https://issues.dlang.org/show_bug.cgi?id=19902

Jonathan M Davis <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[email protected]
                   |                            |m

--- Comment #1 from Jonathan M Davis <[email protected]> ---
Indeed. Without it being updated, any existing code that's testing for whether
it's okay to blit a type is broken when dealing with structs that have copy
constructors (e.g. I expect that std.algorithm's swap is currently broken in
such a fashion). I don't know if there's code that would break because it's
really assuming postblit constructors from hasElaborateCopyConstructor rather
than the inability to blit, but I expect that more code will be broken without
updating the trait than with requiring a new one. Also, there's already been at
least one post in D.Learn where someone was confused, because
hasElaborateCopyConstructor wasn't true for types with copy constructors.

--

Reply via email to