---
name: parked-thin-lane-delivery-emit-residue
description: "PARKED work owed by todlando — convert delivery.rs:149,313 to emit_line_err!, held until doyle confirms the v0.66.0 source-confirmed cut"
metadata: 
  node_type: memory
  type: project
  originSessionId: 880d1285-2246-4429-9b5e-eb4331af141f
  modified: 2026-08-29T20:39:00.833Z
---

Parked 2026-08-29, waiting on an external signal, so it will NOT be obvious from the code.

**The work:** convert two sites in `crates/spt/src/api/delivery.rs` from bare `eprintln!` to
`spt_proto::emit_line_err!` — line 149 (`IO_SINK_WARN:`) and line 313
(`DISPATCH_RESULTS_WARN:`). Nothing else. Behaviourally inert (both write the same line to
stderr), so it rides no release.

**Why they are mine:** I minted both by copying a sibling that was still pre-conversion at the
time — `9ec181e9` (W3/#239, `publish_boundary_io` copied `publish_state_io`) and `2e71652a`
(W2/#238). The #241 emitter conversion (`cfb9d9f8`) converted the three sites its snapshot held
and could not know about the two I added on top. See [[a-snapshot-census-cannot-see-a-site-that-does-not-exist-yet]]
if written, and `EMISSION-RESIDUAL-CENSUS-RECONCILIATION.md` at repo root for the full
attribution.

**Gate ruling (doyle, 2026-08-29):** THIN LANE, mine, dispatched only AFTER the v0.66.0 cut is
source-confirmed — main was frozen for r2 parity (tag==main==r2-tested). Deliberately NOT folded
into the #243 residual wave, which stays scoped to the census-truncation population. doyle pings
when the cut is released.

**Do NOT widen it:** `main.rs:144,193` are #243's, not mine — they are byte-identical at the
census cut `4d6007ac`, so they are census-truncation misses, not my mints. The tree-wide split
is 340 census-mechanism / 2 mine / 20 test-dir files / 1 in a test module.

**Related artifact worth not losing:** `EMISSION-RESIDUAL-CENSUS-METER.py` at repo root (moved
out of a session scratchpad on purpose) — re-runnable meter, `python … <sha>`; #243 wants it
when that request is minted.