Phase 1 / D7: multi-module codegen at verifier parity

[hyperpolymath]

Deliverable 7 of #49. Producer-side multi-module emission matching the verifier's existing multi-module coverage.

Scope

• Emit shared-schema regions across modules + the typedwasm.region-imports carrier.
  • Note: that carrier is proposal 0003 [draft] (docs/proposals/0003-region-imports-carrier.adoc), currently gated on exactly this producer-side multi-module emission. Landing D7 is its unblock; if implementation surfaces a wire gap, file a proposal-0003 amendment rather than improvising.
• Field-IDs must be shared-schema-stable across modules — an importer's field_id numbering must match the exporter's (per the coordination note in docs/developer/handoff-typed-access-sites-codegen.adoc). A per-module counter will fail the cross-module case.

Acceptance

• examples/02-multi-module.twasm round-trips: tw build → .wasm accepted by verify_cross_module.
• A cross-module fixture lands in crates/typed-wasm-verify/tests/fixtures/c5_real/.

Dependencies

• Blocked by the single-module codegen path (F / D1).

Cross-refs

• Verifier verify_cross_module; MultiModule.idr no-spoofing theorem.
• Proposal 0003; L13 isolation: mirror verify_module_isolation in typed-wasm-verify (Rust) + INT-12 C5.1 real AffineScript fixtures #35 (L13 isolation parity).

Deliverable: Phase 1 / deliverable 7 (docs/PRODUCTION-PATH.adoc §Phase 1). Part of #49.

[hyperpolymath]

Closing as completed: the verifier-backed half of this issue — producer-side multi-module emission at parity with the verifier's existing cross-module coverage (the L10 linear-ownership boundary, via extract_exports + verify_cross_module) — landed in #139 (merged 55a105f), with 5 round-trip tests (clean single call accepted; double call rejected).

The remainder — L13 positive-form / typedwasm.region-imports (gated on proposal 0003 [draft]), the c5_real cross-module differential fixture, and wiring examples/02 — is re-filed as #140.

Part of Phase 1 (#49).

---

Generated by Claude Code