feat(v3): add string diagnostic axis vocabulary

[briansrls]

Summary

• Add canonical string-family axis vocabulary in src/v3/std/emit_model.dag for future per-target diagnostic-ordering rows: StringOwnershipAxis, StringLifetimeAxis, StringGrowabilityAxis, and StringEncodingAxis.
• Keep axis values structural declarations rather than strings, with NotApplicable as an explicit growability value.
• Add focused bootstrap ratchets proving the axes live under emit-model authority and retain the intended closed values.
• Regenerate bootstrap snapshots and refresh the handwritten parse corpus manifest.

Boundaries

• Axis vocabulary only.
• Does not extend TypeRealization.
• Does not add std/lifetime.dag.
• Does not populate Rust/string-family target rows or add a Grounding projection reader.

Validation

• CARGO_INCREMENTAL=0 cargo run -p v3-compiler --features bootstrap-regen-fresh --bin regen_bootstrap
• cargo test -p v3-compiler refresh_handwritten_parse_snapshot_manifest -- --ignored
• cargo test -p v3-compiler --test integration string_diagnostic_ordering_axes_live_in_emit_model_authority -- --nocapture
• cargo test -p v3-compiler --test integration string_diagnostic_ordering_axes_are_closed_structural_values -- --nocapture
• cargo test -p v3-compiler --test integration parse_stage4_prep::handwritten_parse_snapshot_matches_manifest -- --nocapture
• cargo fmt --all --check
• git diff --check
• CARGO_INCREMENTAL=0 cargo run -p v3-compiler --features bootstrap-regen-fresh --bin regen_bootstrap -- --verify

[briansrls]

Review metadata

• Provider / model: cursor / composer-2
• Commit: 779d5938 · Trigger: schedule
• Comparison: origin/main @ d25b89a6 ... review/pr-1465-779d5938 @ 779d5938
• Thinking: 40s wall

---

Reviewed the filtered diff: substantive edits are emit_model.dag (four axis types + 🟢 TERMINAL comments), two integration tests, manifest hash lines, and regenerated bootstrap Rust. The large bootstrap_generated*.rs churn matches declaring new substrate types.

Findings: None. The new sum types include 🟢 TERMINAL scope notes and rationale (including distinction from ParameterDisposition), matching docs/modeling-discipline.md coproduct classification. Tests follow the same patterns as special_value_policy_axes_are_closed_sums and existing decl.span.file authority checks in this file (TESTING.md behavior-style pinning of substrate variant labels, not ad hoc message text). Single authority for these axes stays in emit_model.dag, with parse corpus manifest updated — consistent with boundary discipline.

Verdict: APPROVE — Narrow vocabulary + ratchet tests + regen; no invariant or style violations spotted in the substantive diff.

[briansrls]

Review metadata

• Provider / model: codex / unknown
• Commit: f862e64d · Trigger: schedule
• Thinking: 227s wall

---

Non-blocking — Strengths

• src/v3/std/emit_model.dag Mixed PR, primarily .dag model: the new string diagnostic axes are closed emit-model vocabulary with terminal receipts and targeted bootstrap tests.

✅ No blocking concerns found.

[briansrls]

Review metadata

• Provider / model: cursor / composer-2
• Commit: dae39a13 · Trigger: schedule
• Comparison: origin/main @ 31b0a7a2 ... review/pr-1465-dae39a13 @ dae39a13
• Thinking: 24s wall

---

Findings

None. The diff adds closed substrate sums in emit_model.dag with explicit TERMINAL classification and scope notes (aligned with docs/modeling-discipline.md §4), places authority in the emission model with tests that pin declaration site and variant labels (supports P2 single authority and typed carriers vs ad hoc strings — consistent with DB-1 / C-5 direction). Nothing in the diff suggests fail-closed fabrication, duplicate competing authorities, or scaffold without a named forward path (encoding explicitly defers further variants to later LanguageSpec rows).

Verdict

APPROVE — Narrow, coherent slice: substrate vocabulary + regen + inhab tests that lock location and closed sums. No invariant violations identified from the substantive diff (emit_model.dag, m2_substrate_inhabitance_test.rs, manifest tweak; bootstrap churn is expected mechanical fallout).

[briansrls]

Review metadata

• Provider / model: codex / codex-default
• Commit: ed283c8d · Trigger: schedule
• Comparison: origin/main @ 17774325 ... review/pr-1465-ed283c8d @ ed283c8d
• Thinking: 43s wall

---

Verdict: APPROVE — the diff is narrowly scoped to canonical .dag vocabulary plus generated bootstrap updates and regression checks. I don’t see a concrete violation of the pinned invariants, coding discipline, or testing guidance in the changed lines.

I did not run builds or tests, per instruction.

[briansrls]

Review metadata

• Provider / model: codex / unknown
• Commit: ed283c8d · Trigger: schedule
• Thinking: 174s wall

---

✅ Mixed PR, primarily a .dag model vocabulary addition with regenerated bootstrap/tests; no blocking concerns found.

[briansrls]

Review metadata

• Provider / model: codex / codex-default
• Commit: 23a2bb97 · Trigger: schedule
• Comparison: origin/main @ 3bebb824 ... review/pr-1465-23a2bb97 @ 23a2bb97
• Thinking: 50s wall

---

Verdict: APPROVE

The diff adds canonical string diagnostic axis vocabulary in emit_model.dag, with coproduct classifications documented at the declarations and regression coverage confirming authority location and variant shape. I don’t see a concrete violation of the pinned invariants, coding, or testing discipline in the reviewed diff. Tests were not run per instruction.

[briansrls]

Review metadata

• Provider / model: codex / codex-default
• Commit: 6d6f87f0 · Trigger: schedule
• Comparison: origin/main @ 4ba0a6d0 ... review/pr-1465-6d6f87f0 @ 6d6f87f0
• Thinking: 49s wall

---

Verdict: APPROVE

The diff is narrowly scoped to adding canonical string diagnostic axis vocabulary in src/v3/std/emit_model.dag, regenerating bootstrap snapshots, and adding shape/authority assertions. I don’t see a concrete violation of the pinned modeling, coding, or testing discipline in the changed lines.

[briansrls]

Review metadata

• Provider / model: codex / unknown
• Commit: 6d6f87f0 · Trigger: schedule
• Thinking: 192s wall

---

Non-blocking — Strengths

• src/v3/std/emit_model.dag The added string-family axes are typed closed vocabulary in the emit-model authority, with terminal rationale and tests pinning the variant sets.

✅ No blocking concerns.

[briansrls]

Review metadata

• Provider / model: codex / codex-default
• Commit: b3f70668 · Trigger: schedule
• Comparison: origin/main @ 8c745cae ... review/pr-1465-b3f70668 @ b3f70668
• Thinking: 47s wall

---

Verdict: APPROVE

The diff is narrowly scoped: it adds typed string diagnostic axis vocabulary in src/v3/std/emit_model.dag with terminal classification comments, regenerates bootstrap output, and adds focused regression checks for authority location and closed structural values. I don’t see a concrete violation of the pinned modeling, coding, or testing guidance in the changed lines.

[briansrls]

Review metadata

• Provider / model: openai-pro / gpt-5-5-pro
• Commit: 6d6f87f0 · Trigger: manual
• Comparison: main @ 8c745cae ... session/crisp-ibex-569-string-axis-vocabulary @ b3f70668
• Conversation: View conversation

---

1. Story of the diff

This PR adds canonical substrate vocabulary for string-family diagnostic ordering, placing the axes in src/v3/std/emit_model.dag rather than letting future target rows or analyzers normalize string names locally. The new model separates carrier ownership from callable parameter disposition (StringOwnershipAxis at src/v3/std/emit_model.dag:87-95), names shared lifetime roles without Rust identifier leakage (StringLifetimeAxis at src/v3/std/emit_model.dag:97-103), makes growability’s “not relevant” state explicit (StringGrowabilityAxis at src/v3/std/emit_model.dag:105-111), and introduces the current UTF-8 / FreeMonoid<Char> encoding row as a structural value (StringEncodingAxis at src/v3/std/emit_model.dag:113-118). The generated bootstrap fixtures then materialize those declarations, e.g. StringOwnershipAxis through StringEncodingAxis appear as generated TypeConnective::Disj declarations at src/v3/compiler/src/bootstrap_generated.rs:898-901, with the corresponding no-parse-surface fixture regenerated as well. The tests add two substrate-inhabitance checks: one pins the authority file to emit_model.dag (src/v3/compiler/tests/integration/m2_substrate_inhabitance_test.rs:2553-2569), and one pins the closed variant vocabulary (src/v3/compiler/tests/integration/m2_substrate_inhabitance_test.rs:2573-2598).

2. Invariant categories

1. LAYER MODEL (substrate vs implementation).

Compliant — this does touch substrate vocabulary, but it lands the fact at the emission-model authority rather than in implementation-local Rust: type StringOwnershipAxis is declared in src/v3/std/emit_model.dag:93, and the test explicitly requires the axis declarations to remain in src/v3/std/emit_model.dag at src/v3/compiler/tests/integration/m2_substrate_inhabitance_test.rs:2565-2568.

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

Compliant — illegal-states-unrepresentable / fail-closed modeling is handled by explicit closed sums instead of booleans, strings, or absence. The clearest example is StringGrowabilityAxis, where NotApplicable is a real variant at src/v3/std/emit_model.dag:108-111, and the comment states that this prevents the axis from becoming “unknown” by omission at src/v3/std/emit_model.dag:105-107. Coproduct discipline is also addressed: each new multi-variant axis carries a 🟢 terminal classification comment at src/v3/std/emit_model.dag:87, src/v3/std/emit_model.dag:97, and src/v3/std/emit_model.dag:105.

1. CODING.md.

Compliant — the Rust-side additions are small, direct test functions over data, not new hidden state or object behavior. The tests read the generated Dag and ask for declarations by name at src/v3/compiler/tests/integration/m2_substrate_inhabitance_test.rs:2554-2564, then assert on structured declaration facts via disj_variant_labels at src/v3/compiler/tests/integration/m2_substrate_inhabitance_test.rs:2577-2595.

1. TESTING.md.

Compliant — the PR adds behavior-focused regression coverage for the exact substrate contract being introduced. One test checks single-authority placement (src/v3/compiler/tests/integration/m2_substrate_inhabitance_test.rs:2553-2569); the other checks the closed variant sets and the explicit NotApplicable carrier (src/v3/compiler/tests/integration/m2_substrate_inhabitance_test.rs:2573-2598). Using the generated bootstrap Dag is appropriate here because the claim is about generated substrate inhabitance, not an isolated Rust helper.

1. LOCKED DESIGN DECISIONS.

N/A — I do not see a diff line that alters a locked design decision. The only design-facing language is alignment/scoping text for the encoding axis, where StringEncodingAxis names the current .dag String = FreeMonoid<Char> / UTF-8 row and says later byte or target-specific encodings belong in later LanguageSpec rows at src/v3/std/emit_model.dag:113-118.

1. TRACKED vs UNTRACKED DEBT.

N/A — no new TODO, shim, temporary adapter, or scaffold is introduced in the diff. The “Rows land in target specs later” language at src/v3/std/emit_model.dag:90-92 and in the authority test message at src/v3/compiler/tests/integration/m2_substrate_inhabitance_test.rs:2567-2568 reads as scope delimitation for this vocabulary slice, not as a second representation or bridge that consumers can wire into.

3. Verdict

APPROVE. The PR is narrow and structurally consistent: it moves string diagnostic-ordering vocabulary into typed substrate declarations, regenerates the bootstrap fixtures, updates the manifest, and adds targeted regression tests for authority and closed variant shape. I did not find a diff-citable invariant, coding, testing, locked-design, or debt issue.