HASS Event Different to Z2M Event

[D3v01dZA]

What happened?

Using the experimental support in #24233.

I have a four button scene switch https://www.zigbee2mqtt.io/devices/TS0044.html#tuya-ts0044.

The actions for each button in Z2M are (where <button> is a button no like 1_single):

• <button>_single
• <button>_double
• <button>_hold

These all show up correctly in Z2M however in HASS the events for hold specifically do not contain the button number, only the text "hold".

So when clicking button 1 once the event in both is "1_single", when double clicking it its "1_double", but when holding it its "1_hold" in Z2M but "hold" in HASS when means I can't disambiguate which button was actually held.

What did you expect to happen?

When holding a button I expect to know which button was held in the event.

How to reproduce it (minimal and precise)

No response

Zigbee2MQTT version

1.42.0 commit: 861cba6

Adapter firmware version

7.4.2 [GA]

Adapter

TubesZB MGM21

Setup

x86-64 Docker

Debug log

No response

[mundschenk-at]

Have a look at the other attributes (besides event_type). Not all possible action values are parsed in the current experimental implementation, the fallback is to just put everything in event_type.

[D3v01dZA]

Ah I see the button in the attributes, now I can continue. Still it seems like this shouldn't be different based on the event type...

event_types:
  - 1_single
  - 1_double
  - hold
  - 2_single
  - 2_double
  - 3_single
  - 3_double
  - 4_single
  - 4_double
event_type: hold
icon: mdi:gesture-double-tap
friendly_name: Test Quadmote Action
button: "2"

For example this is what happens for single clicks :(

event_types:
  - 1_single
  - 1_double
  - hold
  - 2_single
  - 2_double
  - 3_single
  - 3_double
  - 4_single
  - 4_double
event_type: 1_single
icon: mdi:gesture-double-tap
friendly_name: Test Quadmote Action

[mundschenk-at]

Well, you have just discovered why it is labelled "experimental".

[D3v01dZA]

Yes - hopefully this will be seen as feedback on a new feature

[mundschenk-at]

Sorry if it wasn't clear from my earlier statement: Obviously the parsing logic should (and eventually will) be expanded to cover all actions. There's just a lot of ground to cover and we are currently discussing how this will be done.

[D3v01dZA]

Nice! Let me close this one out then as this is obviously being looked at :)