[Python-ideas] Processing surrogates in

[Stephen J. Turnbull]

Andrew Barnert writes:

 > (I'm not sure if we actually have a UCS-2 codec, but if not, it's
 > trivial to write--it's just UTF-16 without surrogates.)

The PEP 393 machinery knows when astral characters are introduced
because it has to widen the representation.  That might be a more
convenient place to raise an exception on non-BMP characters.