https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83484
Jonathan Wakely <redi at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords|link-failure |accepts-invalid
Status|UNCONFIRMED |NEW
Last reconfirmed| |2017-12-24
Ever confirmed|0 |1
--- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Wilhelm M from comment #3)
> Here ist a simplified version of a MCVE:
>
> template<int>
> struct X {
> static constexpr int x = 0;
> static constexpr int y = 1;
> };
>
> template<>
> constexpr int X<1>::x = 2; // should be ok, but gives duplicate
> initialization error
The compiler is correct.
> template<>
> constexpr int X<1>::y; // should be ill-formed (constexpr must be
> initialized), but gives undefined reference
How do you get an undefined reference if the previous definition gives an
error?
But I think you're right that it should be ill-formed, and that makes the
original testcase ill-formed. So the bug is "accepts-invalid" not
"link-failure".