[Bug c++/31780] [4.2/4.3 regression] ICE with incompatible types for ?: with "complex type" conversion

[gdr at gcc dot gnu dot org]

------- Comment #17 from gdr at gcc dot gnu dot org  2007-12-23 21:22 -------
(In reply to comment #13)
> Gaby --
> 
> Paolo and I would like your input on this issue, please.
> 
> Thanks,
> 
> -- Mark
> 

Sorry for replying late -- this issue escaped by attention; Paolo
kindly sent me a private reminder.

The std::complex<double> constructor taking a __complex__ double exists
primarily so that we can have a simple implementation of std::complex<double>
that is both compatible with C99 _Complex and the GNU extension __complex__,
and beneficiary of the existing machinery supporting that datatype.
I believe that is a plus we all agree we should keep.

I'm very nervous about adding more constructors.
I'd rather distinguish the constructor taking __complex__ by adding
a dummy parameter:

   enum _DummyArg { };
   complex(__complex__ double __z, _DummyArg);


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31780