VoiceXML Interpreter Logging

If the com.inin.securedata property is set to true, then the methods listed below will protect their trace log messages.

Prompt-related traces

Confirmation prompts can contain sensitive data, so the majority of the traces containing prompt text are protected. The following methods contain such trace messages:

VXIPromptImpl::Queue()

PromptResource::queueToSystemLocalCache()

SessionResourceCache::queuePromptToSession()

GlobalResourceCache::getResource()

GlobalResourceCache::handleResFromRemote()

GlobalResourceCache::readResource()

GlobalResourceCache::uploadResToxIC()

GlobalResourceCache::cleanUpResources()

Special cases

cleanUpResources()

This method can trace out prompt file names from the cache, but they are not tied to a particular VoiceXML session. Furthermore, this method does not have access to the com.inin.securedata property, but it is not practical to completely remove these traces. As such, these log messages are only traced when the trace level is set to Verbose or higher.

Recognition-related traces

Speech recognition results can contain sensitive data, so traces containing the results from recognition requests are protected. The following methods contain such trace messages:

VXIRecognitionImpl::RecognizeMain()

VXILogImpl::ContentWrite()

Result-related traces

Information provided as VoiceXML document results can contain sensitive data, so traces containing such information are protected. The following method contains such trace messages:

VXISession::translateDocResultToResults()

Special cases

VXISession::ProcessDocument( url )

VXISession::ProcessDocument( document )

These two methods are only going to trace sensitive data if the handler passes the sensitive data into the document in the input name/value pairs.These traces are done before the VoiceXML document is actually entered, so we can't use the com.inin.securedata property to determine how to treat this data. As such, we will trace the incoming variables at the Verbose trace level and allow the logging trace level determine what gets logged.

BladewareVXML diagnostic traces

BladewareVXML diagnostic messages can contain sensitive data, so traces containing such information are protected. The following method contains such trace:

VXILogImpl::VDiagnostic()

This method will only protect information from BladewareVXML diagnostic messages with the following tagID's:

TagIDs	Where they originate in BladewareVXML
3000, 3001, 3004, 3005	The .inet layer.
3500, 3501, 3502, 3503	The .cache layer.
4000	The .jsi layer.
8002	The .vxi layer.

<log> element traces

We will make no attempt to protect data from any <log> elements, as the inclusion of sensitive data in these logs is a VoiceXML application level decision. This includes Nuance NDM modules, which sometimes <log> sensitive information.