[LLVMdev] [3.5 Release] Release Candidate 4 Now Available

[Larry Evans]

On 09/02/2014 03:32 PM, Larry Evans wrote:
> On 08/28/2014 02:07 PM, Bill Wendling wrote:
>> We had to roll a release candidate 4 for the 3.5 release. It’s up at the normal place:
>>
>> 	http://llvm.org/pre-releases/3.5
>>
>> Please test it and report any major bugs you may find.
>>
>> Thanks!
>> -bw
>>
> Please see https://gist.github.com/cppljevans/73b4e2d357b801be822d
> That gist contains a Makefile, a main program, and an .hpp file,
> and outputs of 3 compilations
> (in the  "problem description and compiler outputs" file,
> A.K.A the .out file, in the gist).
> 
> AFAICT, clang3.5rc3 does require an out-of-class definition when
> the standard does not require one.  The code example shows
> 1 case (using std::pair) where clang behaves as expected; however,
> the 2nd case (using something like std::array) clang does *not*
> behave as expected with no out-of-class definition.
> 
> The .out file in the gist provides more details about the code
> run and the main file contains many References to the standard
> and elsewhere justifying why I think this is a bug.
> 
The stackoverflow entry here:

http://stackoverflow.com/questions/23428684/is-a-constexpr-array-necessarily-odr-used-when-subscripted

had an answer by Richard Smith indicating the array index expression
*was* an odr-use.

Sorry for noise, but as other people's comments on that
stackoverflow page indicate, odr-use is not easily understood :(

Sorry for noise.

-regards,
Larry