Normalize loadAvg to [0,0,0] fallback in both the native host handler
and the dashboard renderer to prevent TypeError when bare-node-os returns
a non-array value for loadavg().
- Add getProcessStats message handler to native host returning process
memory (rss, heap used/total, external), PID, uptime, OS memory,
CPU load averages, core count, platform/arch, and tunnel state counts
- Add stats-tracker.js to background SW tracking cumulative connections,
reconnect count, last disconnect time, and per-minute connection rate
ring buffer (30 buckets) for sparkline display
- Include statsHistory and activeConnections array in all getState responses
- Add Statistics dashboard page with six sections: Native Host Process,
System Resources, Tunnel Health, Proxy Traffic, Peer Connections
(with SVG sparkline), and Extension & Proxy Info
- Preserve checkbox selection across table re-renders on all bulk-action
tables (virtual hosts, servers, service tunnels)
- Add theme-aware custom checkbox styling for bulk selection rows using
design system tokens, with white checkmark override for light theme
Reverts the experimental feature that attempted to serve the dashboard
via a local HTTP server registered as a hardcoded `my.dash.board` virtual
host, intended to satisfy Chrome's PWA "secure origin" installability
requirement. The approach caused proxy connection errors and file-not-found
issues that were not worth resolving.
Removed: native-host/dashboard-server.js, native-host/test-chain.mjs
Reverted: startup.js, message-router.js, virtual-hosts.js, index.js,
build-distributable.js, extension/pages/virtual-hosts.js,
CI workflows (ci.yml, release.yml)
Fix "Proxy error: connection reset by peer" caused by dashboard-server.js
looking for files in native-host/ instead of extension/dashboard/. Remove
the devBasePath parameter and resolve DEV_DASHBOARD_DIR directly from
__dirname so dev and distribution mode both find the correct files.
Fix all manifest warnings (start_url/id/scope/protocol_handlers ignored)
by reverting manifest.webmanifest to relative/extension-origin URLs for
the chrome-extension:// context, and having dashboard-server.js dynamically
generate a manifest at /manifest.webmanifest with absolute
https://my.dash.board/ URLs — satisfying Chrome's same-origin requirement
for PWA installability.
Also serve extension icons at https://my.dash.board/icons/<file> with
correct bundle key mapping, and embed them in the distributable binary
via build-distributable.js.
Serve the dashboard over a local HTTPS virtual host (my.dash.board) so
Chrome treats it as a secure origin and shows the PWA install prompt.
- Add native-host/dashboard-server.js: bare-http1 static file server
backed by bare-bundle assets in distribution mode, disk fallback in dev
- Add setLocalVirtualHost() to virtual-hosts.js; type:local entries are
protected from removal and filtered out of saveState persistence
- Export setLocalVirtualHost from holesail-manager/index.js
- Start dashboard server in startup.js after proxies are ready and
register my.dash.board as a local virtual host
- Stop dashboard server in message-router.js cleanup()
- Update manifest.webmanifest: start_url, id, scope → https://my.dash.board/
- Embed all extension/dashboard/ files as bare-bundle assets in
build-distributable.js patchBundle() — no installer copy step needed
- Hide checkbox, hs:// URL, Reconnect, and Remove controls for built-in
my.dash.board row in the virtual hosts table UI
CRITICAL:
- certificate-authority.js: declare `regenerated` variable in installRootCA Windows path to prevent ReferenceError crash
HIGH:
- virtual-hosts.js/service-tunnels.js: call hs.removeAllListeners() in catch blocks to prevent stale listeners on failed Holesail instances
- https-proxy.js: destroy rawSocket in TLS error handler to prevent file descriptor exhaustion
- message-router.js (native): move setEventEmitter() to module-level init instead of re-calling on every message
- ssh-manager.js: add error handler to WS server to prevent unhandled error crashes
- init.js: store setInterval ID and clear on beforeunload to prevent interval accumulation
- logs.js: store and remove chrome.runtime.onMessage listener on beforeunload; add duplicate-call guard
- events.js: move pending++ before async sendMessage call to fix SSH/RDP-only import showing "Nothing to import"
- ssh.js: store resizeTimer on activeSshSession and clear in disconnectSsh; fix auto-reconnect race with _sshConnecting lock
- native-messaging.js: track retry timer IDs in array and cancel all on disconnect
- rdp.js: reuse single offscreen canvas per session instead of allocating per bitmap
MEDIUM:
- virtual-hosts.js/service-tunnels.js: clear existing.reconnectTimer before replacing tunnel entries
- message-router.js (native): destroy pingTunnel socket on error path; clear 15s fallback timer via finally()
- startup.js: wrap setImmediate body in try/finally to always resolve proxiesReadyPromise
- https-proxy.js: fix pre-connect upstream error handler to avoid writing raw HTTP into piped TLS stream; move HOP_BY_HOP to module-level constant
- connect-proxy.js: destroy upstreamSocket on clientSocket close; track and destroy active sockets in stop()
- ssh-manager.js: call cancelPasswordWatch on WS disconnect during password collection
- rdp-manager.js: remove dead remotePort variable; add error handlers to both WS servers
- backup-manager.js: log cleanupStaging errors and non-zero exit codes
- rdp.js: null out ws callbacks before closing in disconnectRdp; disconnect MutationObserver on beforeunload
- proxy-ca.js: prune stale entries from validationResults Map in renderValidatorTable
- refresh.js: deduplicate in-flight pings per port via Set
- messaging.js: read chrome.runtime.lastError in sendToNative callback
- servers.js (dashboard): add null check for $('serverEditId') element
LOW:
- port-allocator.js: add dedup check before pushing to tunnelPortFreeList
- servers.js (native): add error listener to server-mode Holesail instances
- virtual-hosts.js: remove dead prevReconnectDelay variable
- tab-lifecycle.js: change swarmRefCount fallback from || 1 to || 0 to prevent premature swarm destroy
- ssh.js/rdp.js: disconnect MutationObservers on beforeunload
- Change readyTimeoutMs default from 0 to 30000 in both state files
- Register Alt+Shift+H keyboard shortcut via manifest _execute_action command
- Add severity filter buttons (All/Info/Warn/Error) and Download .txt to Logs page; background logs.js now tags entries with proper level field
- Fire browser notifications on tunnelError events (notifyOnTunnelError setting)
- Add exponential-backoff auto-reconnect for virtual hosts and service tunnels (tunnelAutoReconnect setting, 5s–120s backoff)
- Add backupIntervalHours setting and scheduled auto-backup timer in message-router.js
- Add TCP connect latency badges to Virtual Hosts and Service Tunnels tables via new pingTunnel native message
- Add bytesIn/bytesOut/requests counters to https-proxy.js; expose in Overview status bar via getState
- Add Export/Import connection configs (JSON, no CA key) in Settings
- Add autoReconnect flag and exponential-backoff reconnect to SSH connection cards
- Add light theme CSS variables and theme toggle in Settings (persisted to localStorage)
- Add checkbox column and bulk Stop/Remove actions to Virtual Hosts, Service Tunnels, and Servers tables
- Add Peer Lookup UI card on Overview page using existing lookup message handler
Split host.js (435 lines) into host/{paths,logger,startup,message-router}.js.
Split holesail-manager.js (698 lines) into holesail-manager/{state,settings,
connections,port-allocator,servers,virtual-hosts,service-tunnels,index}.js.
Top-level host.js and holesail-manager.js become thin shims so index.mjs
requires no changes. Deleted dev scratch file test-cp.mjs.
Updated CI with 12 new node --check lines for all sub-modules.
Updated docs/ARCHITECTURE.md with per-file tables for host/ and
holesail-manager/ sub-modules.
No functionality changed. No new dependencies.
Raven Scott