I thought I replied to this, but I can find no record of such now. Anyway...

I agree with going ahead and removing all of these facets in C++26. I'm happy to poll that during the April 9th SG16 meeting if that works for everyone.

Tom.

I think it's a good idea and I was planning to suggest exactly that during the last SG16 telecon but unfortunately had to leave early and forgot to follow-up by email.

- Victor

On Thu, Mar 20, 2025 at 12:56 PM Alisdair Meredith via SG16 <sg16@lists.isocpp.org> wrote:

Earlier in the C++26 cycle, we opted to deprecate the facets:
   codecvt<char16_t, char8_t, mbstate_t>
   codecvt<char32_t, char8_t, mbstate_t>
   codecvt_byname<char16_t, char8_t, mbstate_t>
   codecvt_byname<char32_t, char8_t, mbstate_t>


Now we have agreed to remove the previously deprecated facets:
   codecvt<char16_t, char, mbstate_t>
   codecvt<char32_t, char, mbstate_t>
   codecvt_byname<char16_t, char, mbstate_t>
   codecvt_byname<char32_t, char, mbstate_t>

Should we consider removing all these facets in one hit?
It would seem a nice way to clear the decks for whatever
Unicode support we plan to develop for C++29.

At this stage, removal would mean filing an NB comment once
the CD goes out to ballot, and I would not be prepared to write
such a comment unless this group were already in favor.

AlisdairM
--
SG16 mailing list
SG16@lists.isocpp.org
https://lists.isocpp.org/mailman/listinfo.cgi/sg16