenses/. Corrected to src/v3/std/timing_lens.dag throughout. Worker PR #2360 already placed at correct path; canvas now matches. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): make six gate #55 invariants concrete fields on anchor (P2) Reviewer flagged BLOCKING on PR #2333 inline review (canvas:75): the recommended <Subject, Source> parametric shape only named type params and didn't assign digest/producer/observer/prover/timestamp/run-id/ report-state to a single carrier — gate #55 facts wouldn't flow forward under P2 boundary discipline. Fixed: Q-WAD-S2-Anchor (b) now shows explicit six-invariant field schema as concrete fields (subject/artifact_digest/producer_id/ observer_id/prover_id/attached_at/workflow_run_id/observation_outcome) with Subject + Source as type parameters for variable parts only. Notes worker tidy-raven-610 PR #2360 flat shape is already gate-#55-compliant at the field level; convergence path is adding type parameters to the flat shape (minimal rework). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): correct convergence assessment — invariant 5 split-authority in worker shape Reviewer flagged BLOCKING on PR #2333 inline review (canvas:96): claiming worker's flat anchor shape gate-#55-compliant drops invariant 5 (observation_outcome / report-state) from the anchor. Worker's TimingMeasurement variants carry report-state on the payload, not on the anchor — splits P2 single-authority for invariant 5. Fixed: convergence assessment now correctly identifies 5/6 invariants on anchor + invariant 5 split. Path requires TWO edits not one: (1) add <Subject, Source> type parameters to anchor; (2) add observation_outcome: ObservationOutcome<Source> field to anchor itself. This collapses Q-WAD-S2-Output (c) folded shape back to (a)(i) separate-projection — trade-off documented (single-authority preserved at cost of one extra type). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): reject (a)(i) Budget fabrication per P3/C-8 fail-closed (canvas:118) Reviewer flagged BLOCKING on PR #2333 inline review (canvas:118): recommending projection fabricates violating Budget for non-Observed states violates INVARIANTS P3/C-8 fail-closed discipline — erases typed Missing/Ambiguous/Stale failure evidence into a plausible budget value. Consumers downstream of violates can't distinguish 'budget exceeded by observed value' from 'no observation existed'. Fixed: (a)(i) REJECTED with explicit P3/C-8 rationale. Added (a)(iii) Result-typed projection with auto_violate_on_left bit — smaller substrate change than (ii) full-signature-extension. Director-tier disposition still required for (ii) or (iii); (i) no longer a viable path. Convergence implications for worker PR #2360 + canvas (c) folded shape need re-assessment in next pass — (c) sidesteps the violates signature collision but invariant 5 split-authority (per canvas:96 fix in 960a03f98) means observation_outcome belongs on anchor not payload, which then re-introduces the violates-Output-visibility question via ObservationOutcome typing. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): propagate (a)(i) rejection across all recommendation/status refs Earlier commit 594a4df9c rejected (a)(i) Budget fabrication per P3/C-8 inline review. Updated remaining 3 refs to reflect: - Q-WAD-S2-Anchor convergence trade-off line - Q-WAD-S2-Output recommendation: now (a)(ii)/(a)(iii) pending Director - Q3 (c) Status line: Director call is (a)(ii)/(a)(iii) vs (c) (a)(i) is REJECTED throughout; Director-tier LensEnforcement substrate change unavoidable for fail-closed-correct anchor-side report-state carriage. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): correct invariant-6 claim — fail-closed needs LensEnforcement substrate-extension Reviewer flagged BLOCKING on PR #2333 inline review (canvas:91): claim 'all six invariants concrete fields on anchor' was wrong — invariant 6 (fail-closed) explicitly lives at LensEnforcement.violates layer, not on the anchor. Without Q-WAD-S2-Output (a)(ii)/(a)(iii) substrate- extension landing, gate #55 cannot close on 5/6 — would let consumers attach observation facts without fail-closed evidence (P2/P3 violation). Fixed: clarified 5/6 invariants on anchor + invariant 6 dependency on Director-tier LensEnforcement substrate-extension. Gate #55 closure explicitly cross-linked to canvas:118 escalation thread. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): Output (a) variants wrap Source not TimingMeasurement (P2) Reviewer flagged BLOCKING on PR #2333 inline review (canvas:112): Output (a) Observed{value: TimingMeasurement} would wrap the report-state-bearing TimingMeasurement type, duplicating identity/attestation/report-state with the anchor's invariant 5 authority. Reopens P2 split-authority. Fixed: Observed wraps payload-only Source (e.g., TimingPayload { nanoseconds: Int }), NOT TimingMeasurement. Per Q-WAD-S2-Anchor revised convergence: subject identity / attestation / report-state on anchor; payload carries only observed-value-when-present. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): split gate #55 + canonicalize TimingPayload (parent BLOCKING b526a26f) Reviewer flagged 2 BLOCKING on PR #2333 parent review at sha b526a26f: Finding 1 (gate #55 mixes anchor-fact + fail-closed): added gate #55a/#55b split — #55a = anchor carrier landed (worker scope); #55b = LensEnforcement substrate-extension landed across all lens consumers (canvas-tier separate dispatch, owner: Substrate Mgr). Finding 2 (TimingMeasurement overload): added Naming canonicalization section to Carrier inventory. TimingPayload = observed-value-only; WorkflowObservationAnchor = observation record; ObservationOutcome = report state; TimingObservationSet = lens C (per Q1 ratification). Earlier TimingMeasurement references deprecated. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): rewrite Carrier inventory to reflect Director ratification of worker shape Reviewer flagged BLOCKING on PR #2333 inline review (canvas:35): old carrier inventory still said TimingMeasurement owns subject identity + observer attestation, contradicting revised anchor authority. Same-day Director re-ratification at #828 c#4412301889 reversed the canvas (b) recommendations and ratified worker tidy-raven-610's shape as-shipped: - TimingMeasurement = Observed | Missing | Ambiguous | Stale (carrier-as-report-state; owns invariant 5 via variants) - WorkflowObservationAnchor = timing-specific concrete fields (Q-WAD-S2-Anchor (a) ratified; owns invariants 1-4) - TimingObservationSet = aggregation collection - Lens<TimingMeasurement> per-observation Fixed: rewrote Carrier inventory to reflect Director-ratified shape; explicit invariant ownership (anchor: 1-5; LensEnforcement: 6 via gate #55b separate dispatch). TimingPayload canonicalization (which contradicted Director ratification) removed. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(canvas): resolve option (c) Pro/Con contradiction on signature collision Reviewer flagged BLOCKING on PR #2333 inline review (canvas:137): option (c) Pro line said 'avoids violates signature collision entirely' but lines 135/139 routed back to (a)(ii)/(a)(iii) — internal contradiction; fail-closed obligation under P3/C-8 still ambiguous. Fixed: rewrote (c) Pro/Con honestly — Pro = state-space discipline (per Director's ratification reasoning); Con = does NOT sidestep violates signature collision, still requires substrate-extension path. Status updated to reflect Director ratification at #828 c#4412301889: (c) carrier shape ratified for worker scope (#55a); LensEnforcement substrate-extension disposition (#55b) still pending between (a)(ii) and (a)(iii). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(v3): LensEnforcement.auto_violate_on_left for observation-driven lenses Per Director ratification at gunb-ai/gunbc#828 c#4412884371 + re-confirmation at c#4413159089: extend LensEnforcement<Output, Budget> with additive auto_violate_on_left: Bool field for observation-driven lens classes (T-Workflow-As-Data Slice 2 timing-lens et seq.). Substrate semantics: when Output is Result-typed (e.g., Result<Observed, ObservationFailure> for timing), runtime checks the bit before calling project; on Result-Left + bit=true, enforcement violates without fabricating a Budget value at the projection boundary. Preserves typed failure evidence (Missing/Ambiguous/Stale/ Unobserved) per INVARIANTS P3/C-8 fail-closed discipline. Backward-compatible additive: existing structural-static lenses (complexity, cost, T-LAS) set the bit to false; bit is inert when Output is not Result-typed. Updated complexity_enforcement declaration + hand-Rust LensEnforcement struct mirror. Closes §1.8 ledger #55 sub-gate b (LensEnforcement substrate-extension prerequisite for fail-closed-correct enforcement on non-Observed report states); enables T-Workflow-As-Data Slice 2 worker (tidy-raven-610 #2359) to revise PR #2360 against the ratified Result-typed Output shape. Canvas authority: docs/briefs/r3-substrate-t-wad-slice-2-timing-lens-canvas.md §"Question 3 (Q-WAD-S2-Output)" + Gate #55 closure-predicate split. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * Revert "feat(v3): LensEnforcement.auto_violate_on_left for observation-driven lenses" This reverts commit 5ec59209d1c561788a4f6f686d8d1f62daf086ef. * fix(canvas): Q-WAD-S2-Anchor recommendation aligned to Director ratification of (a) Reviewer flagged BLOCKING on PR #2333 inline review (canvas:99): canvas still recommended (b) generic anchor after Director ratified (a) timing-specific. Competing WorkflowObservationAnchor authorities = P2 violation. Fixed: Q-WAD-S2-Anchor recommendation rewritten to reflect Director RATIFIED (a) per #828 c#4412301889 + c#4413322671. Mgr-tier preliminary (b) was over-engineered (chased hypothetical second-consumer parameterization). Invariant 5 split-authority concern resolved at LensEnforcement layer per (a)(ii) full-signature extension (Director c#4413284764) — violates pattern-matches Output variants directly, no observation_outcome projection wrapper needed. Co-Authored-By: Claude Opus 4.7 …
briansrls
added a commit
that referenced
this pull request
May 9, 2026
…ro 8505) Program plan and module docs now split TypeRealization sequential proof vs CallableRealization structural cost readability; rename callable test and document no sequential pin in that subtest. Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
May 10, 2026
* test(v3): R3 gate #37 cost lens reads target realization Add integration tests that read SymbolicCost via symbolic_cost_of on a compiled program, read TypeRealization/CallableRealization cost from rust.dag bootstrap rows, and compose with Semiring sequential. Mark §1.8 gate cost_lens_reads_target_realization CONSUMER_LANDED and refresh T-CostLens-Composition lane note. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: R3 gate #37: cost lens reads target realization * docs(r3): gate #37 partial ε-slice receipt; keep cost_target_realization meta-only Align program plan + test header with integration proof scope (not emit-time LanguageSpec consumer; #40/#70 follow-on). Revert cost_target_realization.dag to the six declaration_by_name shims after failed generated cost-row walk. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: R3 gate #37: cost lens reads target realization * WIP: R3 gate #37: cost lens reads target realization * WIP: R3 gate #37: cost lens reads target realization * chore: remove session PR body stub from repo (.pr2433-body.md) Keep PR description text on GitHub only; the file is dashboard/session machinery and should not ship on main. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(v3-tests): parse LiteralBits::Int decimal string in realization_row_cost_int LiteralBits::Int now carries a signed decimal String payload (dag.rs); use literal_decimal_i64 so lens_cost_target_realization integration tests compile under RUSTFLAGS=-D warnings (v3 CI). Co-authored-by: Cursor <cursoragent@cursor.com> * docs(test): align gate #37 callable receipt with TESTING.md (openai-pro 8505) Program plan and module docs now split TypeRealization sequential proof vs CallableRealization structural cost readability; rename callable test and document no sequential pin in that subtest. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
May 10, 2026
…n pointer Re-land symbolic_cost × sequential × realization-row assertions in lens_cost_target_realization_test per TESTING.md / INVARIANTS P2; cite tests from §1.8 gate #37 row (partial ε-slice wording). Co-authored-by: Cursor <cursoragent@cursor.com>
6 tasks done
briansrls
added a commit
that referenced
this pull request
May 10, 2026
- Split cron schedule model (`dsl/extdeps/cron_schedule_model.dag`); wire Actions extdeps + tools - Add `t_ci_workflow_as_data_demo.dag` + integration tests; `_ci_wad_seed_*` list Empty rows for opaque Dag harness tags - PB-1 `evaluate_body` harness uses empty substrate Dag carrier (full bootstrap reflection hits unsupported Callable transforms) - Evaluator/compiler bootstrap plumbing + regenerated snapshots (`regen_bootstrap` verify-clean) - Restore R3 gate #37 cost-lens integration assertions + program-plan cite Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
May 10, 2026
- Split cron schedule model (`dsl/extdeps/cron_schedule_model.dag`); wire Actions extdeps + tools - Add `t_ci_workflow_as_data_demo.dag` + integration tests; `_ci_wad_seed_*` list Empty rows for opaque Dag harness tags - PB-1 `evaluate_body` harness uses empty substrate Dag carrier (full bootstrap reflection hits unsupported Callable transforms) - Evaluator/compiler bootstrap plumbing + regenerated snapshots (`regen_bootstrap` verify-clean) - Restore R3 gate #37 cost-lens integration assertions + program-plan cite Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
May 10, 2026
- Split cron schedule model (`dsl/extdeps/cron_schedule_model.dag`); wire Actions extdeps + tools - Add `t_ci_workflow_as_data_demo.dag` + integration tests; `_ci_wad_seed_*` list Empty rows for opaque Dag harness tags - PB-1 `evaluate_body` harness uses empty substrate Dag carrier (full bootstrap reflection hits unsupported Callable transforms) - Evaluator/compiler bootstrap plumbing + regenerated snapshots (`regen_bootstrap` verify-clean) - Restore R3 gate #37 cost-lens integration assertions + program-plan cite Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
May 10, 2026
…arriers) (#2371) * feat(v3): workflow-as-data CI demo, cron carriers, gate #37 receipt - Split cron schedule model (`dsl/extdeps/cron_schedule_model.dag`); wire Actions extdeps + tools - Add `t_ci_workflow_as_data_demo.dag` + integration tests; `_ci_wad_seed_*` list Empty rows for opaque Dag harness tags - PB-1 `evaluate_body` harness uses empty substrate Dag carrier (full bootstrap reflection hits unsupported Callable transforms) - Evaluator/compiler bootstrap plumbing + regenerated snapshots (`regen_bootstrap` verify-clean) - Restore R3 gate #37 cost-lens integration assertions + program-plan cite Co-authored-by: Cursor <cursoragent@cursor.com> * docs(r3): align T-CostLens-Composition lane row with §1.8 gate table §3 claimed #37/#40/#70 “close post-ε-slice” while §1.8 keeps #37 as partial ε-slice receipt and #40/#70 DECLARED with separate follow-on wording. Lane narrative + ETA column now match row #37 (“tracked separately”) and rows Co-authored-by: Cursor <cursoragent@cursor.com> #40/#70. * test(v3): ratchet against include_str pipeline.dag under compiler src Restore automated detection under src/v3/compiler/src while keeping parse-only include_str fixtures in integration tests. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(v3): fail-closed Logical routing + PipelineStageBinding provenance test - Replace unreachable Logical match arm with UnsupportedTransformTarget (CODING.md). - Restore span.file ratchet for PipelineStageBinding rows alongside bootstrap_loads shape checks. Co-authored-by: Cursor <cursoragent@cursor.com> * chore(ci): align slow-test exemption ratchet floor with exemptions file slow-test-exemptions.txt has 45 active rows; default TEST_TIMEOUT_MAX_EXEMPTIONS was 44 and broke v3 per-test ratchet step. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com>
4 tasks done
briansrls
added a commit
that referenced
this pull request
May 10, 2026
…iation per §1.8 Three valid cell-level findings from codex schedule review on sha 1d95e61. All caught the same root issue: §3 cells didn't reconcile against §1.8 ledger + r3-structure.md canonical authority before landing. #2 — T-Numeric-Construction blocker (line 424): Cell said "Float migration + Real/base-carrier convention HELD on proud-raven-495 G2 Phase 2 Substrate S8 ApproximateField<F>" but §1.8 #18 + #24 explicitly say "CONSUMER_LANDED + PASSING for Grounding G2 primitive rows (2026-05-10, PR #2570 squash b96a51a)" — the work landed. Updated cell to: PR #2570 closes the prior HELD; remaining blocker is broader Real<N> emission demonstrations under S9/Shape-A follow-ons per §1.8 #18 close-criterion. #3 — T-Bridge-Retirement count (line 427): Cell said 3 remaining sub-bridges including mark_bootstrap_secret_ nominal_opacity, but §1.8 #32 PASSING + §2.3 explicitly says that bridge is closed. Corrected count: 3/5 sub-bridges retired (gate #32 prior-cycle Secret nominal-opacity + gate #33 this cycle canonical lens + include_str this cycle), 2 remaining (SourceSpan.file participation + patch_lower_helpers residual). #4 — T-Free-Consequences-Demonstration over-attribution (line 430): Cell credited gates #10/#33/#37/#40/#72 to T-Free, but §1.8 assigns those to other lanes: - #10 → T-V-L4-L7-Direct - #33 → T-Bridge-Retirement - #37 + #40 → T-CostLens-Composition - #72 → T-E-P-Producer-Broadening T-Free's canonical demo gate range is #43-#52. Only #43 (auto_parallelism_independent_binds_emit_parallel) MERGED this cycle for T-Free. Updated cell + compile-note to credit each landing only to its canonical-lane row. Compile-note also reconciled per the same §1.8 single-authority pass: T-CostLens-Composition + T-E-P-Producer-Broadening now credited their own gates instead of attributing them to T-Free. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
May 10, 2026
…2583) * docs(r3): §3 lane-status weekly compile (2026-05-11 Monday cadence) PM-derived compile per §9.1 weekly cadence. Updates Status / Current dispatch / Blocker / ETA-to-close columns based on observable PR merge data + worker session activity + silent-ram-834 status report at gunbc#828 c#4414611117. Lanes with substantial movement this cycle: - T-LensProducer-Retirement: gate #5 lens_apply.rs in flight (valiant-otter-715) - T-Numeric-Construction: u128 mirror sync MERGED #2526; gates #17 + #20 active - T-Free-Consequences-Demonstration: 6 gates merged (#10/#33/#37/#40/#43/#72) - T-Bridge-Retirement: 2/5 sub-bridges retired (PR #2459 + #2449) - T-Lens-Behavioral-Parity: #73 + #78 active under Substrate Mgr - T-Debt-Paydown (standing): Mgr re-spawn (gentle-newt-665 → silent-ram-834); Phase 3 fleet 8/10 closed/absorbed; orphan PR #2503 closed - T-Omni-Shape-B: gate #25 salvage path under PB Mgr; #26/#27 mis-parented Lanes with no observable change this cycle: - T-V-L4, T-V-L5-Corpus, T-FixedPoint, T-Anthropic-Wire, T-V2-Retirement, T-Tests-As-Data-Completeness — substrate work continues but no clear gate-level deltas surfaced Mgr canvas refreshes remain formal authority per §3 framing; lane-owning Mgrs may correct/override any PM-derived cell. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): address codex BLOCKING findings on PR #2583 §3 compile 3 valid findings from codex review: 1. T-Lens-Behavioral-Parity status was "RED→YELLOW (PM-derived; Mgr ratification welcome)" — created parallel-representation hedge in a single-authority cell (INVARIANTS P2 violation; per feedback_parallel_representation_debt). Resolved: commit fully to YELLOW as the PM-compiled value (the §3 disclaimer note covers Mgr override authority). The hedge in the cell was worst-of-both-worlds. 2. PM compile note said T-Tests-As-Data-Completeness had "no observable change this cycle" but the table cell records PR #2287 (Verification V1 TC1 first slice) MERGED 2026-05-10. Self-contradicting. Resolved: moved T-Tests-As-Data-Completeness to "lanes with substantial movement" list. Also added T-Anthropic-Wire (PR #2506), T-V2-Retirement (PR #2334), T-V-L7 (gate #10 / PR #2394), T-Tier3-Dissolution (clever-bear-180 active), T-Lens-Application-Surface (crisp-raven-202 active) to the movement list — all had cell-level deltas in the table that the compile note had missed. 3. PR #2394 merge date inconsistency: T-V-L7 cell said "2026-05-09", T-Free-Consequences cell said "2026-05-10". Verified merge timestamp 2026-05-10T00:26:42Z UTC; corrected T-V-L7 cell to 2026-05-10. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): fix T-LensProducer-Retirement blocker (codex BLOCKING #2 on PR #2583) Pre-existing error in §3 cell that prior PM compile preserved instead of correcting. The original cell named "T-FixedPoint + R2-Evaluator" as T-LensProducer-Retirement's blocker, but per the canonical sequence: - r3-structure.md:357: critical path is `R2-Evaluator → T-LensProducer- Retirement → T-FixedPoint → T-V2-Retirement` - r3-program-plan.md:360-363: "T-LensProducer-Retirement comes BEFORE T-FixedPoint, not after; T-FixedPoint depends on SG-0 zero from T-LensProducer" T-LensProducer-Retirement coming AFTER T-FixedPoint creates a circular dependency in the weekly snapshot. Corrected to use the canonical R2-close-dependency from r3-structure.md §"Lane structure": R2-Evaluator (interpreter-as-data; LANDED) + PB-1 generated bin-shim pattern + R2-T-Ground-Lifetime-Analyzer a/b/c basic cases. Also added warm-crab-600's gate #7 work-in-flight signal (regen_lens.rs retirement; the 3rd sub-gate of T-LensProducer-Retirement) per latest subtree status digest. All 3 sub-gates now in flight: #5 valiant-otter- 715, #6 same-cascade, #7 warm-crab-600. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): address codex BLOCKING #2/#3/#4 — single-authority reconciliation per §1.8 Three valid cell-level findings from codex schedule review on sha 1d95e61. All caught the same root issue: §3 cells didn't reconcile against §1.8 ledger + r3-structure.md canonical authority before landing. #2 — T-Numeric-Construction blocker (line 424): Cell said "Float migration + Real/base-carrier convention HELD on proud-raven-495 G2 Phase 2 Substrate S8 ApproximateField<F>" but §1.8 #18 + #24 explicitly say "CONSUMER_LANDED + PASSING for Grounding G2 primitive rows (2026-05-10, PR #2570 squash b96a51a)" — the work landed. Updated cell to: PR #2570 closes the prior HELD; remaining blocker is broader Real<N> emission demonstrations under S9/Shape-A follow-ons per §1.8 #18 close-criterion. #3 — T-Bridge-Retirement count (line 427): Cell said 3 remaining sub-bridges including mark_bootstrap_secret_ nominal_opacity, but §1.8 #32 PASSING + §2.3 explicitly says that bridge is closed. Corrected count: 3/5 sub-bridges retired (gate #32 prior-cycle Secret nominal-opacity + gate #33 this cycle canonical lens + include_str this cycle), 2 remaining (SourceSpan.file participation + patch_lower_helpers residual). #4 — T-Free-Consequences-Demonstration over-attribution (line 430): Cell credited gates #10/#33/#37/#40/#72 to T-Free, but §1.8 assigns those to other lanes: - #10 → T-V-L4-L7-Direct - #33 → T-Bridge-Retirement - #37 + #40 → T-CostLens-Composition - #72 → T-E-P-Producer-Broadening T-Free's canonical demo gate range is #43-#52. Only #43 (auto_parallelism_independent_binds_emit_parallel) MERGED this cycle for T-Free. Updated cell + compile-note to credit each landing only to its canonical-lane row. Compile-note also reconciled per the same §1.8 single-authority pass: T-CostLens-Composition + T-E-P-Producer-Broadening now credited their own gates instead of attributing them to T-Free. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): address codex BLOCKING #5/#6 — PR-merge evidence ≠ gate-PASSING Two valid findings from codex schedule review on sha f6a3a13 (review id 4259176210): #5 — T-Bridge-Retirement count conflated PR-merge with gate-PASSING: Cell said "3/5 retired" but §1.8 truth: #32 PASSING, #33 DECLARED, #34 DECLARED, #35 PASSING. PR #2449 + PR #2459 ARE merged but the gates haven't been promoted from DECLARED → PASSING (separate status drift sweep step, e.g., per PR #2399 cadence). Reframed cell to distinguish PR-merge evidence from canonical §1.8 status: 2/5 gate-PASSING (#32 + #35), 2/5 PR-merged-pending-promotion (#33 + #34), plus SourceSpan.file participation (Substrate-owned hand-Rust audit sites; not in numbered §1.8) + residual semantic patching (`bridge_exact_string_semantic_patching_residual` Open per #35 close-criterion). #6 — T-Free-Consequences over-claim on PR-merge: Cell said "gate #43 MERGED" but §1.8 #43 still DECLARED (PR #2495 is evidence toward promotion, not the promotion event). Same fix: reframe as PR-merge evidence accruing toward §1.8 gate promotion; canonical status authoritative. Compile-note also reframed: explicitly distinguishes PR-merge evidence from §1.8 gate-PASSING promotion. PR-merge events are listed as evidence accruing toward promotion; canonical gate status varies per §1.8. Common root: future Monday compiles must mechanically reconcile each "landed/retired" claim against §1.8 status, NOT PR-merge events. Discipline recorded in feedback_pm_compile_audits_pre_existing_errors (updated to include PR-merge-vs-gate-promotion distinction). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This was referenced May 10, 2026
Merged
briansrls
added a commit
that referenced
this pull request
May 12, 2026
…2744 §1 (#2750) * docs(r3): T-WAD FULL R3 §1.8 ledger sync — +6 gates (#98–#103) per PR #2744 §1 **Authority**: PM scoping doc PR #2744 §1 (T-WAD FULL R3 elevation per operator directive 2026-05-12 + Director ratification msg_5cbdad24 + msg_f9fd669e + (b) ledger-sync disposition msg_2a68a4b5 — follow-up sync PR pattern). **Sync disposition rationale** (per Director msg_2a68a4b5): option (a) bundles too much into PR #2744 mid-review; (c) bakes parallel-authority into scope doc as "temporary" PROPOSED state which calcifies. (b) is operationally clean if sync PR queues for atomic-merge-sequencing alongside PR #2744 — gap window bounded to merge-clock seconds. **6 new §1.8 gate rows** (all T-Workflow-As-Data, NEW 2026-05-12): - **#98** `ci_yml_hand_authority_dissolved` (state-check) — hand-authority NOT file-deletion; (a) absent / (b) emission-artifact / (c) thin-shim per briansrls BLOCKING #PR2744 fix - **#99** `emission_target_open_enum_landed` (substrate-shape) — EmissionTarget sum-type per (c-refined) shape at PR #2749 §7 - **#100** `project_github_actions_landed` (substrate-shape) — projection function declaration in gunbc namespace; consumes extdeps.github.actions.Workflow as codomain + CIWorkflowDag (PR #2736 carrier) as input domain - **#101** `test_cost_dimension_landed` (substrate-shape) — Cost dimension on TEST NODES (distinct from existing compiler-internal cost gates #37/#39/#40/#70/#80 which are about SymbolicCost as the compiler's cost lens; this gate is about Cost-as-Dimension applied to test nodes so slow-test ratchet derives structurally) - **#102** `slow_test_exemptions_dissolved` (state-check) — scripts/slow-test- exemptions.txt deleted; sibling of #101 per kernel-modeling discipline split - **#103** `ci_uses_affected_set_selection` (state-check) — BinaryShim emitter consumes affected-set lens output from PR #2713; Layer 2 path-regex `if:` gates removed; cross-tier co-owned with clever-tern-670 Slice 7 **Count updates** (was 97 enumerated / 96 R3-load-bearing; now 103/102): - §1.5 total enumeration: 97 → 103 - §1.5 R3-load-bearing arithmetic: 96 → 102 (= 103 - 1 canvas-deferred {#11}) - §1.5 composition tally: T-Workflow-As-Data 4 → 10 (+6 NEW) - §1.5 prose: R3 close target 96 → 102 - §1.5 R4-carved-dissolved framing: target 96 → 102 - §1.7 status-taxonomy lead: 96 → 102 R3-load-bearing - §1.8 standing-program note: 96 → 102 load-bearing - §0 R3 close criteria: 97/96 → 103/102 - §1 plan-declared count: 97 → 103 - §1.6 acceptance criteria: 97/96 → 103/102 - §1.8 §1.8 single-canonical-view: 97 → 103 - §1.8 row #11 canvas-deferral arithmetic: 97 → 103 - §Q1 table: 97 → 103 + history pointer **Sequencing discipline** (per Director msg_2a68a4b5): 1. This sync PR sits ready-to-merge until PR #2744 lands 2. As soon as PR #2744 squash-merges, fire this sync PR squash-merge immediately 3. Gap window: bounded to merge-clock seconds (atomic-as-possible without bundle) 4. If reviewer delay accumulates here, fold into PR #2744 retroactively (option (a) escalation path) **Cost-dim distinction note** (per Director verification flag): existing gates #37/#39/#40/#70/#80 are about compiler-internal SymbolicCost (cost lens reading target programs). #101 is structurally distinct — Cost dimension on TEST NODES for slow-test ratchet derivation. Not a duplicate. **§1.9 acceptance-aggregator pilot row** (`t_ci_wad_full_r3_close`): remains in PR #2744 §9 with gate-name references per `feedback_no_snapshot_integers_in_briefs` discipline. No row added to this sync PR per Director msg_2a68a4b5 ("EITHER PR #2744 §9 OR sync PR — your call; either works"). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): cross-doc authority sync — add 6 NEW T-WAD gates to r3-structure.md §Acceptance Per codex BLOCKING review on PR #2750 (2026-05-12T07:34:59Z): scope doc lines 13 and 84 claim r3-structure.md was "updated in this PR" + reference 103-gate total per r3-structure.md §Acceptance, but the PR only changed r3-program-plan.md — r3-structure.md still had no #98–#103 rows. That violated INVARIANTS.md P2 / modeling-discipline.md Practice 5 (single-authority metadata) — the closure ledger became internally inconsistent at the canonical-source level. Fix: add the 6 NEW gate bullets to r3-structure.md §Acceptance T-Workflow-As-Data section (after `ci_workflow_modeled_as_dag`), mirroring r3-program-plan.md §1.8 rows #98–#103. Each bullet carries the full Pass-condition body (single-source authority for Pass conditions per the r3-program-plan.md convention). Now both docs land the same gate set atomically in this PR: - r3-program-plan.md §1.8 rows #98–#103 (commit ef9a140, prior) - r3-structure.md §Acceptance T-WAD bullets (this commit) — matching content The "Documentation Describes Live State" rule + single-authority discipline restored. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): briansrls BLOCKING fix — propagate 103/102 counts through full plan Per briansrls BLOCKING inline review on PR #2750 at line 84 (2026-05-12T07:50:26Z): "The new 103/102 canonical count is not propagated through the full plan, leaving later close criteria at 96/97 and creating competing R3 thresholds (INVARIANTS P2 single authority)." Three stale references found in re-grep + fixed: - Line 114 (§1.7 two-Pass-surfaces context): "**96** R3-load-bearing gates green" → "**102** R3-load-bearing gates green" (post-carve-promotion + T-WAD FULL R3 elevation) - Line 623 (§5.2 R3 close definition): "**96** load-bearing post-carve- promotion" → "**102** load-bearing post-carve-promotion + T-WAD FULL R3 elevation" - Line 1001 (§10 dependency-graph mirror): "lane TestClaim gates (97 total)" → "lane TestClaim gates (103 total; 102 R3-load-bearing post-T-WAD-FULL- R3-elevation 2026-05-12)" Initial sync (commit ef9a140) updated §0 + §1.5 + §1.6 + §1.7 + §1.8 + Q1 table; this commit completes propagation through §1.7 two-Pass-surfaces / §5.2 R3 close definition / §10 dependency-graph mirror. Single-authority discipline (INVARIANTS P2) now consistently asserts 103/102 across the full plan; no competing R3 thresholds remain. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): InlineGunbc DESIGN-ONLY alignment — cross-doc consistency with PR #2744 openai-pro BLOCKING fix Per openai-pro BLOCKING fix on PR #2744 (commit e43aba3): WI-1 / WI-2 briefs + scope doc §2 type sketch all align on InlineGunbc as DESIGN-ONLY (NOT in initial enum, lands when runtime consumer exists). The §1.8 gate bodies in r3-program-plan.md row #99 + r3-structure.md §Acceptance bullet for emission_target_open_enum_landed were stale relative to that alignment. Fix in this sync PR: - r3-program-plan.md §1.8 row #99: "(YamlStatic | BinaryShim | PythonShim | InlineGunbc | ...)" → "3 initial arms (...)" + InlineGunbc DESIGN-ONLY note with PR #2746 §5.4 + openai-pro BLOCKING cross-references - r3-structure.md §Acceptance T-WAD bullet for emission_target_open_enum_landed: same change pattern Single-authority across: - PR #2744 scope doc §0 / §1 gate row / §2 type sketch - PR #2744 WI-1 brief DESIGN-ONLY discipline - PR #2744 WI-2 brief Output / DO-DON'T / Acceptance gates - PR #2750 (this PR) §1.8 ledger row + §Acceptance archive - PR #2746 §5.4 canonical DESIGN-ONLY framing Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): T-WAD ledger sync — lane-definition gate lists pointer-only + scope-doc cross-ref qualifier (codex BLOCKING + non-blocking on PR #2750) Two findings from codex review #10042 on commit 0c08f77 (T-WAD ledger sync PR head): BLOCKING — "T-WAD ledger sync updated the Acceptance archive but skipped the lane-definition gate lists → update those T-Workflow-As-Data gate lists to include #98–#103 or make them pointer-only." The PR #2750 cascade landed the 6 NEW T-WAD FULL gates (#98–#103) in: - r3-structure.md §Acceptance T-Workflow-As-Data bullets (lines 172-177) - r3-program-plan.md §1.8 ledger rows - r3-program-plan.md count propagation (§0/§1.5/§1.6/§1.7/§5.2/§10/§Q1) But the cascade missed two lane-DEFINITION gate lists in r3-structure.md that ALSO enumerate T-WAD closure gates: - Line 41 (numbered lane list, T-WAD entry) - Line 222 (T-WAD row in §"Lane structure" table) Both listed only the original 4 pre-FULL gates (workflow_substrate_carriers_landed / timing_lens_carrier_landed / ci_workflow_modeled_as_dag / shared_external_attachment_pattern_documented). Reviewers reading either list would not see the 6 NEW gates — INVARIANTS P2 single-authority gap. Fix: convert both lane-definition lists to pointer-only references back to §Acceptance T-WAD as the canonical gate list. Rationale: §Acceptance is the authority anchor (per INVARIANTS P2 + sister r3-program-plan.md §1.8); duplication in lane-definition lists would re-introduce drift the cascade is closing. Also augmented both lane-definition entries with FULL R3-close elevation 2026-05-12 framing + multi-Mgr ownership (Substrate Mgr Slices 4-5/8 + Verification Mgr Slice 7 affected-set + Debt-Paydown Mgr Slice 6 sub-component). Non-blocking — "Line 172 cites docs/r3-t-workflow-as-data-full-r3-close-scope.md, but git ls-tree origin/main returned no blob → replace with existing receipt or land the scope doc." The scope doc exists on PR #2744's branch (in flight) but not on origin/main yet. Codex correctly notes the dangling cross-reference against current main. Fix: add explicit "(scope doc landing via in-flight PR #2744)" qualifier so future readers know the citation is forward-referencing a known in-flight PR rather than a typo or missing doc. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): T-WAD ledger sync — remove dangling file refs (briansrls BLOCKING + codex INVARIANTS P2/P5 escalation on PR #2750) briansrls inline BLOCKING at r3-structure.md:172 (2026-05-12T09:28:04Z) escalated the prior non-blocking scope-doc citation issue to BLOCKING: "The new gate cites docs/r3-t-workflow-as-data-full-r3-close-scope.md section 1, but git ls-tree origin/main produced no blob and the reconstructed PR-head test returned 1, so the cited T-WAD scope authority is absent (INVARIANTS P2/P5)." The earlier qualifier fix ("scope doc landing via in-flight PR #2744") acknowledged the dangling reference but didn't resolve the structural P2/P5 violation — the gate description still CITED an authority that doesn't exist on origin/main, which is the merge target. Audit: grepped both docs for refs to files that don't exist on origin/main: - docs/r3-structure.md:172 — `docs/r3-t-workflow-as-data-full-r3-close-scope.md` (PR #2744) - docs/r3-program-plan.md:326 — `docs/briefs/r3-t-wad-full-r3-cidag-scaffold-worker.md` (PR #2744) Fix: replace both file-path references with PR-number anchors. PR numbers are stable references; file paths become valid only post-merge. Gate descriptions are self-contained without the cross-references (the (a)/(b)/(c) enumeration + supporting framing already conveys the gate's substance). - r3-structure.md:172: "per `docs/r3-t-workflow-as-data-full-r3-close-scope.md` §1 — scope doc landing via in-flight PR #2744" → "in-flight scope authority at PR #2744 §1" - r3-program-plan.md:326: "WI-2 implementation: `docs/briefs/r3-t-wad-full-r3-cidag-scaffold-worker.md`" → "WI-2 implementation: in-flight via PR #2744 (brief lands with the scope-doc)" Both gates retain full substantive content; only the file-path crutches are removed. When PR #2744 merges and the files exist on main, future authors may re-add file refs cleanly — but the gate descriptions never needed them as load-bearing authority. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(r3): T-WAD ledger sync — EmissionTarget → WorkflowRuntime rename cascade (warm-wolf-698 PR #2749 commit 575eb7e; SELF_HOSTING.md:609 name collision) Cascade sister-PR rename: PR #2744 branch carries the scope-doc + brief renames in commit 70a49ac; this commit carries the ledger / structure gate-ID renames. warm-wolf-698 surfaced a DECISIVE name-collision finding at PR #2749:666 / :672 (briansrls operator BLOCKING 2026-05-12T09:26:49Z): the canvas EmissionTarget sum-type collides with the canonical Shape-A carrier declared at `src/v3/SELF_HOSTING.md:609`: type EmissionTarget { language: LanguageSpec // what's valid (required) rendering: RenderingSpec? // how to format (optional) } This is the SELF_HOSTING.md emitter-composition authority — INVARIANTS P2 violation. warm-wolf-698 pushed rename to WorkflowRuntime in PR #2749 commit 575eb7e (48 occurrences). All OTHER ratified elements stand per feedback_pre_compaction_framings_self_supersede. This commit cascades the rename through PR #2750 branch: - docs/r3-program-plan.md §1.8 row #99: gate ID emission_target_open_enum_landed → workflow_runtime_open_enum_landed (also EmissionTarget references in row description) - docs/r3-program-plan.md §1.8 row #100: EmissionTarget references in project_github_actions signature - docs/r3-structure.md line 41 (T-WAD lane summary): EmissionTarget references + gate-ID rename in the multi-gate reference - docs/r3-structure.md §Acceptance T-Workflow-As-Data bullets: gate-ID emission_target_open_enum_landed → workflow_runtime_open_enum_landed EmissionTarget references in `project_github_actions_landed` description - docs/r3-structure.md §Lane structure T-WAD row: EmissionTarget references in scope expansion text Variant names unchanged (YamlStatic / BinaryShim / PythonShim). Gate descriptions retain full substantive content; only the type-name and gate-ID identifiers are renamed. Cascade trail across in-flight PRs: - PR #2749 (warm-wolf-698): 575eb7e — substrate canvas rename - PR #2751 (warm-wolf-698): expression-substrate canvas rename (in flight) - PR #2744 (mine): 70a49ac — scope-doc + WI-1 + WI-2 brief rename - PR #2750 (mine): THIS COMMIT — ledger + structure rename - PR #2745 (cool-carp-720): surfaced; WI-2 implementation needs realign - PR #2746 (MERGED): docs/design-ci-workflow-emitter-dispatch.md needs follow-on rename PR (post-cascade-clear) 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 14, 2026
briansrls
added a commit
that referenced
this pull request
May 14, 2026
6 tasks
briansrls
added a commit
that referenced
this pull request
May 14, 2026
…tion) (#3097) * WIP: R3 gate #20: numeric_inherited_bake_ins_dissolved (T-Numeric-Constructio * WIP: R3 gate #20: numeric_inherited_bake_ins_dissolved (T-Numeric-Constructio * docs(r3): promote gate #20 numeric_inherited_bake_ins_dissolved to PASSING Record CONSUMER_LANDED + PASSING in §1.8 with producer/consumer citations (types.dag Int refinements + m2_substrate_inhabitance_test ratchet). Refresh §3 T-Numeric-Construction snapshot, PM compile note, and close-predicate index (HARNESS_NAMED, bucket counts, ledger anchor text). Co-authored-by: Cursor <cursoragent@cursor.com> * docs(r3): align PM compile note header with 2026-05-14 §3 amendments Optional editorial consistency from composer-2 review (cadence stamp vs audit follow-up date). Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: R3 gate #20: numeric_inherited_bake_ins_dissolved (T-Numeric-Constructio * WIP: R3 gate #20: numeric_inherited_bake_ins_dissolved (T-Numeric-Constructio * docs(r3): align §3 + predicate audit with §1.8 for gates #23/#40/#59/#63/#81/#91 §1.8 is canonical (INVARIANTS P2). Re-derive §3 T-Numeric residuals (drop stale #23 DECLARED; ETA lists #23 as landed), fix T-CostLens #40 vs #70 wording, and refresh the predicate-execution index: HARNESS_NAMED rows + verdict/buckets now match mechanical §1.8 Status classification at HEAD (61 harness-eligible). Co-authored-by: Cursor <cursoragent@cursor.com> * docs(r3): sync PM compile T-CostLens clause with §1.8 + §3 lane row §1.8 rows #37/#40/#70 are not both DECLARED; #40 is CONSUMER_LANDED + PASSING and #70 INTEGRATION_RECEIPT (P2 single-authority vs §3 T-CostLens snapshot). Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
May 14, 2026
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
2 participants
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
This PR adds explicit test coverage for coercion behavior in the executor and audits that all three
CoercionKindvariants are correctly implemented at runtime. Four new tests directly verify value shapes for each coercion classification, and generated coercion coverage tests are improved to assert downstream node execution rather than just "didn't crash."Key Changes
Added 4 coercion-kind-mapped tests in
execute.rs:test_coercion_wrap_scalar_single_edge: Verifies scalar[1,1]→ list[0,∞)wraps in a one-elementValue::Listtest_coercion_optional_to_list_absent: Verifies optional[0,1]producingUnit→ empty listtest_coercion_optional_to_list_present: Verifies optional[0,1]with value → one-element listtest_coercion_widen_bounded_to_unbounded: Verifies bounded[2,5]list → unbounded[0,∞)list passes elements through unchangedImproved generated coercion coverage tests in
testgen/codegen.rs:_logbinding tologto enable the assertionUpdated TODO_hacks documentation:
CoercionKindvariants are correctly implementedImplementation Details
The new tests use the existing
TestOpinfrastructure to create minimal DAGs that trigger each coercion kind and directly inspect the executor's output log to verify correct value shapes. This provides definitive verification that the fan-in logic correctly handles:The generated test improvement is forward-looking since current DAG builders use matching cardinalities, but ensures future coercion edges are properly validated.
https://claude.ai/code/session_019RSYYEQsHBWb1TPrbUzSyf