HC3 QuickApps coding - tips and tricks

[Martin_N 143]

1 hour ago, Krikroff said:

Hum... Strange.

i will try to reproduce when sonos are not available on network.

have you any problems with wifi ?

 

 

I solved a similar problem with my QA for iEast player.

Please login or register to see this link.

My conclusion is that workload problems are caused
1) communication failures on w-fi between devices (HC3 - audio)
2) HC3 - LAN communication outages
3) occasional mental problems HC3 own when you can not even save any parameter and the end of a 500/501 Internal Server Error

[Rover 45]

2 hours ago, Krikroff said:

Hum... Strange.

i will try to reproduce when sonos are not available on network.

have you any problems with wifi ?

 

No I have no problems with wifi. At this moment only 5 Sonos players (not playing now), 2 camera's and an iPhone are using wifi (without problems). I deleted one of the two Sonos ZC QA's.

I rebooted the Orbi RBR50 router and the CPU load of the HC3 at the time of this reboot is:

Please login or register to see this spoiler.

I see from time to time a very short time error message at the bottom site, I saw the word "Error" and also "EOF" if I am not mistaken, all in red color.

Could it help to activate a sort of debug mode?

Edited November 24, 2020 by Rover

[Krikroff 199]

There is clearly a communication fault between the HC3 and the network...
Is it possible that the Orbi RBR50 is blocking requests for some reason (ex: excessive polling) ? You also had a network access problem with the QA offered by @10der, didn't you?

Workaround in the next version, I can consider the possibility of completely interrupting the polling for a given time in the event of faulty communication between the HC3 and SONOS devices

[Rover 45]

8 hours ago, Krikroff said:

There is clearly a communication fault between the HC3 and the network...
Is it possible that the Orbi RBR50 is blocking requests for some reason (ex: excessive polling) ? You also had a network access problem with the QA offered by @10der, didn't you?

Workaround in the next version, I can consider the possibility of completely interrupting the polling for a given time in the event of faulty communication between the HC3 and SONOS devices

I do not think it is a network problem of the HC3. I checked in the log and rebooted the router: no difference.

Why is this excessive CPU load only present when Sonos ZC is active?

There was no network problem using @10der's Sonos_Hub, only a problem with the json encoding of the call parameters, what has been solved with help of @cag014.

I have set the Sonos ZC debug mode on:

Please login or register to see this spoiler.

 

Edited November 25, 2020 by Rover

[Momos 101]

Hi All,

 

A quick question please:  I have a weather station data imported via MQTT to a QA with child devices, one child for each sensor.  In the labels of the QA the temperature shows correctly with the - sign.

Please login or register to see this image.

/monthly_2020_11/image.png.ff231357e1d14cdba3924cca4033c2ad.png" />

 

However, in the corresponding child the - sign does not show.

 

I am updating the data to the child device with :

self:updateProperty("value",(data.temperature))

 

The debug shows it  being a negative number.

 

Any idea please ?

 

Thank you

 

 

[10der 830]

9 hours ago, Momos said:

Any idea please ?

search post on the forum by @afendikovwith subj "fibaro be carriful with rounding float values"  

Fibaro... it's fibaro.

[jgab 1,954]

Someone asked for a simple example for a write/read to a socket. Som devices has simple text based protocols over sockets to control them.

This example is similar to the Fibaro TCPSocket example but maybe a bit simpler. This also creates new sockets for each requests (sockets can't be reused after closed, so it requires some extra code to keep a socket alive as long as possible and still cater for errors)

 

Please login or register to see this code.

 

[Momos 101]

1 hour ago, 10der said:

search post on the forum by @afendikovwith subj "fibaro be carriful with rounding float values"  

Fibaro... it's fibaro.

Gaahh ... i have nothing more to say 

[10der 830]

3 hours ago, Momos said:

Gaahh ... i have nothing more to say 

 

[Momos 101]

ahaha ??  My response was targeted at Fibaro ?  I can not belive such a thing can actually exist in a device nowadays.  Especially in one with a high price tag.

Edited November 28, 2020 by Momos

[Rover 45]

How to change size and color of text in self:updateView?

How to change background color of QA buttons?

[jgab 1,954]

9 minutes ago, Rover said:

How to change size and color of text in self:updateView?

How to change background color of QA buttons?

People have managed to put HTML tags in the label text. It may not work in all browsers/apps.

[Rover 45]

47 minutes ago, jgab said:

People have managed to put HTML tags in the label text. It may not work in all browsers/apps.

In Edge as well as Firefox self:updateView("btn_SPH", "text", "<font color=#fd7e14>Setpoint Hoog) shows in the button: font color=#fd7e14>Setpoint Hoog

[jgab 1,954]

12 minutes ago, Rover said:

In Edge as well as Firefox self:updateView("btn_SPH", "text", "<font color=#fd7e14>Setpoint Hoog) shows in the button: font color=#fd7e14>Setpoint Hoog

Labels it works for, buttons no.

[Rover 45]

1 minute ago, jgab said:

Labels it works for, buttons no.

So string formatting or color change of buttons is always a nogo?

[Rover 45]

Is it possible to create a 2/3/4 in a row QA labelset?

[jgab 1,954]

Well, it's possible if you hack the viewLayout property yourself. 

The problem is that the UI can't be edited in the HC3 Web editor anymore. In fact, if you just re-save the QA the viewLayout is "reset". So I would say that's it not really that useful until Fibaro starts to support better layouts.

 

I do all my coding offline in ZBS and the UI spec allows me to add funky UI layouts if I need (but I usually don't)

Please login or register to see this image.

/monthly_2020_12/labels.png.33ceb04c7f804388f89e962da68e5e85.png" />

using the code

Please login or register to see this code.

 

Here is a even more interesting QA with pop-ups etc.

<

Please login or register to see this link.

>

[jayrock 9]

On 4/19/2020 at 6:28 PM, jgab said:

There have been posts about getting triggers/events in QAs using the /refresStates API.

This post is nothing new but wraps the logic in a 'SourceTrigger' class

Please login or register to see this code.

 

If we have this class we can use it like this

Please login or register to see this code.

The functions we define on SourceTrigger has the name <type>_<property>

Please login or register to see this code.

will be called with 'source triggers' of type {type='device', property='value' ...}

If we define

Please login or register to see this code.

we will get all triggers of type {type='device', ...}

For triggers without a 'property' field we only have the type.

The intention is that the triggers should look like the sourceTriggers available for Scenes.

 

If we need different trigger loops we can subclass 'SourceTrigger'

 

Dear all,

 

first of all, thanks for this excellent post. I migrated all of my HC2 stuff to HC3 with basically just this thread as a reference. Kudos to @jgab, @petergebruers and all the other folks.

 

I have some QAs working as MQTT subscribers, 528 and 542. The publisher sends a single update, so two QAs get updated at the same time. The update of 528 triggers the update of another QA (535). These QAs are updated correctly. But the SourceTrigger QA from the post on page 17 gives:

 

Please login or register to see this code.

 

I don't understand why 542 or 528 are missed most of the time. Two events that happen at the same time should be caught by

 

Please login or register to see this code.

 

Is this some crazy race condition? Maybe not, because setTimeout on self:updateProperty in 528 does not really help missed triggers. And it is not a nice solution.

 

Do you have any futher suggestions how to tackle this problem?

 

Cheers, jayrock

 

[jgab 1,954]

2 hours ago, jayrock said:

 

Dear all,

 

first of all, thanks for this excellent post. I migrated all of my HC2 stuff to HC3 with basically just this thread as a reference. Kudos to @jgab, @petergebruers and all the other folks.

 

I have some QAs working as MQTT subscribers, 528 and 542. The publisher sends a single update, so two QAs get updated at the same time. The update of 528 triggers the update of another QA (535). These QAs are updated correctly. But the SourceTrigger QA from the post on page 17 gives:

 

Please login or register to see this code.

 

I don't understand why 542 or 528 are missed most of the time. Two events that happen at the same time should be caught by

 

Please login or register to see this code.

 

Is this some crazy race condition? Maybe not, because setTimeout on self:updateProperty in 528 does not really help missed triggers. And it is not a nice solution.

 

Do you have any futher suggestions how to tackle this problem?

 

Cheers, jayrock

 

 

If you update the property of a QA (ex. the value property) it will only generate an internal trigger (event) if the value differs from the current value.

Ex. self:updateProperty("value",fibaro.getValue(self.id,"value")) will not generate a trigger.

Also, do your MQTT send events if the value changes or with regular intervals?

[jayrock 9]

59 minutes ago, jgab said:

 

If you update the property of a QA (ex. the value property) it will only generate an internal trigger (event) if the value differs from the current value.

 

Thanks, that makes a lot of sense. Indeed the value does't change that often. It also explains why some sensors seemed to work "better".

 

59 minutes ago, jgab said:

Also, do your MQTT send events if the value changes or with regular intervals?

 

Yes, the devices do send on regular intervals, that means they send an update (which can be observed via MQTT) even if the value hasn't changed.

 

Let me do some more monitoring and see if this is the explanation.

 

Cheers,

jayrock

Edited December 20, 2020 by jayrock