Decision spike for the firmware/test/ host harness (split of bf-4ne).
Records why the ESP-IDF --target linux / Unity-host path was rejected:
csi.c is blocked by esp_wifi.h and provision.c by driver/uart.h, and
firmware/main builds as one component whose REQUIRES line names
esp_wifi/bt/driver — none of which have a linux build. Falls back to a
plain gcc harness under firmware/test/ that does not link main/*.c.
Reason must be carried into bf-4ne's test-runner header comment.
Co-Authored-By: Claude <noreply@anthropic.com>
Update docs/plan/plan.md to reflect the actual repo (VERSION 0.1.357):
- Go Module Layout: go.work stitches 3 modules (mothership, cmd/sim,
test/acceptance); no test/integration/ — tests live in test/acceptance/,
mothership/test/acceptance/, and tests/e2e/run.sh
- Dockerfile: espressif/idf:v5.2 firmware stage + golang:1.25-bookworm,
GOOS/GOARCH pinned to linux/amd64 (single-arch, deliberate), image
published as ronaldraygun/spaxel via the spaxel-build WorkflowTemplate
- Quality Gates item 4: document the amd64-only build decision (arm64 is
future work) instead of the contradictory multi-arch gate
- Integration Tests: correct location to the real test dirs
- Open Questions: remove the duplicated Multi-installation coordination bullet
- Bump Last updated to 2026-07-03; Status → maintenance mode
Co-Authored-By: Claude <noreply@anthropic.com>
Detailed IO-1..IO-11 scenarios validating the full new-user journey (fresh install ->
first-run PIN setup -> device onboarding -> operational) entirely via the spaxel-sim
ESP32 simulator, hardware-free and deterministic in CI. IO-1/3/4/6 are release hard-gates.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add BenchmarkFusionLoop and TestTimingBudgetProduction that enforce the fusion loop timing budget as a CI quality gate per plan §Quality Gates / Definition of Done (item 9).
The benchmark runs the full fusion pipeline (phase sanitization → feature extraction → Fresnel accumulation → peak extraction → UKF update) against synthetic CSI data from spaxel-sim output.
Timing constraints:
- Median fusion iteration < 15ms (production target)
- Median fusion iteration < 30ms (CI threshold - 2x allowance for slower CI hardware)
- P99 < 40ms (hard limit)
Typical results on reference hardware:
- Median: ~3-5ms (well under 15ms production target)
- P99: ~14-20ms (well under 40ms hard limit)
Also includes:
- GitHub Actions workflow (.github/workflows/benchmark-ci.yml) for CI
- Documentation (docs/ci-benchmark-integration.md) for Argo Workflows integration
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Verified all REST API endpoints are implemented and tested:
- Settings: GET/POST /api/settings with validation
- Zones: GET/POST/PUT/DELETE /api/zones with history
- Portals: GET/POST/PUT/DELETE /api/portals with crossings
- Triggers: GET/POST/PUT/DELETE /api/triggers with test endpoint
- Notifications: GET/POST /api/notifications/config and test
- Replay: GET/POST sessions, seek, tune, speed control
- BLE Devices: GET/PUT/DELETE /api/ble/devices with aliases
All endpoints include OpenAPI-style godoc comments and return appropriate
JSON with proper HTTP status codes. Settings persist to SQLite across
restarts. Zone/portal changes broadcast via WebSocket for live updates.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Add optional mothership IP override to the provisioning flow so nodes
on networks where mDNS is blocked (enterprise WiFi, mesh, VLANs) can
connect on first boot without manual intervention.
- Add ms_ip field to provisioning Payload and request structs
- Firmware writes ms_ip to both NVS_KEY_MS_IP and NVS_KEY_MS_IP_PROV
- Discovery prefers provisioned IP on first attempt, falls back to
mDNS, then cached IP
- Web Serial wizard adds Mothership IP field in Network Troubleshooting
- Auto-populates IP when browser accesses dashboard by IP address
- Document when/how to use the override in docs/notes/mdns-override.md
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
