On 06/02/2016 10:13 PM, Andrei Alexandrescu wrote:
They do work per spec: find this code point. It would be surprising if 'ö' were found but the string were positioned at a different code point.
The "spec" here is how the range primitives for narrow strings are defined, right? I.e., the spec says auto-decode code units to code points.
The discussion is about whether the spec is good or bad. No one is arguing that there are bugs in the decoding to code points. People are arguing that auto-decoding to code points is not useful.
