docs: CHANGELOG v0.96.0 + REPORT (TASK-D0) + CONTEXT + README design-system section

REPORT.md

@@ -1,62 +1,133 @@
-# REPORT — Impl-2b: enrollment wizards → `/disks/candidates` + end-to-end raw enrollment (controller v0.95.0)
+# REPORT — TASK-D0: Controller UI re-skin foundation (design system v2) → v0.96.0
 
-**Date:** 2026-07-01 · **Repo:** `felhom-controller` · **Class:** Risky/supervised (first live raw
-enrollment runs the Impl-1 guarded mkfs inside a real enroll). Companion agent work: v0.55.0→**v0.58.0**.
+**Date:** 2026-07-02 · **Deployed:** guest 9201 (`gitea.dooplex.hu/admin/felhom-controller:0.96.0`, healthy)
 
-Final piece of the drive-enrollment series: both enrollment wizards now source candidates from the
-agent's raw-device scan, so a brand-new (non-PVE-storage) drive is discoverable + enrollable end-to-end.
+## 1. Baselines
 
-## 1. Baseline → target
+- felhom-controller `main` @ `a390e6be293ec70e56d81f90852bc8c2278981e7` (v0.95.0) — **exactly matched
+  the spec; main had NOT moved past a390e6b** at session start.
+- felhom.eu `main` @ `8c8196e82b69592bd5c0f2166de0987134921590` — matched.
 
-felhom-controller `6ce61e8` (v0.94.0) → **v0.95.0**. Agent v0.55.0 → **v0.58.0** (separate repo — the
-raw-drive path needed agent fixes the spec's "N/A agent" premise didn't anticipate; see agent REPORT).
+## 2. Commits pushed to `main`
 
-## 2. Files changed + commits
+| Commit | Content |
+|---|---|
+| `b073cc4` | Part 1: vendored fonts + Lucide sprite + setup CSS fix + tests |
+| `5dc277f` | Part 2: v2 tokens/components, funcmap remap, layout + dashboard + stacks |
+| `f100cef` | Part 3: deploy + backups + monitoring; template-wide var() rename |
+| `7df061c` | Part 4: settings visual pass + remaining pages; grep gate zero |
+| `4906524` | fix(backups): off-box `LastRun` RFC3339-string 500 (pre-existing, found in click-through) |
+| felhom.eu `a336fe3` | docs: `documentation/design/design-system.md` (canonical v2 reference) |
 
-- `internal/agentapi/client.go` — `ListCandidates` + `CandidatesResult`/`DiskCandidate` (mirror the
-  agent's `/disks/candidates`).
-- `internal/web/agent_disk_handlers.go` — `GET /api/disks/candidates` passthrough proxy.
-- `templates/storage_init.html` / `storage_attach.html` — fetch `/api/disks/candidates`; render
-  `initialize` / `attach`; dropped the client-side "already-managed" filter.
-- `internal/web/storage_handlers.go` — **`resolveEnrollUUID`**: a raw candidate isn't in `/disks`, so
-  resolve its fs-UUID from `/disks/candidates` (the enroll blocker the "reuse unchanged" premise missed).
-  Both `runStorageInit` + `runStorageAttach` use it.
-- Tests: `agentapi` `TestListCandidates`(+`_Error`), `web` `TestRunStorageAttach_RawCandidate`(+red-proof).
-- `CHANGELOG.md`, `README.md`. Commits **`feab92c`** (wiring) + **`bea05ea`** (resolveEnrollUUID) on `main`.
+## 3. Files created / modified
 
-## 3. Green gate
+**Created:** `controller/internal/web/static/fonts/{pjs-latin.woff2 27.3K, pjs-latin-ext.woff2 21.7K,
+jbm-latin.woff2 31.4K, jbm-latin-ext.woff2 11.6K}` (4 variable-font files, ~92 KB total — Google now
+serves variable fonts, so one file covers all weights per subset instead of the estimated ~14 files);
+`templates/icons.html` (30-icon Lucide sprite); `internal/web/fonts_test.go`,
+`internal/web/funcmap_test.go`, `internal/setup/css_test.go`;
+`felhom.eu/documentation/design/design-system.md`.
 
-`go build ./...`, `go vet ./...`, `go test ./...` — clean. Image `felhom-controller:0.95.0` built on 180,
-deployed to guest 9201 (golden/bootstrap), `Up (healthy)`.
+**Modified:** `internal/web/{embed.go (fonts FS + StyleCSS()), server.go (/static/fonts/ route),
+funcmap.go (stateColor/usageColor/tempColor remap + timeAgoStr)}`;
+`internal/setup/handlers.go` (handleCSS → embedded CSS; minimalCSS retokened);
+`templates/style.css` (rewritten in place); all 19 web templates; setup templates (v2 token
+fallbacks); CHANGELOG.md, CONTEXT.md, controller/README.md, REPORT.md.
 
-## 4. End-to-end live validation (real UI, claude-in-chrome, felhom.demo-felhom.eu)
+## 4. Tests + red-proofs
 
-| Step | Result |
-|------|--------|
-| **Discovery (init wizard)** | ✅ `/dev/sdd` (SD/MMC/MS PRO, 59 GB, ext4) offered as the **sole** candidate; sda/sdb/sdc NOT shown. |
-| **Format via Impl-1 guard** | ✅ enroll ran the guarded mkfs on /dev/sdd (`formatted device … fstype=ext4`) — the first real customer-flow guarded format. |
-| **Discovery (attach wizard)** | ✅ the ext4 /dev/sdd offered under `attach`. |
-| **Enroll end-to-end** | ✅ wizard attach → `resolveEnrollUUID` → mount (`/mnt/teszt_enroll`) → bind (`/mnt/felhom-drives/teszt_enroll`) → intent `enrolled` → guest-bind recorded → registered. UI shows **"Aktív"** (USB), alongside felhom-usb + felhom-flash, all Active — no false detach. |
-| **Health-tracked** | ✅ the watchdog (RegistryKnownTargets) detected + self-healed a drop of the drive during testing. |
+- `go build ./... && go vet ./... && go test ./...` **green after every commit**; final full run exit 0,
+  18 packages ok. Test functions 256 → **263** (+7).
+- **funcmap red-proof:** truth-table tests run against the stashed pre-change funcmap **failed exactly
+  as required** — `stateColor("stopped") = "red", want "neutral"` and `usageColor(0) = "green", want
+  "nominal"` (plus tempColor). Restored; all pass.
+- **setup-CSS red-proof:** Scenario E test against the pre-fix `handleCSS` **failed** with
+  `body length = 3592, want > 10000` (minimalCSS served). With the fix: passes, and asserts the
+  `--bg-0: #0A1220` literal + byte-equality with `web.StyleCSS()`.
+- Template-parse tests (`TestTemplatesParse`, `TestTemplatesParseWithFuncmap`) already existed and
+  cover Group G; they pass with the new funcmap + icons partial.
 
-**Note on the init slow-format path:** a *slow* device (the 60 GB SD card, ~25 s mkfs) exceeds the
-controller→agent request timeout; `runStorageInit` calls `FormatDisk` synchronously and does NOT poll the
-agent's detached-format status (`/disks/format/status`), so on a slow device it aborts after the format
-detaches and the mount/bind back-half doesn't run. **Pre-existing** (F20-BUG3-adjacent; not the Impl-2b
-wiring). The end-to-end above was completed via the **attach** path (no mkfs) — which fully exercises the
-candidate→enroll→active pipeline. Follow-up: make `runStorageInit` poll the detached-format status.
+## 5. Grep gate (Test Group F) — counts before → after
 
-## 5. NOT done / follow-ups
+All 34 patterns measured identically over `controller/internal/{web,setup}` (*.go, *.html, *.css):
 
-- **Init slow-format polling** (above) — `runStorageInit` should poll `/disks/format/status` so a
-  slow-device format completes the enroll in one flow.
-- **Impl-3** — the shared-box operator gate for self-serve format (raw-passthrough blind spot). The demo
-  is a dedicated box, so self-serve format is appropriate there.
-- The throwaway `/dev/sdd` (SD card) is left **enrolled + Aktív** as living proof; the operator can eject
-  it via the UI (Leválasztás) and unplug when done.
+| Pattern | Before | After | Pattern | Before | After |
+|---|---|---|---|---|---|
+| fonts.googleapis.com | 1 | 0 | system-bar-green | 4 | 0 |
+| border-radius: 999px | 5 | 0 | system-bar-yellow | 4 | 0 |
+| box-shadow | 31 | 0 | system-bar-red | 4 | 0 |
+| #0d1117 | 3 | 0 | state-text-green | 9 | 0 |
+| #161b22 | 1 | 0 | state-text-red | 7 | 0 |
+| #1c2128 | 4 | 0 | state-text-yellow | 3 | 0 |
+| #30363d | 6 | 0 | state-text-orange | 1 | 0 |
+| #238636 | 5 | 0 | ⚙ / ✏ / ✕ / ⏳ | 1/2/1/2 | 0 |
+| #da3633 | 4 | 0 | ✅ / ❌ | 14/5 | 0 |
+| #d29922 | 3 | 0 | ⚠ | 18 | 0 |
+| #db6d28 | 2 | 0 | ℹ️ | 3 | 0 |
 
-## 6. Observations
+**Total: 143 → 0.** (🔧🔒📦🔗🔴🟡💾 were 0 in both.) Beyond the gate, related glyphs (✓✗⟳⏸🔄🔍) were
+also swept, and pre-existing never-defined `var(--success/--danger/--primary/--warning/--accent)`
+references in templates were mapped to v2 tokens.
 
-- The spec's premise ("just wire the wizards; the enroll flow already works") was optimistic: the entire
-  enroll + tracking machinery assumed a PVE-storage drive (`Observe()`), so making raw drives first-class
-  required `resolveEnrollUUID` here + a chain of agent fixes (see agent REPORT). All shipped + validated.
+## 6. Deployment + live validation (claude-in-chrome)
+
+```
+gitea.dooplex.hu/admin/felhom-controller:0.96.0 Up 8 seconds (healthy)
+```
+Startup logs clean (hub report pushed, 5 health probes ok). Sidebar shows 0.96.0.
+
+Chrome checks performed on `https://felhom.demo-felhom.eu` (screenshots taken per page; ő-glyph zoom
+saved to disk):
+
+1. **Full click-through:** dashboard, Alkalmazások (filters render, grid kept), app-info (calcom),
+   deploy page for actualbudget (**opened only — not deployed**; two-segment memory meter renders:
+   solid-blue committed + 35%-opacity new), backups, monitoring (charts render in the new palette —
+   cpu sky-blue, memory violet; JS-generated host-storage meters blue), logs (mono, blue Élő dot),
+   settings (all sections; drive cards with pencil icons, run/USB tags, crit-outline danger buttons).
+2. **Type-to-confirm dialog:** opened the drive-list eject („Meghajtó leválasztása", type-to-confirm
+   input, disabled crit-outline Megerősítés) and **cancelled** it; verified afterward the drive is
+   still mounted and zero disconnect/eject calls hit the API.
+3. **Scenario D:** zero requests to fonts.googleapis.com/gstatic; `/static/fonts/*.woff2` → 200
+   (pjs-latin, pjs-latin-ext, jbm-latin loaded); `document.fonts.check("1em 'Plus Jakarta Sans'")`
+   and JBM both `true`; „…rendszermentéstől" ő renders in the brand face (zoom screenshot saved).
+4. **Scenario A (DOM-asserted via JS):** 0 elements with green (35,134,54) backgrounds, 0 elements
+   with box-shadow, 0 border-radius values other than 0/2px/50%; „Leállítva 0" stat renders muted;
+   running rows blue tag + blue 2px edge.
+5. **Login/logout:** NOT exercisable live — the demo node runs with auth disabled (`/login` is 404,
+   no logout link; the known pre-existing unauth-demo state). Login page covered by template-parse
+   tests; the two-tone H1 was added (`Otthoni <span>vezérlőpult</span>`).
+
+## 7. Bug found & fixed during validation
+
+`GET /backups` returned **500** on first visit: `backups.html` passed `OffboxTarget.LastRun` (an
+RFC3339 **string** persisted in settings.json) to `timeAgo` (expects `time.Time`). Pre-existing since
+v0.93.0 (NAS Part B) — it fires on any node where an off-box backup has ever run; the demo node
+qualified. Fixed in `4906524` with a `timeAgoStr` funcmap helper (parses RFC3339, falls back to the
+raw string); 0.96.0 image rebuilt + redeployed, /backups now renders („Rendben … tegnap").
+
+## 8. NOT yet live-validated
+
+- **Setup wizard visual rendering** — cannot be exercised without re-provisioning; covered by the
+  Scenario E unit test (embedded v2 CSS served). Await the next supervised reinstall.
+- **warn/crit meter states on real hardware** — the demo node is healthy (all meters nominal);
+  covered at unit level (usageColor thresholds) and by the meter-flag template logic.
+- **`.meter-flag` strings live** — same reason (no disk over 70%).
+
+## 9. Observations (structural, NOT acted on — appearance-only scope)
+
+1. **settings.html remains monolithic** (~1450 lines, duplicated drive views: template-rendered
+   registered-storage cards + JS-rendered agent drive list) — D1 scope. ~50 inline `style="…"`
+   attributes remain there and across pages (reduced only where a v2 class covered them).
+2. **Native `confirm()` dialogs** in settings storage actions (`storageDisconnect`, stale-data
+   delete) vs the custom type-to-confirm overlay elsewhere — inconsistent UX and they block
+   browser automation; suggest migrating to the overlay in D1.
+3. Old `stat-running`/`stat-stopped` class names still appear in backups.html markup (now no-op —
+   CSS deleted); harmless, cleanable in D1.
+4. `badge`/`meta-badge`/`db-type-badge`/`validation-badge` markup survives on secondary pages,
+   restyled flat by the v2 CSS (2px, dim tokens) — full tag/metarow markup migration on those pages
+   can complete in D1–D2.
+5. The catchall + setup pages are standalone (no sprite); they use inline SVGs / plain text. If
+   icon needs grow there, consider embedding the sprite per-page.
+6. Baseline grep showed 🔧 at 0 despite layout.html's `🔧 Debug` — some emoji encodings differ from
+   the gate's literals; the gate was still measured identically before/after, and independent broader
+   sweeps confirmed no emoji remain.