P3471R4 Standard Library Hardening

[StephanTLavavej]

WG21-P3471R4 Standard Library Hardening
WG21-P3697R1 Minor Additions To Standard Library Hardening
WG21-P3878R1 Standard Library Hardening Should Not Use The 'Observe' Semantic

Feature-test macros (note that they're a mix of 202502L and 202506L):

#define __cpp_lib_hardened_array             202502L
#define __cpp_lib_hardened_basic_stacktrace  202506L
#define __cpp_lib_hardened_basic_string      202502L
#define __cpp_lib_hardened_basic_string_view 202502L
#define __cpp_lib_hardened_bitset            202502L
#define __cpp_lib_hardened_common_iterator   202506L
#define __cpp_lib_hardened_counted_iterator  202506L
#define __cpp_lib_hardened_deque             202502L
#define __cpp_lib_hardened_expected          202502L
#define __cpp_lib_hardened_forward_list      202502L
#define __cpp_lib_hardened_inplace_vector    202502L
#define __cpp_lib_hardened_list              202502L
#define __cpp_lib_hardened_mdspan            202502L
#define __cpp_lib_hardened_optional          202502L
#define __cpp_lib_hardened_shared_ptr_array  202506L
#define __cpp_lib_hardened_span              202502L
#define __cpp_lib_hardened_valarray          202502L
#define __cpp_lib_hardened_vector            202502L
#define __cpp_lib_hardened_view_interface    202506L

• This has been substantially implemented by STL Hardening #5274 and Partially implement P3697R1 Minor Additions To Standard Library Hardening #5650, but R4 was specified to use contracts, so this feature is blocked by P2900R14 <contracts> #5286.

Note: We are not yet accepting PRs for C++26 features.

[frederick-vs-ja]

Thanks to WG21-P3878R1, we've almost completed this, except for hardening inplace_vector and FTMs.

Is it OK to move the controlling macros (which should correspond to FTMs) to <yvals_core.h>? I guess it was intended to add them to <yvals.h>, and perhaps core headers shouldn't rely on these macros.