On 14/08/2013 12:05, Brian Gladman wrote:
> On 14/08/2013 11:54, Brian Gladman wrote:
>> On 14/08/2013 11:06, degski wrote:
>>> Hi Brian,
>>>
>>> It does fail for me though:
>>>
>>> Z:\VC\x64\include\mpirxx.h(1596): error : invalid redeclaration of member
>>> function "__gmp_expr<mpz_t, mpz_t>::__gmp_expr(__int64)" (declared at line
>>> 1590)
>>> __gmp_expr(intmax_t l) { mpz_init_set_sx(mp, l); }
>>> ^
>>>
>>> Z:\VC\x64\include\mpirxx.h(1597): error : invalid redeclaration of member
>>> function "__gmp_expr<mpz_t, mpz_t>::__gmp_expr(unsigned __int64)" (declared
>>> at line 1591)
>>> __gmp_expr(uintmax_t l) { mpz_init_set_ux(mp, l); }
>>> ^
>>> This already happens when just including the header mpirxx.h in a project
>>> (VS2012/release/x64).
>>
>> Looking at the tests, there is no test that checks things when <cstdint>
>> is included. I suspect that this is why I don't see these errors.
>
> As I suspected, when I include <cstdint> the errors turn up.
>
> So when <cstdint> is included, we need to:
>
> (a) check if (u)intmax_t is a 'long' and, if it is:
> (b) check if there is an equivalent 'long' overload;
> (c) if there is, don't declare any overloads on (u)intmax_t
>
> (d) check if (u)intmax_t is a 'long long' and, if it is:
> (e) check if there is an equivalent 'long long' overload;
> (f) if there is, don't declare any overloads on (u)intmax_t
>
> If there are no situations where (u)intmax_t is not either 'long' or
> 'long long' we could simply remove all these declarations.
>
> Otherwise I wonder what are the right set of MACROS to achieve the above
> guards?
Answering (partly) my own question, it seems that C++11 has the solution
using <type_traits>:
---------------------------------------------
#include <type_traits>
template< class T, class U > struct is_same;
---------------------------------------------
so we could just remove these defines for 'old' use and add this
solution for 'new' C++ code. But this would mean adding <type_traits>
to mpirxx.h. Would this cause any problems?
>
> Brian
>
--
You received this message because you are subscribed to the Google Groups
"mpir-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to mpir-devel+unsubscr...@googlegroups.com.
To post to this group, send email to mpir-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/mpir-devel.
For more options, visit https://groups.google.com/groups/opt_out.
