From: Gary Wright <gwtmp01@...>
Date: 2007-04-09T08:18:39+09:00
Subject: Re: Trying to GET google with socket....problem


On Apr 8, 2007, at 11:57 AM, Hey You wrote:
> Also putting the Host header or adding the full
> domain to the code such as "GET www.google.ca/index.html" both  
> specifies
> which host we want so I don't see why change them.

The URI provided in the GET request can be an absolute URI only if
the request is going to a proxy server.  In *that* case the GET would
look like:

    GET http://proxy.domain.com/index.html

Otherwise the URI must be an absolute path (i.e., a path starting  
with '/').
In that case the GET would look like:

    GET /index.html

The problem with only having the path is that a web server that is  
hosting
several websites can't determine from the GET request which site the
request pertains to.  The incoming TCP connections only have a  
destination
IP address, not a destination domain name.  The solution to this problem
is the "Host:" header.  By looking at the "Host:" header, the web server
can multiplex several websites at the same IP address.  Without the  
Host:
header you would have to have a separate IP address for every website.

So your request should be sent as:

GET /index.html HTTP/1.0
Host: www.google.ca