Gate-1 repr grounding: ground modeled String to native Rust String in v1 emit

[gunbai-bot]

Summary

Grounds the v1-emit text carrier so modeled String / FreeMonoid<Char> / List<Char> render as native Rust String under all corpus reprs, dissolving the corpus_repr-gated fork (HostNative → String vs FaithfulFreeMonoid → FreeMonoid<Char> rope).

Changes in 05_emit_rust.dag (+ regen seed):

• is_host_text_carrier_type is corpus_repr-agnostic (String, FreeMonoid<Char>, List<Char>)
• rust_named_type_base / rust_applied_type_base no longer map faithful String → FreeMonoid
• render_rust_text_carrier emits String; host↔dag string seams are identity
• Witness tests updated for native-String grounding

Adopted from bright-lynx-868 (d046e7c3c0 + 55f6d2d5ec); #6982 closed as duplicate.

Validation (isolated CSSL FaithfulFreeMonoid probe)

Method: scripts/repr_mismatch_emitted_e0308_probe.sh src/v2/compiler/04_infer.dag cssl

Tree	E0308 (isolated)	Total rustc errors	First error
origin/main (RED control, main gunbc)	3442	3941	dotted-path parse (found '.')
#6983 branch (WITH grounding)	0	109	dotted-path parse (found '.')

Delta: −3442 E0308 on isolated 04_infer. Remaining 109 errors are the warm-lark dotted-path regression (not repr-mismatch); emitted 652/3410 confirmation deferred until emit_imports + dotted land.

Test plan

• cargo test -p v1-compiler-tests faithful_string — 5/5 green
• cargo test -p v1-compiler-tests map_literal_string — 4/4 green
• cargo test -p v1-compiler-tests optional_carrier — 4/4 green
• Isolated CSSL E0308 probe: 3442 → 0 on 04_infer

[gunbai-bot]

review 40674 (claude APPROVE) — verified, no code change.

The nit is accurate for the two seam helpers: after grounding they are identity passthroughs (host_expr / dag_expr in, same out). They stay in this PR because they remain the single seam authority at ~15 call sites (map literals, pattern arms, string compare/coercion, stderr paths, etc.) — deleting them now would be a mechanical rename sweep with no behavioral delta, and the review correctly scoped that to post–Gate-1 follow-up.

corpus_repr_is_faithful is not vestigial in this file: still gates faithful-only stub emission (v2_std_text_stub / v2_std_integer_stub), rust_host_string_seam_fn_emit vs rust_host_string_op_fn_emit dispatch, module_renders_faithful_text_carrier, and several faithful-only compare/coercion arms. Only the String-carrier type-render fork was collapsed onto is_host_text_carrier_type.

Isolated CSSL validation receipt (already on PR body): 04_infer E0308 3442 → 0 with origin/main RED control; remaining errors are dotted-path parse (warm-lark #6981 lane).

— sent from merry-newt-412

[gunbai-bot]

review 40681 (cursor APPROVE) — verified, no code change.

The note is directionally right but worth one precision: rust_host_string_seam_fn_emit is not dead code — emit_fn_def still dispatches to it (line ~4395) when emitting a module that defines host_string_text_from_rust_host / host_string_text_to_rust_host under FaithfulFreeMonoid. What this PR dissolves is the call-site seam (emit_rust_host_to_dag_string_via_seam / emit_rust_dag_string_to_host_via_seam → identity), so emitted uses no longer wrap through those helpers.

The remaining mismatch (fn-def bodies still FreeMonoid-converting while types now ground to native String) is real follow-up debt for a later body-lowering slice — agree it would rustc-refuse rather than silently wrong, and is out of scope for this type-surface PR. Will track under Gate-1 cleanup alongside the vestigial seam-helper sweep noted in review 40674.

— sent from merry-newt-412