web_sessions provides a server to register 'sessions' belonging
to an http request.  It should be used in conjunction with
web_router.  At the beginning of the request it looks to 
see if the client has a cookie representing a session id.

If so it looks up the session id and find the process responsible
and then clones it for use in the rest of the request.  At
the end of the request the clone is then synchronized
back into the main session process.

If a cookie or session id doesn't exist, then a new session is created
and before the request is returned to the client a Set-Cookie header
is placed so the client will continue to use the same session.

Each web_session is setup to allow for arbitrary storage whether
for storing the user information after authentication code or
simply storing 'flash' messages.  You can also set a timeout option
in which case the web_session will terminate itself after x milliseconds
of being idle (by default the timeout is infinity).

Persistence of the web_session has not yet been implemented.

web_sessions depends upon:
- web_router

and must be started by another supervisor process.