Converting 4bytes to signed and unsigned ints

Question

I want to convert a four-byte string to either int32 or uint32 in c++. This answer helped me to write the following code:

#include <string>
#include <iostream>
#include <stdint.h>

int main()
{
    std::string a("\xaa\x00\x00\xaa", 4);

    int u = *(int *) a.c_str();
    int v = *(unsigned int *) a.c_str();
    int x = *(int32_t *) a.c_str();
    int y = *(uint32_t *) a.c_str();
    std::cout << a << std::endl;
    std::cout << "int: " << u << std::endl;
    std::cout << "uint: " << v << std::endl;
    std::cout << "int32_t: " << x << std::endl;
    std::cout << "uint32_t: " << y << std::endl;

    return 0;
}

However the outputs are all the same, where the int (or int32_t) values are correct, but the unsigned ones are wrong. Why does not the unsigned conversions work?

// output
int:      -1442840406
uint:     -1442840406
int32_t:  -1442840406
uint32_t: -1442840406

Pythons struct.unpack, gives the right conversion

In [1]: import struct
In [2]: struct.unpack("<i", b"\xaa\x00\x00\xaa")
Out[2]: (-1442840406,)
In [3]: struct.unpack("<I", b"\xaa\x00\x00\xaa")
Out[3]: (2852126890,)

I would also like a similar solution to work for int16 and uint16, but first things first, since I guess an extension would be trivial if I manage to solve this problem.

Because you're placing them all in (signed) `int` variables. — Rotem, Nov 24 '18 at 19:45
Ahhh! How stupid of me... Been pulling my hair for hours. Thanks. — user10668188, Nov 24 '18 at 19:49

score 0 · Accepted Answer · answered Nov 24 '18 at 19:46

You need to store the unsigned values in an unsigned variables and it will work:

#include <string>
#include <iostream>
#include <stdint.h>

int main()
{
    std::string a("\xaa\x00\x00\xaa", 4);

    int u = *(int *) a.c_str();
    unsigned int v = *(unsigned int *) a.c_str();
    int32_t x = *(int32_t *) a.c_str();
    uint32_t y = *(uint32_t *) a.c_str();
    std::cout << a << std::endl;
    std::cout << "int: " << u << std::endl;
    std::cout << "uint: " << v << std::endl;
    std::cout << "int32_t: " << x << std::endl;
    std::cout << "uint32_t: " << y << std::endl;

    return 0;
}

When you cast the value to unsigned and then store it in a signed variable, the compiler plays along. Later, when you print the signed variable, the compiler generates code to print a signed variable output.

Great! The biggest risk of being lazy and just copy-paste rows... Thanks! — user10668188, Nov 24 '18 at 20:45

Converting 4bytes to signed and unsigned ints

1 Answers1