Thanks for clarifying.
The behavior in our customer's environment is such that when they re-image a machine from their base snapshot, they do not encounter the issue. According to them, after a period of time of use, the application consistently crashes - not random in their environment. We could reproduce the consistency as they sent us their hardware, and narrowed it down to the Listen() method call.
Would the Toshiba Windows Messaging issue still account for that scenario?