Hi,
On 26/04/11 26/04/11 9:07 PM, "Jonathan Wakely" <jwakely.gcc@xxxxxxxxx>
wrote:
> On 26 April 2011 01:52, Anna Ceguerra wrote:
>>
>> Thanks for your reply, it successfully built in one go, and I successfully
>> installed it. However when I try to compile using the new compiler, I get
>> the following error:
>>
>> /usr/local/lib/gcc/i386-apple-darwin10.7.0/4.6.1/../../../../include/c++/4.6
>> .1/tr1/special_function_util.h: In function âÃÃbool
>> std::tr1::__detail::__isnan(_Tp)âÃÃ:
>> /usr/local/lib/gcc/i386-apple-darwin10.7.0/4.6.1/../../../../include/c++/4.6
>> .1/tr1/special_function_util.h:113:19: error: expected unqualified-id before
>> âÃÃ(âÃà token
>>
>> This is the relevant section in the file:
>>
>> #if _GLIBCXX_USE_C99_MATH && !_GLIBCXX_USE_C99_FP_MACROS_DYNAMIC
>>
>> Â Â/// This is a wrapper for the isnan function. Otherwise, for NaN,
>> Â Â/// all comparisons result in false. If/when we build a std::isnan
>> Â Â/// out of intrinsics, this will disappear completely in favor of
>> Â Â/// std::isnan.
>> Â Âtemplate<typename _Tp>
>> Â Âinline bool __isnan(const _Tp __x)
>> Â Â{
>> Â Â Âreturn std::isnan(__x); Â Â Â // it is failing here
>> Â Â}
>>
>> #else
>>
>> Â Âtemplate<typename _Tp>
>> Â Âinline bool __isnan(const _Tp __x)
>> Â Â{
>> Â Â Âreturn __builtin_isnan(__x);
>> Â Â}
>>
>> Â Âtemplate<>
>> Â Âinline bool __isnan<float>(const float __x)
>> Â Â{
>> Â Â Âreturn __builtin_isnanf(__x);
>> Â Â}
>>
>> Â Âtemplate<>
>> Â Âinline bool __isnan<long double>(const long double __x)
>> Â Â{
>> Â Â Âreturn __builtin_isnanl(__x);
>> Â Â}
>>
>> #endif
>>
>> Am I configuring the program incorrectly?
>
> It's impossible to know because you haven't shown your program. It
> might be a bug in the library, I don't think i386-darwin is a very
> well tested configuration, or it could be something you've done wrong.
>
> Please try to reproduce the error with a minimal example to show
> whether you get that error just from including a header, or only under
> certain circumstances.
I managed to track it down to the following test.cpp file:
#include <cmath>
#include <mgl/mgl.h>
#include <algorithm>
Where mgl is from the wxWidgets library, version 2.8.11
When I changed the order to:
#include <cmath>
#include <algorithm>
#include <mgl/mgl.h>
It worked.
Thanks again for your help.
Regards,
Anna.
