CWG2463 Conditions for trivially copyable classes

[jensmaurer]

See CWG2463

CWG suggests to turn the current rule of "if all copy/move operations are trivial" into a rule saying "if any copy/move operation is trivial".

[Dani-Hub]

I'm wondering whether we might to need to distinguish trivially-assignable from trivially-copyable on the core language level (duck!)

[erichkeane]

EWG unanimously agrees that this would be better with a paper. It has been changed frequently, and we need a rational justification for what trivially copyable/triviality in general means, and how we should identify it, in a paper.

[jfbastien]

Discussed in Issaquah. Still needs a paper. We discussed what such a paper might look like.
Ville volunteered to attempt a paper.