It is unclear which semantics were originally intended. CWG should have a look.

From P0593R6 section 3.1, the intent was to "permit implicit object creation for aggregate class types even if the aggregate contains an element with a non-trivial destructor".

How about:

A class S is an implicit-lifetime class if either it is an aggregate or it has at least one trivial eligible constructor and a trivial, non-deleted destructor.

... or switching to bullets:

A class S is an implicit-lifetime class if

• it is an aggregate or
• it has at least one trivial eligible constructor and a trivial, non-deleted destructor.

@jwakely, P0593R6 section 3.1 is pretty straightforward what is meant here.

I've added bullets to clarify the and/or grouping.

@zygoloid: You're certainly right that P0593R6 explicitly mentions "aggregate contains an element with a non-trivial destructor", but that's not the same thing as saying that all aggregates should be allowed, since

struct X {
  Y i;
  ~X();
};

is an aggregate and yet (regardless of the type Y) doesn't satisfy "Destroying an instance of the type runs no code".

CWG telecon 2022-06-17: Approved.

@tkoeppe: CWG has approved this earlier today. Any final thoughts before I merge?

Looks great, merge it! Thanks!