On 2/7/14, Dmitry Olshansky <dmitry.olsh at gmail.com> wrote: > Much simpler - it returns a special dchar to designate bad encoding. And > there is one defined by Unicode spec. A NaN for chars? Sounds great to me! :)