mimicing boost::format()

[metulburr]

I was trying to mimic boost::format(). This was before I found out that boost had this method, but now I would like to know the underlining of it any ways. So i am pretty much trying to indicate the int number of decimal places to cut off at in the second argument.

I either am way off or I confused myself with this and that, but essentially it does not output the expected output.

#include <iostream>
#include <string>
#include <sstream>
#include <iomanip>
#include <boost/format.hpp>


double format(double d, int place){
    double dec = d - (int)d, d_out;
    
    std::stringstream ss;
    ss << std::setprecision(place) << dec;
    ss >> d_out;
    std::cout << "d is: " << d << " dec is: " << dec << " over " << place << " decimal spaces" << std::endl;
    return d_out + (int)d;
}

int main () {
  std::cout << format(.0123456789, 2) << std::endl;
  std::cout << format(11.1123456789, 5) << std::endl;
  std::cout << format(1234.2123456789, 3) << std::endl;
  std::cout << format(1234.3123456789, 2) << std::endl;

  double x = 7.40200133400;
  std::cout << boost::format("%1$.2f")%x << std::endl;
  
}

with the output being:

d is: 0.0123457 stripped is: 0.0123457 over 2 decimal spaces
0.012
d is: 11.1123 stripped is: 0.112346 over 5 decimal spaces
11.1123
d is: 1234.21 stripped is: 0.212346 over 3 decimal spaces
1234.21
d is: 1234.31 stripped is: 0.312346 over 2 decimal spaces
1234.31
7.40

everything but boost is wrong.

[norm b]

Unless I misunderstand, you are way over thinking this:

std::cout << std:: fixed;
std::cout << std::setprecision(2) << 0.0123456789 << std::endl;
std::cout << std::setprecision(5) << 11.2123456789 << std::endl;

Output:

0.01
11.21235

See here: http://www.cplusplus.com/reference/ios/fixed/

[JLBorges]

> I would like to know the underlining of it

This is not how boost::format works, but it should give you a few ideas:

#include <iostream>
#include <string>
#include <sstream>
#include <iomanip>
#include <boost/format.hpp>

std::string format( double d, int place )
{
    std::ostringstream ss;
    ss << std::fixed << std::setprecision(place) << d ;
    return ss.str() ;
}

struct format_a
{
    format_a( double d, int place ) : value(d), prec(place) {}

    const double value ;
    const int prec ;
};

template < typename OSTREAM >
OSTREAM& operator<< ( OSTREAM& stm, const format_a& fmt )
{ return stm << std::fixed << std::setprecision(fmt.prec) << fmt.value ; }

struct format_b
{
    explicit format_b( int place, int w = 0, char f = ' ' )
             : prec(place), width(w), fill(f) {}

    template < typename T > format_b& operator% ( const T& v )
    {
        std::ostringstream ss;
        ss << std::fixed << std::setprecision(prec)
           << std::setw(width) << std::setfill(fill) << v ;
        chars_out += ss.str() ;
        return *this ;
    }

    const int prec ;
    const int width ;
    const char fill ;
    std::string chars_out ;
};

template < typename OSTREAM >
OSTREAM& operator<< ( OSTREAM& stm, const format_b& fmt )
{ for( char c : fmt.chars_out ) stm << stm.widen(c) ; return stm ; }

int main ()
{
  std::cout << format( .0123456789, 2 ) << '\n' ;
  std::cout << format( 11.1123456789, 5 ) << '\n' ;

  std::cout << format_a( 1234.2123456789, 3 ) << '\n' ;
  std::wcout << format_a( 1234.3123456789, 2 ) << '\n' ;

  std::cout << format_b(3) % 1234.2123456789 << '\n' ;
  std::wcout << format_b( 2, 9 ) % "values:" % 999.999 % 1234.3123456789 % '\n' ;

  double x = 7.40200133400;
  std::cout << boost::format("%1$.2f")%x << '\n' ;
}

http://ideone.com/MQAufS

[metulburr]

thanks JLBorges

oh ok returning it as a double instead of string is what is making the hassle.