linearvoid/.github/MEMORY.md

# MEMORY

## 2026-08-07

- Completed runtime validation section from TODO.
- Runtime video normalization tests executed in running `linearvoid-playout` container using ffmpeg/ffprobe.
- Verified normalized outputs for MOV, MKV, VFR, and 5.1 inputs: `h264`, `yuv420p`, `30/1`, `aac`, `48000`, `2 channels`.
- Concat continuity check passed with `CONCAT_OK`; output duration `25.344667` seconds.
- Responsive viewport test on live viewer (`390x844`, `412x915`) detected 8px horizontal overflow from `100vw` usage in viewer container styles.
- Fixed overflow in source by changing `.lv-body` and `.viewer-shell` width from `100vw` to `100%` in `frontend/src/style.css`.
- Redis restart test:
  - Admin session endpoint returned 200 before and after Redis restart.
  - Presence key survived Redis restart in test-lab configuration.
  - Redis health returned to `healthy` post-restart.
- Files changed:
  - `.github/instructions/TODO.md`
  - `.github/instructions/DONE.md`
  - `frontend/src/style.css`
  - `CHANGELOG.md`

## 2026-08-07 (continued)

- Completed documentation half of recovery runbook TODO item.
- Expanded `docs/incident-runbook.md` with quarterly DR drill: prerequisites, outage simulation, restore validation, pass/fail criteria, and evidence template.
- Split TODO item into:
  - documented drill (`[x]`)
  - manual rehearsal execution (`[ ]`)
- Updated `CHANGELOG.md` with incident runbook drill documentation entry.

## 2026-08-07 (continued 2)

- Added manual operations checklist `docs/operations/cloudflare-uptimerobot-checklist.md` for Cloudflare + UptimeRobot task execution and evidence capture.
- Linked checklist from `SETUP.md` Cloudflare section and TODO manual sections.
- Did not mark manual dashboard items complete; checklist is preparation only.

## 2026-08-07 (continued 3)

- Added `docs/operations/scaling-performance-playbook.md` covering manifest TTL tuning, TTFB baseline method, WebSocket/API benchmarking approach, cache ratio tracking, capacity triggers, and Phase 2/3 plans.
- Updated TODO and BACKLOG to mark playbook creation done while keeping execution/measurement tasks pending.
- Linked scaling playbook in `README.md` resources.

## 2026-08-09

- Added WebSocket benchmark summary helper `scripts/loadtest/summarize-artillery.ps1`.
- Updated `scripts/loadtest/run-loadtests.ps1` to auto-generate `ws-chat-summary.txt` for `ws` and `all` scenarios.
- Updated `scripts/loadtest/summarize-k6.ps1` to produce pipeline-friendly output and integrated automatic `admin-api-endpoint-summary.txt` generation.
- Expanded `scripts/loadtest/README.md` with PowerShell-first artifact capture and helper usage for both ws and admin scenarios.
- Updated `.gitignore` to ignore loadtest artifacts (`artifacts/` and `scripts/loadtest/artifacts/`).
- Added artifact-to-worksheet field mapping section in `docs/operations/scaling-performance-playbook.md` so ws/admin worksheet entries map directly to generated summary files.
- Updated `docs/DEVELOPMENT.md` and `.github/instructions/TODO.md` to mark ws/admin benchmark scaffolding complete while live benchmark execution remains pending.
- Added `CHANGELOG.md` entry `1.0.10` summarizing loadtest helper automation and playbook mapping updates.

## 2026-08-09 (continued)

- Completed Phase 1 (test fixtures): documented ffmpeg fixture-generation commands in `playlist_test.go` header comments (valid_720p30, invalid_h265, invalid_24fps).
- Completed Phase 4 (R2 integration):
  - Enhanced `SyncToR2` in `playlist.go` to set CacheControl + ContentType based on file extension:
    - `.m4s`, `.ts`, `.mp4`, `.m4a` -> `public, max-age=31536000, immutable` + `video/MP2T`
    - `.m3u8` -> `public, max-age=2` + `application/vnd.apple.mpegurl`
    - unknown extension -> no cache headers
  - Added three tests: `TestSyncToR2` (segment upload with cache verification), `TestSyncToR2_ManifestCacheHeaders` (manifest upload), `TestSyncToR2_NoCacheHeadersForUnknownExt` (unknown extension).
  - All 17 playout tests pass (16 pass, 1 pre-existing skip).
- Updated TODO.md: Phase 1 marked complete, Phase 4 partially complete (edge cache hit/miss remains manual).
- Updated DEVELOPMENT.md: Test Fixtures marked complete, R2 Integration Test marked partially complete.

## 2026-08-09 (continued 2)

- Completed Phase 2 chat tests in `frontend/src/main.test.ts`:
  - Presence updates: join/leave system messages add and remove users from presence list, sorted rendering.
  - Private DM flow: togglePrivateDM creates correct dm: room name from sorted usernames.
  - Message rendering: system messages, self messages (message-self), and other-user messages (message-other) all render correctly.
  - Total frontend tests: 27 passing (2 test files).
- Marked Phase 4 "Document stubs/mocks" done (putObject injection pattern).
- Updated TODO.md: Phase 2 chat tests checked, Phase 4 stubs checked.
- Updated DEVELOPMENT.md: User Frontend Tests marked partially complete, R2 stubs checked.
- Updated CHANGELOG.md with Phase 2 presence/DM/message tests.

## 2026-08-09 (continued 3)

- Completed Phase 3 admin frontend tests in `frontend/src/admin.test.ts`:
  - Exported `savePlaylist` and `renderPlaylistSortable` from `admin.ts`.
  - Added playlist save test (verifies PUT /api/admin/playlist with JSON body).
  - Added playlist save error handling test (verifies failure status).
  - Added sortable rendering test (verifies index, paths, remove buttons, comment/blank line skip).
  - Total frontend tests: 31 passing (2 test files).
- Updated TODO.md: Phase 3 playlist/dashboard/API integration all checked.
- Updated DEVELOPMENT.md: Admin Frontend Tests marked partially complete.

## 2026-08-10 (continued 2)

- Implemented STREAMING.md production-readiness gaps from plan `.hermes/plans/2026-08-10_164500-production-readiness.md`:
  - **Task 1-2**: Added `runID` (ISO timestamp format) and monotonic `segSequence` for globally unique segment naming (`r-{ISOTIME}/seg-{seq}.m4s`). Replaced `strftime` pattern with run ID-based path.
  - **Task 2**: Added `watchAndPublishSegments` using `fsnotify` for real-time per-segment R2 upload (~7s delay). Previously `syncLatestToR2()` only ran after FFmpeg exit.
  - **Task 3**: Added `publishManifest` with atomic `CopyObject` (temp key → final key) to prevent manifest referencing unconfirmed segments.
  - **Task 4**: Updated `manifestCacheControl` from `"public, max-age=2"` to `"public, max-age=1, s-maxage=2, must-revalidate"` per STREAMING.md.
  - **Task 5**: Added `hls.on(Hls.Events.ERROR)` handler in `main.ts` with bounded retries (5x), exponential backoff (2s→32s), reconnecting overlay (spinner + text). Added `showReconnecting`/`hideReconnecting` functions.
  - **Task 6**: Enriched `HealthHandler` with `run_id`, `active_item`, `playlist_revision`, `last_segment_seq`, `last_r2_pub`, `consecutive_failures`. Added "unhealthy" status when R2 publish is stale (>12s).
- Files changed: `backend/playout/src/playlist.go`, `playlist_test.go`, `go.mod`, `frontend/src/main.ts`, `main.test.ts`, `style.css`, `index.html`.
- Updated `.github/instructions/TODO.md`, `BACKLOG.md`, `MEMORY.md`, `DONE.md`.
- Remaining: CORS on media.linearvoid.tv (Cloudflare R2 config, infrastructure).

## 2026-08-10 (continued 3)

- Completed Phase 2 (viewer) + Phase 3 (admin) responsive/accessibility tests:
  - 5 viewer accessibility tests: video controls/attributes, chat panel aria-labels, chat toggle aria-expanded, message input label, reconnecting overlay aria-live/role.
  - 3 viewer responsive tests: chat-collapsed toggle, CSS grid layout detection, now-playing aria-live.
  - 3 admin accessibility tests: nav aria-label, active tab aria-current, nav button focusability.
  - 2 admin responsive tests: tab content visibility, essential tab presence.
  - Total: 45 frontend tests passing (2 test files).
- Added coverage reporting: `vitest.config.ts` with v8 provider, thresholds at 20/15/15/20 (statements/branches/functions/lines).
- Added `test:coverage` script to `package.json`.
- Fixed CI: bumped Node from 20 to 22 in `ci.yml` and `workers-deploy.yml` (jsdom@30 requires Node 22+). Added `engines` to `package.json`.
- Updated AGENTS.md: added Required tracking files table with 13 docs and update cadence.
- All Go vet, Go tests, frontend test:coverage passing.
- CORS on media.linearvoid.tv remains the only streaming pipeline gap (Cloudflare R2 config, not code).

## 2026-08-10 (continued 4)

- Verified all test suites pass after streaming pipeline implementation:
  - Playout: `go mod tidy` (added fsnotify v1.8.0) → all tests pass
  - Chat: 7 packages pass (go vet + go test)
  - Frontend: 45 tests pass (2 files), test:coverage meets thresholds (22.92% / 18.44% / 19.01% / 25.42%)
- Edge cache validation script `scripts/validate-edge-cache.sh` exists.
- Remaining: CORS on media.linearvoid.tv (Cloudflare R2 bucket CORS config, not a code change). Phases 5-7 are manual/infrastructure only (Cloudflare dashboard, Hetzner provisioning, live benchmarks).

## 2026-08-11

- Consolidated all development tasks from `DEVELOPMENT.md` and `.github/instructions/TODO.md` into `.github/instructions/BACKLOG.md` as the single source of truth. `DEVELOPMENT.md` now obsolete and removed. `TODO.md` references `BACKLOG.md` and `DONE.md`.
- Fixed R2 endpoint validation in playout service (`backend/playout/src/playlist.go`):
  - Added `isR2Configured()` checking all four required env vars (`R2_ENDPOINT`, `R2_ACCESS_KEY_ID`, `R2_SECRET_ACCESS_KEY`, `R2_BUCKET_NAME`).
  - `initR2Client()` now returns a clear error if R2 is not configured instead of creating an S3 client with an empty endpoint.
  - Segment watcher (`watchAndPublishSegments`), `syncLatestToR2()`, and `publishManifest()` are now guarded by `r2Client != nil` — no error spam when R2 is unavailable.
  - Playout gracefully degrades: local HLS via nginx works without R2.
- Updated `.gitea/workflows/cd.yml` "Prepare environment" step to include all R2* variables sourced from Gitea secrets (`R2_ACCESS_KEY_ID`, `R2_SECRET_ACCESS_KEY`, `R2_BUCKET_NAME`, `R2_ENDPOINT`, `R2_ADMIN*\*`), Cloudflare variables (`CLOUDFLARE_API_TOKEN`, `CLOUDFLARE_ACCOUNT_ID`), and frontend public URL variables.
- Updated `docs/DEPLOYMENT.md` "Environment File" section: replaced vague "Cloudflare/R2 variables" bullet with explicit variable lists and added a new "Gitea Secrets" table documenting all 15 required secrets.
- All playout tests pass (20 tests), `go vet` clean.

## 2026-08-13

- Completed all remaining enhancement tasks (only manual/infrastructure phases left):
  - **Playout**: automatic media normalization on upload (`media.NormalizeVideo`, `*.normalized.mp4`, statuses `ready`/`needs_normalization`/`normalizing`/`normalization_failed`, `POST /api/admin/media/{id}/normalize`); R2 segment retention (`cleanupOldRuns`, keep last `R2_KEEP_RUNS` runs, `R2_MAX_RETENTION_BYTES` cap); seamless looping (playlist SHA-256 fingerprint + `-start_number`); single `start-playout.sh` entrypoint + `-validate` mode; loop/cleanup metrics.
  - **Application configuration**: thread-safe JSON runtime config store (`backend/chat/src/config/runtime.go`) shared with playout via `/app/playlist/config.json`; `GET/PUT /api/admin/config`; admin UI config form.
  - **Matomo analytics**: public `GET /api/matomo` + viewer pageview snippet in `index.html`.
  - **Chat user accounts**: register/login/logout/session endpoints, bcrypt hashing, `linearvoid_user_session` cookie, viewer sign-in UI.
  - **DM history & search**: `GET /api/dm/history` (paginated) + `GET /api/dm/search` + DM search UI.
  - **Chat UI**: typing indicator, user-list toggle, media status badges, "Add to playlist" gated to `ready`.
  - **Prometheus TSDB retention**: `--storage.tsdb.retention.time=30d` + `--storage.tsdb.retention.size=15GB` in both compose files; alert rule count corrected 8 → 10.
- Updated `docs/architecture/README.md` to reflect all `[Unreleased]` changes (normalization, looping, retention, runtime config, Matomo, user accounts, DM history/search, new API endpoints, module tables).
- Released **1.0.13**: promoted changelog, bumped version, tagged and pushed.
- Extended frontend test coverage: added `counter.test.ts`, reconnecting/typing/auth/websocket tests, admin failure + status-badge tests → 56 tests; statements 19.89% → 28.58% (CI coverage threshold passing).
- Fixed Go test data race in `TestHandleAdminMediaByID_Normalize` (waits for async normalization goroutine) + shared in-memory SQLite via `SetMaxOpenConns(1)`.
- Fixed viewer crash: `index.html` was missing the `#messages` chat container (stray `</section>`); restored structure + made `addMessage` defensive.
- Set `Doto` as the body font (`--font-body`) and added Google Fonts links to admin + login pages.
- Files changed: across `backend/chat`, `backend/playout`, `frontend`, `docs`, `docker-compose*.yml`, `.github/instructions/*`.