|
Boost Users : |
From: Chris Newbold (Chris.Newbold_at_[hidden])
Date: 2007-09-17 18:50:01
On Mon, 17 Sep 2007 14:33:53 -0400, Doug Gregor wrote:
> On Sep 17, 2007, at 10:25 AM, thomas free wrote:
>> ************************************ //Change here
>>
>> #ifdef _WIN64
>>
>> template <> struct IntegerConcept<__int64> { void constraints()
>> {} };
>>
>> template <> struct IntegerConcept<unsigned __int64> { void
>> constraints() {} };
>>
>> #endif
>>
>> ************************************
>>
>> #endif
>>
>>
>>
>> When I put this two lines, the code compile and link has well has on
>> Win32 platform
>
> Those look fine; I've updated the Boost trunk accordingly.
That's not going to help other 64-bit platforms like Linux, or for the
use of 64-bit types like 'long long' on 32-bit systems.
We've been bitten in the past by large discrepancies between
IntegerConcept and is_integral in the type_traits library. Is there a
reason why IntegerConcept doesn't just use is_integral? At the very
least, it would be nice if the two had consistent answers for all types.
-Chris
Boost-users list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net