p1039 - Flexible Array Members for C++29

[ThePhD]

This paper just need to be updated to fix typos and other stuff in it to be resubmitted. I don't expect it to be seriously considered at all until literally 1 year from now.

Fixes include looking at sized deletion operation and adding a callout to allow for that to delete things in a special manner (http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2017/p0722r1.html). The other is to look at the analogous thing for new, however @griwes has greatly warned me of this and @Morwenn has informed me that operator new's current 18 overloads are very much frowned upon.

I don't know if I have a way forward for this one, so I might can it for Pre-Kona and look to just fix it later in C++23.

[ThePhD]

Supersedes #1.

[Morwenn]

The comments about adding new overloads to new: http://wiki.edg.com/bin/view/Wg21sandiego2018/P0894R0

[ThePhD]

@Leandros @ubsan The feedback was strongly positive from Today's C++ Standard Meeting about solving the Flexible Array Member paper. They want far more thought put into doing things like:

• deleting special members
• making sure to retain C compatbility
• spelling out that there are cases where derivation are fine

An implementation of this would likely make this sail through EWG Incubator and EWG.

[Leandros]

That's good news!

[ThePhD]

Mostly good: they voted to solve Flexible Array Members, but definitely wanted the paper to really be put to the grinder and heavily improved. There are a lot of corner cases and things we need to prove out.