Issaquah (Feb 2023) P2630 LEWG feedback

Customization point design does matter even for experts, because there are performance implications.

Add a new feature test macro (don't just bump the number of the old one).

LWG has a different way to say "strided_slice is an aggregate type." See my comment below for the words of power.

In the wording for "submdspan function," we lost the definition of sub_map (the result of submdspan_mapping). Also, we probably shouldn't call the result of submdspan_mapping "sub_map_offset"?

Do we want to permit anything with a value member that's a constant expression, not just integral_constant? Just need exposition-only concept or trait. Has a constexpr static value data member. "static-integral-constant-like" etc.

Polled changes:

1. Add [a new] feature test macro
2. Relax requirements to "integral constant like"

Link to feedback: https://github.com/cplusplus/papers/issues/1293

Start with the linked feedback in papers issue 1293 and the wording for the “submdspan function” mentioned here. Review the proposed feature-test macro, the strided_slice aggregate wording, the sub_map definition, and the integral-constant-like requirement; done means the two polled changes and the related wording issues are resolved in the paper.