Thanks for your reply.
-std=c++98 or -std=c++03 doesn't help.
There is also one more line of code with an error, something like:
complex.real()=0;
I'm not sure what that does. It looks like it just sets the real value to 0.
I have this short test code:
#include <iostream>
#include <complex>
using namespace std;
int main(int argc, const char *argv[]){
complex<double> x;
x.real(1);
cout<<x.real()<<'\n';
x.real()=0;
cout<<x.real()<<'\n';
x.real(2);
cout<<x.real()<<'\n';
double* y=&x.real();
cout<<*y<<'\n';
}
It has the same problem as the source code I'm trying to build.
I could try the changes you proposed but the member variables are private
so I would have to modify the library file.
Here is the declaration in the library file:
template<>
class _LIBCPP_TYPE_VIS_ONLY complex<double>
{
double __re_;
double __im_;
public:
typedef double value_type;
_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR complex(double __re = 0.0,
double __im = 0.0)
: __re_(__re), __im_(__im) {}
_LIBCPP_CONSTEXPR complex(const complex<float>& __c);
explicit _LIBCPP_CONSTEXPR complex(const complex<long double>& __c);
_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR double real() const {return
__re_;}
_LIBCPP_INLINE_VISIBILITY _LIBCPP_CONSTEXPR double imag() const {return
__im_;}
_LIBCPP_INLINE_VISIBILITY void real(value_type __re) {__re_ = __re;}
_LIBCPP_INLINE_VISIBILITY void imag(value_type __im) {__im_ = __im;}
_LIBCPP_INLINE_VISIBILITY complex& operator= (double __re)
{__re_ = __re; __im_ = value_type(); return *this;}
_LIBCPP_INLINE_VISIBILITY complex& operator+=(double __re) {__re_ +=
__re; return *this;}
_LIBCPP_INLINE_VISIBILITY complex& operator-=(double __re) {__re_ -=
__re; return *this;}
_LIBCPP_INLINE_VISIBILITY complex& operator*=(double __re) {__re_ *=
__re; __im_ *= __re; return *this;}
_LIBCPP_INLINE_VISIBILITY complex& operator/=(double __re) {__re_ /=
__re; __im_ /= __re; return *this;}
template<class _Xp> _LIBCPP_INLINE_VISIBILITY complex& operator= (const
complex<_Xp>& __c)
{
__re_ = __c.real();
__im_ = __c.imag();
return *this;
}
template<class _Xp> _LIBCPP_INLINE_VISIBILITY complex& operator+=(const
complex<_Xp>& __c)
{
__re_ += __c.real();
__im_ += __c.imag();
return *this;
}
template<class _Xp> _LIBCPP_INLINE_VISIBILITY complex& operator-=(const
complex<_Xp>& __c)
{
__re_ -= __c.real();
__im_ -= __c.imag();
return *this;
}
template<class _Xp> _LIBCPP_INLINE_VISIBILITY complex& operator*=(const
complex<_Xp>& __c)
{
*this = *this * complex(__c.real(), __c.imag());
return *this;
}
template<class _Xp> _LIBCPP_INLINE_VISIBILITY complex& operator/=(const
complex<_Xp>& __c)
{
*this = *this / complex(__c.real(), __c.imag());
return *this;
}
};
It apears to be the same library file from llvm. I'm using the fastcomp
compiler. So if there is no other solution I will go ahead and try to
modify the library file source and then build the project.
On Thu, Jul 10, 2014 at 3:16 AM, Jukka Jylänki <[email protected]> wrote:
> You could try building with -std=c++03 to test if that solves the problem.
>
> I am assuming that the code is something like this:
>
> double COMPLEX::real() const { return this->real_value; }
> double COMPLEX::imag() const { return this->imag_value; }
>
> in which case the two assert()s you have are broken assumptions, in both
> emscripten and native: the assert()s won't hold if you take the address of
> a temporary, they'll be pointing to something else. If you change your code
> to something like
>
> const double &COMPLEX::real() const { return this->real_value; }
> const double &COMPLEX::imag() const { return this->imag_value; }
> double &COMPLEX::real() { return this->real_value; }
> double &COMPLEX::imag() { return this->imag_value; }
>
> or
>
> COMPLEX x;
> COMPLEX* px = &x;
> double* prx = &x.real_value;
> double* pix = &x.imag_value;
>
> assert(reinterpret_cast<void*>(prx) == reinterpret_cast<void*>(px));
> assert(reinterpret_cast<void*>(pix-1) == reinterpret_cast<void*>(px));
>
> then that should fix the code up. The first form returns by reference
> rather than by value, and the second form directly takes in addresses of
> the actual COMPLEX member vars proper, and not some temporaries.
>
>
> 2014-07-10 6:08 GMT+03:00 nunya biznez <[email protected]>:
>
>> I'm trying to compile some c++ code and I have some issues with using the
>> complex library.
>> The code does some funny pointer stuff with the complex class like this:
>>
>> COMPLEX x;
>> COMPLEX* px = &x;
>> double* prx = &x.real();
>> double* pix = &x.imag();
>> assert(reinterpret_cast<void*>(prx) == reinterpret_cast<void*>(px));
>> assert(reinterpret_cast<void*>(pix-1) == reinterpret_cast<void*>(px));
>>
>> and this:
>>
>> tr_load_point(_n[ii], _n[jj], &(_matrix[nii][njj].real()),
>> &(_matrix[nii][njj].imag()));
>>
>> The error I get is cannot take the address of an rvalue of type 'double'.
>>
>> I get the same errors when I use clang by itself with "-std=c++11 " ,but
>> it compiles fine without that setting.
>> Same thing with gcc.
>>
>> I was hoping a compiler setting for emscripten would solve the problem.
>>
>> Thanks.
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "emscripten-discuss" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected].
>>
>> For more options, visit https://groups.google.com/d/optout.
>>
>
> --
> You received this message because you are subscribed to a topic in the
> Google Groups "emscripten-discuss" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/emscripten-discuss/OnzpRJAxzO0/unsubscribe
> .
> To unsubscribe from this group and all its topics, send an email to
> [email protected].
> For more options, visit https://groups.google.com/d/optout.
>
--
You received this message because you are subscribed to the Google Groups
"emscripten-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
