int main(void)
{
int sockfd;
char filler[16] = {0};
/*
* this is the container for socket's address that contains
* address family, ip address, port
*/
struct sockaddr serv_addr;
/*
* creates a socket of family Internet sockets (AF_INET) and
* of type stream. 0 indicates to system to choose appropriate
* protocol (eg: TCP)
*/
sockfd = socket(AF_INET, SOCK_STREAM, 0);
/*
* Address represented by struct sockaddr:
* first 2 bytes: Address Family,
* next 2 bytes: port,
* next 2 bytes: ipaddr,
* next 8 bytes: zeroes
*/
/*
* htons() and htonl() change endianness to
* network order which is the standard for network
* communication.
*/
filler[0] = (unsigned short)AF_INET;
filler[2] = (unsigned short)htons(port);
filler[4] = (unsigned long)htonl(INADDR_ANY);
/* copy data in the filler buffer to the socket address */
memcpy(&serv_addr, filler, sizeof(serv_addr));
/* makes connection per the socket address */
connect(sockfd, &serv_addr, sizeof(serv_addr));
interact_with_server(sockfd);
return 0;
}
int main(void)
{
int sockfd;
/*
* struct sockaddr defines a socket address.
* A socket address is a combination of address family,
* ip address and port.
* For IP sockets, we may use struct sockaddr_in which is
* just a wrapper around struct sockaddr.
* Funtions like bind() etc are only aware of struct sockaddr.
*/
struct sockaddr_in serv_addr;
/*
* creates a socket of family Internet sockets (AF_INET) and
* of type stream. 0 indicates to system to choose appropriate
* protocol (eg: TCP)
*/
sockfd = socket(AF_INET, SOCK_STREAM, 0);
/*
* Socket adddress represented by struct sockaddr:
* first 2 bytes: Address Family,
* next 2 bytes: port,
* next 4 bytes: ipaddr,
* next 8 bytes: zeroes
*/
/*
* htons() and htonl() change endianness to
* network order which is the standard for network
* communication.
*/
serv_addr.sin_family = AF_INET;
serv_addr.sin_port = htons(port);
serv_addr.sin_addr.s_addr = htonl(INADDR_ANY);
/*
* The above achieves what could be done using the following
* on a little endian machine.
* This breaks if the structure has padding
char filler[16] = {0};
filler[0] = AF_INET & 0xFF;
filler[1] = AF_INET >> 8 & 0xFF;
filler[2] = htons(port) & 0xFF;
filler[3] = htons(port) >> 8 & 0xFF;
filler[4] = htonl(INADDR_ANY) & 0xFF;
filler[5] = htonl(INADDR_ANY) >> 8 & 0xFF;
filler[6] = htonl(INADDR_ANY) >> 16 & 0xFF;
filler[7] = htonl(INADDR_ANY) >> 24 & 0xFF;
memcpy(&serv_addr, filler, sizeof(serv_addr));
*/
/* makes connection per the socket address */
connect(sockfd, (struct sockaddr*) &serv_addr, sizeof(serv_addr));
interact_with_server(sockfd);
/* close socket since we are done */
close(sockfd);
return 0;
}
int main(void)
{
int sockfd;
/*
* struct sockaddr defines a socket address.
* A socket address is a combination of address family,
* ip address and port.
* For IP sockets, we may use struct sockaddr_in which is
* just a wrapper around struct sockaddr.
* Funtions like bind() etc are only aware of struct sockaddr.
*/
struct sockaddr_in serv_addr;
/*
* creates a socket of family Internet sockets (AF_INET) and
* of type stream. 0 indicates to system to choose appropriate
* protocol (eg: TCP)
*/
sockfd = socket(AF_INET, SOCK_STREAM, 0);
/*
* Socket adddress represented by struct sockaddr:
* first 2 bytes: Address Family,
* next 2 bytes: port,
* next 4 bytes: ipaddr,
* next 8 bytes: zeroes
*/
/*
* htons() and htonl() change endianness to
* network order which is the standard for network
* communication.
*/
serv_addr.sin_family = AF_INET;
serv_addr.sin_port = htons(port);
serv_addr.sin_addr.s_addr = htonl(INADDR_ANY);
/*
* The above achieves what could be done using the following
* on a little endian machine.
* This breaks if the structure has padding
char filler[16] = {0};
filler[0] = AF_INET & 0xFF;
filler[1] = AF_INET >> 8 & 0xFF;
filler[2] = htons(port) & 0xFF;
filler[3] = htons(port) >> 8 & 0xFF;
filler[4] = htonl(INADDR_ANY) & 0xFF;
filler[5] = htonl(INADDR_ANY) >> 8 & 0xFF;
filler[6] = htonl(INADDR_ANY) >> 16 & 0xFF;
filler[7] = htonl(INADDR_ANY) >> 24 & 0xFF;
memcpy(&serv_addr, filler, sizeof(serv_addr));
*/
/*
* Note that we do not bind() our socket to any socket adddress here.
* This is because on the client side, you would only use bind() if you want
* to use a particular client side port to connect to the server.
* When you do not bind(), the kernel will pick a port for you.
* Read here how kernel gets you a port: https://idea.popcount.org/2014-04-03-bind-before-connect
* There are a few protocols in the Unix world that expect clients to connect from a particular port.
* Create a new socket address definition and bind it to socket in such cases:
struct sockaddr_in client_addr;
client_addr.sin_family = AF_INET;
client_addr.sin_port = htons(CLIENT_PORT);
client_addr.sin_addr.s_addr = htonl(INADDR_ANY);
bind(sockfd, (struct sockaddr*) &client_addr, sizeof client_addr);
*/
/* makes connection per the socket address */
connect(sockfd, (struct sockaddr*) &serv_addr, sizeof(serv_addr));
interact_with_server(sockfd);
/* close socket since we are done */
close(sockfd);
return 0;
}
