1.0.1 continuously warning [STATE CHANGE] wrong value about iobroker.statistics

[foxthefox]

this is solved with 1.0.4

[foxthefox]

the only thing what was recently updated was the rounding to 4 digits.
but this was 1.0.0 to 0.2.2
the major update to 1.0.0 cam from bluefox, I have to dig in the code

Edit: what should be the statistical case?
If it is a count value, then only the the true state is used. false values are not considered.

[Diginix]

Strange. Ive done a rollback to 0.2.2 and all is fine again.
Maybe I'll try the update again tomorrow.

It is this settings for all devices:

[fuchs-1978]

Same problem here:
STATE CHANGE] wrong value => false on sonoff.0.Gaszaehler.POWER2 => check the other adapter where value comes from

With 0.2.2 all is fine

[foxthefox]

no need to retry upgrade to 1.0.1, it will happen again.

The changed code in 1.0.0 is:
`
if( (state.val === null) || (state.val === undefined) || (state.val === NaN) ){

to

if (!state.val && state.val !== 0) {

`
and throws the same message.
before it was checking for invalid only.
may be @GermanBluefox can say something

As I said before, the adapter does not count on negative edges, only 0->1. So from working it is right, from logging I have no opinion.

[passuff]

I have the same issue here with true/false logging. My log is spamed all over with it. Could we somehow turn it off or is a rollback the only solution currently?

[foxthefox]

you can roll back, if you want.

[massimo1971]

Strange. Ive done a rollback to 0.2.2 and all is fine again.
Maybe I'll try the update again tomorrow.

It is this settings for all devices:

Same Problem here, I had to roll back to 0.2.2

[watcherkb]

I have the same issue and no values are count.

[Diginix]

Since no developer has fixed the issue in 1.x so far, it seems 0.2.2 is the only suitable version for many users.

[maeb3]

I can confirm what @foxthefox said in the post above:
I'm running version 1.0.2 with the line
if (!state.val && state.val !== 0) {
manually changed in the main.js back to the original one (from version 0.2.2). Running for some weeks now. No problems, everything works fine. Time count calculation is now correct again (was wrong before because adapter did not recognize 1->0 edges, so it's not only a logging issue but also wrong values for online-time/offline-time).

Has anybody any ideas wether there might be negative impact when switching back to the old line of code?

[Diginix]

Updated yesterday from 0.2.2 to 1.0.4 without log entries so far. Seems to be solved. Thanks!

[Stefan-83]

Now I have the same issue 😒
Version 1.0.9

[Stefan-83]

Please re-open
@foxthefox

[Stefan-83]

Update: After I've downgraded to 1.0.4., it's working again.
Is there a known error in 1.0.9 (latest version)