Problem with duplicate GA

[TheMax74]

Hello everyone.
I have a simple routing function. Each received GA is forwarded to my PC via TCP.

Part of code:
[highlight=epc] if event(readknx(adr,info)) and TCP_ConnectionState == 0 then {
sendtcp(Name_Port, 192.168.1.78,convert(adr,$$)+$=$+convert(info,$$)) ;
}endif [/highlight]
Everythings goes well but i've a strange behaviour when i query a GA (via ETS for example)
If i query for read GA 2/4/2, for example, in the Eibstudio console everythings goes fine:

% 2013-11-30 21:23:51 | Sender: 1.1.49 | GA: "Wh AC Essential BUS-2/4/9" | Value: 428.86 | Type: 32-Bit Floating Point | Write
% 2013-11-30 21:23:55 | Sender: 15.15.0 | GA: "Wh Consumo TOTALE-2/4/2" | Value: ?% | Type: ?% | Read
% 2013-11-30 21:23:55 | Sender: 1.1.49 | GA: "Wh Consumo TOTALE-2/4/2" | Value: 714.3 | Type: 32-Bit Floating Point | Reply

...but...

The readknx function is triggered twice (with the GA received before the queried GA, then with the replied GA)

For example, in the case above, i have:
2/4/9=428.86
2/4/2=714.3

instead of only the
2/4/2=714.3

I need your help because i cannot figure out why this happens.

[enertegus]

The readknx function is triggered twice (with the GA received before the queried GA, then with the replied GA)
For example, in the case above, i have:
2/4/9=428.86
2/4/2=714.3

That seems absolutely correct or I donot understand your post.

[TheMax74]

Yes it's a bit hard to explain, sorry.
I receive telegrams as follow:

...
...
...
2/4/9=428.86 <---- last ga from bus
here i click READ on ETS on Ga 2/4/2
2/4/9=428.86 <--- this is not from bus
2/4/2=714.3 <---- this is the correct reply to the read

Seems that the last arrived GA frum bus, is sent again after i ckick READ on ETS.
The duplicated ga came not from bus and may vary, depending on what ga was sent from bus just before the click on the READ button.
i hope it's more clear now.

[enertegus]

The duplicated ga came not from bus and may vary, depending on what ga was sent from bus just before the click on the READ button.
i hope it's more clear now.

Ah, I see. I think, the problem comes from the readknx: As an read on the bus contains no data, the readknx-Function just fills the last data into the arguments.
You should try
if eventwrite(readknx) ...
or
if eventwrite(readknx) or eventresponse(readknx) ...

[TheMax74]

if eventwrite(readknx) ...

No success. The GA 2/4/9 is duplicated and the response to the READ 2/4/2 is not trapped by the eventwrite

if eventwrite(readknx) or eventresponse(readknx) ...

Same as readknx.

Is there a reason to re-transmit the last readknx buffer if the telegram has no data?
As i can see, the Eibstudio messages console can distinguish between "write", "read" and "reply" telegrams. Is there a way to do this, maybe with readrawknx ?

[enertegus]

No success. The GA 2/4/9 is duplicated and the response to the READ 2/4/2 is not trapped by the eventwrite

I made a simple example:

[highlight=epc]
adr=0u16
info=$$
if event(readknx(adr,info)) and adr!=getaddress("TextDebug-0/2/100") then {
write("TextDebug-0/2/100",convert(adr,$$c14));
write("TextDebug-0/2/100",convert(info,$$c14));
}endif
[/highlight]

I have e.g.

13-12-01 13:04:05 | Sender: 1.1.43 | GA: "BrennerTemperatur-3/3/2" | Wert: 54.0 | Typ: 16-Bit-Fliesskommazahl |  Schreiben
% 2013-12-01 13:04:05 | Sender: EibPC | GA: "TextDebug-0/2/100" | Wert: 6914 | Typ: Text |  Schreiben
% 2013-12-01 13:04:05 | Sender: EibPC | GA: "TextDebug-0/2/100" | Wert: 54.00 | Typ: Text |  Schreiben
[B]% 2013-12-01 13:04:08 | Sender: 15.15.11 | GA: "RkBadTemp-3/2/13" | Wert: ?%  | Typ: ?%  |  Lesen
% 2013-12-01 13:04:08 | Sender: 1.1.33 | GA: "RkBadTemp-3/2/13" | Wert: 20.3 | Typ: 16-Bit-Fliesskommazahl |  Antworten
% 2013-12-01 13:04:08 | Sender: EibPC | GA: "TextDebug-0/2/100" | Wert: 6669 | Typ: Text |  Schreiben
% 2013-12-01 13:04:08 | Sender: EibPC | GA: "TextDebug-0/2/100" | Wert: 20.30 | Typ: Text |  Schreiben[/B]
% 2013-12-01 13:04:17 | Sender: 1.1.43 | GA: "WarmwasserIstWert-3/3/10" | Wert: 52.0 | Typ: 16-Bit-Fliesskommazahl |  Schreiben

This is correct, I think.
Hm, which FW do you use? Here I only have the actual beta Software.

[TheMax74]

My firmware is the latest stable.

Thanks for your help. I've been modified my code as follows

[highlight=epc]
if event(readknx(adr,info)) and TCP_ConnectionState == 0 and adr!=getaddress('0/0/0'c14) then {
sendtcp(Name_Port, 192.168.1.78,convert(adr,$$)+$=$+convert(info,$$)) ;
// Send a value to a dummy GA to replace readknx buffer after a READ
// GA Dummy!
write('0/0/0'b01,AUS)

}endif [/highlight]

This seems to work but the traffic on the bus is doubled.
It was wonder if you can fix the readknx function.
Where can i download your beta?

[enertegus]

Where can i download your beta?

The next update is only for free, if you bought the EibPC after 1.1.2013. If you have bought before, it will cost 98 Euro.

[TheMax74]

Ok, thanks for your kindly support, it's very great!

[enertegus]

This seems to work but the traffic on the bus is doubled.

Why doubled? Do you mean because of the write?
In my system, an read-event does not trigger any output.
If this is the case, an possible solution is to write the dummy only, after a read.
if eventread() then write(dummy...) endif

[TheMax74]

Good idea.
But unfortunately doesn't work.
Sometimes "event" is trapped before "eventread" = before the write of the dummy GA