https://issues.dlang.org/show_bug.cgi?id=19931

Simen Kjaeraas <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|INVALID                     |---

--- Comment #2 from Simen Kjaeraas <[email protected]> ---
Well, duh. What I'm saying it's just as wrong to define all three as it is to
define two of them, and an error message should be present in both cases.

We could perfectly well define that if there's an rvalue constructor and a copy
constructor, one of them takes precedence and the other is ignored. Instead, we
have an error message. The fact that we allow all three to be defined with no
error message is confusing, and leads the programmer to believe there are no
errors even though his carefully-crafted copy constructor is ignored.

--

Reply via email to