Merge branch 'master' into fix/update-actions-node24

The flat 'deploy' concurrency group caused ALL PRs to share one queue,
so pushing to any PR would cancel CI runs on other PRs.

Changed to deploy-${{ github.event.pull_request.number || github.ref }}
so each PR gets its own concurrency group while re-pushes to the same
PR still cancel the previous run.

.github/workflows/deploy.yml

@@ -17,7 +17,7 @@ on:
       - 'docs/**'
 
 concurrency:
-  group: deploy
+  group: deploy-${{ github.event.pull_request.number || github.ref }}
   cancel-in-progress: true
 
 env:
@@ -41,10 +41,10 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v5
 
       - name: Set up Go 1.22
-        uses: actions/setup-go@v5
+        uses: actions/setup-go@v6
         with:
           go-version: '1.22'
           cache-dependency-path: |
@@ -124,7 +124,7 @@ jobs:
 
       - name: Upload Go coverage badges
         if: always()
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v5
         with:
           name: go-badges
           path: .badges/go-*.json
@@ -139,12 +139,12 @@ jobs:
     runs-on: self-hosted
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v5
         with:
           fetch-depth: 2
 
       - name: Set up Node.js 22
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v5
         with:
           node-version: '22'
 
@@ -265,7 +265,7 @@ jobs:
 
       - name: Upload Node.js test badges
         if: always()
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v5
         with:
           name: node-badges
           path: .badges/
@@ -282,10 +282,10 @@ jobs:
     runs-on: self-hosted
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v5
 
       - name: Set up Node.js 22
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v5
         with:
           node-version: '22'
 
@@ -307,7 +307,7 @@ jobs:
     runs-on: self-hosted
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v5
 
       - name: Start staging on port 82
         run: |
@@ -352,18 +352,18 @@ jobs:
     runs-on: self-hosted
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v5
 
       - name: Download Go coverage badges
         continue-on-error: true
-        uses: actions/download-artifact@v4
+        uses: actions/download-artifact@v5
         with:
           name: go-badges
           path: .badges/
 
       - name: Download Node.js test badges
         continue-on-error: true
-        uses: actions/download-artifact@v4
+        uses: actions/download-artifact@v5
         with:
           name: node-badges
           path: .badges/

RELEASE-v3.1.0.md

@@ -0,0 +1,144 @@
+# v3.1.0 — Now It's CoreScope
+
+MeshCore Analyzer has a new name: **CoreScope**. Same mesh analysis you rely on, sharper identity, and a boatload of fixes and performance wins since v3.0.0.
+
+48 commits, 30+ issues closed. Here's what changed.
+
+---
+
+## 🏷️ Renamed to CoreScope
+
+The project is now **CoreScope** — frontend, backend, Docker images, manage.sh, docs, CI — everything has been updated. The URL, the API, the database, and your config all stay the same. Just a better name for the tool the community built.
+
+---
+
+## ⚡ Performance
+
+| What | Before | After |
+|------|--------|-------|
+| Subpath analytics | 900 ms | **5 ms** (precomputed at ingest) |
+| Distance analytics | 1.2 s | **15 ms** (precomputed at ingest) |
+| Packet ingest (prepend) | O(n) slice copy | **O(1) append** |
+| Go runtime stats | GC stop-the-world on every call | **cached ReadMemStats** |
+| All analytics endpoints | computed per-request | **TTL-cached** |
+
+The in-memory store now precomputes subpaths and distance data as packets arrive, eliminating expensive full-table scans on the analytics endpoints. The O(n) slice prepend on every ingest — the single hottest line in the server — is gone. `ReadMemStats` calls are cached to prevent GC pause spikes under load.
+
+---
+
+## 🆕 New Features
+
+### Telemetry Decode
+Sensor nodes now report **battery voltage** and **temperature** parsed from advert payloads. Telemetry is gated on the sensor flag — only real sensors emit data, and 0°C is no longer falsely reported. Safe migration with `PRAGMA` column checks.
+
+### Channel Decryption for Custom Channels
+The `hashChannels` config now works in the Go ingestor. Key derivation has been ported from Node.js with full AES-128-ECB support and garbage text detection — wrong keys silently fail instead of producing garbled output.
+
+### Node Pruning
+Stale nodes are automatically moved to an `inactive_nodes` table after the configurable retention window. Pruning runs hourly. Your active node list stays clean. (#202)
+
+### Duplicate Node Name Badges
+Nodes with the same display name but different public keys are flagged with a badge so you can spot collisions instantly.
+
+### Sortable Channels Table
+Channel columns are now sortable with click-to-sort headers. Sort preferences persist in `localStorage` across sessions. (#167)
+
+### Go Runtime Metrics
+The performance page exposes goroutine count, heap allocation, GC pause percentiles, and memory breakdown when connected to a Go backend.
+
+---
+
+## 🐛 Bug Fixes
+
+- **Channel decryption regression** (#176) — full AES-128-ECB in Go, garbage text detection, hashChannels key derivation ported correctly (#218)
+- **Packets page not live-updating** (#172) — WebSocket broadcast now includes the nested packet object and timestamp fields the frontend expects; multiple fixes across broadcast and render paths
+- **Node detail page crashes** (#190) — `Number()` casts and `Array.isArray` guards prevent rendering errors on unexpected data shapes
+- **Observation count staleness** (#174) — trace page and packet detail now show correct observation counts
+- **Phantom node cleanup** (#133) — `autoLearnHopNodes` no longer creates fake nodes from 1-byte repeater IDs
+- **Advert count inflation** (#200) — counts unique transmissions, not total observations (8 observers × 1 advert = 1, not 8)
+- **SQLite BUSY contention** (#214) — `MaxOpenConns(1)` + `MaxIdleConns(1)` serializes writes; load-tested under concurrent ingest
+- **Decoder bounds check** (#183) — corrupt/malformed packets no longer crash the decoder with buffer overruns
+- **noise_floor / battery_mv type mismatches** — consistent `float64` scanning handles SQLite REAL values correctly
+- **packetsLastHour always zero** (#182) — early `break` in observer loop prevented counting
+- **Channels stale messages** (#171) — latest message sorted by observation timestamp, not first-seen
+- **pprof port conflict** — non-fatal bind with separate ports prevents Go server crash on startup
+
+---
+
+## ♿ Accessibility & 📱 Mobile
+
+### WCAG AA Compliance (10 fixes)
+- Search results keyboard-accessible with `tabindex`, `role`, and arrow-key navigation (#208)
+- 40+ table headers given `scope` attributes (#211)
+- 9 Chart.js canvases given accessible names (#210)
+- Form inputs in customizer/filters paired with labels (#212)
+
+### Mobile Responsive
+- **Live page**: bottom-sheet panel instead of full-screen overlay (#203)
+- **Perf page**: responsive layout with stacked cards (#204)
+- **Nodes table**: column hiding at narrow viewports (#205)
+- **Analytics/Compare**: horizontal scroll wrappers (#206)
+- **VCR bar**: 44px minimum touch targets (#207)
+
+---
+
+## 🏗️ Infrastructure
+
+### manage.sh Refactored (#230)
+`manage.sh` is now a thin wrapper around `docker compose` — no custom container management, no divergent logic. It reads `.env` for data paths, matching how `docker-compose.yml` works. One source of truth.
+
+### .env Support
+Data directory, ports, and image tags are configured via `.env`. Both `docker compose` and `manage.sh` read the same file.
+
+### Branch Protection & CI on PRs
+- Branch protection enabled on `master` — CI must pass, PRs required
+- CI now triggers on `pull_request`, not just `push` — catch failures before merge (#199)
+
+### Protobuf API Contract
+10 `.proto` files, 33 golden fixtures, CI validation on every push. API shape drift is caught automatically.
+
+### pprof Profiling
+Controlled by `ENABLE_PPROF` env var. When enabled, exposes Go profiling endpoints on separate ports — zero overhead when off.
+
+### Test Coverage
+- Go backend: **92%+** coverage
+- **49 Playwright E2E tests**
+- Both tracks gate deploy in CI
+
+---
+
+## 📦 Upgrading
+
+```bash
+git pull
+./manage.sh stop
+./manage.sh setup
+```
+
+That's it. Your existing `config.json` and database work as-is. The rename is cosmetic — no schema changes, no API changes, no config changes.
+
+### Verify
+
+```bash
+curl -s http://localhost/api/health | grep engine
+# "engine": "go"
+```
+
+---
+
+## ⚠️ Breaking Changes
+
+**None.** All API endpoints, WebSocket messages, and config options are backwards-compatible. The rename affects branding only — Docker image names, page titles, and documentation.
+
+---
+
+## 🙏 Thank You
+
+- **efiten** — PR #222 performance fix (O(n) slice prepend elimination)
+- **jade-on-mesh**, **lincomatic**, **LitBomb**, **mibzzer15** — ongoing testing, feedback, and issue reports
+
+And to everyone running CoreScope on their mesh networks — your real-world data drives every fix and feature in this release. 48 commits since v3.0.0, and every one of them came from something the community found, reported, or requested.
+
+---
+
+*Previous release: [v3.0.0](RELEASE-v3.0.0.md)*

cmd/server/routes.go

@@ -33,6 +33,11 @@ type Server struct {
 	memStatsMu   sync.Mutex
 	memStatsCache runtime.MemStats
 	memStatsCachedAt time.Time
+
+	// Cached /api/stats response — recomputed at most once every 10s
+	statsMu      sync.Mutex
+	statsCache   *StatsResponse
+	statsCachedAt time.Time
 }
 
 // PerfStats tracks request performance.
@@ -380,6 +385,17 @@ func (s *Server) handleHealth(w http.ResponseWriter, r *http.Request) {
 }
 
 func (s *Server) handleStats(w http.ResponseWriter, r *http.Request) {
+	const statsTTL = 10 * time.Second
+
+	s.statsMu.Lock()
+	if s.statsCache != nil && time.Since(s.statsCachedAt) < statsTTL {
+		cached := s.statsCache
+		s.statsMu.Unlock()
+		writeJSON(w, cached)
+		return
+	}
+	s.statsMu.Unlock()
+
 	var stats *Stats
 	var err error
 	if s.store != nil {
@@ -392,7 +408,7 @@ func (s *Server) handleStats(w http.ResponseWriter, r *http.Request) {
 		return
 	}
 	counts := s.db.GetRoleCounts()
-	writeJSON(w, StatsResponse{
+	resp := &StatsResponse{
 		TotalPackets:       stats.TotalPackets,
 		TotalTransmissions: &stats.TotalTransmissions,
 		TotalObservations:  stats.TotalObservations,
@@ -411,7 +427,14 @@ func (s *Server) handleStats(w http.ResponseWriter, r *http.Request) {
 			Companions: counts["companions"],
 			Sensors:    counts["sensors"],
 		},
-	})
+	}
+
+	s.statsMu.Lock()
+	s.statsCache = resp
+	s.statsCachedAt = time.Now()
+	s.statsMu.Unlock()
+
+	writeJSON(w, resp)
 }
 
 func (s *Server) handlePerf(w http.ResponseWriter, r *http.Request) {

cmd/server/store.go

@@ -98,6 +98,11 @@ type PacketStore struct {
 	// computed during Load() and incrementally updated on ingest.
 	distHops  []distHopRecord
 	distPaths []distPathRecord
+
+	// Cached GetNodeHashSizeInfo result — recomputed at most once every 15s
+	hashSizeInfoMu    sync.Mutex
+	hashSizeInfoCache map[string]*hashSizeNodeInfo
+	hashSizeInfoAt    time.Time
 }
 
 // Precomputed distance records for fast analytics aggregation.
@@ -3722,8 +3727,26 @@ type hashSizeNodeInfo struct {
 	Inconsistent bool
 }
 
-// GetNodeHashSizeInfo scans advert packets to compute per-node hash size data.
+// GetNodeHashSizeInfo returns cached per-node hash size data, recomputing at most every 15s.
 func (s *PacketStore) GetNodeHashSizeInfo() map[string]*hashSizeNodeInfo {
+	const ttl = 15 * time.Second
+	s.hashSizeInfoMu.Lock()
+	if s.hashSizeInfoCache != nil && time.Since(s.hashSizeInfoAt) < ttl {
+		cached := s.hashSizeInfoCache
+		s.hashSizeInfoMu.Unlock()
+		return cached
+	}
+	s.hashSizeInfoMu.Unlock()
+	result := s.computeNodeHashSizeInfo()
+	s.hashSizeInfoMu.Lock()
+	s.hashSizeInfoCache = result
+	s.hashSizeInfoAt = time.Now()
+	s.hashSizeInfoMu.Unlock()
+	return result
+}
+
+// computeNodeHashSizeInfo scans advert packets to compute per-node hash size data.
+func (s *PacketStore) computeNodeHashSizeInfo() map[string]*hashSizeNodeInfo {
 	s.mu.RLock()
 	defer s.mu.RUnlock()
 

docker-compose.yml

@@ -5,9 +5,12 @@
 
 services:
   prod:
+    build: .
     image: corescope:latest
     container_name: corescope-prod
     restart: unless-stopped
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
     ports:
       - "${PROD_HTTP_PORT:-80}:${PROD_HTTP_PORT:-80}"
       - "${PROD_HTTPS_PORT:-443}:${PROD_HTTPS_PORT:-443}"
@@ -26,9 +29,12 @@ services:
       retries: 3
 
   staging:
+    build: .
     image: corescope:latest
     container_name: corescope-staging
     restart: unless-stopped
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
     ports:
       - "${STAGING_HTTP_PORT:-81}:${STAGING_HTTP_PORT:-81}"
       - "${STAGING_MQTT_PORT:-1884}:1883"
@@ -57,6 +63,8 @@ services:
     image: corescope-go:latest
     container_name: corescope-staging-go
     restart: unless-stopped
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
     ports:
       - "${STAGING_GO_HTTP_PORT:-82}:80"
       - "${STAGING_GO_MQTT_PORT:-1885}:1883"

docs/rename-migration.md

@@ -0,0 +1,101 @@
+# CoreScope Migration Guide
+
+MeshCore Analyzer has been renamed to **CoreScope**. This document covers what you need to update.
+
+## What Changed
+
+- **Repository name**: `meshcore-analyzer` → `corescope`
+- **Docker image name**: `meshcore-analyzer:latest` → `corescope:latest`
+- **Docker container prefixes**: `meshcore-*` → `corescope-*`
+- **Default site name**: "MeshCore Analyzer" → "CoreScope"
+
+## What Did NOT Change
+
+- **Data directories** — `~/meshcore-data/` stays as-is
+- **Database filename** — `meshcore.db` is unchanged
+- **MQTT topics** — `meshcore/#` topics are protocol-level and unchanged
+- **Browser state** — Favorites, localStorage keys, and settings are preserved
+- **Config file format** — `config.json` structure is the same
+
+---
+
+## 1. Git Remote Update
+
+Update your local clone to point to the new repository URL:
+
+```bash
+git remote set-url origin https://github.com/Kpa-clawbot/corescope.git
+git pull
+```
+
+## 2. Docker (manage.sh) Users
+
+Rebuild with the new image name:
+
+```bash
+./manage.sh stop
+git pull
+./manage.sh setup
+```
+
+The new image is `corescope:latest`. You can clean up the old image:
+
+```bash
+docker rmi meshcore-analyzer:latest
+```
+
+## 3. Docker Compose Users
+
+Rebuild containers with the new names:
+
+```bash
+docker compose down
+git pull
+docker compose build
+docker compose up -d
+```
+
+Container names change from `meshcore-*` to `corescope-*`. Old containers are removed by `docker compose down`.
+
+## 4. Data Directories
+
+**No action required.** The data directory `~/meshcore-data/` and database file `meshcore.db` are unchanged. Your existing data carries over automatically.
+
+## 5. Config
+
+If you customized `branding.siteName` in your `config.json`, update it to your preferred name. Otherwise the new default "CoreScope" applies automatically.
+
+No other config keys changed.
+
+## 6. MQTT
+
+**No action required.** MQTT topics (`meshcore/#`) are protocol-level and are not affected by the rename.
+
+## 7. Browser
+
+**No action required.** Bookmarks/favorites will continue to work at the same host and port. localStorage keys are unchanged, so your settings and preferences are preserved.
+
+## 8. CI/CD
+
+If you have custom CI/CD pipelines that reference:
+
+- The old repository URL (`meshcore-analyzer`)
+- The old Docker image name (`meshcore-analyzer:latest`)
+- Old container names (`meshcore-*`)
+
+Update those references to use the new names.
+
+---
+
+## Summary Checklist
+
+| Item | Action Required? | What to Do |
+|------|-----------------|------------|
+| Git remote | ✅ Yes | `git remote set-url origin …corescope.git` |
+| Docker image | ✅ Yes | Rebuild; optionally `docker rmi` old image |
+| Docker Compose | ✅ Yes | `docker compose down && build && up` |
+| Data directories | ❌ No | Unchanged |
+| Config | ⚠️ Maybe | Only if you customized `branding.siteName` |
+| MQTT | ❌ No | Topics unchanged |
+| Browser | ❌ No | Settings preserved |
+| CI/CD | ⚠️ Maybe | Update if referencing old repo/image names |