Getting random/weird data using Winsock

Question

I'm new to using Winsock and to see what it could do I made a quick HTTP client (not really) that simply requested the index page of a website. However, when trying to read the data received from the server, it is not what I would expect. Here's some example output:

ÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌ¤÷

This happens when connecting to any website I try and I get this from trying to output the contents of recvbuf, where the data that is received is supposed to be stored. Is this how data is normally returned from winsock? Also, I don't think it's my code because I don't get any errors. I'm pretty sure this is normal and I'm just missing a step, but searching didn't turn up anything so I'm asking here.

EDIT: Sorry, can't believe I forgot my code:

#ifndef WIN32_LEAN_AND_MEAN
#define WIN32_LEAN_AND_MEAN
#endif

#include "stdafx.h"
#include <winsock2.h>
#include <windows.h>
#include <ws2tcpip.h>
#include <iphlpapi.h>
#include <iostream>

#pragma comment(lib, "Ws2_32.lib")

using namespace std;

int _tmain(int argc, _TCHAR* argv[])
{
    addrinfo *result, *ptr, hints;
ZeroMemory(&hints, sizeof(hints));
hints.ai_family = AF_UNSPEC;
hints.ai_socktype = SOCK_STREAM;
hints.ai_protocol = IPPROTO_TCP;

WSADATA wsaData;
int iResult = WSAStartup(MAKEWORD(2, 2), &wsaData);
iResult = getaddrinfo("google.com", "80", &hints, &result);
if (iResult != 0) {
    cout << "Error in getaddrinfo()!\n";
    cin.get();
    return 1;
} else {
    cout << "Success!\n";
}

SOCKET ConnectSocket = INVALID_SOCKET;

ptr = result;

ConnectSocket = socket(ptr->ai_family, ptr->ai_socktype, ptr->ai_protocol);

iResult = connect(ConnectSocket, ptr->ai_addr, int(ptr->ai_addrlen));

freeaddrinfo(result);

int recvbuflen = 512;
char *sendbuf = "GET / http/1.1";
char recvbuf[512];

send(ConnectSocket, sendbuf, (int)strlen(sendbuf), 0);

shutdown(ConnectSocket, SD_SEND);

do {
    iResult = recv(ConnectSocket, recvbuf, recvbuflen, 0);
    cout << recvbuf << endl;
} while (iResult > 0);

cin.get();
return 0;
}

What kind of data are you expecting to receive? cout expects a nul terminated ASCII string. — ScottMcP-MVP, May 31 '14 at 23:34
I was thinking I'd get normal HTML code back. Sorry if that's somewhat stupid to expect, but I'm a complete noob to winsock. — diodepain, May 31 '14 at 23:36
*sigh*. Check the value returned by `recv` before printing the buffer. — Captain Obvlious, May 31 '14 at 23:43
@CaptainObvlious, after checking it seems like the connection closes immediately. I'm pretty sure I'm getting nothing from the server. — diodepain, May 31 '14 at 23:48
The request is invalid, use `char *sendbuf = "GET / HTTP/1.1\r\n\r\n";` instead. Notice the capitalization of HTTP and the required trailing CRLF pairs. — Retired Ninja, Jun 01 '14 at 07:42

score 1 · Answer 1 · edited Oct 07 '21 at 06:15

Your HTTP request is faulty. Read RFC 2616. Try this instead:

char *sendbuf = "GET / HTTP/1.1\r\nHost: google.com\r\n\r\n";

That is the bare minimum needed for an HTTP 1.1 request.

Since you are closing the send portion of your socket, you should also include a Connection: close header, since you will not be reusing the socket for subsequent requests:

char *sendbuf = "GET / HTTP/1.1\r\nHost: google.com\r\nConnection: close\r\n\r\n";

If you send an HTTP 1.0 request instead, you can omit the Host and Connection headers (Host is not used in 1.0, and close is the default behavior in 1.0):

char *sendbuf = "GET / HTTP/1.0\r\n\r\n";

That being said, your reading loop is too simple, that is not how HTTP responses should be read. You need to read the HTTP headers line-by-line until you encounter a blank line, then parse the headers and read the rest of the HTTP body, if any, based on how the headers tell you to read (see RFC 2616 Section 4.4 for details). See this other answer I wrote for a proper workflow.

score -1 · Answer 2 · edited May 09 '16 at 17:34

-1

I know little about "GET / http/1.1", but Google use HTTPS.

So I replace "google.com" by "web.mit.edu", and goto refresh "http://web.mit.edu/".

And it really works.

edited May 09 '16 at 17:34

halfer

19,824
17
99
186

answered Jun 01 '14 at 00:26

loverszhaokai

127
10

Google primarily uses HTTPS, but it will respond to HTTP and redirect to HTTPS as needed. – Remy Lebeau Jun 01 '14 at 18:35
But i failed by "google.com" which in fact is "https ://www.google.com" and succeeded by "web.mit.edu". – loverszhaokai Jun 03 '14 at 07:07

Getting random/weird data using Winsock

2 Answers2