Compilation errors with Clang and C++11 or C++14 (in func_trigonometric.inl)

[kankaristo]

I'm getting compilation errors when compiling with Clang using C++11 or C++14. With GCC and C++11/14 everything compiles fine. I'm compiling on Ubuntu 15.10.

Adding std:: in front of the asinh, acosh and atanh on the lines which give the errors seems to fix this.

I assume the #if GLM_HAS_CXX11_STL isn't working correctly here?

    // asinh
#   if GLM_HAS_CXX11_STL
        this seems to be removed // by the preprocessor, even if I compile with C++11 or C++14
        using std::asinh;
#   else
        template <typename genType> 
        GLM_FUNC_QUALIFIER genType asinh(genType const & x)
        {
            GLM_STATIC_ASSERT(std::numeric_limits<genType>::is_iec559, "'asinh' only accept floating-point input");

            return (x < static_cast<genType>(0) ? static_cast<genType>(-1) : (x > static_cast<genType>(0) ? static_cast<genType>(1) : static_cast<genType>(0))) * log(std::abs(x) + sqrt(static_cast<genType>(1) + x * x));
        }
#   endif

    template <typename T, precision P, template <typename, precision> class vecType>
    GLM_FUNC_QUALIFIER vecType<T, P> asinh(vecType<T, P> const & v)
    {
        return detail::functor1<T, T, P, vecType>::call(asinh, v);  // <- error here, if I don't change `asinh` -> `std::asinh`
    }

Full error log:

In file included from Release/swig_glm_wrap.cxx:2810:
In file included from ../thirdparty/glm/glm.hpp:110:
In file included from ../thirdparty/glm/trigonometric.hpp:35:
In file included from ../thirdparty/glm/detail/func_trigonometric.hpp:206:
../thirdparty/glm/detail/func_trigonometric.inl:181:51: error: address of overloaded function 'asinh' does not match required type 'float (float)'
  return detail::functor1<T, T, P, vecType>::call(asinh, v);
                                                  ^~~~~
Release/swig_glm_wrap.cxx:109070:17: note: in instantiation of function template specialization 'glm::asinh<float, glm::precision::highp, tvec2>' requested here
  result = glm::asinh((glm::tvec2< float,glm::highp > const &)*arg1);
                ^
../thirdparty/glm/detail/func_trigonometric.inl:170:18: note: candidate template ignored: could not match 'const type-parameter-0-0 &' against 'float'
  inline genType asinh(genType const & x)
                 ^
../thirdparty/glm/detail/func_trigonometric.inl:179:23: note: candidate template ignored: could not match '<type-parameter-0-0, P>' against 'float'
 inline vecType<T, P> asinh(vecType<T, P> const & v)
                      ^
../thirdparty/glm/detail/_vectorize.hpp:58:38: note: passing argument to parameter 'Func' here
  inline static tvec2<R, P> call(R (*Func) (T x), tvec2<T, P> const & v)
                                     ^
In file included from Release/swig_glm_wrap.cxx:2810:
In file included from ../thirdparty/glm/glm.hpp:110:
In file included from ../thirdparty/glm/trigonometric.hpp:35:
In file included from ../thirdparty/glm/detail/func_trigonometric.hpp:206:
../thirdparty/glm/detail/func_trigonometric.inl:181:51: error: address of overloaded function 'asinh' does not match required type 'float
 (float)'
  return detail::functor1<T, T, P, vecType>::call(asinh, v);
                                                  ^~~~~
Release/swig_glm_wrap.cxx:109097:17: note: in instantiation of function template specialization 'glm::asinh<float, glm::precision::highp,
 tvec3>' requested here
  result = glm::asinh((glm::tvec3< float,glm::highp > const &)*arg1);
                ^
../thirdparty/glm/detail/func_trigonometric.inl:170:18: note: candidate template ignored: could not match 'const type-parameter-0-0 &' ag
ainst 'float'
  inline genType asinh(genType const & x)
                 ^
../thirdparty/glm/detail/func_trigonometric.inl:179:23: note: candidate template ignored: could not match '<type-parameter-0-0, P>' again
st 'float'
 inline vecType<T, P> asinh(vecType<T, P> const & v)
                      ^
../thirdparty/glm/detail/_vectorize.hpp:67:38: note: passing argument to parameter 'Func' here
  inline static tvec3<R, P> call(R (*Func) (T x), tvec3<T, P> const & v)
                                     ^
In file included from Release/swig_glm_wrap.cxx:2810:
In file included from ../thirdparty/glm/glm.hpp:110:
In file included from ../thirdparty/glm/trigonometric.hpp:35:
In file included from ../thirdparty/glm/detail/func_trigonometric.hpp:206:
../thirdparty/glm/detail/func_trigonometric.inl:181:51: error: address of overloaded function 'asinh' does not match required type 'float
 (float)'
  return detail::functor1<T, T, P, vecType>::call(asinh, v);
                                                  ^~~~~
Release/swig_glm_wrap.cxx:109124:17: note: in instantiation of function template specialization 'glm::asinh<float, glm::precision::highp,
 tvec4>' requested here
  result = glm::asinh((glm::tvec4< float,glm::highp > const &)*arg1);
                ^
../thirdparty/glm/detail/func_trigonometric.inl:170:18: note: candidate template ignored: could not match 'const type-parameter-0-0 &' ag
ainst 'float'
  inline genType asinh(genType const & x)
                 ^
../thirdparty/glm/detail/func_trigonometric.inl:179:23: note: candidate template ignored: could not match '<type-parameter-0-0, P>' again
st 'float'
 inline vecType<T, P> asinh(vecType<T, P> const & v)
                      ^
../thirdparty/glm/detail/_vectorize.hpp:76:38: note: passing argument to parameter 'Func' here
  inline static tvec4<R, P> call(R (*Func) (T x), tvec4<T, P> const & v)
                                     ^
In file included from Release/swig_glm_wrap.cxx:2810:
In file included from ../thirdparty/glm/glm.hpp:110:
In file included from ../thirdparty/glm/trigonometric.hpp:35:
In file included from ../thirdparty/glm/detail/func_trigonometric.hpp:206:
../thirdparty/glm/detail/func_trigonometric.inl:202:51: error: address of overloaded function 'acosh' does not match required type 'float
 (float)'
  return detail::functor1<T, T, P, vecType>::call(acosh, v);
                                                  ^~~~~
Release/swig_glm_wrap.cxx:109241:17: note: in instantiation of function template specialization 'glm::acosh<float, glm::precision::highp,
 tvec2>' requested here
  result = glm::acosh((glm::tvec2< float,glm::highp > const &)*arg1);
                ^
../thirdparty/glm/detail/func_trigonometric.inl:189:18: note: candidate template ignored: could not match 'const type-parameter-0-0 &' ag
ainst 'float'
  inline genType acosh(genType const & x)
                 ^
../thirdparty/glm/detail/func_trigonometric.inl:200:23: note: candidate template ignored: could not match '<type-parameter-0-0, P>' again
st 'float'
 inline vecType<T, P> acosh(vecType<T, P> const & v)
                      ^
../thirdparty/glm/detail/_vectorize.hpp:58:38: note: passing argument to parameter 'Func' here
  inline static tvec2<R, P> call(R (*Func) (T x), tvec2<T, P> const & v)
                                     ^
In file included from Release/swig_glm_wrap.cxx:2810:
In file included from ../thirdparty/glm/glm.hpp:110:
In file included from ../thirdparty/glm/trigonometric.hpp:35:
In file included from ../thirdparty/glm/detail/func_trigonometric.hpp:206:
../thirdparty/glm/detail/func_trigonometric.inl:202:51: error: address of overloaded function 'acosh' does not match required type 'float
 (float)'
  return detail::functor1<T, T, P, vecType>::call(acosh, v);
                                                  ^~~~~
Release/swig_glm_wrap.cxx:109268:17: note: in instantiation of function template specialization 'glm::acosh<float, glm::precision::highp,
 tvec3>' requested here
  result = glm::acosh((glm::tvec3< float,glm::highp > const &)*arg1);
                ^
../thirdparty/glm/detail/func_trigonometric.inl:189:18: note: candidate template ignored: could not match 'const type-parameter-0-0 &' ag
ainst 'float'
  inline genType acosh(genType const & x)
                 ^
../thirdparty/glm/detail/func_trigonometric.inl:200:23: note: candidate template ignored: could not match '<type-parameter-0-0, P>' again
st 'float'
 inline vecType<T, P> acosh(vecType<T, P> const & v)
                      ^
../thirdparty/glm/detail/_vectorize.hpp:67:38: note: passing argument to parameter 'Func' here
  inline static tvec3<R, P> call(R (*Func) (T x), tvec3<T, P> const & v)
                                     ^
In file included from Release/swig_glm_wrap.cxx:2810:
In file included from ../thirdparty/glm/glm.hpp:110:
In file included from ../thirdparty/glm/trigonometric.hpp:35:
In file included from ../thirdparty/glm/detail/func_trigonometric.hpp:206:
../thirdparty/glm/detail/func_trigonometric.inl:202:51: error: address of overloaded function 'acosh' does not match required type 'float
 (float)'
  return detail::functor1<T, T, P, vecType>::call(acosh, v);
                                                  ^~~~~
Release/swig_glm_wrap.cxx:109295:17: note: in instantiation of function template specialization 'glm::acosh<float, glm::precision::highp,
 tvec4>' requested here
  result = glm::acosh((glm::tvec4< float,glm::highp > const &)*arg1);
                ^
../thirdparty/glm/detail/func_trigonometric.inl:189:18: note: candidate template ignored: could not match 'const type-parameter-0-0 &' ag
ainst 'float'
  inline genType acosh(genType const & x)
                 ^
../thirdparty/glm/detail/func_trigonometric.inl:200:23: note: candidate template ignored: could not match '<type-parameter-0-0, P>' again
st 'float'
 inline vecType<T, P> acosh(vecType<T, P> const & v)
                      ^
../thirdparty/glm/detail/_vectorize.hpp:76:38: note: passing argument to parameter 'Func' here
  inline static tvec4<R, P> call(R (*Func) (T x), tvec4<T, P> const & v)
                                     ^
In file included from Release/swig_glm_wrap.cxx:2810:
In file included from ../thirdparty/glm/glm.hpp:110:
In file included from ../thirdparty/glm/trigonometric.hpp:35:
In file included from ../thirdparty/glm/detail/func_trigonometric.hpp:206:
../thirdparty/glm/detail/func_trigonometric.inl:223:51: error: address of overloaded function 'atanh' does not match required type 'float
 (float)'
  return detail::functor1<T, T, P, vecType>::call(atanh, v);
                                                  ^~~~~
Release/swig_glm_wrap.cxx:109412:17: note: in instantiation of function template specialization 'glm::atanh<float, glm::precision::highp,
 tvec2>' requested here
  result = glm::atanh((glm::tvec2< float,glm::highp > const &)*arg1);
                ^
../thirdparty/glm/detail/func_trigonometric.inl:210:18: note: candidate template ignored: could not match 'const type-parameter-0-0 &' ag
ainst 'float'
  inline genType atanh(genType const & x)
                 ^
../thirdparty/glm/detail/func_trigonometric.inl:221:23: note: candidate template ignored: could not match '<type-parameter-0-0, P>' again
st 'float'
 inline vecType<T, P> atanh(vecType<T, P> const & v)
                      ^
../thirdparty/glm/detail/_vectorize.hpp:58:38: note: passing argument to parameter 'Func' here
  inline static tvec2<R, P> call(R (*Func) (T x), tvec2<T, P> const & v)
                                     ^
In file included from Release/swig_glm_wrap.cxx:2810:
In file included from ../thirdparty/glm/glm.hpp:110:
In file included from ../thirdparty/glm/trigonometric.hpp:35:
In file included from ../thirdparty/glm/detail/func_trigonometric.hpp:206:
../thirdparty/glm/detail/func_trigonometric.inl:223:51: error: address of overloaded function 'atanh' does not match required type 'float
 (float)'
  return detail::functor1<T, T, P, vecType>::call(atanh, v);
                                                  ^~~~~
Release/swig_glm_wrap.cxx:109439:17: note: in instantiation of function template specialization 'glm::atanh<float, glm::precision::highp,
 tvec3>' requested here
  result = glm::atanh((glm::tvec3< float,glm::highp > const &)*arg1);
                ^
../thirdparty/glm/detail/func_trigonometric.inl:210:18: note: candidate template ignored: could not match 'const type-parameter-0-0 &' ag
ainst 'float'
  inline genType atanh(genType const & x)
                 ^
../thirdparty/glm/detail/func_trigonometric.inl:221:23: note: candidate template ignored: could not match '<type-parameter-0-0, P>' again
st 'float'
 inline vecType<T, P> atanh(vecType<T, P> const & v)
                      ^
../thirdparty/glm/detail/_vectorize.hpp:67:38: note: passing argument to parameter 'Func' here
  inline static tvec3<R, P> call(R (*Func) (T x), tvec3<T, P> const & v)
                                     ^
In file included from Release/swig_glm_wrap.cxx:2810:
In file included from ../thirdparty/glm/glm.hpp:110:
In file included from ../thirdparty/glm/trigonometric.hpp:35:
In file included from ../thirdparty/glm/detail/func_trigonometric.hpp:206:
../thirdparty/glm/detail/func_trigonometric.inl:223:51: error: address of overloaded function 'atanh' does not match required type 'float (float)'
  return detail::functor1<T, T, P, vecType>::call(atanh, v);
                                                  ^~~~~
Release/swig_glm_wrap.cxx:109466:17: note: in instantiation of function template specialization 'glm::atanh<float, glm::precision::highp,
 tvec4>' requested here
  result = glm::atanh((glm::tvec4< float,glm::highp > const &)*arg1);
                ^
../thirdparty/glm/detail/func_trigonometric.inl:210:18: note: candidate template ignored: could not match 'const type-parameter-0-0 &' ag
ainst 'float'
  inline genType atanh(genType const & x)
                 ^
../thirdparty/glm/detail/func_trigonometric.inl:221:23: note: candidate template ignored: could not match '<type-parameter-0-0, P>' again
st 'float'
 inline vecType<T, P> atanh(vecType<T, P> const & v)
                      ^
../thirdparty/glm/detail/_vectorize.hpp:76:38: note: passing argument to parameter 'Func' here
  inline static tvec4<R, P> call(R (*Func) (T x), tvec4<T, P> const & v)
                                     ^
9 errors generated.

[Groovounet]

I think I fixed this issue in GLM 0.9.7 branch for GLM 0.9.7.6 release.

I don't have a system to reproduce it, so it would be great if you could confirm the fix. Also I am concerned that there is actually 2 bugs here: 1/ Bad C++11 STL detection and 2/ Somehow GLM built-in asinh scalar implementation is not recognized...

I only fixed the first bug, if the second exist.

Thanks for reporting,
Christophe

[kankaristo]

Seems to be fixed now, thanks!

However, after updating to the new version (I was using a version that was a few months old), I get a bunch of new errors and warnings... I'll create a new issue, because the error logs are quite long.

[Groovounet]

ahh, it's very possible this fix is actually causing it... :/

[kankaristo]

They seem unrelated, I was using an older version when I reported this issue, and started getting the errors in #524 after updating to the latest master.