使用stdout stdin将数组从C ++ exe传递给Python

Question

I am struggling with a problem. I have Python program which send array to C++ exe . However I can't receive array from C++. My python code:

import struct
import subprocess
from cStringIO import StringIO

stdin_buf = StringIO()
array = [1.0 for _ in range(10)]
for item in array:
stdin_buf.write(struct.pack('<f', item))

proc = subprocess.Popen(['Comsol1.exe'], stdin=subprocess.PIPE, stdout = subprocess.PIPE)
out, err = proc.communicate(stdin_buf.getvalue())

# assuming the result comes back the same way it went in...
item_len = struct.calcsize('<f')
stdout_buf = StringIO(out)
stdout_buf.seek(0)
for i in range(len(out)/item_len):
   val = struct.unpack('<f', stdout_buf.read(4))
   print (val)

C++ code:

// Comsol1.cpp : Defines the entry point for the console application. //

#include "stdafx.h"
#include <streambuf>
#include "stdafx.h"
#include <stdio.h>
#include <fcntl.h>
#include <io.h>
#include <iostream>

int main(void)
{
int result;

// Set "stdin" to have binary mode:
result = _setmode(_fileno(stdin), _O_BINARY);
if (result == -1)
    perror("Cannot set mode");
else
    fprintf(stderr, "'stdin' successfully changed to binary mode\n");

// Set "stdout" to have binary mode:
result = _setmode(_fileno(stdout), _O_BINARY);
if (result == -1)
    perror("Cannot set mode");
else
    fprintf(stderr, "'stdout' successfully changed to binary mode\n");

int i = 0;
while (!std::cin.eof())
{
    float value;
    std::cin.read(reinterpret_cast<char*>(&value), sizeof(value));
    if (std::cin.gcount() > 0)
    {
        std::cerr << "Car " << i << ": " << value << std::endl;
        i++;
    }
}
      }

Thank you.

Answer 1

So you have two problems here:

1. You are printing to stderr instead of stdout, and because stderr is not piped you'll actually get the messages printed to the console when you run your python script.

2. You are printing more than just floats to stdout and not in raw binary mode. If you expect to read a list of floats from out (in python) you have to print only the floats (in c++) and in binary mode:

   std::cout.write(reinterpret_cast<const char*>(&value), sizeof(value));

I tried the above approach on Ubuntu and it's working properly. You can find my source code here . I had to tweak the code to work on unix, but you can get the idea.