Hexfloat manipulator and precision

问题

How come output using the hexfloat manipulator ignores any precision on ostream?

#include <iostream>
#include <cmath>
#include <iomanip>

using namespace std;

int main(){

    cout << setw(17) << left << "default format: " << setw(20) << right << 100 * sqrt(2.0) << " " <<  cout.precision() << '\n'
         << setw(17) << left << "scientific: " << setw(20) << right << scientific << 100 * sqrt(2.0) << " " <<  cout.precision() <<  '\n'
         << setw(17) << left << "fixed decimal: " << fixed << setw(20) << right << 100 * sqrt(2.0) << " " <<  cout.precision() <<  '\n'
         << setw(17) << left << "hexadecimal: " << hexfloat << setw(20) << right << uppercase << 100 * sqrt(2.0) << nouppercase << " " <<  cout.precision() <<  '\n'
         << setw(17) << left << "use defaults: " << defaultfloat << setw(20) << right << 100 * sqrt(2.0) << " " <<  cout.precision() <<  "\n\n";

}

Despite a default precision of 6 this seems to get ignored when outputting the double in hexadecimal format (coliru) (gcc 5.2.0):

default format:               141.421 6
scientific:              1.414214e+02 6
fixed decimal:             141.421356 6
hexadecimal:     0X1.1AD7BC01366B8P+7 6
use defaults:                 141.421 6

Is it possible to ensure a decimal precision of 6 using hexadecimal format?

回答1:

The std::hexfloat format is intended to dump out the exact repesentation of a floating point value. The string represention isn't for human consumption but primarily to restore the exact same representation again. Thus, it does not make sense to truncate the format in any way based on any stream setting.

来源：https://stackoverflow.com/questions/34323580/hexfloat-manipulator-and-precision

标签

c++

c++11

hex

precision