# R04 evidence — [#25697](https://git.cafe/pingdotgg/t3code/pull/25697)

Before [3d369d5596](https://git.cafe/pingdotgg/t3code/commit/3d369d55968b09861bf61ba8ebb2eeb6f02973ba); after [a77de508c2](https://git.cafe/pingdotgg/t3code/commit/a77de508c2d34983d2bf07dd64e075df582554dc). Updated R07 goal expectation applies.

## R04 — server runtime (viewport/theme N/A)

Environment: fs-19 • Linux x86_64 • Node 26.13.0 • vite-plus 1.4.0 / Vitest 5.0.5 • in-memory SQLite • Effect TestClock at Unix epoch; synthetic Git seed timestamps fixed. H: real runtime/orchestrator/event sink/projection + registered fake Codex provider; real t3_queue toolkit/service/access context. No model execution. Viewport/theme/DPR/emulation/font scaling: N/A (server-only runtime transcript).

| Setup | Before ([3d369d5596](https://git.cafe/pingdotgg/t3code/commit/3d369d55968b09861bf61ba8ebb2eeb6f02973ba)) | After ([a77de508c2](https://git.cafe/pingdotgg/t3code/commit/a77de508c2d34983d2bf07dd64e075df582554dc)) | Differs as expected? |
| --- | --- | --- | --- |
| **R04-absent-drain** — H. EventSink legacy q injection U=1,C1=2,C2=3, marker absent; drain untouched. | Legacy absent marker stays completion-first. Actual starts: **C1, C2, U**. | Same legacy completion-first behavior. Actual starts: **C1, C2, U**. | Yes; required differences/preservation observed. |
| **R04-absent-rebuild** — H. Legacy q injection, marker absent; prioritize U at unchanged q1; rebuild; replay command ID; move U to end; rebuild. | Priority reorder rejects; rebuild completion-first; replay rejected receipt; zero replay run updates. Actual starts: **C1, C2, U**. | Marker-only U=true q1 survives rebuild; accepted replay adds 0 run updates; move-back false survives second rebuild. Actual starts: **C1, C2, U**. | Yes; required differences/preservation observed. |
| **R04-false-drain** — H. EventSink legacy q injection U=1,C1=2,C2=3, explicit false; drain untouched. | Unknown false field is stripped by baseline codec; completion-first. Actual starts: **C1, C2, U**. | Explicit false is retained by codec; same completion-first behavior. Actual starts: **C1, C2, U**. | Yes; required differences/preservation observed. |
| **R04-false-rebuild** — H. Legacy q injection, explicit false; prioritize U at unchanged q1; rebuild; replay command ID; move U to end; rebuild. | False precondition reads absent; priority/replay rejected; completion-first. Actual starts: **C1, C2, U**. | False→true at unchanged q1 survives rebuild; accepted replay adds 0 updates; false survives move-back rebuild. Actual starts: **C1, C2, U**. | Yes; required differences/preservation observed. |

PR-ready capture links (all named variants above are present in each paired log):

| Before<br>[3d369d5596](https://git.cafe/pingdotgg/t3code/commit/3d369d55968b09861bf61ba8ebb2eeb6f02973ba) | After<br>[a77de508c2](https://git.cafe/pingdotgg/t3code/commit/a77de508c2d34983d2bf07dd64e075df582554dc) | Diff0<br>[3d369d5596](https://git.cafe/pingdotgg/t3code/commit/3d369d55968b09861bf61ba8ebb2eeb6f02973ba) → [a77de508c2](https://git.cafe/pingdotgg/t3code/commit/a77de508c2d34983d2bf07dd64e075df582554dc) | DiffT<br>[3d369d5596](https://git.cafe/pingdotgg/t3code/commit/3d369d55968b09861bf61ba8ebb2eeb6f02973ba) → [a77de508c2](https://git.cafe/pingdotgg/t3code/commit/a77de508c2d34983d2bf07dd64e075df582554dc) |
| --- | --- | --- | --- |
| [Runtime capture + full run.updated events](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/c347defd849ce198/R04.before.raw.log) | [Runtime capture + full run.updated events](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/7570016df57dc625/R04.after.raw.log) | [Zero-threshold unified diff](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/72289d767b464635/R04.Diff0.diff) | [No unexpected differences](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/bf29448016c2ba5e/R04.DiffT.txt) |

[Normalized before](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/a76ed3c6bb26cb4a/R04.before.normalized.txt) · [Normalized after](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/dcfb0641c6391328/R04.after.normalized.txt) · [One-line inventory of every difference](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/6a0b960be35230cd/R04.difference-inventory.txt) · [All raw persisted domain events](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/39570f588e6fb493/R04.raw-events.tar.gz)

[driver + all synthetic seeds](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/3203c71e7e07a194/queue-evidence.test.ts) · [reproduction commands](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/8310722205aaf986/reproduce.sh) · [original regression driver](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/fd526b4ceced7e37/queue-regression.test.ts) · [normalization and packaging](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/d5ee54a963c65213/package-evidence.py) · [payload comparison audit](https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/c0895aa7b3577b15/audit-transcripts.py).

Unverified variants: none.