Getting error code when creating a socket in Linux

Question

I'm doing some socket programming in Linux and am wondering how to get the error code when the function socket(...); fails.

for example for the "getaddrinfo" function i can do this:

//Resolve the server address and port
    result = (struct addrinfo *) calloc(1, sizeof(struct addrinfo));
    iResult = getaddrinfo("google.com", DEFAULT_PORT, &hints, &result);
    if (iResult != 0){
        printf("%d\n", iResult);
        fprintf(stderr, "getaddrinfo failed: %s\n", gai_strerror(iResult));
        getchar();
        exit(EXIT_FAILURE);
    }

However I want to do a similar thing using socket(...) function.

According to this: http://linux.die.net/man/2/socket

the function returns -1 on failure, and sets errno to the appropriate error number. How do i access this "errno" though? This is my code so far:

int connectSocket = 0;
connectSocket = socket(AF_INET, SOCK_STREAM, 0);
printf("%d\n", connectSocket);

if (connectSocket == -1){
    printf("socket failed with error: %s\n", error_string); //TODO: HELP DECLARING error_string
    getchar();
    exit(EXIT_FAILURE);
}

Answer 1

errno is a thread-local global variable, defined in <errno.h>. The man page for many library functions will indicate that they return -1 on error, and set errno.

You can convert an errno value to a useful string with the strerror function.

In general, you should code like this:

#include <stdio.h>
#include <errno.h>

int main(void) {
    int s;

    s = socket(...);
    if (s < 0) {
        fprintf(stderr, "socket() failed: %s\n", strerror(errno));
        exit(1);
    }
}

Alternatively, glibc's printf and friends support a %m format specifier, which is replaced with strerror(errno) (no argument is needed). So the above example could be replaced with:

    if (s < 0) {
        fprintf(stderr, "socket() failed: %m\n");
        exit(1);
    }

And to make it all simpler, there is the perror function, which prints out a message similar to above.

    if (s < 0) {
        perror("socket");
        exit(1);
    }

Wrapping it all up - error handling need not be complex and verbose. Putting the socket call and the test for < 0 in one statement, the above code could look like this, and you'll be a real UNIX pro:

#include <stdio.h>
#include <errno.h>

int main(void) {
    int s;

    if ((s = socket(...)) < 0) {
        perror("socket");
        exit(1);
    }
}