CWG3150 Incomplete consteval-only class types

[jensmaurer]

There is ongoing, active discussion both on-list and off-list about CWG3150.

Some of the ideas being floated (not yet materialized in wording suggestions) alter the design around consteval-only types sufficiently to require EWG approval.

This issue serves to reserve a scheduling slot in EWG for that topic; any discussions in EWG should have CWG representation.

[hanickadot]

Note for scheduling: see it on Tuesday at earliest based on CWG's request.

[jensmaurer]

Related papers:

• https://isocpp.org/files/papers/D4101R0.html Consteval-only Values for C++26 (Barry Revzin et al) P4101 R0 Consteval-only Values for C++26 #2672
• https://isocpp.org/files/papers/D4135R0.pdf Consteval-only Types (Wyatt Childers) P4135 R0 Consteval-only Types #2677
• https://isocpp.org/files/papers/D4132R0.html

[jensmaurer]

CWG 2026-03-23 in Croydon: Discussed D4135R0 (consteval-only types) and D4101R0 (consteval-only values). CWG is tied on whether to address this issue for C++26, or postpone to C++29. CWG considers the approaches in both papers viable (implementable and solving the problem space), with slightly more confidence in the constexpr-only values approach. CWG prefers the constexpr-only values approach, with a majority favoring including support for pointers to immediate functions (also for template arguments).

[hanickadot]

EWG in Croydon:

EWG wants to solve the CWG-3150 issue for C++26

SF	F	N	A	SA
4	20	14	19	6

Result: not consensus

Encourage more work in direction P4101r0 "consteval-only values for C++26" as a solution to CWG-3150

SF	F	N	A	SA
36	18	4	1	1

Result: strong consensus

Encourage more work in direction P4135r0 "consteval-only types" as a solution to CWG-3150

SF	F	N	A	SA
6	11	31	6	8

Result: not consensus