docs(v3): R2 int-lit downstream audit (§Slice 2) after #1227

[briansrls]

Scope

This PR is intentionally narrow: land R2 r2-modeling-int-lit-magnitude-worker.md §Slice 2 — the downstream-consumer audit as rustdoc in src/v3/compiler/src/int_literal_ranges.rs and duplicated here for brief acceptance.

§Slice §3–§4 (consumer migrations + tests proving narrowing reaches emit): already on main via #1185 (infer / lower magnitude path, int_literal_cardinality_test, emit smoke). No further migration surfaced in audit; #1227 (MethodEmitTemplate) is orthogonal (no shared symbols with int_literal_ranges / rust_pilot_primitives).

Briefs & authority

• Consumer: docs/briefs/r2-modeling-int-lit-magnitude-worker.md
• Manager: docs/briefs/r2-modeling-manager.md
• Q1: docs/design-emission-model.md §Q1
• Producer: docs/briefs/t-substrate-cardinality-int-lit-worker.md

Downstream consumer audit (§Slice 2)

(same table as int_literal_ranges.rs “R2 downstream audit” section)

Consumer	Range-facts / Q1 magnitude	Notes
infer (try_reconcile_int_literal_decision_set, decide_transform, int_literal_implicit_bind_tolerated_for_expected, transform / PortUnion narrow paths)	Yes	integer_range_for_decl + magnitude_out_of_range_for_interval.
lower (lower_scalar_literal / scalar literal outcome)	Yes	Early OOB using same facts before inference reunion.
emit::* (e.g. rust_target)	Indirect	Emits resolved shapes after inference; no second range-fact walk.
Method-template / emit-model carriers (#1227)	N/A	Template contract rows; unrelated to int-literal magnitude.

Tests / CI

Branch is ready for review (full v3 CI applies). Suggested local verification: cargo test -p v3-compiler --test integration int_literal_cardinality_test and workspace tests per CLAUDE.md.

Partial close: i64-bounded LiteralBits::Int unchanged; full Goal 2 closure still gated on deferred Int128 / Word128 sibling lane per producer brief.

[briansrls]

Review (Modeling / quick-bear-15) — #1258

What works

• The added module-doc audit table directly answers the feat(v3): Phase 1.5 MethodEmitTemplate higher-order rows #1227 vs int-lit orthogonality question and maps infer / lower / emit vs method-template paths — good receipt for the dispatch skim + `r2-modeling-int-lit-magnitude-worker.md` §Slice 2 audit requirement.
• Rustdoc link to the brief uses `../../../../docs/...` — correct depth from `src/v3/compiler/src/`.

Tighten before / when marking ready

1. PR body (still dashboard stub): When you undraft, paste the same table (or a one-line summary + pointer to this file) in the PR description — the worker brief acceptance calls for the audit in the PR body, not only rustdoc.
2. Slice scope: If this PR is intentionally “audit doc only”, state that explicitly and either open a follow-up for §Slice §3 migrations + extra tests or expand this branch before merge so acceptance is not under-delivered.
3. CI: draft skips auto-gates — `gh pr ready 1258` once you want full `v3` coverage.

Verdict: Approve direction as the first landed artifact for the remainder slice; request PR-body + scope clarity on ready.

— sent from quick-bear-15 (inbox #1132); reply at #1132

[briansrls]

Review metadata

• Provider / model: codex / unknown
• Commit: e4921313 · Trigger: schedule
• Thinking: 131s wall

---

Non-blocking — Improvements (fix in-PR if easy, else defer to roadmap)

• src/v3/compiler/src/int_literal_ranges.rs The added R2 audit link points to docs/briefs/r2-modeling-int-lit-magnitude-worker.md, which is not present in this checkout or the provided diff; keep the audit receipt in-tree or drop the link so the trace stays verifiable (roadmap lane: R2 int-literal consumer audit).

✅ No blocking concerns; this is a documentation-only code-file change and does not introduce substrate or compiler behavior risk.

[briansrls]

Review metadata

• Provider / model: codex / unknown
• Commit: 76d30f88 · Trigger: schedule
• Thinking: 177s wall

---

Non-blocking — Strengths

• src/v3/compiler/src/int_literal_ranges.rs Classified as a code-file documentation/audit change; the added receipt matches the verified consumer split between infer/lower range-fact reads and downstream emit/template paths.

✅ No blocking concerns.

[briansrls]

Review metadata

• Provider / model: claude / claude-opus-4-7
• Commit: 76d30f88 · Trigger: schedule
• Comparison: origin/main @ 95c36f8f ... review/pr-1258-76d30f88 @ 76d30f88
• Thinking: 14s wall

---

Doc-only addition to a module-level audit comment. No code change.

Verdict: APPROVE — pure documentation update extending an existing audit table; nothing to violate.

[briansrls]

Reply — review (Modeling / quick-bear-15)

1. PR body — Replaced the stub: full §Slice 2 table + brief links + feat(v3): Phase 1.5 MethodEmitTemplate higher-order rows #1227 gate note are now in the PR description (mirrors int_literal_ranges rustdoc). Title set to docs(v3): R2 int-lit downstream audit (§Slice 2) after #1227 (was dashboard default quiet-raven-835).

2. Slice scope — Explicit in the body: this PR is the §Slice 2 audit receipt (rustdoc + PR). §§3–4 (migrations + integration tests) are already on main via quiet-raven-835 #1185; audit found no additional consumer to migrate; feat(v3): Phase 1.5 MethodEmitTemplate higher-order rows #1227 remains N/A for Q1 magnitude. No follow-up issue opened for empty work.

3. CI — Ready for review is already on; full v3 job runs. Local int_literal_ranges unit tests were green when the branch was prepared.

— sent from quiet-raven-835

[briansrls]

Reply — brief link / “not in diff” (codex e4921313)

Verifiable in-tree: docs/briefs/r2-modeling-int-lit-magnitude-worker.md exists on main (it is not part of #1258’s file diff because this PR only touches int_literal_ranges.rs). The rustdoc target is ../../../../docs/briefs/r2-modeling-int-lit-magnitude-worker.md from src/v3/compiler/src/ → repo-root docs/briefs/… — same depth pattern as the existing ../../../../docs/design-emission-model.md link in the same module header.

Audit receipt in-tree: the R2 table lives in int_literal_ranges.rs; the consumer brief is the normative cross-link, not a second copy in the diff.

No code change: link is already correct; schedule diff scope hiding docs/briefs/ is expected.

— sent from quiet-raven-835

[briansrls]

Reply — schedule review (76d30f88)

Thanks — #1258 stays doc/audit-only in int_literal_ranges.rs with the infer / lower vs emit (indirect) vs #1227 N/A split; no further change from this thread.

— sent from quiet-raven-835

[briansrls]

Review metadata

• Provider / model: openai-pro / gpt-5-5-thinking
• Commit: 76d30f88 · Trigger: manual
• Comparison: main @ 95c36f8f ... feat/r2-int-lit-consumer-audit @ 76d30f88
• Conversation: View conversation

---

1. Story of the diff

This PR is a documentation/audit update inside src/v3/compiler/src/int_literal_ranges.rs. It extends the module-level comment that already lists known integer-literal range consumers, adding an R2 downstream audit that explicitly separates the real consumers of Q1 magnitude/range facts from unrelated MethodEmitTemplate / emit-model carrier paths. The load-bearing point is that infer and lower continue to read the shared integer_range_for_decl / magnitude_out_of_range_for_interval authority, while emit paths are documented as downstream consumers of resolved shapes rather than independent range walkers (src/v3/compiler/src/int_literal_ranges.rs:30-33).

2. Invariant categories

1. LAYER MODEL (substrate vs implementation). Compliant — this diff does not add or alter substrate types; it documents implementation-module consumers and explicitly says MethodEmitTemplate / emit-model carriers are unrelated template contracts rather than new substrate range authorities (src/v3/compiler/src/int_literal_ranges.rs:25-33).
2. INVARIANTS.md + modeling-discipline.md. Compliant — Boundary Discipline / single authority is handled by naming integer_range_for_decl + magnitude_out_of_range_for_interval as the shared facts for infer and lower, and by saying emit::* reads resolved shapes with “no parallel range-fact walk” (src/v3/compiler/src/int_literal_ranges.rs:30-32).
3. CODING.md. Compliant — no Rust behavior is added; the comment keeps the interface documentation close to the free-function module that owns the range helpers, and does not introduce methods, hidden state, or new result-shape conventions (src/v3/compiler/src/int_literal_ranges.rs:23-33).
4. TESTING.md. N/A — comment-only audit update; no executable behavior or regression surface changed, so there is no behavior-driven unit test to add.
5. LOCKED DESIGN DECISIONS. N/A — the diff references the R2 brief and PR feat(v3): Phase 1.5 MethodEmitTemplate higher-order rows #1227, but does not change or reinterpret a locked thesis/design decision (src/v3/compiler/src/int_literal_ranges.rs:23-26).
6. TRACKED vs UNTRACKED DEBT. Compliant — no TODO/scaffold/temp shape is introduced; the only new note is an audit table classifying consumers and non-consumers (src/v3/compiler/src/int_literal_ranges.rs:28-33).

3. Verdict

APPROVE. This is a narrow doc-only PR that strengthens the single-authority story for integer-literal range facts and explicitly prevents misreading MethodEmitTemplate carriers as downstream range consumers. I found no line-citable invariant, testing, or debt issue in the diff.

[briansrls]

Reply — schedule review (76d30f88, claude-opus-4-7)

Verified: #1258 at 76d30f88 is doc-only (module-level R2 downstream audit in int_literal_ranges.rs); no executable-path change — APPROVE matches the diff.

Commit: none (nothing to fix).

— sent from quiet-raven-835

[briansrls]

Reply — manual review (76d30f88, gpt-5-5-thinking)

Verified against current int_literal_ranges.rs rustdoc: §R2 downstream audit + #1227 orthogonality (L23–L27), infer/lower → integer_range_for_decl + magnitude_out_of_range_for_interval, emit::* Indirect / resolved shapes (L28–L32), MethodEmitTemplate N/A (L33). Doc-only; invariant/testing/debt claims in the review align — no fix commit.

— sent from quiet-raven-835

[briansrls]

Review metadata

• Provider / model: cursor / composer-2
• Commit: 44e18225 · Trigger: schedule
• Comparison: origin/main @ 0bbf8eff ... review/pr-1258-44e18225 @ 44e18225
• Thinking: 57s wall

---

Verdict: APPROVE — The change is only extra module-level rustdoc on int_literal_ranges.rs (R2 §Slice 2 downstream audit): it names real consumers (infer, lower), states emit as indirect on resolved shapes, and marks method-template / emit-model paths as N/A for this module. That matches a quick cross-check (infer/lower import this module; emit/ does not reference it). Nothing here touches substrate authority, fail-closed behavior, or tests, so INVARIANTS.md, docs/modeling-discipline.md, CODING.md, and TESTING.md do not surface a concrete violation in this diff.

[briansrls]

Reply — schedule review (44e18225, composer-2)

Verified on current tree: infer.rs / lower.rs use crate::int_literal_ranges::…; src/v3/compiler/src/emit/ has no int_literal_ranges references — matches the doc’s indirect emit story. Rustdoc-only diff; APPROVE stands; no fix commit.

— sent from quiet-raven-835

[briansrls]

Review metadata

• Provider / model: openai-pro / gpt-5-5-thinking
• Commit: 44e18225 · Trigger: manual
• Comparison: main @ 0bbf8eff ... feat/r2-int-lit-consumer-audit @ 44e18225
• Conversation: View conversation

---

1. Story of the diff

This PR adds a focused rustdoc audit note to src/v3/compiler/src/int_literal_ranges.rs documenting the post-#1227 downstream picture for integer-literal range facts. The new section points back to the R2 Slice 2 brief at src/v3/compiler/src/int_literal_ranges.rs:23, then records that MethodEmitTemplate / emit-model template-contract paths do not share symbols with this module at src/v3/compiler/src/int_literal_ranges.rs:25-26. The load-bearing part is the consumer matrix: inference and lowering are the direct users of integer_range_for_decl / magnitude_out_of_range_for_interval, emission only sees already-resolved shapes, and method-template carriers are explicitly out of scope at src/v3/compiler/src/int_literal_ranges.rs:28-33.

2. Invariant categories

1. LAYER MODEL (substrate vs implementation).

N/A — the diff is documentation-only inside a Rust implementation module; it does not add or alter substrate types, Dag-resident fields, variants, or cross-pass carriers.

1. INVARIANTS.md + modeling-discipline.md.

Compliant — Boundary Discipline / single authority is the relevant principle: the audit says inference/lowering read the shared range facts, while emit::* does not add a parallel range walk and only emits resolved shapes at src/v3/compiler/src/int_literal_ranges.rs:30-32. That aligns with the reference rule that boundaries should carry facts once and avoid parallel authority. chatgpt-review-98e72d8f-8eaf-43…

1. CODING.md.

Compliant — the change adds module-level rustdoc only; it introduces no method, helper, hidden state, global dependency, builder, panic surface, or object-style API. The documented consumer split is also interface-oriented: it names who consumes which fact rather than encoding behavior in another path at src/v3/compiler/src/int_literal_ranges.rs:28-33. chatgpt-review-15cecad7-0283-46…

1. TESTING.md.

N/A — no behavior changes and no testable interface changes are introduced. The PR documents an audit result; it does not promote a behavioral-complete lens claim, change inference/lowering behavior, or add a new boundary that would need hermetic regression coverage. chatgpt-review-94d87da2-038d-43…

1. LOCKED DESIGN DECISIONS.

N/A — the diff references #1227 MethodEmitTemplate context, but only to state non-overlap with int-literal narrowing paths at src/v3/compiler/src/int_literal_ranges.rs:25-26 and src/v3/compiler/src/int_literal_ranges.rs:33; it does not alter or reinterpret a locked design surface.

1. TRACKED vs UNTRACKED DEBT.

Compliant — no new scaffold, TODO, bridge, fallback, or temporary representation is introduced. The added audit actually narrows ambiguity by documenting that method-template / emit-model carriers are not another consumer of these range facts at src/v3/compiler/src/int_literal_ranges.rs:33.

3. Verdict

APPROVE. This is a small documentation-only PR that clarifies the authority boundary around integer-literal range facts and avoids inventing a second consumer path. I found no changed line in the diff that violates the project invariant categories.

[briansrls]

Reply — manual review (44e18225, gpt-5-5-thinking)

Verified on int_literal_ranges.rs L23–L33: R2 brief link + #1227 non-overlap (L25–L26); consumer matrix infer/lower direct, emit::* indirect / no parallel range walk, MethodEmitTemplate N/A (L28–L33). Doc-only; invariant categories in the review match the diff — APPROVE stands, no fix commit.

— sent from quiet-raven-835

[briansrls]

Review metadata

• Provider / model: codex / unknown
• Commit: 44e18225 · Trigger: schedule
• Thinking: 101s wall

---

✅ The PR is a documentation-only downstream audit update and introduces no new substrate shape, compiler behavior, or authority split that would block landing.

[briansrls]

Reply — schedule review (44e18225, codex)

Verified: #1258 at 44e18225 remains rustdoc / audit-only in int_literal_ranges.rs — no new substrate, behavior, or parallel authority. No blocking finding — no fix commit.

— sent from quiet-raven-835