W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2007

RE: [RFC] Optional header negotitation

From: Robert Brewer <fumanchu@aminus.org>
Date: Sat, 18 Aug 2007 16:52:50 -0700
Message-ID: <9BBC2D2B2CCF7E4DA0E212D1BD0CC6FA1FEEB7@ex10.hostedexchange.local>
To: "Stefanos Harhalakis" <v13@priest.com>
Cc: <ietf-http-wg@w3.org>
Stefanos Harhalakis wrote:
> There is no indention at all to restrict HTTP! Just
> to provide a way for extending it with headers that
> aren't required everywhere.

It already exists: SEND THE HEADERS.

Right now, go write your script to send "Header-Request:
Timezone". I'll even write one for you in CherryPy 3:

    import cherrypy

    class App:
        
        def index(self):
            tz = cherrypy.request.headers.get("Timezone")
            if tz:
                return "Supplied timezone was: %r" % tz
            else:
                cherrypy.response.headers["Header-Request"] = "Timezone"
                return "No timezone supplied."
        index.exposed = True

    cherrypy.quickstart(App())

Run that and inspect the traffic with a tool like TamperData
for Firefox. Did the "Header-Request" header make it to the
client? (Yes.) Pretend you're a client that understands the
received "Header-Request", and send a "Timezone" header in
your next request. Did it make it to the server? (Yes.) Did
the server understand it? (Yes.) Are we violating the HTTP
spec? (No.) Did we make any changes to the HTTP spec? (No.)
Would there be any difference in the messages if we added
"Header-Request" to the spec? (No.) So why do it? Masochism?


Robert Brewer
fumanchu@aminus.org
Received on Saturday, 18 August 2007 23:53:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:15 GMT