We use tassert to report logic errors that aren't severe enough to warrant shutting down the server. The server returns an error to the client and keeps running. This is similar to uassert, but the difference is that tassert is supposed to make it clear that the error is the server's fault, not the client's fault.

Is there an easy way for clients to tell the difference between uassert and tassert? Would that be useful? Would error labels be a good mechanism?