Proposed resolution:

This wording is relative to N5054.

1. Modify [version.syn], header <> synopsis, as indicated:

   […]
   #define __cpp_lib_char8_t                           201907L
     // freestanding, also in <atomic>, <filesystem>, <iosfwd>, <istream>, <limits>, <locale>, <ostream>,
     // <string>, <string_view>
   
   #define __cpp_lib_charconv_results                  202306L // also in <charconv>
   
   #define __cpp_lib_chrono                            202306L // also in <chrono>
   […]
   

It is mentioned in P3505R4 that the paper should be implemented as a defect report against C++17. However, P3505R4 bumps `__cpp_lib_to_chars`, while P2497R0 already bumped the same feature-test macro in C++26, and currently there is no way to indicating that the implementation has implemented P3505R4 but not P2497R0.

P2497R0 looks like a pure new feature which eases use of result types and thus doesn't seem suitable for DR. Additionally, libc++ has fully implemented the functional changes in P2497R0 but can't bump the FTM because libc++ still lacks long double support in <charconv>. So, maybe it would be better to make use another FTM for indicating P2497R0.