Xiaomi Mijia round single key wireless WXKG01LM - Deconz 86 issue

SAOPP commented 3 years ago

Describe the bug

In version 86 this button has a issue with long press / event. The release event is now the same as the single click event.

Long click --> 1000 --> 1001 --> 1002

1002 is a single click event, long click it should be 1001 and release will be 1003. But this does not happen, it is a mistake.The release now works like a single click.

Environment

Raspbee / Conbee 2 / 2.05.86 / 15.10.2020 / 26350500 / 264A0700 / 26580700

Host system: (PC)
Running method: (Marthoc Docker container)
Firmware version: (264A0700)
deCONZ version: (2.05.86)
Device: (ConBee II / RaspBee I)
Do you use an USB extension cable: (no)

deCONZ Logs

9:44:
{
    "event_type": "deconz_event",
    "data": {
        "id": "hall_sofa_mi_button",
        "unique_id": "00:15:8d:00:01:6d:86:28",
        "event": 1002
    },
    "origin": "LOCAL",
    "time_fired": "2020-10-23T06:44:41.543487+00:00",
    "context": {
        "id": "3a7b419f14fb11ebb2f0d92301fc637c",
        "parent_id": null,
        "user_id": null
    }
}
9:44:
{
    "event_type": "deconz_event",
    "data": {
        "id": "hall_sofa_mi_button",
        "unique_id": "00:15:8d:00:01:6d:86:28",
        "event": 1001
    },
    "origin": "LOCAL",
    "time_fired": "2020-10-23T06:44:41.440208+00:00",
    "context": {
        "id": "3a6b7fb614fb11eb8bd8c382bdb861cb",
        "parent_id": null,
        "user_id": null
    }
}
9:44:
{
    "event_type": "deconz_event",
    "data": {
        "id": "hall_sofa_mi_button",
        "unique_id": "00:15:8d:00:01:6d:86:28",
        "event": 1000
    },
    "origin": "LOCAL",
    "time_fired": "2020-10-23T06:44:40.587226+00:00",
    "context": {
        "id": "39e957aa14fb11ebbc63afd4b24c1713",
        "parent_id": null,
        "user_id": null
    }
}

SwoopX commented 3 years ago

You might be right on this, however, that must be ancient then. Assuming, the zigbee modelID starts with "lumi.sensor_switch", single and long press have the same parameters and that has been in there presumably for years.

[1, "0x01", "ONOFF", "0x0a", "1", "S_BUTTON_1", "S_BUTTON_ACTION_SHORT_RELEASED", "Normal release"],
[1, "0x01", "ONOFF", "0x0a", "1", "S_BUTTON_1", "S_BUTTON_ACTION_LONG_RELEASED", "Long release"],

To actually verify this, someone owning the device should remove that button map temporarily and share the corresponding debug output.

ebaauw commented 3 years ago

Edit added mentioned (already) supported quadruple press.

The Mija button would be lumi.sensor_switch; the 2016 Aqara button is lumi.sensor_switch.aq2; the 2018 is lumi.remote.b1acn01. There's also a variant with built-in vibration sensor, I think lumi.sensor_switch.aq3 - I ordered those, but the freaking webshop on AliExpress delivered the wrong model (or rather: advertised the wrong description) and wouldn't even acknowledge that.

All of these support different button actions and use different ways of reporting these.

I only have the lumi.sensor_switch.aq2. It reports single press (1002), double press (1004), triple press (1005), and quadruple press (1006). It uses attribute reporting on the server On/Off cluster:

On release it reports the OnOff attribute (0x0000) - the report actually contains two records in the same message, one for Off followed by one for On. It doesn't matter how long the button was held before release;
On release after double press, it reports value 2 for attribute 0x8000;
On release after triple press, it reports value 3 for attribute 0x8000.
On release after quadruple press, it reports value 4 for attribute 0x8000.

Note to self: looks like 0x8000 (u8) is missing from general.xml.

As I said, the other models support different actions and use a different way of reporting these. The Mija does not report long press; @manup hacked the code to measure the time between press and release and invalidate the match for the first entry, so it would fall back to the second entry. That probably broke with the move to external button maps. https://github.com/dresden-elektronik/deconz-rest-plugin/blob/cbef8cc540c5f7e4258b45f396fbf48a5cb21852/de_web_plugin.cpp#L3964-L3990

kooskaspers commented 3 years ago

Same issue here. Since Deconz 86 I can't get a '1003' anymore (Button long release), and indeed, the release now works as a single click. Kinda frustrating since we're using these buttons as switches for our bedlamps :cry:

SwoopX commented 3 years ago

Tbh, I don't see what the issue should be here. Other than changing the variable name and type, the code has remained untouched. As I see it, it even gets entered, otherwise you wouldn't see the hold event.

I wonder if it could be a matter of changed timing...

manup commented 3 years ago

Fixed in the PR. Note that the out-commented check for an empty button map might cause other issues as well since updateSensorNode(const deCONZ::NodeEvent &event) and the button maps work in parallel.

https://github.com/dresden-elektronik/deconz-rest-plugin/blob/master/de_web_plugin.cpp#L7575-L7576

kooskaspers commented 3 years ago

Tbh, I don't see what the issue should be here. Other than changing the variable name and type, the code has remained untouched. As I see it, it even gets entered, otherwise you wouldn't see the hold event.

I wonder if it could be a matter of changed timing...

The issue seems to be resolved here. And I haven’t changed the container image of deconz. Very, very strange. I guess home-assistant plays a role here. I’ve upgraded hass a few days ago and that seem to have solved the problem.

stale[bot] commented 3 years ago

As there hasn't been any response in 21 days, this issue has been automatically marked as stale. At OP: Please either close this issue or keep it active It will be closed in 7 days if no further activity occurs.

stale[bot] commented 3 years ago

As there hasn't been any response in 28 days, this issue will be closed. @ OP: If this issue is solved post what fixed it for you. If it isn't solved, request to get this opened again.

dresden-elektronik / deconz-rest-plugin