E_WEB Server Ver. 1.2c
by The Wizards and hard workers of E_MOO
Documentation by GE@PythonMOO [In progress]

Intro
-----
E_WEB is a web server for your moo that lets objects return html codes to a port
on the server from inside the MOO.

Features
--------
Basic Install : Main web page, all players/objects can return html text to make
their objects readable to the world.
Extra Install (Basic + Generic Web Request Handler) : VRML, who list, view
object code, view objects, read mailing lists, insert CGI.

Installation
------------
   1. Basic Install: Insert this text into your MOO
   2. Extras: Insert this text into your MOO
   3. @set $httpd.server_name to "your moo server"
   4. @set $httpd.server_port to <port # you want to run the server on>
      NOTE: Do not use the default port 80 if you are on a system with a real
web server.
   5. Edit #0:server_started and insert this line:
      $httpd:start_listening()
      This will start the http object with the MOO.
   6. Type: ;$httpd:start_listening()
      Starts the http object now.
   7. Edit $httpd.home_page_body to return your default home page html.
   8. Make :http_request(tnt)(+x) verbs on objects to allow them to return html
code.

Port 80 Frolics
---------------
You may wish to enable people to use the http service through the default port
80. If a conventional web server is already running, this can be difficult. This
can be worked around quite effectively with the Apache webserver, and possibly
others.
This solution proxies certain requests through to the moo, an advantage to this
is that you can also serve non-moo content (eg images and static documents)
without adding extra load.
You should probably talk to your sysop about doing this, since it requires
server-wide changes.

In httpd.conf, add:
LoadModule proxy_module	<path to>/libproxy.so
LoadModule rewrite_module  <path to>/mod_rewrite.so
<IfModule mod_proxy.c>
  ProxyRequests Off
</IfModule>

The following line should be added to the MOO virtual host definition in
httpd.conf
ProxyPassReverse / http://127.0.0.1:E_WEBport

This will enable the proxy module, but disallow the conventional proxy support.
Both are recommended for security reasons. The last line would be defined in
.htaccess, but is not allowed there, it will rewrite any REDIRECT headers sent
by the E_WEB to the proper counterpart.
That is all that needs changing in the main server config, you should restart
apache so it re-reads the config file.

Create a .htaccess file in your designated document root:
RewriteEngine On
RewriteCond $1 !apache(.*)
RewriteCond $1 !manual(.*)
RewriteRule (.*) http://127.0.0.1:E_WEBport/$1 [P]

The above will catch all requests BUT the ones to /apache /manual or /mirror, or
any items below them. The final line will redirect any that are not one of the
above and pass the request through to the E_WEB. In the case of this moo, the
apache stores images and other non-moo files, manual contains the Lambda
programmers manual.

Verbs
-----
$httpd Core

:start_listening()
=> port or error
Starts the MOO server listening on $httpd.server_port. Run by default from
$server_started()

:do_login_command()
=> 0
Called by MOO server when first line of data is sent down a connection.
Arguments to this are ignored, instead, input is immediately held, and then
read. It will try to log, debug, and handle_request (those verbs in that order)
and if failing, will return a html traceback via the connection.

:log_connection()
Unprogrammed

:debug(msg, description)
=> msg
When .debug==1, will tell the room it's in: description (or caller): msg

:handle_request(list lines)
=> 0
This block of code handles the request, parsing the url and passing it to the
appropriate handler. Text recieved back from the handler is checked for headers,
and if there are none, then applies the default ok set. Text then passed to
:notify_connection.

:complete_url(path)
=> url
Returns full URL of given path, using $httpd.server_name and .server_port
Eg: "/test" => "http://domain.name:port/test"

:parse_url(path)
=> list
Returns a URL path in the form of a list
Example: $httpd:parse_url("/users/Fabuley?Planet=Neptune")
           => {"/users", "/Fabuley", "?Planet=Neptune"}

:GET(list url) (aka :HEAD)
=> list page_content
GET and HEAD method handler, matches object, if not valid it returns error page,
if valid will call the request verb (usually http_request) on that object.
Returns whatever the request verb returns. If a HEAD request, and the request
verb does not return headers, will return default ok.

:POST

:match_object(path)
=> object
This verb takes the first element of a list of the form from :parse_url. If
empty, it will return $web_objects.default ($httpd); if it's a number, and an
object with that number exists and has a http_request (or whatever your
$httpd.request_method_name is) verb, then it will be returned; if is a property
on $web_objects, then that value returned; finally other objects such as players
and mailing lists are checked.