On Thursday, 29 August 2024 at 16:44:46 UTC, Nick Treleaven wrote: > Should we make S() for any struct an error in the next edition? I meant any struct *with a constructor*.