arjenhiemstra/ithowifi Version-3.2.0-beta1

Version 3.2.0-beta1

on GitHub

Version 3.2.0-beta1

First beta of the 3.2 line. Carries forward the work from the (skipped) 3.1.5-beta line and adds active RF status polling for RF-standalone setups.

Features

• Periodic 31DA + 31D9 RF status request in RF standalone mode. When itho_rf_standalone == 1, the add-on now actively requests 31DA and 31D9 from the Itho over RF on the same interval as the existing Itho status update frequency setting. Each request is addressed from the first bi-directional Send remote that's been joined to the unit, so an HRU 400 / RF-only setup can see live status without an I2C connection. Skipped silently when no qualifying remote is configured, when the CC1101 is absent, or when the update frequency is 0.
• Debug page: Request 31DA / Request 31D9 buttons with an RF remote selector and an optional 3-byte destination-ID override. Lets you probe an arbitrary Itho's RF address from an arbitrary slot — handy when a slot was never actually joined to a real unit. Override is one-shot: the slot's stored destinationID is snapshotted and restored after the request.
• OTA progress over MQTT so Home Assistant's firmware-update entity shows a progress bar instead of silently downloading and flashing (#371). The HA Discovery value template emits in_progress and update_percentage, and the firmware republishes deviceinfo ~once per second during OTA so the bar moves. No HA integration update needed — the new template ships in the discovery payload itself. End-to-end visibility kicks in on the next OTA after a 3.2.0-beta1 device republishes Discovery.
• Multi-zone HRU support. parseRF31DA / parseRF31D9 accept any payload[0] "domain" byte instead of dropping everything non-zero, so Itho DuoZone HRUs no longer look offline on the RF Status page (#366). The latest seen domain byte is surfaced as zone31DA / zone31D9 on /api/v2/rfstatus. Single-zone units stay at 0 and behave exactly as before. Proper per-zone storage is a follow-up.

Bug fixes

• POST /api/v2/rfremote/config with setrfdevicedestid now persists to flash, not just the runtime CC1101 state. The value previously only lived in the in-memory CC1101 layer, so it was lost on reboot and on the next Update click on the RF Devices page.
• destinationID survives the Update click on the RF Devices page. The web-UI Update was calling updateRFDevice, which resets the in-memory destinationID to the slot's own sourceID and silently wiped anything a join handshake or the setrfdevicedestid API had put there. The persisted destID is now re-applied right after updateRFDevice.
• Paired 31DA + 31D9 RF status request actually sends both opcodes. Stuffing them into one ISR-disabled critical section left the CC1101 in an intermediate state for the second TX, so it silently didn't go out. Each frame now gets its own ISR cycle with a 50 ms gap.
• Debug page buttons no longer reload the page. Every button on the Debug page was inside a <form> with no type="button" attribute, so the browser treated each click as a form submit and navigated away before the websocket message could land. Marked them type="button".
• TX Power dropdown on the RF Devices page activates the correct slot. The dropdown is only rendered for SEND remotes, so the forEach iteration index didn't match the slot index — selecting one Send remote enabled the previously-selected remote's TX Power dropdown.