How to create a simple Proxy to access web servers in C

Question

I’m trying to create an small Web Proxy in C. First, I’m trying to get a webpage, sending a GET frame to the server. I don’t know what I have missed, but I am not receiving any response. I would really appreciate if you can help me to find what is missing in this code.

int main (int argc, char** argv) {
   int cache_size,     //size of the cache in KiB
       port,
       port_google = 80,
       dir,
       mySocket,
       socket_google;

   char google[] = "www.google.es", ip[16];
   struct sockaddr_in socketAddr;
   char buffer[10000000];

   if (GetParameters(argc,argv,&cache_size,&port) != 0)
       return -1;

   GetIP (google, ip);
   printf("ip2 = %s\n",ip);

   dir = inet_addr (ip);
   printf("ip3 = %i\n",dir);

   /* Creation of a socket with Google */
   socket_google = conectClient (port_google, dir, &socketAddr);
   if (socket_google < 0) return -1;
   else printf("Socket created\n");

   sprintf(buffer,"GET /index.html HTTP/1.1\r\n\r\n");
   if (write(socket_google, (void*)buffer, MESSAGE_LENGTH+1) < 0 )
       return 1;
   else printf("GET frame sent\n");

   strcpy(buffer,"\n");
   read(socket_google, buffer, sizeof(buffer));

   // strcpy(message,buffer);
   printf("%s\n", buffer);

   return 0;
}

And this is the code I use to create the socket. I think this part is OK, but I copy it just in case.

int conectClient (int puerto, int direccion, struct sockaddr_in *socketAddr) {
   int mySocket;
   char error[1000];

   if ( (mySocket = socket(AF_INET, SOCK_STREAM, 0)) == -1) {
       printf("Error when creating the socket\n");
       return -2;
   }

   socketAddr->sin_family = AF_INET;
   socketAddr->sin_addr.s_addr = direccion;
   socketAddr->sin_port = htons(puerto);

   if (connect (mySocket, (struct sockaddr *)socketAddr,sizeof (*socketAddr)) == -1) {
       snprintf(error, sizeof(error), "Error in %s:%d\n", __FILE__, __LINE__);
       perror(error);
       printf("%s\n",error);
       printf ("-- Error when stablishing a connection\n");
       return -1;
   }
   return mySocket;
}

Thanks!

Answer 1

Since you didn't post the GetIP routine, I am not certain that your hostname lookup is correct, as from the looks of it, I am not sure that you are using inet_addr function correctly.

Nikolai has pointed out some very good points (and I fully agree). In fact you GET request is actually broken, and while I was testing it on my own local Apache web server on my system, it didn't work.

sprintf(buffer,"GET /index.html HTTP/1.1\r\n\r\n");
if (write(socket_google, (void*)buffer, LONGITUD_MSJ+1) < 0 )
    return 1;
else printf("GET frame sent\n");
...

strcpy(buffer,"\n");
read(socket_google, buffer, sizeof(buffer));

should be replaced with

  snprintf(buffer, sizeof(buffer), 
      "GET / HTTP/1.1\r\nHost: %s\r\nUser-Agent: TEST 0.1\r\n\r\n", 
      google);

  if (write(socket_google, buffer, strlen(buffer)+1) < 0 ) {
      close(socket_google);
      return 1;
  } else 
      printf("GET frame sent\n");
  ...

  buffer[0] = '\0';
  /* Read message from socket */
  bytes_recv = read(socket_google, buffer, sizeof(buffer));
  if (bytes_recv < 0) {
       fprintf(stderr, "socket read error: %s\n", strerror(errno));
       close(socket_google);
       exit(10);
  }

  buffer[bytes_recv] = '\0';    /* NUL character */

  /* strcpy(message,buffer); */
  printf("%s\n", buffer);

  ...

You should also close the socket before exiting the program. Enable standard C89/90 or C99 mode of your compiler (e.g. -std=c99 for gcc) and enable warnings (e.g. -Wall for gcc), and read them. And #include the necessary header files (assuming Linux in my case) for function prototypes:

 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <unistd.h>
 #include <errno.h>
 #include <sys/types.h>
 #include <sys/socket.h>
 #include <netinet/in.h>
 #include <arpa/inet.h>
 #include <netdb.h>      /* for gethostbyname() */

There is some casting of pointers and structs in regards to the hostname / IP address resolving, which can be confusing and easy place to make a mistake, so verify that is working as you expect it is.

 in_addr_t ip;
 ...

 GetIP(google, &ip);   /* I changed the parameters */
 printf("IP address = %x (%s)\n", 
     ip, 
     inet_ntoa(*((struct in_addr*)&ip)));

Answer 2

First, you're not checking how many bytes the write(2) call actually wrote to the socket. The return value of the call tells you that. Same for the read(2). TCP socket is a bi-directional stream, so as a rule always do both in a loop until expected number of bytes is transferred, EOF is read (zero return from read(2)), or an error occurred (which you are not checking for when reading either).

Then HTTP is rather complex protocol. Make yourself familiar with RFC 2616, especially application level connection management and transfer encodings.

Edit 0:

Hmm, there's no such thing as "simple" proxy. You need to manage multiple connections (at least client-to-proxy and proxy-to-server), so it's probably best to look into select(2)/poll(2)/epoll(4)/kqueue(2) family of system call, which allow you to multiplex I/O. This is usually combined with non-blocking sockets. Look into helper libraries like libevent. Look at how this is done in good web-servers/proxies like nginx. Sound like it's a lot for you to discover, but don't worry, it's fun :)

Answer 3

Actually, I've been implementing a small web proxy using my library called rzsocket link to it.

One of the most difficult things I've found when implementing the web proxy, perhaps this might also be your problem, was that, in order to make the proxy work properly, I had to set keep-alive settings false. One way of doing this in FireFox, is accessing about:config address, and setting the value of network.http.proxy.keep-alive to false.