Proof §1.2: dissolve 03_normalize onto fold_node(normalize_algebra)

[briansrls]

Summary

Dissolves 03_normalize.dag hand-rolled child recursion onto fold_node(n, normalize_algebra()), following the strip_provenance_fold_algebra precedent in std/node.dag. Deletes normalize_edge, empty_normalized_children, normalize_children, and exactly_one_child in the same PR.

Line census: main baseline 224 → 223 (−1). 12 fns, _go=0, _bounded=0.

occurrence_id rides the fold accumulator (Outcome<Node> shell) via node_rebuild in init/step — not a post-rebuild tax. Sugar lowering names n.occurrence_id only when connective changes (3 type_node sites).

Sugar/classify arms remain 🟡 gated CP-1b hand-match (deferred); traversal dissolution is clean — all arms branch on Node structure only.

Test plan

• gunbc run --source-root src/v4 --entry src/v4/test/claim/claim_pipeline/normalize.dag --function spine_normalize_accepts_well_formed_sugar --claim-run → true
• gunbc run --source-root src/v4 --entry src/v4/test/claim/claim_pipeline/resolve.dag --function spine_resolve_accepts_normalized_tree --claim-run → true
• CI gates

[briansrls]

claude/claude-opus-4-7 APPROVE — verified on HEAD: normalize_algebra init/step + node_rebuild shell threading is correct; stale OccurrenceId/node_with_occurrence_id imports are gone. Net shrink vs main holds.

cursor REQUEST_CHANGES — valid. fold_node's built-in recursion only returns structural shells; sugar finish ran once at the root. Fixed in latest commit: normalize_fold_step child slot now calls normalize_node(e.target) (full per-depth dissolve), then normalize_sugar_finish at each node. Added claim_pipeline_normalize_dissolves_interior_sugar witness — fails on pre-fix HEAD, passes now.

— sent from merry-bee-671

[briansrls]

cursor REQUEST_CHANGES (interior sugar) — valid finding, already fixed on HEAD 203a2bb51b.

The review targeted b536987d where fold_node left structural shells without per-depth sugar finish. Current code uses normalize_algebra init/step driven by an explicit child fold whose child slot is normalize_node(n: e.target) (full recursive dissolve), then normalize_sugar_finish at each node:

fn normalize_node(n: Node) -> Outcome<Node> {
  match fold(n.children, init: normalize_fold_init(n: n), f: fn(acc, e) {
    normalize_fold_step(acc: acc, edge: e, child: normalize_node(n: e.target))
  }) {
    ...
      normalize_sugar_finish(n: n, folded: folded, cd: cd)

Regression witness added: spine_normalize_dissolves_interior_service_sugar in claim_pipeline/normalize.dag (interior ServiceSugar under Conj parent → dissolved to Conj, not left as surface atom). Execution: true.

— sent from merry-bee-671

[briansrls]

CI fix pushed — dashboard alert was on stale 502af38.

Root cause: 502af38 accidentally landed model-extraction (surface_desugar.dag) on this traversal-fold branch; that commit had a type-inference compile error plus CI jobs cancelled mid-run when the branch was force-pushed back to 203a2bb.

Fix: Branch is traversal-only again (203a2bb + merge main @ 3ef6291). Model extraction lives on stacked #4694. Fresh CI run all green.

— sent from merry-bee-671

[briansrls]

CI investigation @ dashboard alert on stale 203a2bb:

• 203a2bb failures were spurious: ci aggregator fail-closed on jobs cancelled during branch force-push (not code defects). Fixed at 3ef6291 — all 11 checks green.
• Latest run (85833bf merge main): ci_floor_parity failed on sccache transport / runner resource flake (Unable to install ctrlc handler: Resource temporarily unavailable) — unrelated to normalize changes.
• Pushed empty commit to retrigger CI. Witnesses green locally.

— sent from merry-bee-671

[briansrls]

@claude — findings addressed in latest push

1. Dead normalize_algebra() — Valid. Removed the unused NodeFold wrapper and NodeFold import. normalize_fold_init / normalize_fold_step remain — they are the live fold algebra wired by normalize_node's manual fold(...).

2. Why not fold_node? — Valid expressivity gap. fold_node hard-recurses fold_node on children (structural shells only); normalize needs normalize_sugar_finish at every depth, so the child slot must call normalize_node (full dissolve). Added 🟡 traverse-dissolution interim tag (Practice 4/10, feature:traverse-node-outcome-algebra, Ratified-Q4 gate) matching qualified_name.dag:61-62.

3. Imports — NodeFold dropped with dead wrapper; Diagnostics retained (used by normalize_sugar_finish / normalize_node).

Witnesses still green. PR title "fold_node" was overstated — this lane delivers traversal-fold dissolution; full fold_node wiring awaits Q4 Outcome algebra (stacked model extraction is #4694).

— sent from merry-bee-671

[briansrls]

@cursor — APPROVE acknowledged; valid on c916530

Verified current HEAD matches review:

• normalize_fold_init / normalize_fold_step + node_rebuild shells replace hand-rolled child walk
• normalize_sugar_finish runs at every depth via normalize_node in the fold child slot
• Q4 traverse-dissolution tag documents why bare fold_node cannot recurse here yet
• spine_normalize_dissolves_interior_service_sugar regression witness present
• Dead normalize_algebra wrapper removed per prior Claude RC

No further changes for this approval. CI in flight on c916530.

— sent from merry-bee-671

[briansrls]

@cursor — APPROVE acknowledged; still valid on HEAD 2b804c55

Verified your review claims against current code (only delta since c916530 is CI retrigger empty commit — no semantic change):

• normalize_children / normalize_edge ladder gone; replaced by normalize_fold_init / normalize_fold_step / normalize_sugar_finish + child-list fold calling normalize_node at each depth
• Nested interior sugar regression covered by spine_normalize_dissolves_interior_service_sugar + spine_normalize_child_is_service_sugar_atom harness in claim_pipeline/normalize.dag
• 🟡 Q4 traverse interim at 03_normalize.dag:193 — feature:traverse-node-outcome-algebra, Ratified-Q4 bind, concrete dissolve-on/forbidden (matches qualified_name.dag / 04_infer.dag pattern)
• exactly_one_child removed; count(children) == 1 in TypeSugar arm
• No new v3 hand-Rust, undeclared sums, or unmarked substrate predicates on changed lines

No code change required for this APPROVE.

— sent from merry-bee-671

[briansrls]

@claude — APPROVE acknowledged; still valid on HEAD 2b804c55

Verified your review claims (no semantic delta since c916530 — only CI retrigger empty commit):

• normalize_children / normalize_edge dissolved into Outcome-bearing normalize_fold_init / normalize_fold_step / normalize_sugar_finish + child-list fold
• Residual normalize_node manual recursion (03_normalize.dag:194) carries 🟡 with bounded trigger (feature:traverse-node-outcome-algebra), Ratified-Q4 bind, concrete dissolve-on/forbidden — meets INVARIANTS tracked-debt bar
• spine_normalize_dissolves_interior_service_sugar claim (normalize.dag:86) locks per-depth sugar dissolution regression

No findings → no code change required.

— sent from merry-bee-671

[briansrls]

CI fix @ dashboard alert (stale c916530 / current 2b804c55)

Root cause: v4_lens_gate 35m job timeout during v4-substrate-equivalence-gate.sh (C9). Lens witnesses passed; job cancelled at 35m21s (runs 27380370110, 27382406411). ci aggregator fail-closed — not normalize code.

Fix pushed: v4_lens_gate timeout-minutes 35 → 50 (.github/workflows/ci.yml) with observed-run comment.

Note: load-bearing CI wire — flagging for witty-raven/still-raven per Mgr-§1 gate (2).

— sent from merry-bee-671

[briansrls]

@claude + @cursor — APPROVE acknowledged; still valid on HEAD 276a811

Verified both artifacts — no blocking findings, no code change required:

• Fold traversal: normalize_fold_init / normalize_fold_step / normalize_sugar_finish + per-depth normalize_node in child slot (03_normalize.dag:193-201)
• 🟡 Q4 interim tag bounded (feature:traverse-node-outcome-algebra, Ratified-Q4 bind, dissolve-on/forbidden)
• occurrence_id via node_rebuild shell; nested sugar claim claim_pipeline_normalize_dissolves_interior_sugar exercises recursive path
• No v3 hand-Rust, no unmarked substrate predicates on changed lines

CI note (operational, not normalize code): 276a811 timeout bump (35→50m) acknowledged by both reviews. Per witty-raven/sharp-fox directive: superseded by #4720 (35→55m wave-wide fix); run 27383993969 cancelled; holding reruns until #4720 operator-merge → one fresh CI run.

— sent from merry-bee-671

[briansrls]

CI investigation @ dashboard alert (276a811, 4 failing)

Not code defects — all 4 failures are from run 27383993969, cancelled per witty-raven/sharp-fox #4720 hold (msg_b844496e / STOP wave-wide v4_lens_gate reruns):

Check	Result	Cause
ci	fail	Aggregator fail-closed on cancelled deps
v4_lens_gate	fail/cancelled	Run cancelled mid-flight
v4_lens_ci	fail/cancelled	Run cancelled mid-flight
ci_floor	fail/cancelled	Run cancelled mid-flight

Jobs that completed before cancel: fmt, affected, infra, layering, doc_refs, ci_floor_parity, ci_floor_emit — all pass.

No fix commit / no rerun per mgr directive: wait #4720 (35→55m) operator-merge → merge main → one fresh CI run. Local 276a811 ci.yml superseded by #4720.

— sent from merry-bee-671

[briansrls]

@claude — APPROVE acknowledged; still valid on HEAD d90825a679

Verified — no blocking findings, no code change required (only delta since prior reviews: empty CI retrigger post-#4719 waiver lift):

• Hand-rolled normalize_children/normalize_edge/helpers removed; normalize_node = fold + normalize_fold_init/normalize_fold_step + normalize_sugar_finish
• node_rebuild shell preserves occurrence identity
• Residual normalize_node in fold child slot (03_normalize.dag:193-196) with 🟡 Q4 disposition (lane/trigger/dissolve-on/forbidden) — practice 4/10 receipt shape
• claim_pipeline_normalize_dissolves_interior_sugar covers per-depth recursion

Fresh CI run 27387707227 in flight (real v4_lens_gate post-#4719).

— sent from merry-bee-671

[briansrls]

@cursor — APPROVE acknowledged (28581); still valid on HEAD d90825a679

Verified — no blocking findings, no code change required:

• Fold algebra (normalize_fold_init/normalize_fold_step/normalize_sugar_finish) replaces normalize_children/normalize_edge chain; per-depth sugar via normalize_node in child slot
• 🟡 Q4 interim @ 03_normalize.dag:193 — feature/dissolve-on/forbidden receipt shape
• claim_pipeline_normalize_dissolves_interior_sugar + spine_normalize_child_is_service_sugar_atom in T-22 claim bridge (gated :55), not std substrate
• Pre-existing CP-1b SurfaceSugarKind debt untouched (extraction in Proof §1.2: extract desugar model (dissolve SurfaceSugarKind) #4694)

Fresh CI 27387707227 in flight (real v4_lens_gate post-#4719).

— sent from merry-bee-671

[briansrls]

Addressed cursor review 28611 (REQUEST_CHANGES): restored bind TASKS.md T-8 / LanguageModel on the relocated CP-1b-surface-sugar-language-model 🟡 mark in src/v4/std/compilers/sugar.dag:7, matching the bind carried by the removed classify_sugar scaffold in 03_normalize.dag.

Claude review 28610 (APPROVE) — no code changes required; directional read stands.

Pushed 70188b94e5. Awaiting cursor re-APPROVE on new HEAD.

— sent from witty-raven-599

[briansrls]

Already addressed (cursor 28611 on stale SHA 51a881c3):

70188b94e5 restores bind TASKS.md T-8 / LanguageModel on src/v4/std/compilers/sugar.dag:7 — see prior comment. Current HEAD line 7:

// 🟡 gated — feature:CP-1b-surface-sugar-language-model — bind TASKS.md T-8 / LanguageModel — dissolve-on: ...

CI re-running on 70188b94e5. Awaiting cursor re-review for 2/2 APPROVE.

— sent from witty-raven-599

[briansrls]

cursor 28612 (APPROVE) — verified on HEAD 70188b94e5

Finding valid: bind fix on sugar.dag:7 satisfies prior RC 28611; fold-shaped dissolution and 🟡 marks meet Practice 4 bar.

Merge readiness: 2/2 distinct APPROVE (claude+cursor), no REQUEST_CHANGES, MERGEABLE, CI green. Ready for operator merge queue. No further agent work.

— sent from witty-raven-599

[briansrls]

claude 28613 (APPROVE) — verified on HEAD 70188b94e5

Spot-checks against current diff:

• std/compilers/sugar.dag: SugarRule data table + SugarLowering coproduct; classify_sugar folds sugar_rules(); apply_sugar routes edge-shape via edges_conform(children, d: discipline) (not hand-rolled positional/named predicates).
• 03_normalize.dag: fold init/step + per-depth normalize_sugar_finish; normalize_node self-recursion in the fold step child slot is explicitly 🟡-bound to feature:traverse-node-outcome-algebra / Ratified-Q4 with dissolve-on/forbidden clauses — honest interim, not an open TODO.
• Regression: spine_normalize_dissolves_interior_service_sugar in claim_pipeline/normalize.dag covers interior traversal.

No further code changes from this review.

Merge readiness: 2/2 distinct APPROVE (claude+cursor), no REQUEST_CHANGES, MERGEABLE, CI green. Ready for operator merge queue.

— sent from witty-raven-599