How to use output of C++ program as variables in a Bash script

Question

I'm trying to come up with a work around to perform a few math functions for a script since bash apparently cannot do anything other than integer math (or can it even do that?).

The script I'm coming up with needs to write a series of macros which will eventually be used for a simulation program. I'm currently just trying to output positions of a particle source which are used as parameters of the macro.

The C++ Program I wrote is very simple, it takes in i and outputs x and y as such:

#include <iostream>
#include <math.h>

using namespace std;

int main()
{
  double i;
  double theta = 11 * (3.1415926535897932384626/180);

  cin >> i;

  double b = 24.370;

  double y = (b/i)*sin(theta);
  double x = (b/i)*cos(theta);

  cout << x << " " <<  y << endl;

  return 0;
}

The script I'm writing outputs a bunch of stuff that has to do with the macros i'm trying to create, but the line I'm stuck on (labled as (1) ) needs to do something like this...

for i in {1..100}
do
  echo "./a.out" #calling the C program
  echo "$i" #entering i as input

  x = first output of a.out, y = second output a.out #(1) Confused on how to do this part!

  echo -n "/gps/position $x $y 0 27.7 cm" > mymacro.mac

done

I know there has to be a really easy way to do this but I'm not exactly sure what to do. I basically just need to use the output of a c program as variables in the script.

Answer 1

You should probably consider passing $i to your c++ program as a variable and using argv to read it in (this might be helpful). That's most likely the "proper" way to do it. Then you can use this for bash:

#!/bin/bash
IFS=' ';
for i in {1..100}
do
    read -ra values <<< `./a.out $i`
    x=${values[0]}
    y=${values[1]}
    echo -n "/gps/position $x $y 0 27.7 cm" > mymacro.mac
done

And are you sure you want > mymacro.mac instead of >> mymacro.mac (if the former is inside of a loop, only the last value will be written to the file mymacro.mac.

Answer 2

You can use cegfault's answer, or more simply:

read val1 val2 <<< $(./a.out $i)

Which executes a.out and stores the two numbers in $val1 and $val2.

You may find it even easier to use awk, which does handle floating-point numbers and most math functions. Here's an arbitrary example:

bash> read x y <<< $(echo 5 | awk '{ printf "%f %f\n", cos($1), sin($1) }')
bash> echo $x
0.283662
bash> echo $y
-0.957824

Answer 3

If your script will be long-lived or has a lot of data to process, and especially since you're writing part of the functionality in a C++ program anyway, you'll probably do a lot better to just write the whole thing in C++, and the end result will be a whopping lot faster... and more centralized & easier to see what's going on.

#include <iostream>
#include <math.h>
#include <sstream>
#include <fstream>

int main()
{
  const double theta = 11 * (3.1415926535897932384626/180);
  const double b = 24.370;

  for (int n=1; n<=100; ++n)
  {
    double i = n;
    double y = (b/i)*sin(theta);
    double x = (b/i)*cos(theta);

    // Two ways of sending the output to mymacro.mac......

    // 1. Via the system() call.....
    std::stringstream ss;
    ss << "echo -n \"/gps/position " << x << ' ' << y << " 0 27.7 cm\" > mymacro.mac";
    system(ss.str().c_str());

    // 2. Via C++ file I/O.....
    std::ofstream ofile("mymacro.mac");
    ofile << "/gps/position " << x << ' ' << y << " 0 27.7 cm";
    ofile.close(); //...not technically necessary; ofile will close when it goes out of scope.
  }
}

Note that this solution replicates your example with exquisite fidelity, including the part about overwriting the file 'mymacro.mac' on every loop iteration.