Yes, it works with Java 17 too.

The only way you can do is searching this string "com.avaya.sce.runtime.tracking.ReportInfo". It will give you all the java file that has this code which is corresponding a report item in the node that matches the Java file.

My understanding of OAuth2.0 is you have to first obtain the access key using the credential. Then you can use the access key to call the desired REST API. OD does not support this process automatically. You have to be creative about stringing the steps in your app.

So it looks like EP did process the EventPrompt and try to send the event to Dialogflow. Maybe it is because of the existence of the <prompt> element, it is not sending the event. Can you try removing the prompt item from EventPrompt node in the node editor. It would be like the picture I attached

20/03/2024 11:00:11:535 INFO - 1EBBBA1BE98997434F5D0DFB162520CA:/DefaultDialogflowAppV3 : Using SCESession 1EBBBA1BE98997434F5D0DFB162520CA:/DefaultDialogflowAppV3 servlet : EventPrompt
20/03/2024 11:00:11:535 INFO - 1EBBBA1BE98997434F5D0DFB162520CA:/DefaultDialogflowAppV3 : [condition] Expression [InputPrompt] Not Empty [null]
20/03/2024 11:00:11:535 INFO - 1EBBBA1BE98997434F5D0DFB162520CA:/DefaultDialogflowAppV3 : [condition] Expression [eventprops:name] String Not Equal [DF_SYSTEM_NO_INPUT]
.....
20/03/2024 11:00:11:550 DEBUG - 1EBBBA1BE98997434F5D0DFB162520CA:/DefaultDialogflowAppV3 : *** Reply for [/DefaultDialogflowAppV3/EventPrompt] ***
....
19:<prompt bargein="false" timeout="8000ms">
20:</prompt>

It is not the eclipse process that access the file. It is the simulator (pavb.exe) process, a separate windows process. I am not sure how you can grant access particular to that process as I have never seen such problem.

A successful log would look like this (please ignore the fact my log shows OD 8.1.2. Nothing has changed). This shows you a cache file would have been created by the simulator in the eclipse/plugin/com.avaya.sceavb<version>/AVB/cache/com.avaya.OSBinet directory on your desktop. So your error indicate simulator couldn't do that. Pretty likely a permission issue.

Nov 08 10:35:49.769|64|0064-00000000|15004|FINE|Prompt|BufferTranscoder.cpp:187:Transcoding from Wave-PCM to Wave-mulaw
Nov 08 10:35:49.769|64|0064-00000000|15004|FINE|Prompt|BufferTranscoder.cpp:623:Sucessfully transcoded 3200 bytes from PCM to mulaw
Nov 08 10:35:49.769|64|0064-00000000|15004|FINE|Cache|SBinetChannel.cpp:1269:SBinetChannel.cpp:1269:SBinetHttpCacheStream::staticWriteBufferToCache called
Nov 08 10:35:49.770|64|0064-00000000|15004|FINE|Cache|*** CACHE ENTRY FILE gets created at: *** CACHE ENTRY FILE gets created at: C:\\AOD-8.1.2\\eclipse\\plugins\\com.avaya.sceavb_8.1.2.1101\\AVB/cache/com.avaya.OSBinet/0/1.wav
Nov 08 10:35:49.771|64|0064-00000000|15004|FINE|Platform|ChannelImpl.cpp:3518:CChannelImpl::ReportURIFetchStatus: Fetched Document C:\\AOD-8.1.2\\eclipse\\plugins\\com.avaya.sceavb_8.1.2.1101\\AVB/data/beep.wav, Type=Prompt, FetchTime=15, parseTime=0, from cache
Nov 08 10:35:49.771|64|0064-00000000|15004|FINE|Prompt|Queuing Audio - URL: C:\\AOD-8.1.2\\eclipse\\plugins\\com.avaya.sceavb_8.1.2.1101\\AVB/data/beep.wav Local file: C:\\AOD-8.1.2\\eclipse\\plugins\\com.avaya.sceavb_8.1.2.1101\\AVB/cache/com.avaya.OSBinet/0/1.wav

Like I said before I didn't really have experience of this functionality. I just thought it would have worked the way I described. Now I've looked deeper into what is happening, I don't see how the VXML code generated by OD is getting any value returned matching the input grammar. Below is the sample vxml in the log I am looking at. That's how the variable is filled by OD. I don't think we ever fully implemented the functionality for some reason. However, if we know that EP does return a value matching the grammar, you maybe able to develop a customer VXML Servlet instead of using the Record node to handle this.

29:<filled>
30:<assign name="untitledRecord___value" expr="try {untitledRecord} catch (errMsg) {'undefined'}"/>
31:<assign name="untitledRecord___confidence" expr="try {untitledRecord$.confidence} catch (errMsg) {'undefined'}"/>
32:<assign name="untitledRecord___utterance" expr="try {untitledRecord$.utterance} catch (errMsg) {'undefined'}"/>
33:<assign name="untitledRecord___duration" expr="try {untitledRecord$.duration} catch (errMsg) {'undefined'}"/>
34:<assign name="untitledRecord___size" expr="try {untitledRecord$.size} catch (errMsg) {'undefined'}"/>
35:<assign name="untitledRecord___termchar" expr="try {untitledRecord$.termchar} catch (errMsg) {'undefined'}"/>
36:<assign name="untitledRecord___maxtime" expr="try {untitledRecord$.maxtime} catch (errMsg) {'undefined'}"/>
37:</filled>

Once you embed the grammar in the record element, the variable field would have been created for the grammar slot. The recognition would be sent back and stored in the variable field the same way as other input elements

Vasily, Ross has been long gone.
This only thing that you might need is couple of samples. I've attached them here.

So you only want "#" to work. You would need to set the dtmfTerm property to false then. It would then use the grammar embedded in the Record element. The recognition result would probably come back to the column0 field instead. I have never tried this myself. Hope this works.

Not sure why you need to grammar for this. By default, the dtmfTerm property is set to true. The log you show doesn't seem to be from the record result which is expected to have the termchar field returned. When user, presses any key, it would terminate the recording and send back the termchar and this would be stored in the variable's termchar field. You can even test it out in OD simulation.
You would see something like this in the log:
31/08/2022 08:58:15:793 INFO - 4FC850F7AA708A2ECB781CCFA264F5AE:/SpeechApp : Storing :untitledRecord___termchar to complex: untitledRecord:termchar as [1]

try/catch is the only option right now. You can suppress the exception in the log by setting the showcaughtexceptions property to disabled in the ddrt.properties file or in the report settings of EPM

Check this out
https://www.devconnectprogram.com/forums/posts/list/22183.page#p154184

I am not sure. You have to verify with the platform documentation.

The older version of weblm.jar would not have this issue.