Creating a raw HTTP request with sockets

问题:

I would like to be able to construct a raw HTTP request and send it with a socket. Obviously, you would like me to use something like urllib and urllib2 but I do not want to use that.

It would have to look something like this:

import socket

tcpsoc = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
tcpsoc.bind(('72.14.192.58', 80)) #bind to googles ip
tcpsoc.send('HTTP REQUEST')
response = tcpsoc.recv()

Obviously you would also have to request the page/file and get and post parameters

回答1:

Most of what you need to know is in the HTTP/1.1 spec, which you should definitely study if you want to roll your own HTTP implementation: http://www.w3.org/Protocols/rfc2616/rfc2616.html

回答2:

import socket
import urlparse


CONNECTION_TIMEOUT = 5
CHUNK_SIZE = 1024
HTTP_VERSION = 1.0
CRLF = "\r\n\r\n"

socket.setdefaulttimeout(CONNECTION_TIMEOUT)


def receive_all(sock, chunk_size=CHUNK_SIZE):
    '''
    Gather all the data from a request.
    '''
    chunks = []
    while True:
        chunk = sock.recv(int(chunk_size))
        if chunk:
            chunks.append(chunk)
        else:
            break

    return ''.join(chunks)



def get(url, **kw):
    kw.setdefault('timeout', CONNECTION_TIMEOUT)
    kw.setdefault('chunk_size', CHUNK_SIZE)
    kw.setdefault('http_version', HTTP_VERSION)
    kw.setdefault('headers_only', False)
    kw.setdefault('response_code_only', False)
    kw.setdefault('body_only', False)
    url = urlparse.urlparse(url)
    sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
    sock.settimeout(kw.get('timeout'))
    sock.connect((url.netloc, url.port or 80))
    msg = 'GET {0} HTTP/{1} {2}'
    sock.sendall(msg.format(url.path or '/', kw.get('http_version'), CRLF))
    data = receive_all(sock, chunk_size=kw.get('chunk_size'))
    sock.shutdown(socket.SHUT_RDWR)
    sock.close()

    data = data.decode(errors='ignore')
    headers = data.split(CRLF, 1)[0]
    request_line = headers.split('\n')[0]
    response_code = request_line.split()[1]
    headers = headers.replace(request_line, '')
    body = data.replace(headers, '').replace(request_line, '')


    if kw['body_only']:
        return body
    if kw['headers_only']:
        return headers
    if kw['response_code_only']:
        return response_code
    else:
        return data


print(get('http://www.google.com/'))

回答3:

Yes, basically you just have to write text, something like :

GET /pageyouwant.html HTTP/1.1[CRLF]
Host: google.com[CRLF]
Connection: close[CRLF]
User-Agent: MyAwesomeUserAgent/1.0.0[CRLF]
Accept-Encoding: gzip[CRLF]
Accept-Charset: ISO-8859-1,UTF-8;q=0.7,*;q=0.7[CRLF]
Cache-Control: no-cache[CRLF]
[CRLF]

Feel free to remove / add headers at will.

回答4:

For a working example to guide you, you might want to take a look at libcurl, a library written in the C language that:

1. does what you want and much more;

2. is a snap to use;

3. is widely deployed; and

4. is actively supported.

It's a beautiful thing and one of the best examples of what open source can and should be.