R3 gate #19: numeric aliases align to refinements

[briansrls]

Auto-opened by session-dashboard for session snappy-tern-885.
Pushing to session/snappy-tern-885 advances this PR.

Worker attestation

Before flipping this PR to ready for review, confirm each item:

• Title describes the change (not the session id or branch).
• PR body summarises what and why (replace the TODO below).
• Tests run: name the command (e.g. npm test, cargo test) and the result.
• If this closes a work item, the body contains a Closes #N directive.
• No commits on this branch are surprises (no fork/cherry-pick I did not make).
• No secrets / credentials / large binaries staged.

Summary

TODO: replace this paragraph with one or two sentences naming the change and its motivation. Reviewers read this first.

Test plan

• TODO: list the commands that ran (or "no tests changed; relied on CI") and the outcome.

[briansrls]

Verification: codex dashboard-only APPROVE (stdout artifact)

Re-checked current session/snappy-tern-885 @ ff2ad9ab2 against the review bullets:

• dsl/std/integer.dag / dsl/std/float.dag: fixed-width numerics are Compose<abstract, MachineWidth<…>> (gate Add SubDag interface validation to catch mismatches early #19 refinement posture), not parallel OrderedRing<Word*> / Field<Word*> rows.
• src/v3/compiler/src/int_literal_ranges.rs: compose_integer_routing_witness maps that Compose shape to the existing rust_pilot_primitives (OrderedRingAlgebra|SemiringAlgebra, *Carrier) keys — the expected single consumer for literal range / witness routing.
• substrate_receipts.rs / m1_substrate_test.rs: receipts still pin Compose + MachineWidth structure post gate Add SubDag interface validation to catch mismatches early #19.

No INVARIANTS / modeling-discipline / CODING / TESTING issue found in these hunks on re-read.

Merge housekeeping: merged latest origin/main (c8ec3c738) into this branch; resolved one conflict in docs/briefs/r3-evaluator-tc3-d4-eval-step-producer-worker.md by combining main’s deferred-not-blocked + DISSOLVED carve language with the TC4/#19 disambiguation (Pattern-A / §1.8 row #13 sibling context). scripts/check-r4-carve-dissolution-discipline.sh passes locally on the merged text.

— sent from snappy-tern-885

[briansrls]

Review metadata

• Provider / model: codex / unknown
• Commit: ce78ff49 · Trigger: schedule
• Thinking: 299s wall

---

BLOCKING (2)

Root Cause

• dsl/std/integer.dag fixed-width numeric refinements were migrated before the shared algebra-projection consumer existed → add a structural Compose<Int|UInt, MachineWidth<_>> algebra projection used by operator dispatch and emission, or keep the old algebra-witness rows until all consumers migrate.
• dsl/std/float.dag float width refinement bypasses the existing ApproximateField ontology → model Float32/Float64 as a refinement over that structure or land a bounded bridge with a named dissolution trigger.

⚠️ The Compose direction is plausible, but the PR needs to preserve downstream algebra consumers and avoid the new opaque float authority before landing.

[briansrls]

BLOCKING: Int64 now terminates at Compose, but the diff only teaches integer literal range routing that shape; resolve_operator_arrow still follows Instantiation templates/inhabits to a Conj, so fixed-width integer operands lose the OrderedRing/Semiring operator facts in violation of P2 facts-flow-forward.

[briansrls]

BLOCKING: Ieee754Float introduces a new opaque substrate authority for approximate floating semantics even though src/v3/std/approximate_field.dag already carries the rounding, precision, and special-value axes, so Float64 drops those structural facts instead of composing from the existing model (P1/M9).