Markus Scherer wrote:
Right, but as I said, those downstream tasks shouldn't be consumers
of UTF-16 code units anyway. They should be consumers of Unicode
code points, which by definition excludes loose surrogates.
Code points include surrogates. Maybe you mean "UTF-32 code units" or
"Unicode scalar values".
You're right, I meant Unicode scalar values.
I don't see the difference between allowing loose UTF-16 code units in
what purports to be a character stream and allowing loose UTF-8 code
units.
--
Doug Ewell | Thornton, Colorado, USA | http://www.ewellic.org
RFC 5645, 4645, UTN #14 | ietf-languages @ is dot gd slash 2kf0s
