What, the encoder function doesn't have a way to configure the error behavior? Sounds so very Microsoft. (I've had to avoid using the built-in Win32 functions for reasons related to that)