[llvm] r210920 - Finishing touch for the std::error_code transition.

[Hal Finkel]

----- Original Message -----
> From: "Jeffrey Yasskin" <jyasskin at google.com>
> To: "Chandler Carruth" <chandlerc at google.com>
> Cc: "Hal Finkel" <hfinkel at anl.gov>, "Marshall Clow" <mclow.lists at gmail.com>, "Rafael Espindola"
> <rafael.espindola at gmail.com>, "Commit Messages and Patches for LLVM" <llvm-commits at cs.uiuc.edu>, "Richard Smith"
> <richard at metafoo.co.uk>
> Sent: Friday, June 13, 2014 3:40:59 PM
> Subject: Re: [llvm] r210920 - Finishing touch for the std::error_code transition.
> 
> On Fri, Jun 13, 2014 at 7:45 PM, Chandler Carruth
> <chandlerc at google.com> wrote:
> >
> > On Fri, Jun 13, 2014 at 6:42 PM, Hal Finkel <hfinkel at anl.gov>
> > wrote:
> >>
> >> > +// * std::errc is a namespace in some implementations. That
> >> > meas
> 
> I'm not sure exactly what Rafael means here, but it sounds like those
> implementations don't match the standard. [syserr] requires std::errc
> to be an enum class, not a namespace.
> 
> >> > that ADL
> >> > +//   doesn't work and it is sometimes necessary to write
> >> > std::make_error_code
> >> > +//   or in templates:
> >> > +//   using std::make_error_code;
> >> > +//   make_error_code(...);
> >> > +//
> >> > +//   with this enum it is safe to always just use
> >> > make_error_code.
> >>
> >> Should we file a defect report?
> >
> >
> > Marshall or Jeffrey would know better.
> 
> Defect reports about ADL working badly are either a lost cause, or a
> commitment to a huge amount of work.

That's not what I meant; I meant to constrain std::errc from being a namespace. But it sounds like that is already done ;)

 -Hal

> 

-- 
Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory