P0009 : Address 2017-11 LEWG feedback to advance to LWG

[hcedwar]

Address 2017-11 LEWG feedback to advance to LWG.

[dsunder]

• nullptr constructor removed
• constexpr added to operator()( ... )
• fixed example typos
• moved examples to appendix
• Changed note in 3.2.2 to a see appendix
  @hcedwar I changed note to see_below, but not sure if it is in the exact form they want, please review

---

TODO

@hcedwar small group requested that the oral explanation that you and Christian gave in section 3.5.2 about "layout mapping of indices ... to codomain" be specified in detail

Small group requested ability to index by template arguments, but large group showed no interest,
should we mention that in the Appendix?

[dhollman]

@dsunder why did we remove the nullptr constructor again? Don’t we want this to be a nullable type?

[hcedwar]

Feedback from LEWG on span and mdspan From: David Hollman <notifications@github.com> Reply-To: kokkos/array_ref <reply@reply.github.com> Date: Sunday, February 11, 2018 at 8:22 AM To: kokkos/array_ref <array_ref@noreply.github.com> Cc: "Edwards, Harold C" <hcedwar@sandia.gov>, Mention <mention@noreply.github.com> Subject: [EXTERNAL] Re: [kokkos/array_ref] P0009 : Address 2017-11 LEWG feedback to advance to LWG (#48) @dsunder<https://github.com/dsunder> why did we remove the nullptr constructor again? Don’t we want this to be a nullable type? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#48 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AHQObYWVcAizcD7VCgAHVWWuZvYmUs8Dks5tTwWngaJpZM4R-qHR>.

[dhollman]

That was feedback on span, and it was pretty controversial. I don't think span intends to remove the nullptr constructor. I think it was made by someone who wasn't familiar with the importance of a nullptr constructor (their comment was "why not use the default constructor" and there's a really good reason for that: implicit convertibility from nullptr). I went through this in #46

[hcedwar]

From LEWG notes (included in #46): Poll to remove nullptr_t constructor: 4-5-3-0-0

[dhollman]

I guess I was mixing it up with other discussion; sorry.