Re: [hybi] workability (or otherwise) of HTTP upgrade

[Zhong Yu <[e-mail]>]

The two arguments are mutually strengthening each other. If cookies
are good, we better use HTTP syntax. If HTTP syntax is good, we better
as well ship cookies. I thought there is some circularity there, and
we better rethink state management in isolation. But I give up,
dependency on cookies is too strong, and HTTP syntax is probably
unavoidable.

On Fri, Dec 10, 2010 at 3:39 AM, Greg Wilkins <[e-mail]> wrote:
> On 9 December 2010 20:42, Zhong Yu <[e-mail]> wrote:
>> I believe HTTP cookies shouldn't be included in WS, and I'll make that
>> case in time.
>
> If WS starts with a normal HTTP request, then I think that request
> should fully respect the current state of the art with regards to
> cookies.  It is a HTTP request and will be routed and handled by
> existing HTTP mechanisms, many of which use Cookies as part of their
> security, routing or other value add. Whatever you think of cookies as
> a design, adding additional exceptions and provisions for special HTTP
> requests is just going to add to confusion and make worse designs.
>
> If WS does not use a normal HTTP request as the handshake, then I
> don't see any need to use cookies.   I think CONNECT is not a "normal"
> HTTP request and handling of it is treated specially, so I see less
> (or zero) need for cookies with CONNECT.
>
> cheers
>