v2.0.22 — HA Docker bridge discovery fallback
Fix: Home Assistant integration discovers devices in Docker bridge mode
Many HA installs run inside a Docker container with bridge networking. SSDP multicast (239.255.255.250:1900) does not traverse the bridge, so both HA's own discovery framework and Dibby Wemo's internal M-SEARCH returned zero devices — users were stuck with a "No devices found" dialog.
What changed (custom_components/dibby_wemo/wemo_client.py)
- New
_local_subnet_base()detects the container's outbound interface IP and derives the local/24base - New
_probe_wemo_ip_sync(host)does a unicastGET /setup.xmlagainst every Wemo port candidate (49152–49156) with a short timeout, accepting only responses containing "Belkin" - New
_unicast_subnet_scan_sync()sweeps the full/24with a 32–64 worker thread pool — finishes in ~5 seconds discover_devices()runs SSDP first as before; if SSDP returns nothing, automatically falls back to the unicast scan
Host-network and bare-metal HA installs still use the SSDP path first and are unaffected.
Upgrade
- Home Assistant (HACS): Dibby Wemo → ⋮ → Redownload → restart HA → Add Integration
- Homebridge:
npm install -g homebridge-dibby-wemo@2.0.22 - Node-RED:
npm install -g node-red-contrib-dibby-wemo@2.0.22 - Desktop app: no functional change in this release (Windows installer assets from v2.0.21 remain current)