A common issue is that users fail to configure the ThinLinc setup properly for a NAT environment (mostly forgetting agent_hostname). Unfortunately, it is not terribly obvious that this is the issue if you aren't experienced with those types of issues. The client could probably report things in a better way, where it might be more obvious what is going on. E.g.: * Clearly specify what address it tried to talk to * Suggest that NAT might be an issue * Detect that the agent address is a RFC1918 address and complain a bit extra
A related issue is also that the client will generally lock up whilst waiting for a timeout for the unreachable NAT address. We have bug 1197 for this.
