Re: New Version Notification for draft-momoka-httpbis-settings-enable-websockets-00.txt

[Momoka Yamamoto <[e-mail]>]

Hi Ilari,

- Does setting SETTINGS_ENABLE_WEBSOCKETS=1 imply support for necressary
>   parts of extended connect, or does server also need to send
>   SETTINGS_ENABLE_CONNECT_PROTOCOL=1?

The server also needs to send SETTINGS_ENABLE_CONNECT_PROTOCOL=1.
(although SETTINGS_ENABLE_WEBSOCKETS=1 imply support for extended connect)
This is following the discussion in WebTransport wg at IETF115 [1] and the
Editor's Copy of the draft WebTransport over HTTP/3 [2]



- I don't think the default for SETTINGS_ENABLE_WEBSOCKETS should be 0,
>   as I think absence of SETTINGS_ENABLE_WEBSOCKETS is different from
>   SETTINGS_ENABLE_WEBSOCKETS=0. Later signals that websockets is not
>   supported, the former together with SETTINGS_ENABLE_CONNECT_PROTOCOL=1
>   signals that websockets may be supported (as there are currently
>   servers that support websockets and behave that way).

This makes sense. Thank you.
So you are saying there should be no default value,
as the absence of the SETTINGS parameter has it's own meaning.

Thank you,
Momoka

[1] https://datatracker.ietf.org/doc/minutes-115-webtrans-202211101300/
[2]
https://ietf-wg-webtrans.github.io/draft-ietf-webtrans-http3/draft-ietf-webtrans-http3.html#name-extended-connect-in-http-3



On Wed, Mar 1, 2023 at 3:31 AM Ilari Liusvaara <[e-mail]>
wrote:

> On Tue, Feb 28, 2023 at 04:12:21PM +0800, Momoka Yamamoto wrote:
> > Hello,
> > I have submitted a new revision of this
> > draft draft-momoka-httpbis-settings-enable-websockets.
> >
> >
> > The proposed parameter will be useful if an active HTTP/2 (or HTTP/3)
> > connection to the server already exists when a document with a wss:/ URL
> is
> > loaded. The browser can then choose whether to try opening the WebSocket
> > stream over the connection, as the proposed SETTINGS_ENABLE_WEBSOCKETS
> has
> > likely been received by that point.
> >
> > However, the proposed parameter won't be useful when a browser tries to
> > establish a new connection because the SETTINGS will not be received
> when a
> > decision to use HTTP/2 or HTTP/1 is made.
>
> Some comments:
>
> - Does setting SETTINGS_ENABLE_WEBSOCKETS=1 imply support for necressary
>   parts of extended connect, or does server also need to send
>   SETTINGS_ENABLE_CONNECT_PROTOCOL=1?
>
> - I don't think the default for SETTINGS_ENABLE_WEBSOCKETS should be 0,
>   as I think absence of SETTINGS_ENABLE_WEBSOCKETS is different from
>   SETTINGS_ENABLE_WEBSOCKETS=0. Later signals that websockets is not
>   supported, the former together with SETTINGS_ENABLE_CONNECT_PROTOCOL=1
>   signals that websockets may be supported (as there are currently
>   servers that support websockets and behave that way).
>
>
>
> -Ilari
>
>