trappings of stof()

technologist (607)
Please tell me how to keep stof. Ive read it to be problematic. I am getting an error
 
\src\main.cpp:18:9: error: 'stof' is not a member of 'std'

I've made sure that my compiler flags correctly:
-std=c++11

source code:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
 
#include <iostream>
#include <string>
#include <sstream>
#include <string>
#include <cstdlib>

using namespace std;

float get_float(string message)
{
	return std::stof(message);
}
int main()
{
    std::string str_value;
    std::cout << "Enter a numeric string:";
    std::cin >> str_value;

    float f1Value = 0.0f;
    f1Value = get_float(str_value);
    std::cout<<"literal converted value (float):"<<f1Value;

}


If you can tell me how to keep using it I would be very appreciative!
jlb (4973)
What compiler are you using? And don't forget to tell us the version of that compiler as well.


Cubbi (4774)
From the use of backslashes in your path I am guessing you're on Windows, and from the fact it does not work, I am guessing you're using a very old distribution of mingw. Try a modern one from https://nuwen.net/mingw.html or http://mingw-w64.org or just use Visual Studio (it's also free, and is native to your platform)
Last edited on
whu coder (1)
If you use the Code::Blocks IDE, you have to overwrite your three header files before using these functions. Try this http://tehsausage.com/mingw-to-string
Last edited on
technologist (607)
I am using Eclipse IDE, yes windows. Surely there is some solution that doesn't require such a seismic shift for such a small problem. : )

Oh, how do I find out my minGW version?
jlb (4973)
Surely there is some solution that doesn't require such a seismic shift for such a small problem. : )

If your compiler doesn't support stoX() then there is always the always available stringstream approach to convert to strings.

I believe the stoX() functions require a 5.0 or higher version of the g++ compiler and if your compiler doesn't support this function then you will need to either upgrade the compiler or find a compiler that does support this feature.
Thomas1965 (4571)
One way to do it:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
 
#include <sstream>
#include <stdexcept>

float get_float(std::string& message)
{
  std::stringstream ss(message);
  float value;

  ss >> value;
  if (!ss)
    throw std::invalid_argument(message);

  //TODO check for ot of range error

  return value;
}


Alternativ you can use the good old atof, though it returns a double
lastchance (6980)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
 
#include <iostream>
#include <string>
#include <cstdlib>

double get_double(std::string message)           // decide whether to qualify string with std:: or not
{
// return std::stod(message);                    // update your compiler
   return std::strtod( message.c_str(), 0 );     // sledgehammer job for older compilers
}


int main()
{
    std::string str_value;
    std::cout << "Enter a numeric string: ";
    std::cin >> str_value;

    double f1Value = get_double(str_value);
    std::cout<<"literal converted value (double): " << f1Value;
}
Last edited on
JLBorges (13770)
> how do I find out my minGW version?

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
 
#include <iostream>

int main()
{
    #ifdef __MINGW32__
        std::cout << "mingw32  : " << __MINGW32_MAJOR_VERSION << '.' << __MINGW32_MINOR_VERSION << '\n' ;
    #endif // __MINGW32__

    #ifdef __MINGW64__
        std::cout << "mingw64  : " << __MINGW64_VERSION_MAJOR << '.' << __MINGW64_VERSION_MINOR << '\n' ;
    #endif // __MINGW64__

    #ifdef __GNUC__
        std::cout << "g++      : " << __GNUC__ << '.' << __GNUC_MINOR__ << '.' << __GNUC_PATCHLEVEL__ << '\n'
                  << "libstdc++: " << _GLIBCXX_RELEASE << " (release date " << __GLIBCXX__ << ")\n" ;
    #endif // __GNUC__
}
technologist (607)
> how do I find out my minGW version?

getting an error with the version type algorithm above.



..\src\main.cpp:21:39: error: '_GLIBCXX_RELEASE' was not declared in this scope
                   << "libstdc++: " << _GLIBCXX_RELEASE << " (release date " << __GLIBCXX__ << ")\n" ;


I'll continue to work on that and with the other comments.

my includes
1
2
3
4
5
 

#include <iostream>
#include <string>
#include <sstream>
#include <cstdlib>


Lets see whether a compiler change is necessary.

For everyone suggesting a workaround, here is a sample code that needs a change:

1
2
 
\src\main.cpp:573:82: error: 'stof' was not declared in this scope
                vertex.y = stof(dataLine.substr(foundStart, foundEnd - foundStart));

I'm not familiar with atof, but tried it anyway:
1
2
 
.\src\main.cpp:569:87: error: cannot convert 'std::__cxx11::basic_string<char>' to 'const char*' for argument '1' to 'double atof(const char*)'
                vertex.x = std::atof(dataLine.substr(foundStart, foundEnd - foundStart));
Last edited on
technologist (607)
mingw32 : 3.22
g++ : 5.3.0
libstdc++:

I was only able to get this by commenting out (see below) -- for part of the answer.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
 
#include <iostream>
#include <string>
#include <sstream>
#include <cstdlib>

int main()
{
    #ifdef __MINGW32__
        std::cout << "mingw32  : " << __MINGW32_MAJOR_VERSION << '.' << __MINGW32_MINOR_VERSION << '\n' ;
    #endif // __MINGW32__

    #ifdef __MINGW64__
        std::cout << "mingw64  : " << __MINGW64_VERSION_MAJOR << '.' << __MINGW64_VERSION_MINOR << '\n' ;
    #endif // __MINGW64__

    #ifdef __GNUC__
        std::cout << "g++      : " << __GNUC__ << '.' << __GNUC_MINOR__ << '.' << __GNUC_PATCHLEVEL__ << '\n'
                  << "libstdc++: "; //<< _GLIBCXX_RELEASE << " (release date " << __GLIBCXX__ << ")\n" ;
    #endif // __GNUC__
}
JLBorges (13770)
1
2
 
// vertex.x = std::atof(dataLine.substr(foundStart, foundEnd - foundStart));
vertex.x = std::atof( dataLine.substr(foundStart, foundEnd - foundStart).c_str() );

Ideally switch to a current 64-bit version of MinGW (See Cubbi's post for links).
technologist (607)
Will it be backwards compatible?

What will it do to my 32bit installation, and how does my Eclipse IDE find it?

I guess my questions orbit around upgrading my compiler.

Better worded: how do I tell my Eclipse IDE to stop paying attention to the current 32 bit installation and to use the new 64bit?
Last edited on
JLBorges (13770)
> What will it do to my 32bit installation

To keep the existing 32bit installation intact, install the new version of MinGW in to a different directory.


> how do I tell my Eclipse IDE to stop paying attention to the current 32 bit installation and to use the new 64bit?

Consider posting this question on an Eclipse/CDT form.
technologist (607)
Right, thank you.
technologist (607)
Well, I'm back! I've installed MingW64 as per above and have been compiling 64bit programs for a while now. Very cool. The successful compiler change didn't fix the pesky atof error though:

error

1
2
3
 
..\src\main.cpp: In function 'void readObj(std::__cxx11::string)':
..\src\main.cpp:569:87: error: cannot convert 'std::__cxx11::basic_string<char>' to 'const char*' for argument '1' to 'double atof(const char*)'
           vertex.x = std::atof(dataLine.substr(foundStart, foundEnd - foundStart));

Any thoughts?
Cubbi (4774)
You're trying to feed a C++ string to the C function atof. It's either std::stof(dataLine.substr(foundStart, foundEnd - foundStart)); or std::atof(dataLine.substr(foundStart, foundEnd - foundStart).c_str());
Last edited on
technologist (607)
Yeah, I'm weak on c strings/functions.
technologist (607)
That worked, thank you.
Topic archived. No new replies allowed.