HDMI acceptance and timing deferral — 2026-10-09

Evidence class: operator acceptance linked to previously committed serial, sysfs and build identity. No new board read, reboot, camera capture or latency measurement was performed for this closeout.

The operator observed HDMI → panel → HDMI with the expected input modes:

came up on hdmi. worked fine, touchpad worked. unpliugged hdmi, panel came up, touch worked

After the faster worker trial:

that’s great it’s perfect. replugging hdmi is slower than i’d like (couple of seconds) but unplugging -> display was almost instant

The contextual qualification records the requested cable cycle and real-touch All Apps, Overview and Terminal → Home checks, then the response:

it works great land it

Source records: fast-runtime-operator-report.json, fast-runtime-qualification.json, fast-runtime-qualified-state.json and fast-runtime-serial-boot.json in this directory. They retain trial boot ID cff529ad-6aa6-4a31-b9af-337bbc783521, HDMI portrait transform and virtual trackpad. This is physical operator acceptance; the named navigation checks use contextual affirmation, without invented per-gesture traces.

The operator then directed:

k what’s next ignore latency measurement good enough for now what proposals can we close

Precise sampled ≤1s HPD-to-connector and ≤3s HPD-to-enabled targets are therefore deferred, not performed or passed. fast-runtime-latency.json remains an empty transition capture with INCOMPLETE_OR_SLOW; it establishes neither a measured failure nor success. The approximate visible observations support the original 30-second ceiling for the accepted transitions. Monitor lock time is not separately measured. Independent performance proposals retain their own gates.

Installed artifact

Implementation/default promotion landed in 6b4fe555645ac553914549fcd926d1f0902b65d1. The normal installed system/profile is /nix/store/yl3si5ak6yi709yg1fqsnwgq0zn4xfcs-nixos-system-nixos-26.11.20260919.20b1ddd. Matching kernel: /nix/store/f3rnipvbc5vqam8kwdmn3wgcl5rbz9yx-linux-riscv64-unknown-linux-gnu-7.3.0-rc5/Image, SHA256 9d94102c65c8e4fd6b8880ef7b19b8bee5fa91c85596384319f0104bd89509cb. Normal bundle is /nix/store/jsqi00cgac7v773xi3r1zc6ppib3sc7p-k230-coherent-shell-boot-files; built default image is /nix/store/gw9p4yhc2yyp026bgmn635m26bhlrzhv-k230-sd-image.img.

normal-hotplug-install-serial.json proves recoverable normal installation and ordinary autoboot, retaining protected stage 1, DT selectors and rollback roots. normal-hotplug-runtime-state.json identifies installed boot 4bf73b24-a16a-4786-96fb-f1288244d96f, enabled HDMI and active shell/input services. default-hotplug-image-check.json proves the built image matches the qualified bundle. No whole-image flash is claimed. The volatile trial and subsequent normal autoboot remain distinct evidence classes.

Integration ownership

Worktree /home/jadams/tmp/k230-hdmi-close-final, branch closeout/hdmi-accepted-2026-10-09, base ba5217ca72f05fbc49ba818418737e8549ae5685. Owned paths: the HDMI change/archive, this evidence, its work-board entry and necessary references, and CLI-generated display/hdmi and display/touch specs. No source changes, build slot or board reservation. The operator subsequently canceled Settings reboot scope and authorized the landscape successor. See scope-decision-2026-10-09.md for the exact decision and task mapping. No canceled or transferred gate is reported as performed.

Narrow closeout checks: openspec validate plugging-in-hdmi-moves-the-display --strict, then (only after scope preservation) openspec archive plugging-in-hdmi-moves-the-display --yes --json, openspec validate --all --strict, python3 scripts/build_site.py and python3 tools/blob-scan.py --no-vendor. Record exact successor/task mapping and sync verification before final integration.