once more: Cannot create vDev based on: 38-0-113

[piet66]

It's an old problem that out of the blue zway starts to print those error messages:

Code: Select all

[I] [core] Notification: error (core): Cannot create vDev based on: 38-0-113: TypeError: Cannot read property 'forEach' of null
[I] [core] TypeError: Cannot read property 'forEach' of null

This is caused by additional parts in interview, which were not there before. A new force interview stops this printing.

Today it was an Aeotec Multisensor Gen 6.
The additional parts in interview results were:
-instance 0, command 113 (Alarm), 3 (CO2)
-instance 2 (there was only instance 0 before)

Code: Select all

                113: {
                    name: "Alarm",
                    data: {
                        3: {
                            value: null,
                            type: "empty",
                            typeString: {
                                value: "CO2",
                                type: "string",
                                invalidateTime: 1574135876,
                                updateTime: 1574135877
                            },
                            status: {
                                value: 255,
                                type: "int",
                                invalidateTime: 1574135876,
                                updateTime: 1574135877
                            },
                            eventMask: {
                                value: null,
                                type: "empty",
                                invalidateTime: 1574135876,
                                updateTime: 1574135877
                            },
                            event: {
                                value: 8,
                                type: "int",
                                invalidateTime: 1574135876,
                                updateTime: "1574135877 2019-11-19 04:57:57"
                            },
                            eventString: {
                                value: "",
                                type: "string",
                                invalidateTime: 1574135876,
                                updateTime: 1574135877
                            },
                            eventParameters: {
                                value: [],
                                type: "binary",
                                invalidateTime: 1574135876,
                                updateTime: 1574135877
                            },
                            eventSequence: {
                                value: null,
                                type: "empty",
                                invalidateTime: 1574135876,
                                updateTime: 1574135877
                            },
                            invalidateTime: 1574135876,
                            updateTime: 1574135877
                        },
                        7: {

In past I had this effect also with other kind of devices (Sensative Strips) and when I remember right, it was also class 113.
What I don't understand:
- where do these wrong data come from? There was no entry in log (debug switched off). But an interview of class Alarm would have produced an info message (Interview done, interview finished). I don't think that the device has sent anything like that.
- why does ZWay enter these data in interview result. The interview was completed. It shouldn't be changed.
- why tries ZWay to create a new device for this additional part (instance 0, command 113 (Alarm), 3 (CO2)) every time it gets a message from device for (instance 0, command 113 (Alarm), 7 (Burglar) = Movement detected)? As long as the device doesn't send any message for Alarm/CO2 it's not necessary from my point of view.

[PoltoS]

Need to check it. Can we access the system to reproduce it. Remote support would help.

[piet66]

How do you want to reproduce it? Can I do it by myself?
Have you tried to reproduce it on an own system? What's the result?

[PoltoS]

We were not able to reproduce it so we need to access some system that is affected by this bug and we check where is the problem in the data structure.