On 09/01/2024 00.55, Eddie Nolan via SG16 wrote:
> Hi,
>
> I wanted to announce that I’m taking over ownership of Zach Laine’s Unicode papers: P2728, P2729, and P2745.
>
> I’m currently working on a new revision of P2728, “Unicode in the Library, Part 1: UTF Transcoding.” The main update will be a different approach to error handling, replacing the previous |transcoding_error_handler| mechanism with a new member function in |utf_iterator| that specifies whether the code unit that |operator*| would return is part of a replacement character that was substituted for an ill-formed subsequence.
>
> Please let me know if you have any feedback on P2728

I think I've voiced that before, and I haven't looked into a recent update
(if any), but continue to believe the facility should be designed from
a purely std::ranges perspective, not from an iterator perspective.
(If you need iterators for some reason, you can get those from a range.)

If something about UTF transcoding doesn't work in a pure ranges world,
I'd like to understand why. Maybe std::ranges concepts are still lacking
in some areas.

Jens