3782. Should <math.h> declare ::lerp?

Section: 17.14.7 [support.c.headers.other] Status: C++23 Submitter: Jiang An Opened: 2022-09-17 Last modified: 2023-11-22 15:47:43 UTC

Priority: Not Prioritized

View other active issues in [support.c.headers.other].

View all other issues in [support.c.headers.other].

View all issues with C++23 status.


According to 17.14.7 [support.c.headers.other]/1, <math.h> is required to provide ::lerp, despite that it is a C++-only component. In P0811R3, neither <math.h> nor C compatibility is mentioned, so it seems unintended not to list lerp as an exception in 17.14.7 [support.c.headers.other].

Currently there is implementation divergence: libstdc++ provide ::lerp in <math.h> but not in <cmath>, while MSVC STL and libc++ don't provide ::lerp in <math.h> or <cmath> at all.

I'm not sure whether this should be considered together with LWG 3484. Since nullptr_t has become a part of the C standard library as of C23 (see WG14-N3042 and WG14-N3048), I believe we should keep providing ::nullptr_t in <stddef.h>.

[2022-10-12; Reflector poll]

Set status to Tentatively Ready after six votes in favour during reflector poll.

[2022-11-12 Approved at November 2022 meeting in Kona. Status changed: Voting → WP.]

Proposed resolution:

This wording is relative to N4917.

  1. Modify 17.14.7 [support.c.headers.other] as indicated:

    -1- Every C header other than <complex.h> (17.14.2 [complex.h.syn]), <iso646.h> (17.14.3 [iso646.h.syn]), <stdalign.h> (17.14.4 [stdalign.h.syn]), <stdatomic.h> (33.5.12 [stdatomic.h.syn]), <stdbool.h> (17.14.5 [stdbool.h.syn]), and <tgmath.h> (17.14.6 [tgmath.h.syn]), each of which has a name of the form <name.h>, behaves as if each name placed in the standard library namespace by the corresponding <cname> header is placed within the global namespace scope, except for the functions described in 28.7.6 [sf.cmath], the std::lerp function overloads (28.7.4 [c.math.lerp]), the declaration of std::byte (17.2.1 [cstddef.syn]), and the functions and function templates described in 17.2.5 [support.types.byteops]. […]