Lane A E0308 re-baseline (Route-A last mile): CONTINUE emitter/seed-green-integration (do NOT start fresh). Assemble the REAL crate — emitted seed closure + the ~20 hand-written periphery modules (v1_interpreter/cli_run/v1_rt/lens kernel) — cargo-build it, and produce a CATEGORIZED E0308 count by va

[briansrls]

Summary

E0308 re-baseline for the Route-A last mile. Assembles the REAL faithful emitted
seed crate — the emitted seed closure (108 modules) + the 22 hand-written periphery
modules (v1_interpreter/cli_run/v1_rt/lens-kernel project modules) — and
cargo-builds it to measure the latent emitter-completeness gaps that the committed
hand-synced seed hides (it deliberately leaves the std-tower modules unwired in
lib.rs). Produces a categorized E0308 count by variety — the input the carrier-mark
dissolution trigger (node://adhoc-80af9ff8-40f) was waiting on.

Per DESIGN §6 (no parallel-ledger docs), the categorized baseline lives in this PR body /
lane node, not in a revived docs/ file. The committed change is the reusable
apparatus: a non-destructive harness mode + a verified reproducer.

Result — the "~150 gaps" estimate was stale by ~4×

RUSTC_WRAPPER="" cargo build --lib on the assembled crate: 73 total errors, 38
E0308. The 38 reduce to two emitter root causes covering 89%:

n	family	root cause	fix locus
24 (+1 wrapped-label = 25)	A — map-literal &str key	emitter writes map keys as bare &str literals (__m.insert("true", …)) without .to_string(); values do get .to_string(), so the map infers HashMap<&str, V> vs the declared HashMap<String, V>	map-literal emission (one fix → −25 E0308, 66%)
10	B — Box/Rc/Option wrap	auto-wrap/unwrap coercion not inserted: Box<String>↔String (7), Rc<X> vs Option<Rc<X>> missing Some() (3)	coercion-insertion at call/construct sites
3	C — scalar &str/String	string literal not .to_string() in scalar position	same coercion path, scalar arm

Worst files: v1_compiler_stage0_crates.rs (6), std_types.rs (5),
extdeps_languages_dag_syntax.rs (4).

Non-E0308 residue (the other 35) — distinct classes, separately owned

n	code	class	where
12	E0560	measure-tower field emission (Rc<Measure<Q,U,i64>> has no field count)	std_measure.rs
8	E0433	unwired periphery (inert_carrier_project / non_fold_residue_project not declared in emitted lib.rs — regen patch-list stale vs the 3 newest HAND_MAINTAINED modules)	v1_interpreter.rs
5	E0425	unresolved fn/type (v1_rt::utf8_decode_bytes, SemVerConstraint)	extdeps_cargo.rs, …
2+2+2+1×4	E0573/E0599/E0091/E0277/E0252/E0107/E0063	std-tower generic-emission residue	std_realization_schedule / std_machine_constraints / std_types / std_termination

Honesty caveat (§5): the 38 is a lower bound. E0433 (8) and E0425 (5) are
resolution failures — rustc aborts type-checking in those regions, so they mask an
unknown count of further E0308. Families A and B are independent of those regions and
are the trustworthy, actionable core. (One confound was removed for this measurement: the
emitted Cargo.toml omitted im-rc, used by hand-written v1_interpreter → 2 spurious
E0432; the harness now adds it, mirroring the committed stage0 deps.)

Changes (the apparatus)

• regen_stage0.rs: additive, opt-in --emit-fresh <dir> — runs the normal regen
  assembly phases (emit closure → copy periphery → patches → rustfmt) into a caller-named
  dir and stops (no copy-back, no temp cleanup). The committed seed is never mutated.
  write / --verify behaviour is untouched. Also adds im-rc to the assembled
  Cargo.toml.

Reproduce

cargo build -p v1-compiler --bin regen_stage0
./target/debug/regen_stage0 --emit-fresh <dir>          # ~20 min: full v1-seed self-compile
cd <dir>
RUSTC_WRAPPER="" cargo build --message-format=json --lib > build.json 2>&1   # ="" bypasses flaky sccache

# total + E0308 share:
grep -oE '"code":"E[0-9]{4}"' build.json | sort | uniq -c | sort -rn

# E0308 by expected->found pair (one per diagnostic; sums to 38):
jq -r 'select(.reason=="compiler-message") | .message | select(.code.code=="E0308")
        | ([.rendered | scan("expected `[^`]+`, found `[^`]+`")][0]) // "<pair-on-note>"' build.json \
  | sort | uniq -c | sort -rn

Implication for the regen-fixpoint lane

E0308-to-green is not 150 independent gaps — it is two emitter root causes
(map-literal key coercion; Box/Rc/Option auto-wrap) covering 34/38, plus a small scalar
arm. The remaining green work splits into separately-owned classes: the measure-tower
generic emission (12 E0560 + residue) and regen-harness wiring (8 E0433).

Test plan

• cargo build -p v1-compiler --bin regen_stage0 — green; cargo clippy … -D warnings — green.
• ./target/debug/regen_stage0 --emit-fresh <dir> — assembles 108 emitted + 22 periphery;
  verified the emitted lib.rs wires the full std-tower (committed wires only std_algebra).
• RUSTC_WRAPPER="" cargo build --lib in <dir> — 73 errors / 38 E0308; the jq pipeline
  above reproduces the per-family pair counts (sums to 38).
• No .dag authority or load-bearing pipeline stage modified; measurement + non-destructive
  harness only. No docs/ file added (avoids the doc-reachability orphan gate).